1 Understanding the role of parents’ attitudes in children’s alcohol use: Evaluation of a universal parent-oriented alcohol consumption prevention programme Mariliis Tael-Öeren Department of Public Health and Primary Care School of Clinical Medicine Murray Edwards College University of Cambridge November 2019 This dissertation is submitted for the degree of Doctor of Philosophy 2 Preface This dissertation is the result of my own work and includes nothing which is the outcome of work done in collaboration except as declared in the Preface and specified in the text. It is not substantially the same as any that I have submitted, or, is being concurrently submitted for a degree or diploma or other qualification at the University of Cambridge or any other University or similar institution except as declared in the Preface and specified in the text. I further state that no substantial part of my dissertation has already been submitted, or, is being concurrently submitted for any such degree, diploma or other qualification at the University of Cambridge or any other University of similar institution except as declared in the Preface and specified in the text. The length of this dissertation does not exceed the word limit of 60,000 words, excluding figures, tables, appendices, and references, as prescribed by the Degree Committee of the Faculty of Clinical Medicine. 3 Abstract Understanding the role of parents’ attitudes in children’s alcohol use: Evaluation of a universal parent-oriented alcohol consumption prevention programme Alcohol use is considered as one of the risk factors that contribute to social and physical harm and disease development globally. Alcohol use onset usually takes place long before adulthood and the surrounding environment has a central role in the development of alcohol use behaviour among children. Thus, targeting parents in prevention programmes has become more common in the past 20 years. This thesis aimed to investigate the relationship between parental attitudes towards children’s alcohol use and their children’s alcohol use and to evaluate a universal parent-oriented alcohol use prevention programme. A systematic review and meta-analysis (k=29 studies) of observational and experimental studies was conducted to assess the former. The findings showed that children whose parents had less restrictive attitudes towards children’s alcohol use had 1.45 (95% CI=1.17–1.80) times higher odds of initiating alcohol use, 1.52 (95% CI=1.24–1.86) times higher odds of drinking alcohol frequently, and 1.58 (95% CI=1.35–1.85) times higher odds of getting drunk. A similar relationship was present between perceived parental attitudes and children’s alcohol use frequency (OR=1.76, 95% CI=1.29–2.40), but no evidence was found for a relationship between perceived attitudes and alcohol use initiation (OR=1.65, 95% CI=0.93–2.94). When data from the Effekt trial was included, the change in pooled effect sizes was mostly marginal. The only exception involved the relationship between perceived parental attitudes and children’s alcohol use initiation, which became statistically significant (OR=1.55, 95% CI=1.14–2.11). Parental attitudes were only weakly positively correlated with perceived parental attitudes (r=0.27, p≤0.001). These findings confirmed and extended the earlier findings presented by Swedish researchers 20 years ago who developed a prevention programme (ÖPP, later called Effekt) to delay and reduce alcohol use among 13–16-year-old children by targeting parental attitudes. The programme was adapted and implemented in Estonia in 2012–2015 among the parents of 11–13-year-old children. The format (i.e. two meetings and two newsletters twice a year, agreements between parents) and the main messages focusing on parental attitudes, alcohol supply and communication were kept unchanged; but more emphasis was put on authoritative parenting, other risk behaviours besides alcohol use and active communication with parents. A matched-pair cluster- randomised controlled trial was carried out among children (n=985) and their parents (n=790) in 66 schools (34 intervention, 32 control) in 2012–2015. While no evidence for an 4 intervention effect was found for students’ alcohol use initiation, past year use or drunkenness, the parents in the intervention group had double the odds (OR=1.92, 95% CI=1.31–2.83) of having restrictive attitudes by the end of the programme than the parents in the control group. A qualitative study at the end of the programme with the facilitators (n=8; focus group), teachers (n=12; individual interviews) and parents (n=24; individual interviews) indicated that the participants perceived the programme to have a possible effect on parents’ attitudes and behaviour, but were unsure as to whether this would have an impact on children’s drinking behaviour. One of the main perceived reasons behind reduced effectiveness was the issue of low participation rates at the meetings throughout the programme. Suggested reasons behind that included the general practice of not attending meetings in higher grades, the parents’ perception of the importance of their role decreasing as children get older, having enough knowledge on the topic, having minimal information about the programme before it started, perceiving that their parenting skills were questioned and having liberal views on the matter. In general, participants supported the idea of the long- term continuation of the programme; although they recommended changes in the content (e.g. reducing the repetitive content, including additional relevant topics, more practical examples, tailored content) and format (e.g. more emphasis on engaging teachers, including children). The findings of the thesis will provide valuable insights for the development of family-based programmes that target substance use prevention by focusing on a broader set of factors that influence behaviour and by conducting qualitative assessment in parallel with quantitative to better understand potential values and weaknesses and participants’ experiences and attitudes towards the programmes. 5 Acknowledgements I am deeply grateful to all the people who gave input to my PhD and supported me during the challenging four-year journey. None of this would have been possible without the Archimedes Foundation who financed my studies. Thank you for making a dream come true! Dr. Felix Naughton and Professor Stephen Sutton, my supervisors, thank you for seeing potential in my research and taking me under your wing. Your ideas, constructive feedback and support have been indispensably valuable. Charlotte, James B., Jo E., Jo P., Johanna, Kate, Katerina, Mel, Miranda, Vicky, thank you for making me feel at home. Our seminars, pub visits and Christmas dinners have been just awesome. I would like to thank everyone who contributed to adapting, delivering and evaluating Effekt in Estonia: Anneken Metsoja, Eha Nurk, Erle Kuusik, Esta Kaal, Helve Saat, Kadri Järv- Mändoja, Kai Klandorf, Kristel Aimee Laur, Krystiine Liiv, Külli Tatter, Laura Aaben, Lembi Posti, Liisa Maistrov, Liisa Salak, Maie Angerjas, Mart Felding, Meelis Kivisild, Merike Kull, Monika Siemann, Riina Raudne, Sandra Mägi, Tanel Kreek, Tiia Pertel, Triin Vilms and Triinu Tikas. Special thanks goes to Eve-Mai Rao and Mari Raudsepp for being the best on-site support team. I could not have done it without your optimism, enthusiasm and valuable ideas and practical advice. Maris Jesse, Tiia Pertel and Annika Veimer, thank you so much for seeing the potential in me to lead Effekt and supporting me throughout the doctoral studies. Lastly, I am deeply grateful to all the children and parents who participated in the survey and to all the parents and teachers who our team met at the meetings. Even if our world views did not always match, I hope we all had a chance to learn from each other. Dr. Kersti Pärna, nine years ago, when we started doing research on adolescents’ alcohol consumption, I could not dream of pursuing the path the way I have. Thank you for your advice and for showing me the direction. Dr. Silja-Riin Voolma, your persistence amazes me, and without your continuous pep talks, I would not be writing this right now. Thank you for planting the “PhD in Cambridge” idea into my mind and empowering me along the journey. Soon to be Doctors, Esther Rivas Adrover and Karin Streimann, we feel each other’s “pain”. Thank you for your support and wise thoughts on enjoying this journey. 6 Moving to Cambridge was the first time being away from family and friends, and I am grateful for the countless calls and postcards. Kudos to the ones who even made it to Cambridge – mom, dad, Liina, Nele, Daniel, Triin K., Mirtel, Kristjan, Raigo, Liina A., Allar, Armin, Triin V., Mari-Liis, Liisa, Laura, Raul T., Kalev, Riho, Raul R. and Raphael. Helen and Triin V., thank you for being the best cheerleading squad a PhD student could have. But new places bring new people into our lives. Markus and Petra – my (our) second family, our trips and playdates have made the last four years a bliss and kept me sane! Johanna and Stephen, I love our talks and cooking evenings and these rare occasions when we actually manage to play some board games. Optibrium family – thank you for being so supportive and cheery, and letting me take part in all the fun things you do. My eternal gratitude goes to the best husband in the world. Mario, you never stopped believing in me (even when I did) – it is difficult to describe my gratitude in words but thank you to the moon and back! Remember, you said that doing a PhD is a bumpy ride? Well, I finally admit it. But you made me feel it more like conquering a mountain, and I love doing that! Nännu and Taadu, oh how I wish I could hug you and celebrate my lifelong dream with you. You made it possible by igniting curiosity and enthusiasm in me when I was just a little girl, and you taught me to be persistent (26 years persistent!). Thank you! “Under a silent light Screaming biology Queen Gemini The world in her sights” − Foo Fighters 7 Publications Papers This research has led to the following publications: Tael-Öeren, M., Naughton, F., & Sutton, S. (2019a). The relationship between parental attitudes and children’s alcohol use: A systematic review and meta-analysis. Addiction, 114, 1527–1546. doi:10.1111/add.14615 (see Appendix 1) Tael-Öeren, M., Naughton, F., & Sutton, S. (2019b). A parent-oriented alcohol prevention programme “Effekt” had no impact on adolescents’ alcohol use: Findings from a cluster- randomised controlled trial in Estonia. Drug and Alcohol Dependence, 194, 279–287. doi:10.1016/j.drugalcdep.2018.10.024 (see Appendix 2) The following manuscript has been submitted for publishing: Tael-Öeren, M., Kaal, E., M., Sutton, S., & Naughton, F. (2020). Preventing adolescents’ alcohol use among adolescents by targeting parents: A qualitative study of the views of facilitators, parents and teachers on a universal prevention programme Effekt. 8 Presentations This research has led to the following oral presentations: Tael-Öeren, M. (2018, February). Parents' role in children's alcohol consumption prevention. Presentation at the Murray Edwards College’s Graduate Symposium, Cambridge, UK. Tael-Öeren, M. (2017, September). Targeting parents’ attitudes to prevent and reduce adolescents’ alcohol use: Piloting the "Effekt" programme in Estonia. Presentation at the Annual Conference of the European Society for Prevention Research, Vienna, Austria. Tael-Öeren, M. (2016, June). Parents' role in children's substance use - evaluation of an alcohol consumption prevention program in Estonia. Presentation at the First Year Presentations Seminar, Cambridge, UK. 9 Table of contents Preface........................................................................................................................................ 2 Abstract ...................................................................................................................................... 3 Acknowledgements .................................................................................................................... 5 Publications ................................................................................................................................ 7 Presentations .............................................................................................................................. 8 Table of contents ........................................................................................................................ 9 Tables ....................................................................................................................................... 14 Figures...................................................................................................................................... 16 Abbreviations ........................................................................................................................... 18 Chapter 1. Background ............................................................................................................ 19 1.1 Alcohol in modern society ............................................................................................. 21 1.2 Alcohol use prevalence and consequences among adolescents ..................................... 22 1.3 The developing brain and alcohol .................................................................................. 26 1.4 Risk and protective factors of alcohol use among children .......................................... 27 1.5 The role of the home environment in children’s alcohol use ......................................... 29 1.6 Preventing and reducing alcohol use among children by targeting the home ................ 34 1.7 Evaluation – an essential part of the programme ........................................................... 47 1.8 Conclusions .................................................................................................................... 49 Aim of the thesis ...................................................................................................................... 51 Chapter 2. The relationship between parents’ attitudes and children’s alcohol use: A systematic review and meta-analysis ....................................................................................... 52 2.1 Introduction .................................................................................................................... 53 2.2 Methods .......................................................................................................................... 54 2.2.1 Inclusion and exclusion criteria ............................................................................... 54 2.2.2 Search strategy and study selection ......................................................................... 55 10 2.2.3 Outcomes ................................................................................................................. 55 2.2.4 Exposures................................................................................................................. 56 2.2.5 Data extraction ......................................................................................................... 56 2.2.6 Quality assessment................................................................................................... 56 2.2.7 Meta-analyses .......................................................................................................... 57 2.3 Results ............................................................................................................................ 59 2.3.1 Study characteristics ................................................................................................ 60 2.3.2 Quality of included studies ...................................................................................... 61 2.3.3 Meta-analyses findings ............................................................................................ 70 2.3.4 Subgroup analyses and meta-regression .................................................................. 76 2.3.5 Sensitivity analyses.................................................................................................. 80 2.4 Discussion ...................................................................................................................... 80 2.4.1 Summary of main findings ...................................................................................... 80 2.4.2 Key considerations................................................................................................... 81 2.4.3 Comparability with previous studies ....................................................................... 82 2.4.4 Limitations and strengths......................................................................................... 83 2.5 Conclusions .................................................................................................................... 84 Chapter 3. Adaptation and delivery of a universal parent-oriented alcohol use prevention programme in Estonia .............................................................................................................. 85 3.1 Development of the parent-oriented prevention programme ......................................... 86 3.2 Previous evaluation of the programme........................................................................... 86 3.3 Bringing Effekt to Estonia.............................................................................................. 89 3.4 Preparations to deliver the programme .......................................................................... 90 3.5 Meetings I–II .................................................................................................................. 93 3.6 Meetings III–VI .............................................................................................................. 96 3.7 The results from the internal audit – feedback from the parents.................................. 103 3.8 Discussion .................................................................................................................... 105 11 3.9 Conclusions .................................................................................................................. 107 Chapter 4. Results from a universal parent-oriented alcohol use prevention programme Effekt in Estonia................................................................................................................................ 108 4.1 Introduction .................................................................................................................. 109 4.2 Methods ........................................................................................................................ 111 4.2.1 Recruitment, allocation and participants ............................................................... 111 4.2.2 Intervention ............................................................................................................ 114 4.2.3 Measures ................................................................................................................ 116 4.2.4 Sample size and power analysis ............................................................................ 118 4.2.5 Data analysis .......................................................................................................... 118 4.3 Results .......................................................................................................................... 121 4.3.1 Baseline characteristics.......................................................................................... 121 4.3.2 Attrition analysis.................................................................................................... 121 4.3.3 Primary outcome.................................................................................................... 125 4.3.4 Secondary and intermediate outcomes .................................................................. 127 4.3.5 Dose-response relationship .................................................................................... 135 4.3.6 Programme-related measures................................................................................. 136 4.4 Discussion .................................................................................................................... 138 4.4.1 Summary of main findings .................................................................................... 138 4.4.2 Comparability with previous studies ..................................................................... 138 4.4.3 Key considerations................................................................................................. 139 4.4.4 Limitations and strengths....................................................................................... 140 4.5 Conclusions .................................................................................................................. 142 Chapter 5. Preventing children’s alcohol use by targeting parents: A qualitative study of facilitators delivering the Effekt programme ......................................................................... 144 5.1 Introduction .................................................................................................................. 145 5.2 Methods ........................................................................................................................ 146 5.2.1 Epistemology and ontology ................................................................................... 146 12 5.2.2 Study design........................................................................................................... 147 5.2.3 Study sample and recruitment ............................................................................... 147 5.2.4 Data collection and analysis .................................................................................. 147 5.3 Results .......................................................................................................................... 149 5.3.1 The perceived value of the programme ................................................................. 149 5.3.2 Encountered challenges ......................................................................................... 150 5.3.3 Long-term continuation of the programme............................................................ 153 5.4 Discussion .................................................................................................................... 156 5.4.1 Summary of main findings and key considerations............................................... 156 5.4.2 Limitations and strengths....................................................................................... 159 5.5 Conclusions .................................................................................................................. 159 Chapter 6. Preventing adolescents’ alcohol use by targeting parents: A qualitative study of parents’ and teachers’ views on the Effekt programme ......................................................... 161 6.1 Introduction .................................................................................................................. 162 6.2 Methods ........................................................................................................................ 164 6.2.1 Epistemology and ontology ................................................................................... 164 6.2.1 Study design........................................................................................................... 164 6.2.3 Study sample and recruitment ............................................................................... 164 6.2.4 Data collection and analysis .................................................................................. 165 6.3 Results .......................................................................................................................... 166 6.3.1 Perceiving and tackling the problem ..................................................................... 167 6.3.2 Experience of the first contact ............................................................................... 174 6.3.3 Low participation rates and meeting the objective ................................................ 176 6.3.4 Views on the content and delivery......................................................................... 179 6.3.5 The perceived value of the programme ................................................................. 184 6.3.6 Long-term continuation of the programme............................................................ 188 6.4 Discussion .................................................................................................................... 192 13 6.4.1 Summary of main findings and key considerations............................................... 192 6.4.2 Limitations and strengths....................................................................................... 197 6.5 Conclusions .................................................................................................................. 198 Chapter 7. Discussion ............................................................................................................ 200 7.1 Overview of the main findings..................................................................................... 201 7.2 Findings in the context of the literature ....................................................................... 203 7.3 Limitations ................................................................................................................... 206 7.4 Strengths ....................................................................................................................... 208 7.5 Practical implications ................................................................................................... 209 7.6 Conclusions .................................................................................................................. 213 References .............................................................................................................................. 214 Appendices............................................................................................................................. 259 Appendix 1. Systematic review and meta-analysis on the relationship between parental attitudes and children’s alcohol use published in Addiction .............................................. 260 Appendix 2. Quantitative evaluation of the Effekt trial published in Drug and Alcohol Dependence ........................................................................................................................ 280 Appendix 3. Supplementary material from the meta-analysis ........................................... 289 Appendix 3.1 Search strategy ......................................................................................... 289 Appendix 3.2 Quality assessment ................................................................................... 290 Appendix 4. Supplementary material from the Effekt programme.................................... 291 Appendix 4.1 An overview of the assessments carried out during the programme ....... 291 Appendix 4.2 Feedback form ......................................................................................... 292 Appendix 4.3 Example of a newsletter ........................................................................... 293 Appendix 4.4 Case study scenarios and a problem-solving model ................................ 294 Appendix 4.5 Alcohol advertisement analysis ............................................................... 297 Appendix 4.6 Fictional stories used in the roleplay ....................................................... 298 Appendix 4.7 Agreements made between parents.......................................................... 299 14 Tables Table 1. Reviews of family-oriented interventions to prevent and reduce alcohol use among children..................................................................................................................................... 38 Table 2. Methodological quality assessment criteria ............................................................... 57 Table 3. Participants’ age (final assessment) used in meta-regression .................................... 59 Table 4. Description of characteristics of the included studies ............................................... 62 Table 5. Methodological quality assessment of included studies ............................................ 69 Table 6. Parents who participated in the meetings and gave feedback.................................. 103 Table 7. Parents’ feedback on attending the meetings and receiving materials, meetings II–VI ................................................................................................................................................ 104 Table 8. Parents’ feedback on facilitators, meetings II–VI.................................................... 104 Table 9. Adolescents’ and parents’ participation rates at the intervention and control schools at T1, T2 and T3..................................................................................................................... 114 Table 10. Number of participants (schools, classes and parents) and parents’ participation rates in the meetings............................................................................................................... 114 Table 11. Topics covered in the meetings and/or newsletters ............................................... 115 Table 12. Logical imputation of missing values and inconsistencies of adolescents’ alcohol use initiation and drunkenness ............................................................................................... 120 Table 13. Inconsistencies between alcohol use initiation and past year use and parental alcohol supply at T1, T2 and T3 ............................................................................................ 121 Table 14. Missing values among primary, secondary and intermediate outcomes and background characteristics at T1, T2 and T3 ......................................................................... 122 Table 15. Baseline characteristics of the sample ................................................................... 123 Table 16. Baseline rates of primary, secondary and intermediate outcomes and background characteristics among follow-up completers and non-participants by intervention condition ................................................................................................................................................ 124 Table 17. Baseline rates of primary, secondary and intermediate outcomes by intervention condition................................................................................................................................. 126 Table 18. Unadjusted two-level logistic regression models on the effect of intervention condition on all outcomes at T2 and T3................................................................................. 128 Table 19. Adjusted two-level logistic regression models on the effect of intervention condition on all outcomes at T2 and T3................................................................................. 129 15 Table 20. Unadjusted and adjusted two-level logistic regression models on the effect of attitude related measures on adolescents’ alcohol use initiation at T3 among students who had not initiated alcohol at T1 ...................................................................................................... 132 Table 21. Unadjusted and adjusted two-level logistic regression models on the effect of intervention condition on the original parental attitudes measure at the first and second follow-ups .............................................................................................................................. 134 Table 22. Unadjusted two-level logistic regression models on the effect of the number of meetings on primary, secondary and intermediate outcomes at T2 and T3 at the intervention schools.................................................................................................................................... 135 Table 23. Adjusted two-level logistic regression models on the effect of the number of meetings on primary, secondary and intermediate outcomes at T2 and T3 at the intervention schools.................................................................................................................................... 136 Table 24. Perceived strengths and weaknesses related to the programme and its delivery ... 153 Table 25. Perceived values and weaknesses of the programme among parents and teachers ................................................................................................................................................ 188 16 Figures Figure 1. Alcohol use trends among 15–16-year-old students in 1995–2015 in 25 European countries. .................................................................................................................................. 22 Figure 2. Lifetime alcohol use trends among 15–16-year-old students in 1995–2015 in 25 European countries................................................................................................................... 24 Figure 3. PRISMA flow diagram of the study selection process. ............................................ 60 Figure 4. Forest plot for meta-analysis of parental attitudes towards children’s alcohol use and children’s alcohol use initiation. ....................................................................................... 70 Figure 5. Forest plot for meta-analysis of parental attitudes towards children’s alcohol use and children’s alcohol use frequency. ...................................................................................... 72 Figure 6. Funnel plot of standard error by log OR for alcohol use frequency. ........................ 73 Figure 7. Forest plot for meta-analysis of parental attitudes towards children’s alcohol use and children’s drunkenness. ..................................................................................................... 74 Figure 8. Forest plot for meta-analysis of children’s perception of parental attitudes towards children’s alcohol use and children’s alcohol use initiation. ................................................... 75 Figure 9. Forest plot for meta-analysis of children’s perception of parental attitudes towards children’s alcohol use and children’s alcohol use frequency. .................................................. 76 Figure 10. Forest plot for meta-analysis of children’s perception of parental attitudes towards children’s alcohol use and parental attitudes reported by parents. .......................................... 76 Figure 11. Forest plot for meta-analysis of parental attitudes towards children’s alcohol use and children’s alcohol use frequency by study design. ............................................................ 77 Figure 12. Forest plot for meta-analysis of parental attitudes towards children’s alcohol use and children’s alcohol use frequency by sample s ize. ............................................................. 78 Figure 13. Forest plot for meta-analysis of parental attitudes towards children’s alcohol use and children’s alcohol use frequency by study location. ......................................................... 78 Figure 14. Forest plot for meta-analysis of parental attitudes towards children’s alcohol use and children’s alcohol use frequency by frequency (lifetime vs last year). ............................. 79 Figure 15. Meta-regression of the effect of age on the association between parental attitudes and alcohol use frequency across studies................................................................................. 79 Figure 16. Meta-regression of the effect of age on the association between parental attitudes and drunkenness across studies................................................................................................ 80 Figure 17. The expected programme’s impact pathways to delayed and reduced alcohol use among children....................................................................................................................... 101 17 Figure 18. The logic model of the Effekt programme and expected outcomes ..................... 102 Figure 19. Description of the allocation and randomisation process in the trial. .................. 112 Figure 20. Consolidated Standards of Reporting Trials 2010 flow diagram. ........................ 113 Figure 21. Alcohol use initiation rates at baseline (T1), first (T2) and second (T3) follow-ups at the intervention and control schools among adolescents who had not initiated alcohol use at T1. .......................................................................................................................................... 127 Figure 22. Alcohol use in the past 12 months at baseline (T1), first (T2) and second (T3) follow-ups among adolescents at the intervention and control schools................................. 127 Figure 23. Lifetime drunkenness rates at baseline (T1), first (T2) and second (T3) follow-ups at the intervention and control schools among adolescents who had not been drunk at T1. . 130 Figure 24. Parental alcohol supply reported by adolescents at baseline (T1), first (T2) and second (T3) follow-ups at the intervention and control schools. ........................................... 130 Figure 25. Perception of restrictive parental attitudes towards adolescents’ alcohol use among adolescents at baseline (T1), first (T2) and second (T3) follow-ups at the intervention and control schools. ...................................................................................................................... 131 Figure 26. Restrictive parental attitudes towards adolescents’ alcohol use among parents at baseline (T1), first (T2) and second (T3) follow-ups at the intervention and control schools. ................................................................................................................................................ 131 Figure 27. Restrictive parental attitudes towards adolescents’ alcohol use among parents at baseline (T1), first (T2) and second (T3) follow-ups at the intervention and control schools. ................................................................................................................................................ 133 Figure 28. Description of the allocation process in the qualitative study. ............................. 165 18 Abbreviations BCT – behaviour change technique cRCT – cluster-randomised controlled trial CI – confidence interval ESPAD – The European School Survey Project on Alcohol and Other Drugs HBSC – The Health Behaviour in School-aged Children HED – heavy episodic drinking ISFP – Iowa Strengthening Families Program NHPS – Network of Health Promoting Schools NIHD – National Institute for Health Development OR – odds ratio ÖPP – Örebro Prevention Programme PDFY – Preparing for the Drug-Free Years Program PRISMA – Preferred Reporting Items for Systematic Reviews and Meta-Analyses RCT – randomised controlled trial SFP – Strengthening Families Programme SMD – standardised mean difference WHO – World Health Organization 19 Chapter 1. Background 20 Girl: Beer [identifying a sample of alcohol by its smell]. Interviewer: Beer, okay. So where have you seen that one before? Girl: Everywhere.… Everywhere in the house, apart from upstairs. Interviewer: And who drinks that one? Girl: Daddy. Interviewer: Have you tasted it ever? Do you know what it tastes like? Girl: Yeah, I do like it but daddy doesn’t let me have a lot ... because it’s too fizzy. [edit later] Interviewer: [showing a picture of a brand of beer] And who drinks that one? Girl: Daddy and me. Interviewer: And you? How many times have you tasted beer do you think? Girl: Five or six. (Mel, aged 5) (Valentine, Jayne, Gould, & Keenan, 2010) 21 1.1 Alcohol in modern society Alcohol is considered the most harmful of all drugs due to its combined physical, psychological and social harms on individual and societal levels (Nutt, King, & Phillips, 2010; Bonomo et al., 2019). British neuropsychopharmacologist David Nutt has stated: “If alcohol was discovered today it would be controlled as an illegal drug…” (Nutt, 2010). Yet, in 2016, 2.4 billion people worldwide were considered as current drinkers (Griswold et al., 2018). According to the World Health Organization (WHO), the average annual alcohol intake in the world in 2016 was 6.4 litres of pure alcohol per person aged 15 and over (World Health Organization, 2018). The lowest alcohol use was reported in the Eastern Mediterranean region (0.6 litres per person) and the highest in the European region (9.8 litres per person). When the “Health 21 – The health for all policy framework for the WHO European Region” was introduced in 1999, one of the many goals was to reduce alcohol consumption from 12 litres per person per year in 1990 to six litres by the year 2015 (World Health Organization, 1999; Shield, Rylett, & Rehm, 2016). As of 2016, only seven countries out of 51 in the WHO European region had consumption below six litres – Armenia, Azerbaijan, Israel, Tajikistan, Turkey, Turkmenistan and Uzbekistan (World Health Organization, 2018). Highest intake was reported in the Republic of Moldova (15.2 litres), Lithuania (15.0 litres) and Czechia (14.4 litres). The latest findings show that to minimise alcohol-related harm, one should not consume alcohol at all (Griswold et al., 2018). Alcohol use has a detrimental impact on a myriad of health conditions, such as neuropsychiatric disorders, cardiovascular diseases, cancers, infectious diseases and gastrointestinal diseases (Griswold et al., 2018; Rehm et al., 2017). It is related to intentional and unintentional injuries (Cherpitel, 2014; Hingson, Edwards, Heeren, & Rosenbloom, 2009; Hingson, Heeren, Jamanka, & Howland, 2000; Macdonald et al., 2013; Taylor et al., 2010) and socioeconomic harm (e.g. unemployment, loss of quality of life, health care related costs) to the users themselves, but also to other individuals (Bellis et al., 2015; Navarro, Doran, & Shakeshaft, 2011). Already at the beginning of the 1970s it was suggested that limiting alcohol intake would result in better health and lower mortality (Belloc, 1973; Belloc & Breslow, 1972), but fast-forwarding 50 years and alcohol is the seventh leading risk factor for death, globally accountable for 2.8 million deaths – 2.2% of all female and 6.8% of all male deaths (Griswold et al., 2018). Based on the data from six countries (France, USA, Scotland, Canada, Korea and Thailand), social and economic costs 22 attributable to alcohol use have been estimated to cover 1.3–3.3% of the gross domestic product (Rehm et al., 2009). Despite the long list of adverse outcomes, alcohol use is highly prevalent, and goals set up 20 years ago in the WHO European region are still hard to reach. Alcohol use and related harm are also prevalent among young people, and in 2014, the Committee on National Alcohol Policy and Action (2014) that supports the coordination of European Union alcohol policy, approved the action plan to tackle youth drinking and binge drinking in 2014–2016. 1.2 Alcohol use prevalence and consequences among adolescents In 1995, the European School Survey Project on Alcohol and Other Drugs (ESPAD) (Hibell, Andersson, Bjarnason, Kokkevi, Morgan, & Narusk, 1997) was initiated among 15–16-year- old students in 23 European countries; thereafter it has been conducted every four years. The latest results from 2015 show a clear decline in lifetime alcohol use among adolescents in the last 20 years (Kraus et al., 2016) (Figure 1). The lifetime alcohol use rate stayed stable around 90% between 1995–2003 and started to decrease thereafter, reaching 81% in 20151. *Five or more drinks on one occasion Figure 1. Alcohol use trends among 15–16-year-old students in 1995–2015 in 25 European countries. Despite the decrease, there are still countries where the lifetime alcohol use rate was above 90% in 2015 – Croatia, Czechia, Denmark, Greece, Hungary and Slovakia, and countries that did not witness a decline as sharp as in Ireland, the Netherlands, Sweden, Norway and Iceland, with the latter two crossing 40% lifetime abstinence rate (Figure 2). For example, in Bulgaria, Czechia, Denmark and Hungary, the situation has stayed somewhat unchanged over 1 Based on the data from 25 countries: Bulgaria, Croatia, Cyprus, Czechia, Denmark, Estonia, Faroe Islands, Finland, France, Greece, Hungary, Iceland, Ireland, Italy, Lithuania, Malta, the Netherlands, Norway, Poland, Portugal, Romania, Slovakia, Slovenia, Sweden, Ukraine. 89 90 90 88 86 81 56 61 63 60 58 47 36 39 40 42 41 35 0 20 40 60 80 100 1995 1999 2003 2007 2011 2015 % Lifetime use Use in the last month HED* in the last month 23 20 years, and in Estonia, Faroe Islands, France, Malta, Slovakia and Slovenia, the decline took place only between 2011–2015. Compared to other countries in Europe, Northern European countries (with few exceptions) show a steeper decline in lifetime alcohol use over the years. At the same time, current and heavy episodic drinking (HED; drinking five or more drinks on one occasion) in the last month showed an increase between 1995–2003 and then dropped back to 1995 levels by 2015. While alcohol use rate among boys and girls over the years has been similar regarding lifetime and past month alcohol use, the 10% difference in past month HED in 1995 has shrunk to a few per cent difference in 2015, due to HED rate increase among girls in 2003–2007 and thus, narrowing the gap (Kraus et al., 2016). There is no single reason why people consume alcohol; instead, it is a complex mix of factors that influence the individual to engage in the behaviour. Although, it has been suggested that the decision to drink alcohol is strongly related to motives that also act as mediators for other factors (e.g. general coping strategies, alcohol use expectancies) (Anderson, Grunwald, Bekman, Brown, & Grant, 2011; Cooper, 1994; Cooper, Frone, Russell, & Mudar, 1995; Cox & Klinger, 1988; Cronin, 1997; Hasking, Lyvers, & Carlopio, 2011; Kuntsche, Knibbe, Engels, & Gmel, 2007). Often the reasons behind consumption do not differ much between adults and adolescents. For example, when adults were asked about the reasons why they drank alcohol in the past 12 months, the main three reasons were related to improving parties and celebrations, liking the feeling and alcohol use being fun (Lemmens, Vandlik, & Elekes, 2016); all three related to enhancing the social motives. Social and enhancement motives have been reported as the most common motives among adolescents and young adults (up to the age of 25) in various studies, and the higher the number of motives, the higher the likelihood of increased alcohol use (Kuntsche, Knibbe, Gmel, & Engels, 2005; Taylor, Babineau, Keogan, Whelan, & Clancy, 2016). One of the reasons behind this might be that younger people are more prone to experiment, take risks and look for new experiences (Steinberg, 2010) while focusing more on reward and excitement, and less on (negative) consequences (Casey, Jones, & Hare, 2008; Steinberg, 2008). Adolescents who consume alcohol are more likely to engage in other risky behaviours, such as smoking and drug use (Brain, Parker, Carnwath, 2000; Jackson, Sher, Cooper, & Wood, 2002), violence (Bellis et al., 2010), anti-social behaviour (Hammerton et al., 2017) and risky sexual behaviour (Bellis et al., 2010; Stueve & O’Donnell, 2005). 24 Figure 2. Lifetime alcohol use trends among 15–16-year-old students in 1995–2015 in 25 European countries. 30 40 50 60 70 80 90 100 1995 1999 2003 2007 2011 2015 % Eastern Europe Czechia Slovakia Hungary Ukraine Poland Bulgaria Romania 30 40 50 60 70 80 90 100 1995 1999 2003 2007 2011 2015 % Southern and Western* Europe Greece Croatia Slovenia Cyprus Malta France* Italy Netherlands* Portugal 30 40 50 60 70 80 90 100 1995 1999 2003 2007 2011 2015 % Northern Europe Denmark Lithuania Estonia Faroe Islands Finland Ireland Sweden Norway Iceland 25 Alcohol use can lead to injuries and health conditions and have a negative impact on academic performance and future consequences (e.g. due to breaking the law and getting a criminal record) (Newbury-Birch et al., 2009; Public Health Agency, 2017). Several studies have shown that the earlier the use is initiated, the higher the likelihood of adverse outcomes in adulthood (Grant & Dawson, 1997; Hingson, Heeren, & Winter, 2006; Pitkänen, Lyyra, & Pulkkinen, 2005; Winters & Lee, 2008). The adolescents in the ESPAD survey in 2015 were asked about the negative consequences they had experienced in the past 12 months while under the influence of alcohol (Kraus et al., 2016). Out of the 14 listed problems2, 30% of respondents had had at least one of these problems in the past year, with the average number of problems being 2.8. The three highest- rated consequences were property loss (14%), serious arguments (11%) and deep-water swimming (10%). Least likely were participants hospitalised due to intoxication (1%) or injury (2%) or had had an accident while driving (1%). When asked about the harm caused by excessive alcohol use by family members or friends, 11% reported that they knew someone who was an excessive drinker and whose drinking had caused them harm or problems. When investigating the relationship between risk perceptions and own substance use, it was found that participants who had never used alcohol or had drunk alcohol more than a year ago perceived daily binge drinking at greater risk of harm than participants who had consumed alcohol in the past year (Substance Abuse and Mental Health Services Administration, 2017). Sanchez-Ramirez, Franklin, and Voaklander (2018) who conducted a study among the general population in Alberta and Queensland found that people who were hazardous users themselves did not perceive alcohol as a contributor to health problems and injuries. The relationship between own use and risk perceptions was also assessed in a study among social science undergraduates in the United Kingdom and Norway (Pedersen, Fjaer, Gray, & von Soest, 2016). The harm related to alcohol, tobacco and cannabis use was assessed among five domains on a six-point scale: physical harms, mental health conditions, dependence, injuries and social consequences. The strongest relationship between substance use frequency and harm score was found among cannabis users, where an increase in frequency was associated with a decrease in the perceived harm score. A similar relationship, although weaker, was 2 Physical fighting, accident or injury, property loss, serious arguments, having been robbed, trouble with police, hospitalised (intoxication), hospitalised (injury), unprotected sexual intercourse, unwanted sexual advance, self- injury, impaired driving, accident while driving, deep-water swimming. 26 present among alcohol and tobacco users. In general, alcohol was perceived as the most harmful substance, followed by cannabis and tobacco. Individual assessment of domains resulted in alcohol being perceived as the most harmful substance regarding injuries and social consequences, tobacco regarding physical harms and dependence and cannabis regarding mental health conditions. As we can see, people regard alcohol as harmful and understand that there is a connection between alcohol use and negative outcomes. Yet, there is a gap between knowledge and behaviour, and next, I will try to explain why it is crucial to close this gap with regard to children. 1.3 The developing brain and alcohol Transitioning from childhood to adolescence is a critical time, as children become physically and psychologically more mature (Giedd et al., 1999; Keating, 1990; Windle et al., 2008), experience changes in their surrounding environment (Windle et al., 2008) and express higher need for independence (Spear & Kulbok, 2004). One of the most significant changes takes place in the brain, where ineffective and weak connections between neurons will be eliminated to increase the brain’s efficiency and energy conservation (Spear, 2010). The human brain matures from back to front, and the frontal area, prefrontal cortex, that is related to various functions and skills used on regular basis, e.g. self-control, attention and information processing, decision-making, planning, abstract reasoning, continues to develop until young adulthood (Crews, He, & Hodge, 2007; Giedd, 2004; Giedd et al., 1999; Mills, Lalonde, Clasen, Giedd, & Blakemore, 2014). The reward and sensation seeking areas in the brain (e.g. nucleus accumbens, amygdala – parts of the limbic system) (Gupta, Koscik, Bechara, & Tranel, 2011; Van den Bulk et al., 2013) show higher activity during adolescence, and thus may be the reason behind young people being prone to try out new things and take risks (Casey et al., 2008; Ernst, Pine, & Hardin, 2006; Meyer & Bucci, 2016; Steinberg, 2008), making them more vulnerable to negative consequences related to risk behaviour (Bava & Tapert, 2010; Schulenberg & Maggs, 2002; Stautz & Cooper, 2013). The onset of alcohol use often takes place in adolescence (Kraus et al., 2016), but due to the ongoing development of the brain, children are more susceptible to the negative impact of alcohol (Bava & Tapert, 2010). Alcohol has a strong effect on all areas mentioned above – it 27 inhibits the prefrontal cortex that controls the executive functions (Abernathy, Judson Chandler, & Woodward, 2010) and activates the reward and sensation seeking areas (National Institute on Drug Abuse, 2018). Thus, consuming alcohol while the brain is still under development makes it more difficult to make well-considered decisions (e.g. outweighing short-term benefits with long-term goals) and maximally use different skills (Arain et al., 2013). Also, alcohol use has been shown to be related to poorer attention, verbal learning, verbal and arithmetic memory, and visuospatial skills (Spear, 2018). Using insights from animal studies, it is suggested that human adolescents are less sensitive to the intoxicating effects of alcohol, due to the brain’s ability to quickly adapt to short exposure of alcohol (also known as acute tolerance, a process that is modest among adults) (Morales, Varlinskaya, Spear, 2011; Silveri & Spear, 1998), but more sensitive to the neurotoxic effects of alcohol (Monti et al., 2005). This may lead to increased consumption and a higher likelihood of alcohol-related harms. If alcohol is used more frequently and in large quantities, the impact on the brain may be more extensive and, in some cases, also permanent (Bava & Tapert, 2010; Geibprasert, Gallucci, & Krings, 2010). Although the brain is plastic and regions can compensate for each other and restore functions after damage has occurred (Pascual-Leone, Amedi, Fregni, & Merabet, 2005), it is more difficult to reach the expected brain capacity, when the harm done has occurred during brain development (Bava & Tapert, 2010). Alcohol can have a wide-ranging influence on the brain, varying from short-term toxic impact to long-term impact on cognitive functions. The earlier children initiate alcohol use, the more adverse the consequences can be. Therefore, it is important to understand how different factors increase and decrease the risk of alcohol use. 1.4 Risk and protective factors of alcohol use among children Alcohol use is a complex behaviour, influenced by a combination of different individual and environmental factors (Windle et al., 2008). The influence of these factors may vary by their strength and combined effect with other factors, individual characteristics (National Institute on Drug Abuse, 2003) and stage of use (Ellickson & Hays, 1991). In general, two different perspectives are assessed – risk (i.e. increased likelihood of a negative outcome) and protection (i.e. reduced likelihood of a negative outcome) (Hawkins, Catalano, & Miller, 1992; National Institute of Drug Abuse, 2003), and over the years, different approaches have 28 been applied when categorising these two perspectives regarding substance use among children. Hawkins and colleagues (1992) were among the first to summarise frequent substance use risk factors, by categorising 17 factors into two broad categories: 1) societal and cultural factors – laws and norms favourable toward behaviour, availability, extreme economic deprivation, neighbourhood disorganization, and 2) individual and interpersonal (i.e. family, school, friends) factors – physiological factors, family’s attitudes towards drug use, poor family management practices (e.g. involvement, discipline, communication, quality of the parent- child relationship), family conflict, low bonding with the family, antisocial behaviour, low school performance, poor commitment to school, peer rejection in early grades, substance use among peers, alienation and rebelliousness, own attitude towards substance use, and early onset of substance use. The influence of risk factors can be buffered by protective factors that indirectly enhance the resiliency to manage risk exposure (Hawkins et al., 1992; Rutter, 1985). Potential protective factors include strong parent-child attachment, positive personal characteristics and supportive internal (i.e. home) and external (i.e. school, community) environments (Hawkins et al., 1992). Donovan (2004) emphasised the need to treat the substances separately and distinguish between initiation and frequent use. Thus, in his review of longitudinal studies, he focused solely on alcohol and assessed the impact of psychosocial risk factors on alcohol use initiation. The following risk factors in five domains predicted increased risk of alcohol use onset: 1) sociodemographic – higher age, 2) family – living with a stepparent, parental/sibling alcohol use and approval of use (including perceived approval), perceived higher parental permissiveness, lower parental support and parent-child relationship, 3) peers – peer involvement in delinquent or drug-using behaviours, perceived friends’ approval of alcohol use, 4) personality – temperament factors (increased sensation seeking, impulsivity), lower values on academic achievement, lower expectancies on academic achievement, lower school motivation, tolerant attitudes towards deviant behaviour, lower levels of religiosity, less of an orientation to hard work, higher rebelliousness, higher rejection of parental authority, pro- alcohol personal beliefs, 5) behavioural – prior involvement in delinquent behaviour, lower grades at school. Schulte, Ramo, and Brown (2009) offer another perspective by assessing how different factors affect alcohol use from early adolescence to early adulthood among boys and girls 29 separately. While genetic (e.g. polymorphism of specific enzymes) and neurobiological factors (e.g. reduced amygdala volume), positive alcohol expectancies and personality characteristics (e.g. sensation and novelty-seeking, poorer behavioural and emotional control) were considered as shared risk factors, lower sensitivity to alcohol, later development of specific brain regions, alcohol use and poor monitoring by parents, higher perceived alcohol use by peers, higher acceptance of traditional gender roles and use of other substances were more likely to have a greater impact on alcohol use among males. All the above-mentioned reviews show that the surrounding environment has an important role in the development of substance use behaviour. Children typically spend a considerable amount of time in their childhood with their family and are influenced by their parents’ attitudes and behaviours. This emphasises the importance of the home environment, as it can have a direct influence on child’s alcohol use, but also attenuate or amplify other factors that are related to risk behaviours. 1.5 The role of the home environment in children’s alcohol use Alcohol is strongly embedded in our culture, being part of our daily lives, celebrations and traditions (Long & Mongan, 2014; Room, 2013). According to the findings from various studies, the majority of the parents find it acceptable to drink alcohol in front of their children (Foster, Bryant, & Brown, 2017; Valentine et al., 2010; Wolf & Chávez, 2015), with some assuming that the children are not paying attention to it (Foster et al., 2017). According to social learning theory (Bandura, 1977), children observe and learn from influential people in their social environments (e.g. parents), and it is easier to see the direct influence when they are toddlers, e.g. mimicking parents’ moves, repeating phrases. This influence still exists with older children but is less marked. Perceiving parents’ attitudes and beliefs and seeing their behaviour might explain why by the age of six, children can identify different alcoholic beverages, understand cultural norms related to adults’ alcohol use, and have developed positive and negative explicit expectancies towards alcohol (Jones & Gordon, 2017; Kuntsche, 2017; Kuntsche, Le Mével, & Zucker, 2016; Voogt, et al., 2017a; Voogt, Otten, Kleinjan, & Engels, 2017b). Additionally, parental alcohol use has been shown to predict alcohol use initiation and later use among children (Rossow, Keating, Felix, & McCambridge, 2016; Ryan, Jorm, & Lubman, 2010; Yap, Cheong, Zaravinos-Tsakos, Lubman, & Jorm, 2017). Handley and Chassin (2013) found that higher alcohol consumption among mothers, but not among fathers, was related to more frequent situations where they 30 shared stories of their own negative experiences with their children, which in turn increased the likelihood of the children initiating alcohol use. The authors suggested that despite the mothers portraying their experiences as negative, the children might have perceived them as a reflection of normality. Compared to tobacco products and illegal drugs, parents perceive alcohol to be less harmful and its use among children as inevitable (Carver, Elliott, Kennedy, & Hanley, 2017; Eadie et al., 2010; Jones, 2016). For example, studies among parents in the UK revealed that 68% of their 5–12-year-old children had been allowed to taste alcohol (Valentine et al., 2010) and 17% of parents allowed their 14-year-old children to drink alcohol (Maggs & Staff, 2018). When asked about the source of alcohol on the last occasion of use, 33% of 12-year-olds said that they received alcohol from their parents (Hearst, Fulkerson, Maldonado-Molina, Perry, & Komro, 2007). There is a common belief among parents that, in order to minimise alcohol-related harm among children, parents should offer them alcohol, thereby teaching them to drink responsibly and have control over the situation (Donovan, 2007; Graham, Ward, Munro, Snow, & Ellis, 2006; Jackson, Ennett, Dickinson, & Bowling, 2012; Jones, 2016; Jones, Magee, & Andrews, 2015; Kypri, Dean, & Stojanovski, 2007; Lundborg, 2007). However, the evidence from various review papers show that providing alcohol to children increases the likelihood of alcohol use (both initiation, frequent use and drunkenness) and related harms (Kaynak, Winters, Cacciola, Kirby, & Arria, 2014; Ryan et al., 2010; Sharmin et al., 2017a; Yap et al., 2017). For example, Kaynak and colleagues (2014) suggested that offering children alcohol in a supervised or unsupervised context increases alcohol use and on some occasions also heavy use and alcohol-related problems and increases the likelihood of children consuming alcohol outside the home. But offering alcohol and drinking alcohol in front of the children are not the only parental factors that have shown to have a clear impact on children’s alcohol use. The way parents apply their parenting skills influences their children’s behaviour, including alcohol use. Baumrind (1967) and Maccoby and Martin (1983) suggest that, based on the level of parental responsiveness (i.e. being supportive and warm) and demandingness (i.e. applying control, establishing boundaries), parents can be divided into groups that represent four different parenting styles: authoritarian, neglectful, permissive and authoritative. The authoritative parenting style is considered the most supportive by combining high demandingness and high 31 responsiveness. It is also characterised by the ability to find a balance between parents’ and child’s needs, by supporting the parent-child relationship and the development of child’s skills and independence. Čablová, Pazderková, and Miovský (2014) researched the relationship between parenting styles and children’s alcohol use and suggested that the authoritative parenting style can be considered as a protective factor against children’s alcohol use, although other individual and environmental factors should be taken into account. Thus, involving the child in the family’s decision-making processes, actively communicating with the child, being persistent, knowing the child’s closest friends and the parents of friends, supporting the child’s self-esteem, showing interest in the child’s activities and expressing clearly restrictive attitudes towards inappropriate behaviours (Baumrind, 1967; Maccoby & Martin, 1983), should make it less likely for the child to drink alcohol. Different parenting factors that are related to the authoritative parenting style have also been assessed separately, and the findings from review papers show a negative relationship with children’s alcohol use (Ryan et al., 2010; Yap et al., 2017). For example, higher parental monitoring (e.g. being aware of the child’s whereabouts, activities and friends), increased parental support (e.g. providing emotional and practical support), higher parental involvement (e.g. doing activities together) and higher general discipline (e.g. establishing and enforcing rules) were related to delayed alcohol use initiation, lower alcohol use or both among children (Ryan et al., 2010; Yap et al., 2017). A good quality relationship between the parent and the child that is of great importance from the child’s psychological and social development perspective (Maccoby & Martin, 1983; Stafford, Kuh, Gale, Mishra, & Richards, 2016) has also been shown to be related to postponed alcohol use onset and reduced frequent use (Carver et al., 2017; Ryan et al., 2010; Visser, de Winter, & Reijneveld, 2012; Yap et al., 2017), although the views on the strength of the relationship vary among researchers. Visser and colleagues (2012) suggest that a weak causal relationship between the parent-child relationship and alcohol use exists, although the findings from the included studies in their literature review are somewhat equivocal and in some cases, the relationship was present only in specific subgroups (e.g. dependent on age or gender). The authors also say that the relationship might operate in the opposite direction, with alcohol use causing a lower quality relationship. 32 The assessment of the relationship between the level of family conflict, e.g. tensions in the home environment, has resulted in equivocal findings. While Ryan and colleagues (2010) did not find evidence of the relationship between the level of family conflict and alcohol use among children, Yap and colleagues (2017) showed that higher family conflict was related to alcohol use initiation. Similar findings have been reported on communication, both general and alcohol-specific. Ryan and colleagues (2010) suggested that good general communication helped to delay alcohol use initiation and decrease subsequent alcohol use. Yap and colleagues (2017), who combined a higher number of longitudinal studies, showed that there was no evidence that good-quality communication would predict delayed or lower alcohol use among children. Carver and colleagues (2017) suggest that better communication is related to lower alcohol use (especially among younger children) and might be mediated by factors such as academic motivation and non-substance using friends. Regarding alcohol- specific communication (e.g. discussions about alcohol-related topics), while both Ryan and colleagues (2010) and Yap and colleagues (2017) did not find evidence for a relationship between communication and children’s alcohol use, Carver and colleagues (2017) found that the outcome is dependent on the content of the communication. They suggested that high- quality two-sided discussions about substance use (e.g. health risks) were related to decreased use of alcohol, but frequent discussions about parental use, lenient messages and consequences were related to increased use or did not show evidence of effectiveness. It might be that the perceived feeling of invincibility (e.g. bad things would not happen to them) and expected beneficial short-term outcomes (e.g. having a good feeling) among youth outweigh their understanding of alcohol-related risks (Wickman, Anderson, & Greenberg, 2008). While it might be expected that some parental factors that seem to directly prevent alcohol use among children – not offering alcohol, having rules about alcohol use, having a restrictive attitude towards children’s alcohol use – are more effective in reducing alcohol use, the findings from review studies on the latter two are mixed. Sharmin and colleagues (2017b) suggested that having clear rules concerning alcohol use was related to lower risky drinking among children, but no evidence was found for a similar relationship between rules and alcohol in other reviews (Ryan et al., 2010; Yap et al., 2017). Findings from a study among university students and their parents showed that compared to the students, parents perceived children’s compliance with the rules to prevent alcohol use to be higher, communication regarding the rules more straightforward and reasoned (Baxter, Bylund, Imes, & Routsong, 33 2009). When establishing the rules, it is important to involve the child in the process and make him/her aware of the consequences if the rules are not complied with (Ryan et al., 2011). Additionally, being persistent is another essential aspect regarding rules, as when parents are inconsistent and send out mixed messages (e.g. alcohol use is not allowed, but children can drink it on New Year’s Eve), they show inconsistent discipline (Bourdeau, Miller, Vanya, Duke, & Ames, 2012; Glowacki, 2016; Tildesley & Andrews, 2008). While not all parents might have established rules about children’s alcohol use, their attitudes and the way they express these attitudes might affect children. Some researchers have suggested that attitudes might play an even more vital role than behaviours, especially among younger children (Koning, Engels, Verdurmen, & Vollebergh, 2010a). There has been a long debate about how attitudes impact behaviour, and while the former seems to predict the latter (Glasman & Albarracin, 2006), the relationship is dependent on various factors. For example, attitudes based on direct experience are more strongly related to future behaviour than those based on indirect influence. Also, when people perceive the behaviour relevant and their attitudes correct, have stable attitudes over time or develop their attitudes based on behaviour-related knowledge (including what would be the consequences of performing the behaviour), attitudes are more likely to predict behaviour (Glasman & Albarracin, 2006). The strength of the relationship can also be influenced by the level of social pressure (i.e. subjective norm) and perceived difficulty to perform the behaviour, i.e. increased levels of both weaken the relationship (Wallace, Paulson, Lord, & Bond Jr, 2005). Regarding the relationship between parental attitudes and children’s alcohol use, the findings are mixed. While Ryan and colleagues (2010) showed that parental disapproval of children’s drinking predicted only later alcohol use, Yap and colleagues (2017) showed it was related to both initiation (r=0.11, P<0.01) and later use (r=0.18, P<0.001). Sharmin and colleagues (2017b) used different measures regarding attitudes and showed that while parental approval of alcohol use was related to risky alcohol use and related problems (OR=2.44, 95% CI=2.00–2.94), no evidence was found for a relationship regarding disapproval and permissiveness. These reviews had included a variety of studies, but none treated the data on parental attitudes from children and parents separately. Children's perception of their parents' attitudes may be different from parents’ own reports, and they may perceive it somewhat less restrictive than parents themselves (Gerrard, Gibbons, Reis-Bergan, & Russell, 2000; Yu, 2003). Also, parents’ attitudes might not always be in accordance with their behaviour 34 (Valentine et al., 2010) and parents have been shown to become less restrictive as children get older (Glatz, Stattin, & Kerr, 2012; Özdemir & Koutakis, 2016; Zehe & Colder, 2014). This knowledge on the relationship between parental factors and children’s alcohol use has been used in a myriad of programmes, which have been developed over the years to tackle the alcohol use issue among children. 1.6 Preventing and reducing alcohol use among children by targeting the home Preventive interventions and approaches that aim to prevent and reduce substance use can be divided into three categories – universal, selective and indicated – based on the level of vulnerability and exposure to risk among the target group (Brotherhood & Sumnall, 2011; Mrazek & Haggerty, 1994). Although this framework covers a variety of disorders and behaviours and is not explicitly developed for alcohol use prevention, it can be applied to alcohol use and thus, will be explained through that perspective. Universal prevention focuses on whole-group (e.g. population, community, school) approaches, which aim to prevent and delay alcohol use, irrespective of the level of risk of alcohol use (Brotherhood & Sumnall, 2011; Mrazek & Haggerty, 1994). Examples of universal prevention initiatives include programmes that increase the target group’s awareness of alcohol-related consequences, develop attitudes and values that do not support alcohol use, develop social skills, but also support the target group via environmental changes (e.g. limiting alcohol availability) (Brotherhood & Sumnall, 2011; Stockings et al., 2016 ). Selective prevention focuses on groups of more vulnerable children (e.g. juvenile offenders, ethnic minorities, children from vulnerable families) who are at higher risk of drinking alcohol (Brotherhood & Sumnall, 2011; Mrazek & Haggerty, 1994). It is more likely that these children are or become socially excluded, which in turn might lead to (increased) alcohol use (Brotherhood & Sumnall, 2011). An important aspect of selective prevention is to identify at-risk subpopulations early on and offer solutions (e.g. targeted programmes for children and/or parents) to reduce the risk (Cuijpers, 2003). Indicated prevention focuses on individuals and targets children who are at highest risk of developing alcohol use disorder (e.g. children with mental health problems, showing signs of antisocial behaviour) (Brotherhood & Sumnall, 2011; Mrazek & Haggerty, 1994). Indicated 35 prevention initiatives may aim to prevent alcohol use, but the main aim is to reduce use and prevent the development of alcohol use disorders among those who have already started alcohol use (Brotherhood & Sumnall, 2011). Examples of indicated prevention include mentoring, normative feedback programmes, therapy and counselling (Brotherhood & Sumnall, 2011; Cuijpers, 2003; Foxcroft, 2014a). Foxcroft (2014a) suggested an additional dimension of functionality within each level of prevention – environmental, developmental and informational. While environmental prevention approaches focus more on the legislative aspect and reducing the availability of alcohol, developmental approaches focus on the development of social skills, and informational approaches focus on increasing awareness. Regarding the effectiveness of interventions, Foxcroft (2014b) suggests that environmental interventions tend to be more effective than developmental interventions, and informational interventions the least effective, as people tend to act based on the cues in the environment rather than apply high- level cognitive thinking (Clark, 2013). Although considered the least effective, informational interventions add a valuable perspective, but mostly when complementing other approaches (Oncioiu et al., 2018). Despite knowing what works and what does not in preventing alcohol use among children, there are still many ineffective approaches commonly used, e.g. focusing only on information provision, using one-off activities, scare tactics and non-interactive methods (United Nations Office on Drugs and Crime, 2018). As one of the environments where children spend a big part of their time is home, parents have an important role in influencing children’s alcohol use (see Section 1.5). There have been a vast number of studies conducted to evaluate family-based prevention programmes that aim to prevent and reduce alcohol use among children, and of which the majority are universal or combine universal and selective prevention (Blueprints for Healthy Youth Development, 2019a). To give a more thorough overview of the findings, a systematic search was conducted to find reviews that summarised the findings from these programmes. Search strategy and inclusion/exclusion of the studies To find eligible reviews, a literature search was carried out in Medline. The following search criteria were applied to titles and abstracts: (intervention OR interventions OR program*) AND (parent* OR family) AND (alcohol) AND (review OR meta anal* OR meta-anal*). The search included all available peer-reviewed articles with an abstract in English published by 8 36 March 2019, resulting in 317 articles in total. The titles and abstracts were screened against the inclusion and exclusion criteria. All review studies (e.g. systematic reviews, meta- analyses, literature reviews) that reported on the effectiveness of universal parent- and/or family-based prevention programmes that targeted children’s alcohol use were included. Reviews were eligible if parent- and/or family-based programmes were defined as programmes that included parents as one of the target groups (or the only one) and if the programmes aimed to prevent or reduce alcohol use among children up to the age of 17 (or higher if a younger age group was included, e.g. 10–24). Reviews that were not in English and/or reported on solely treatment or selective or indicated prevention programmes were excluded. Also, reviews that did not separately look at the programmes mentioned above were excluded. In total, 290 papers were excluded, and full texts were obtained for the remaining 27 papers that either reported on universal parent-based alcohol use prevention programmes or indicated the inclusion of such programmes in the full text. Fifteen papers were excluded in the screening process, and five additional papers were identified from the remaining 12 review papers through reference searching. In addition, two papers published after the initial search were subsequently included, resulting in 19 reviews published between 2003 and 2019. Characteristics of the included reviews The 19 identified review papers that draw together the results from studies that have assessed the effectiveness of family-based universal prevention programmes included three reviews of systematic reviews and/or meta-analyses (Martineau, Tyner, Lorenc, Petticrew, & Lock, 2013; Mewton et al., 2018; Stockings et al., 2016), four systematic reviews and meta- analyses (Bo, Hai, & Jaccard, 2018; Gilligan et al., 2019; MacArthur et al., 2018; Smit, Verdurmen, Monshouwer, & Smit, 2008; Van Ryzin, Roseth, Fosco, Lee, & Chen, 2016), nine systematic reviews (Allen et al., 2016; Cairns, Purves, & McKell, 2014; Foxcroft & Tsertsvadze, 2011a, 2011b; Garcia-Huidobro, Doty, Davis, Borowsky, & Allen, 2018; Hurley, Dietrich, & Rundle-Thiele, 2019; Kuntsche & Kuntsche, 2016; Newton et al., 2017; Petrie, Bunn, & Byrne, 2007) and two narrative literature reviews (Cuijpers, 2003; Jones, Bates, Downing, Sumnall, & Bellis, 2010) (Table 1). Out of 19 reviews, five solely focused on parent/family-based programmes (Bo et al., 2018; Foxcroft & Tsertsvadze, 2011a; Gilligan et al., 2019; Hurley et al., 2019; Smit et al., 2008) and two on multi-component programmes that targeted children’s alcohol use (Cairns et al., 37 2014; Foxcroft & Tsertsvadze, 2011b). The rest distinguished between different settings (e.g. family, school) and risk behaviours (e.g. alcohol use, tobacco use, sexual risk behaviour), with the primary aim being broader than assessing only parent/family-based programmes. Similarly, while all reviews included universal parent/family-based programmes that targeted children’s alcohol use, twelve reviews included selective and/or indicated programmes as well (Allen et al., 2016; Bo et al., 2018; Cairns et al., 2014; Cuijpers, 2003; Garcia-Huidobro et al., 2018; Gilligan et al., 2019; Jones et al., 2010; Kuntsche & Kuntsche, 2016; Martineau et al., 2013; MacArthur et al., 2018; Petrie et al., 2007; Van Ryzin et al., 2016). Types of interventions included in the reviews In total, the articles included in the review papers covered 70 prevention programmes, of which 51 were universal, 16 selective and one indicated; additionally, one programme included universal and selective prevention measures and one selective and indicated prevention measures. The majority of the studies have been conducted in the United States, with the rest carried out in Europe (e.g. Croatia, Italy, the Netherlands, Russia, Sweden), Australia and India, and included students between the ages of six and 19. The most frequently evaluated (e.g. several follow-up studies, replication studies) programmes were the Strengthening Families Programme (SFP), Preparing for the Drug-Free Years Program (PDFY), the Parent-Based Intervention (PBI), Project Northland and the Örebro Prevention Programme (ÖPP). A large number of studies that have assessed the effectiveness of PDFY and SFP programmes used the data from a trial named the Project Family that started in 1993 among 6th graders (~11-year-old) and their parents in rural schools in Iowa, where participants were allocated into three groups – PDFY, Iowa SFP (ISFP), control with minimal contact (Spoth, Redmond, & Shin, 1998). Both programmes are universal, theory-based, address risk and protective factors that are related to substance use among children and focus on improving skills that improve the parent-child relationship and help children to refuse substances when offered (Catalano & Hawkins, 1996; DeMarsh & Kumpfer, 1986; Kumpfer, Molgaard, & Spoth, 1996). While PDFY was carried out in five consecutive weeks and mostly targeted parents (parents attended four two-hour sessions and children one two-hour session), ISFP sessions took place in seven consecutive weeks and provided six one-hour weekly sessions for parents and children separately and a combined session during the last week (Spoth et al., 1998). 38 Table 1. Reviews of family-oriented interventions to prevent and reduce alcohol use among children Study Study design Types of interventions Time period Number and origin of included studies Results Allen et al., 2016 SR of RCTs Parent-oriented universal and selective interventions that aim to decrease the use of alcohol, tobacco and illicit drugs among 10–19- year-olds Up to 1 Mar 2015 34 studies on alcohola Interventions that are delivered at home and/or school settings, use sessions and computer-based approach, and consist of fewer than 12 hours of training may result in reduced alcohol use. Bo et al., 2018 SR & MA of RCTs Parent-oriented universal, selective and indicated interventions that aim to prevent and decrease alcohol use among 10–18-year-olds Up to Mar 2017 21 studies from USA (17), NLD (3), SWE (1) Interventions that addressed general and alcohol-specific parenting or only general parenting were found more effective than alcohol-specific parenting programmes in preventing and reducing alcohol use. Cairns et al., 2014 SRb Universal and selective combined school/family alcohol education programmes that aim to prevent and reduce alcohol use among 11–18- year-olds Nov 2000– 2010 34 studiesa Interventions that addressed restrictive parental attitudes, increasing awareness and family communication, and developing skills were found more likely to be effective in reducing alcohol use. Cuijpers, 2003 LR Universal and selective prevention programmes that aim to prevent and reduce alcohol, tobacco and illicit drug use in different settings (school, family, community) Up to 2002 Four studies from the USA Three studies (from the same trial) on universal family-based drug prevention programmes showed effectiveness in reducing alcohol use; one other study did not find any difference between intervention and control groups regarding motivation not to use alcohol. Foxcroft & Tsertsvadze, 2011a SR of RCTs Universal family-based programmes that aim to prevent alcohol misuse among youth up to the age of 18 2002 – July 2010 12 studies from USA (11), NLD (1) Interventions showed small positive effects in reducing alcohol use in short- and long-term. Foxcroft & Tsertsvadze, 2011b SR of RCTs Universal multi-component programmes that aim to prevent alcohol misuse among youth up to the age of 18 2002 – July 2010 20 studies from USA (17), AUS (1), IND (1), NLD (1) Interventions tend to show positive effects in reducing alcohol use in short- and long-term, but the evidence on multi-component programmes being more effective than single-component programmes is inconclusive. Garcia- Huidobro et al., 2018 SR of RCTs Parent-oriented universal and selective interventions that aim to decrease the use of alcohol, tobacco and illicit drugs among 10–19- year-olds Up to Oct 2016 31 studies on alcohola Interventions showed effectiveness in reducing alcohol use. Gilligan et al., 2019 MA of RCTs Universal, selective and indicated family-based programmes that aim to prevent (heavy) alcohol use among youth up to the age of 18 Up to June 2018 27 studies from USA (20), NLD (2), SWE (2), GER (1), IND (1), POL (1) Universal family-based interventions (versus no intervention or standard care) were not found to be effective in reducing alcohol use prevalence and number of occasions of use but showed a small effect in reducing the number of drinks. Universal interventions that combined family-based and adolescent interventions (versus adolescent only) were not found to be effective in reducing alcohol use prevalence and frequency. 39 Study Study design Types of interventions Time period Number and origin of included studies Results Hurley et al., 2019 SR of TSs Universal school-based alcohol education programmes that include a parent component and aim to prevent and reduce alcohol use among 10–18-year-olds Up to August 2019 34 studies from USA (17), SWE (5), NLD (5), NOR (3), AUS (2), EST (1), RUS (1) The aim of this SR was to assess the effectiveness of programmes regarding parental outcomes. Out of 13 programmes, ten showed positive effects on at least one parent-related outcome (i.e. parental attitudes, parental monitoring, parent-child communication, alcohol-specific rule-setting). Jones et al., 2010 LR of SRs, MAs, RCTs, nRCTs, CBAs Educational universal and selective interventions that address alcohol use and sexual health among 5–19-year-olds Up to 2009 17 studies from USA (16), SWE (1) Interventions delivered to families showed mixed findings regarding alcohol use decrease. Kuntsche & Kuntsche, 2016 SR of RCTs, RTs, QESs Parent-oriented universal and selective interventions that aim to prevent and decrease the use of alcohol, tobacco and cannabis among 9–18-year-olds 2003 – 2014 21 studies on alcohol from USA (14), SWE (3), NLD (2), AUS (1), ITA (1) Interventions that improved general parenting measures showed some positive effects in reducing alcohol use. MacArthur et al., 2018 SR & MA of RCTs Universal and selective prevention programmes that aim to prevent multiple risk behaviours up to the age of 18 Up to 14 Nov 2016 Two studies on alcohol from the USA No evidence was found on the effectiveness of universal family-oriented prevention programmes on children’s alcohol use. Martineau et al., 2013 SR of SRs Interventions that aim to reduce alcohol use and use-related harm at the population level 2002 – 09 Oct 2012 Three SRs Based on the findings from SRs (Foxcroft & Tsertsvadze, 2011a, Jones et al., 2010, Petrie et al., 2007), family-based universal and selective interventions tend to have a small effect in reducing alcohol use, but the evidence is mixed. Mewton et al., 2018 SR of SRs, MAs Universal prevention programmes that aim to prevent alcohol and illicit drug use in different social settings 2006 – July 2016 Three SRs and one MA on alcohol Based on the findings from SRs (Foxcroft & Tsertsvadze, 2011a, Kuntsche & Kuntsche, 2016, Petrie et al., 2007) and MA (Smit et al., 2008), family - oriented interventions have found to be effective in delaying alcohol use initiation and reducing use. Newton et al., 2017 SR of RTs Combined student- and parent-oriented universal interventions that aim to prevent and reduce alcohol, tobacco and illicit drug use among 11–18-year-olds 2000 – Apr 2015 13 studies from USA (8), NLD (2), AUS (1), HRV (1), IND (1) Interventions tend to be effective in reducing alcohol use. Petrie et al., 2007 SR of RCTs, CTs, CBAs Parent-oriented universal and selective programmes that aim to increase parenting skills, and prevent alcohol, tobacco and drugs misuse among youth up to the age of 17 Up to Oct 2003 14 studies (5 on alcohol, 9 on combined substances) from USA (13), RUS (1) Interventions that focused on active parental involvement and the development of social and parenting skills and self-regulation tend to be effective in reducing alcohol use. 40 Study Study design Types of interventions Time period Number and origin of included studies Results Smit et al., 2008 MA of RCTs Family-oriented universal interventions that aim to prevent and decrease alcohol use among youth up to the age of 15 1995 – Sept 2006 Nine studies from the USA Interventions resulted in delayed alcohol use onset and frequency of use, but there was no evidence found on the intervention effect on past month use. Stockings et al., 2016 SR of Rs Programmes and approaches that aim to prevent and reduce alcohol use and use-related harm among 10–24-year-old young people 1990 – 23 Apr 2015 Three SR-s Based on the findings from SR-s (Foxcroft & Tsertsvadze 2011a, 2011b; Martineau et al., 2013), solely educational family-oriented interventions showed insufficient findings in reducing alcohol use. Skills-based interventions showed small effects in preventing and reducing alcohol use, but there are mixed findings regarding reducing problematic use and use- related harm. Van Ryzin et al., 2016 MAb Universal and selective family-oriented prevention programmes that aim to prevent alcohol, tobacco and illicit drug use among 11– 21-year-old youth Up to Sept 2012 105 studies on alcohola,c Interventions had a small (more likely among larger samples) to moderate effect in reducing children’s alcohol use, with no statistically significant difference between substances and prevention levels. Including youth- focused components predicted an increase in interventions’ effectiveness. aCountries not specified, bStudy design of included studies not specified, cNumber obtained from the study author; AUS – Australia, CAN – Canada, CBA – controlled before/after studies, CHE – Switzerland, CHL – Chile, CI – confidence interval, CT – controlled trials, ES – experimental studies, EST – Estonia, HRV – Croatia, IND – India, ITA – Italy, ITS – interrupted time series studies, LR – literature review, MA – meta-analysis, N/A – not available, NLD – Netherlands, NOR – Norway, nRCT – nonRCT, OR – odds ratio, OV – overview, QES – quasi-experimental studies, R – review, RT – randomised trials, RUS – Russia, SR – systematic review, SMD – standardised mean difference, SWE – Sweden, SWZ – Swaziland, TS – trial studies, USA – United States of America 41 While the findings from both programmes indicated their effectiveness (e.g. delayed alcohol use initiation, reduced alcohol use frequency and drunkenness) over short- and long-term periods, the impact on the outcome seemed to be stronger among children who participated in the ISFP intervention arm (Allen et al., 2016; Foxcroft & Tsertsvadze 2011a; Jones et al., 2010). In 1997, ISFP was modified to take into account the ethnic background of the participants and was renamed the Strengthening Families Program: For Parents and Youth 10–14 (SFP 10–14) (Molgaard, Spoth, & Redmond, 2000). Despite its success in the United States, the findings from RCTs in Germany, Poland and Sweden do not seem to support previous findings of delayed and/or reduced alcohol use (Baldus et al., 2016; Foxcroft, Callen, Davies, & Okulicz-Kozaryn, 2017; Skärstrand, Sundell, & Andréasson, 2014). Additionally, three quasi-experimental studies have been carried out in the United Kingdom and Ireland (Allen, Coombes, & Foxcroft, 2008; Coombes, Allen, Marsh, & Foxcroft, 2009; Kumpfer, Xie, & O’Driscoll, 2012). The two pre- and post-test exploratory pilot studies carried out among small number of families in four cities in the UK in the early 2000s had mixed findings – the youth in Barnsley reported lower alcohol and drug use (Coombes et al., 2009), but no significant changes in alcohol use were found among the youth in the other three cities (Allen et al., 2008). The Irish programme SFP 12–16 that is by its nature similar to SFP 10–14, but consists of more sessions, 14 respectively, was assessed in a pre- and post- test retrospective study among 250 high-risk families and resulted in reduced alcohol and drug use among youth (Kumpfer et al., 2012). Similarly to ISFP, PDFY went through a revision process to increase participants’ active engagement in sessions, and since 2003 it is known as Families That Care: Guiding Good Choices (Blueprints for Healthy Youth Development, 2019b). Currently, no additional studies on the programme’s potential effectiveness after revisions seem to have been published. Project Northland was developed in the United States at the beginning of the 1990s to reduce alcohol use and related problems among 6–8th graders (Perry et al., 1993). By using a community-wide approach and focusing on parent-child and peer communication, social norms, improving skills (among children and parents) and active community engagement regarding alcohol use topics, the programme provides a holistic approach that offers support in individual and environmental levels. The initial assessment of the programme was carried out among 6–8th graders in schools in Minnesota in 1991–1994, and the results showed that while there was no significant difference between intervention and control group students in the short term (1- and 2-year follow-ups), by the end of the eighth grade, students in the 42 intervention group had lower alcohol use tendency, past week and month use (Perry et al., 1996). As alcohol use rates tend to increase during high school, it was decided to develop the programme further and deliver an additional intervention in grades 11–12 (Perry et al., 2000). The two-year intervention had a positive effect on the target group’s alcohol use, although the effect was smaller compared to earlier years (Perry et al., 2002). Komro and colleagues (2008) tested the programme in Chicago among 6–8th grade students and showed that the students’ alcohol use and intentions to initiate use did not differ between intervention and control groups at all follow-ups (end of 6th, 7th and 8th grade). The programme has also been carried out in Croatia, Poland and Russia. In Moscow, Russia, the programme was delivered during the fifth grade, and while the knowledge of intervention group students showed improvement, alcohol use rates (lifetime, past year and month, binge drinking) did not significantly differ between the intervention and control groups by the end of the fifth grade (Williams et al., 2001). The Polish version of the programme was carried out among 4–5th graders in Warsaw and the results at 27-month follow-up indicated that the programme effectively delayed the initiation of drunkenness, but no effect was found regarding reduced alcohol use (Bobrowski, Pisarska, Ostaszewski, & Borucka, 2014). The Croatian version of Project Northland was carried out among 6–8th graders in Split (West, Abatemarco, Ohman- Strickland, Zec, Russo, & Milic, 2008). Alcohol use among students was assessed by a measure that combined intention to use (e.g. if somebody offers) and actual alcohol use (i.e. lifetime, past year, month, week), and the results showed that by the end of the 7th grade, tendency to use alcohol was lower among the intervention group, but the same finding was not present a year later. The PBI focuses on alcohol use among first-year college students. The intervention was initially carried out among 18-year-old students and their parents in the states of Idaho and New York in the United States (Turrisi, Jaccard, Taki, Dunnam, & Grimes, 2001). The content of the intervention (i.e. a handbook for parents) was developed based on the theory and empirical findings regarding students’ decision making, parent-youth communication and alcohol use and related norms in college. Between students’ high-school graduation and the start of the college, parents received a handbook that included information on college drinking and parent-youth communication and practical examples on how to enhance the latter. The initial assessment of the intervention indicated that the students in the intervention group had lower drunkenness rates and decreased incidence of negative consequences in the past month due to binge drinking; no significant difference between intervention and control 43 groups was found regarding the number of times of binge drinking in the past two weeks (Turrisi et al., 2001). While several later studies that have assessed the intervention’s effectiveness have supported previous positive findings (Cleveland, Lanza, Ray, Turrisi, & Mallett, 2012; Doumas, Turrisi, Ray, Esp, & Curtis-Schaeffer, 2013; Turrisi et al., 2009), some have shown that there was no significant difference between the intervention and control group or the use was even increased in the intervention group. For example, the findings from a study among freshmen in a university in California showed that while the number of weekly drinks consumed in the past month increased in both groups, the growth over eight months was smaller among intervention group students (Ichiyama et al., 2009). At the same time, the intervention did not seem to affect binge drinking or alcohol-related problems. Similar effects regarding binge drinking were reported by Wood and colleagues (2010) and Testa, Hoffman, Livingston, and Turrisi (2010). Mallett and colleagues (2010), who investigated how the age of alcohol use onset moderates alcohol use among college students, found that intervention group students who had initiated alcohol use at the age of 14 or younger showed an increased number of weekly drinks and alcohol-related negative consequences. Based on the studies presented in the included review articles, the PBI seems to have been delivered only in the United States. The ÖPP, a universal prevention programme that originates from Sweden, targets the parents of 13–15-year-old students to delay and reduce alcohol use among children (Koutakis, Stattin, & Kerr, 2008). The main focus of the programme is on maintaining parents’ restrictive attitudes as the children get older and supporting children’s engagement in extracurricular activities. Over the three years that the programme takes place, parents attend six meetings and receive additional six topic-related newsletters. The programme was initially carried out at the end of the 1990s in one Swedish county, and the results from a quasi-experimental study showed that by the end of the programme parents’ attitudes had become more restrictive and children’s drunkenness rates had decreased in the intervention group. Between 2007 and 2010, the programme was carried out on a broader scale (participants from 13 Swedish counties), and the results from a cluster-randomised controlled trial (cRCT) did not fully support previous findings – there were more parents with restrictive attitudes in the intervention group by the end of the programme, but no difference between groups regarding drunkenness among children (Bodin & Strandberg, 2011; Strandberg & Bodin, 2011). A shorter version of the programme took place in the Netherlands, where parents attended two meetings between 2006–2007 (Koning et al., 2009). The results from a 44 cRCT showed that when only the parent-oriented programme was delivered, there was no significant difference between intervention and control group regarding children’s alcohol use (i.e. weekly and monthly use, binge drinking), but in combination with student intervention, the programme managed to reduce use rates over time (Koning et al., 2009; Koning, Van den Eijnden, Engels, Verdurmen, & Vollebergh, 2010b; Koning, Van den Eijnden, Verdurmen, Engels, & Vollebergh, 2013). In 2012, the programme was revised and renamed Effekt (Özdemir & Koutakis, 2016), and several researchers explored the possibility of carrying out the programme in their countries (e.g. Belarus, Finland, Iceland, Norway, Russia) (Koutakis, 2011). Findings from the reviews Two reviews (Foxcroft & Tsertsvadze 2011a, 2011b) updated a previously published review, which was broader in its scope (i.e. setting-, age-, study design- and prevention level-wise) and assessed the effectiveness of universal prevention programmes that targeted alcohol use among young people up to the age of 25 (Foxcroft, Ireland, Lister-Sharp, Lowe, & Breen, 2003). While both of the reviews included randomised controlled trials (RCTs) that reported findings from universal prevention programmes that addressed alcohol use among youth up to the age of 18, one focused on family-based programmes (Foxcroft & Tsertsvadze, 2011a) and the other on multi-component programmes (i.e. included more than one setting) (Foxcroft & Tsertsvadze, 2011b). The review on family-based programmes included 12 studies of which nine showed small positive effects in reducing alcohol use in both short- (two months) and long-term (eight years) (Foxcroft & Tsertsvadze, 2011a). Based on the findings from 20 studies included in their other review on multi-component programmes, twelve studies showed positive effects in reducing alcohol use in short- (three months) and long-term (three years) (Foxcroft & Tsertsvadze, 2011b). However, the authors pointed out that the findings are inconclusive to conclude that multi-component programmes are more effective than single-component programmes. Stockings and colleagues (2016) who published a systematic review of review articles brought together the findings from the reviews by Foxcroft and Tsertsvadze (2011a, 2011b) and a review of systematic reviews (Martineau et al., 2013), and concluded that family/parent-based universal programmes that focus only on educating parents show insufficient evidence in reducing alcohol use among children, but skills-based programmes show small effects in preventing and reducing alcohol use in the short- and long-term (from 45 two months to 10 years), with mixed findings regarding reducing problematic use and use- related harm. A similar conclusion was reached by Mewton and colleagues (2018) who implied based on the findings from three systematic reviews (Foxcroft & Tsertsvadze, 2011a; Kuntsche & Kuntsche, 2016; Petrie et al., 2007) and one meta-analysis (Smit et al., 2008), that family- based programmes are effective in delaying alcohol use initiation and reducing use. Although the number of studies included in the meta-analysis studies varied, the direction of the effect and conclusions were mostly similar. Smit and colleagues (2008) showed that alcohol use initiation (odds ratio (OR)=0.71, 95% confidence interval (CI)=0.54–0.94) and frequency of use (Cohen’s d=-0.25, 95% CI=-0.37–-0.12) were lower in intervention groups compared to control groups. Bo and colleagues (2018) showed that interventions had a small positive effect on drinking intention (Hedge’s g=-0.30, 95% CI=-0.52–-0.08), binge drinking (g=- 0.16, 95% CI=-0.25–-0.06) and combined alcohol use outcome (g=-0.23, 95% CI=-0.35–- 0.10), but there was no clear evidence on the intervention effect on drinking frequency, quantity and drunkenness. Van Ryzin and colleagues (2016) included universal and selective interventions that targeted children’s substance use in their analysis and concluded that despite the variation in programmes’ effectiveness, there was a small to moderate effect on reducing children’s alcohol use (d=0.31), larger samples showing smaller effects and the effect size being similar across prevention levels (universal/selective). A meta-analysis by MacArthur and colleagues (2018) included two subgroups from the same intervention, and the results showed no evidence of the programme being effective in reducing alcohol use in the long term (OR=0.86, 95% CI=0.47–1.55). Gilligan and colleagues (2019) reached a similar conclusion in their meta-analysis. They showed that compared to standard care or no intervention, no evidence was found of effectiveness for universal family-based interventions reducing alcohol use prevalence [standardised mean difference (SMD)=0.02, 95% CI=-0.06– 0.11] or the number of occasions of use (SMD=0.18, 95% CI=-0.40–0.75). However, a small effect was found regarding the reduced number of drinks (SMD=-0.21, 95% CI=-0.32–-0.10). When comparing adolescent only interventions with universal interventions that combined family-based and adolescent interventions, no evidence of effectiveness was found regarding reduced alcohol use prevalence (SMD=-0.44, 95% CI=-1.08–0.20) and frequency (SMD=- 0.30, 95% CI=-0.68–0.07). 46 As there is considerable overlap between studies included in the review articles, it might explain why most of the review authors suggest that family-based programmes show promising results in delaying and reducing alcohol use. While Hurley and colleagues (2019) assessed the effectiveness of 13 parent-based prevention programmes, they focused only on parent-related outcomes. Out of the four outcomes looked at, including restrictive parental attitudes towards children’s alcohol use, parental monitoring, parent-child relationship and alcohol-specific rule-setting, the majority of the programmes showed a positive impact on at least one of these. In addition, they showed that most of the programmes lacked extensive stakeholder engagement and only approximately half of the programmes were described as theory-driven. Several studies have identified characteristics that have been shown to increase programmes’ effectiveness in reducing alcohol use: addressing general and alcohol-specific parenting measures (e.g. parental monitoring, parent-child communication, establishing and reinforcing rules, parental attitudes), development of social and parenting skills and self-regulation, focusing on active parental involvement, delivery at home and/or school settings, using sessions and computer-based approach, less than 12 hours of training, more intensive programme (frequent sessions), using group approach, including youth-focused components (e.g. positive family relationships and future orientation) (Allen et al., 2016; Bo et al., 2018; Cairns et al., 2014; Kuntsche & Kuntsche, 2016; Petrie et al., 2007; Smit et al., 2008; Van Ryzin et al., 2016). When considering target groups, Garcia-Huidobro and colleagues (2018) showed that family-based interventions are effective across genders, but also when targeting only girls, among younger adolescents, but also when younger and older adolescents were both targeted, and across ethnicities. Kuntsche and Kuntsche (2016) emphasised that the positive effects of the programmes are dependent on the child’s age and the parents’ background, meaning that families with younger children and disadvantaged conditions are more likely to benefit from the programmes. Implications for further research To summarise the conclusions from the reviews, we can see that family-based prevention programmes show potential in influencing children’s alcohol use, but there is a need for further research to gain a more in-depth understanding of how the delivery methods, programmes’ components and parental factors influence the programmes’ effectiveness, 47 while taking into account that the results may differ by age groups. This also means that sharing a more detailed description of the programmes, making the study results available (irrespective of programme’s effectiveness) and using more similar evaluation methodology, outcome measures and follow-up times should become more common practice. Babor and Robaina (2013) emphasise the importance of being transparent regarding funding, as the alcohol industry has rapidly increased its involvement in public health matters since the beginning of 2000s (Hill, 2017) and putting great effort in lobbying to minimise the evidence, shape the social narrative by its own interest and oppose whole population approaches (Martino, Miller, Coomber, Hancock, & Kypri, 2017; McCambridge, Mialon, & Hawkins, 2018; Mialon & McCambridge, 2018). Several industry-funded studies have highlighted the effectiveness of educational programmes. For example, Cairns and colleagues (2014), whose review was funded by Drinkaware Trust, pointed out in their review of effective educational programmes, that these interventions are commonly used. At the same time, while increasing awareness about alcohol and related consequences is common, in order to change behaviour, the focus should be more on developmental approaches that help to reduce risk factors and enhance protective factors (Foxcroft & Tsertsvadze, 2011a; Kumpfer, Alvarado, & Whiteside, 2003; Robertson, David, & Rao, 2003; Stockings et al., 2016). 1.7 Evaluation – an essential part of the programme Although new interventions are developed from time to time, it can be argued that more focus should be placed on delivering interventions that have already shown effective results (Mihalic & Elliott, 2015). While several family-based prevention programmes have shown some effectiveness in delaying and/or reducing children’s alcohol use, replication studies in other contexts (e.g. different countries) have not always achieved similar results. When delivering an intervention that has been developed in another country, adapting the programme to the local context is a common approach, as there may be differences in language, culture, norms and alcohol use patterns, but also in ethnicity, gender, and class (Elliott & Mihalic, 2004; World Health Organization, 2018). At the same time, Elliott and Mihalic (2004) suggest that the need to locally adapt prevention programmes is strongly overrated and this practice results in lower implementation fidelity, which is defined as “the degree to which teachers and other program providers implement programs as intended by the program developers” (Dusenbury, Brannigan, Falco, & Hansen, 2003). Although implementation refers to scaling up an already existing evidence-based approach (Eccles & 48 Mittman, 2006), the following five implementation fidelity dimensions suggested by Dane and Schneider (1998) can be applied to assess an intervention’s fidelity before wider uptake:  adherence – refers to the extent of programme delivery as intended in the original programme,  dose – reflects the amount (e.g. number of meetings) of interaction with the programme’s components among participants,  quality – defined as the level of identical content delivery to all participants,  responsiveness – refers to the degree of participants’ involvement in the programme’s activities and content,  differentiation – refers to the possibility to distinguish the distinctive programme’s components from each other. When Dusenbury and colleagues (2003) assessed all five dimensions of implementation fidelity among drug use prevention studies, they found that the first three were frequently reported, but for example, none of the studies reported on programme’s differentiation. Similar findings were reported by Durlak and DuPre (2008), who assessed the level of implementation fidelity among ~500 studies that reported on the evaluation of children- oriented prevention and health promotion programmes and showed that most studies report on programme’s adherence and dosage. Additionally, the findings indicated that higher fidelity is more likely to lead to positive outcomes among the target group. For example, the adaptation of two youth-oriented violence and drug prevention programmes showed that the adaptation process was typically initiated outside the managerial level and consisted of changing the content by adding or excluding elements, reducing the dosage and lowering the required level of training, thus reducing the likelihood of reaching the programme’s objective (Elliott & Mihalic, 2004). Assessing the intervention’s fidelity is one of the components of a more holistic process evaluation, which focuses also on the mechanisms of impact and how external factors (i.e. contextual barriers and facilitators) influence the interventions, and which together complement the outcome evaluation (Moore et al., 2015). Moore and colleagues (2015) developed further the existing guidelines on process evaluation created by Craig and colleagues (2008) and introduced a framework on how to plan, design and carry it out, but also on how to analyse and report the findings. 49 In 2011, the European Monitoring Centre for Drugs and Drug Addiction published a manual on European drug prevention quality standards to increase the quality of preventive interventions across Europe (Brotherhood & Sumnall, 2011). The guide gives a thorough overview of how to plan, deliver and evaluate an intervention, and how to disseminate the results. Organisations in several countries (e.g. Austria, Belgium, Poland, Lithuania, Sweden) have incorporated the quality standards in the development/improvement of their prevention standards (Liverpool John Moores University, 2019), but also to support the development of interventions, e.g. cannabis prevention in Sweden, school-based drug prevention and health promotion in Germany and Life Skills Training programme in Italy (based on personal communication with Prof. Harry Sumnall, 21.08.2019). A year after the quality standards were published, the second edition of drug prevention evaluation guidelines was published, providing a more thorough description of the assessment process (European Monitoring Centre for Drugs and Drug Addiction, 2012). Thus, there is a need to incorporate rigorous planning and process evaluation as an integral part of intervention programmes, to: 1) ensure that intervention fidelity and quality are kept high when making adaptations, 2) refer to programmes’ applicability in real-world settings, 3) make strong inferences on programmes’ effectiveness or ineffectiveness, 4) increase the transparency of their delivery, 5) have a better understanding of the underlying mechanisms of how and why the programmes work (Brotherhood & Sumnall, 2011; Durlak & DuPre, 2008; Lendrum & Humphrey, 2012; Moore et al., 2015). 1.8 Conclusions The current literature overview shows that extensive research has been done to better understand how parents influence children’s alcohol use. However, the findings show that some of the parent-related factors may have a stronger impact than others. So far, the review studies have combined the data on parental factors irrespective of the target source (i.e. parent, child), and this might explain why the results relating to some factors are equivocal as children’s and parents’ perception might differ (besides including newer studies and using different inclusion/exclusion criteria). Thus, there is a need to investigate how the results differ by target source, as it can provide additional input on which factors to address among specific groups. This thesis addresses a part of that knowledge gap by systematically reviewing the evidence on the relationship between self-reported parental attitudes and alcohol use among children. 50 Similarly, many studies have been undertaken to evaluate programmes that target parents, with findings showing that targeting parents can result in reduced alcohol use among children. However, replication studies of programmes initially found to be effective have not always confirmed previous findings, and there is no clear understanding of the reasons behind this. One of the widely promoted parent-oriented prevention programme in Europe, Effekt, focuses on targeting parental attitudes to delay and reduce children’s alcohol use. As the programme was developed in Sweden, where children’s alcohol use rates are significantly lower compared to many other European countries, evaluation of the programme’s effectiveness in countries with higher rates could provide valuable insight of its applicability in a different context. This thesis addresses that issue and provides a thorough description of Effekt programme’s delivery and participants’ views on the programme in addition to a traditional quantitative outcome evaluation in Estonia, where alcohol use rates are among the highest in Europe. 51 Aim of the thesis The thesis aims to examine the relationship between parents’ attitudes and children’s alcohol use and to evaluate a universal parent-oriented alcohol prevention programme delivered in Estonia. The thesis addresses four questions: 1) What is the relationship between parents’ self-reported attitudes towards children’s alcohol use and alcohol use among children? (Chapter 2) 2) How was the parent-oriented alcohol prevention programme adapted and delivered in Estonia? (Chapter 3 ) 3) Does the prevention programme delay alcohol use initiation and reduce alcohol use among children? (Chapter 4) 4) What are the views and attitudes of parents, teachers and facilitators on the prevention programme’s content and delivery process? (Chapter 5, Chapter 6) 52 Chapter 2. The relationship between parents’ attitudes and children’s alcohol use: A systematic review and meta-analysis 53 This chapter has been published in the journal “Addiction“. The unpublished data from the Effekt trial (see Chapter 4) have been included in the meta-analyses presented in this chapter to complement the results. The citation Tael-Öeren et al., 2019b is used when describing the study. 2.1 Introduction Alcohol use is one of the biggest risk factors for social and physical harm and disease development globally (World Health Organization, 2014). According to the ESPAD study, approximately 47% of 15–16-year-old students had consumed alcohol, and 8% had been drunk by the age of 13 (Kraus et al., 2016). Exposure to alcohol starts from early on; children as young as 2–6 years old become aware of alcohol and related norms (Voogt et al., 2017a; Voogt, Otten, Kleinjan, Engels, & Kuntsche, 2017b). Additionally, positive and/or negative explicit expectancies towards alcohol use have been shown to develop from the age of four (Voogt et al., 2017b). The parent’s role stays important throughout the child’s development (Collins & Laursen, 2004; Moretti & Peled, 2004), and home environment plays a crucial role in alcohol use prevention (National Institute on Drug Abuse, 2018), as parents are one of the main sources when learning norms, values and behaviours (Oetting & Donnermeyer, 1998). A wide range of studies have been conducted to investigate the relationship between children’s alcohol use and parent-related indicators. Parents’ own alcohol use (Rossow, Keating, Felix, & Mccambridge, 2016; Ryan et al., 2010; Yap et al., 2017) and providing alcohol to children (Kaynak et al., 2014; Ryan et al., 2010; Sharmin et al., 2017a; Yap et al., 2017) are consistently associated with increased risk of children’s alcohol use in longitudinal studies. Better quality of the parent-child relationship (Carver et al., 2017; Ryan et al., 2010; Visser et al., 2012; Yap et al., 2017) and higher levels of monitoring (Ryan et al., 2010; Yap et al., 2017) are considered as protective factors against alcohol use. Contradictory findings have been presented regarding parental support, involvement, general discipline, family conflict (Ryan et al., 2010; Yap et al., 2017), communication (Carver et al., 2017), rules about alcohol and attitudes towards adolescents’ alcohol use (Ryan et al., 2010; Sharmin et al., 2017b; Yap et al., 2017). While most of the aforementioned indicators are an example of establishing an action, attitudes are considered as one of the precursors of behaviour (Ajzen & Fishbein, 2005). It has been suggested that parental attitudes might be even more important than their behaviour in influencing children’s alcohol use (Koning et al., 2010a). 54 The relationship between parental attitudes and children’s alcohol use has been previously assessed in three reviews (Ryan et al., 2010; Sharmin et al., 2017b; Yap et al., 2017). While all state that attitudes are related to alcohol use, none distinguished between attitudes reported by parents and perceived by children. Koning and colleagues (2010a) point out that both parents’ and children’s responses should be taken into account when investigating parent- related factors and children’s alcohol use, as studies have shown that children’s reports on their perceptions of parental attitudes and behaviour can differ from parents’ own reports, and vice versa (Aas, Jakobsen, & Anderssen, 1996; Smith, Miller, Kroll, Simmons, & Gallen, 1999; Van der Vorst, Engels, Meeus, Deković, & Van Leeuwe, 2005; Williams, McDermitt, Bertrand, & Davis, 2003). Similar discrepancies have been shown in other studies reporting on parent-related indicators (Abar, Jackson, Colby, & Barnett, 2015; Cohen & Rice, 1997; Reynolds, MacPherson, Matusiewicz, Schreiber, & Lejuez, 2011). Therefore, this review applies a new approach by including only studies with parent-child dyads. Additionally, as this review focuses on dyads, cross-sectional studies were included to increase the breadth and statistical power of meta-analysis. The main objective of the current review is to assess the relationship between parental attitudes towards children’s alcohol use and children’s alcohol use, with the former reported by parents and the latter by children. Included articles will also be used to address secondary aims, where data allow assessment of the relationship between attitudes reported by parents and perceived by children, and between perceived parental attitudes and children’s alcohol use. 2.2 Methods The reporting of the review follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement (Moher, Liberati, Tetzlaff, & Altman, 2009). The review protocol is available on PROSPERO (registration number: CRD42017076694). 2.2.1 Inclusion and exclusion criteria All studies reporting on associations between self-reported parental attitudes and children’s alcohol use were included. Participants were considered as children if they were aged less than 18 years. When children’s data included participants older than 17 (e.g. 16–19), the study was included if most participants were under 18 or if the results were presented for separate age groups. Neither time nor setting restrictions were applied. The search included 55 all available peer-reviewed articles in English published until April 2018. Studies using only qualitative methods were excluded. 2.2.2 Search strategy and study selection To find eligible studies, literature searches were conducted in Medline, PsycINFO, EMBASE, Scopus and Web of Science. MeSH terms were included in addition to text words in the first three. A full description of search terms is presented in Appendix 3.1 Table S 1. The following information had to be presented in the abstract for the article to be included for full screening: (a) quantitative study design, (b) both children and parents described as participants and (c) information implying that parental attitudes and children’s alcohol use were assessed. M.TÖ and F.N independently screened the titles and abstracts of ~5% of all articles against the inclusion criteria [agreement rate 98.9% (κ=0.74)]. Afterwards, M.TÖ screened the titles and abstracts of all other articles. Full texts were obtained for all articles that met the inclusion criteria and where it was unclear, or there was reason to think the study might have included parents and children as participants and/or parental attitudes as an exposure. Thereafter, all three authors independently screened full texts [agreement rate of 97.1% (κ=0.93)]. Disagreements were resolved through discussion. Additional information was sought from study authors where it was necessary to resolve questions about eligibility. Reasons for exclusion were documented. 2.2.3 Outcomes The main outcome was children’s alcohol use (initiation, frequency, drunkenness). Initiation is typically assessed in studies by asking participants to indicate if they have consumed alcohol at least once [the amount can vary from a sip to full drink(s)] and/or the age of alcohol use onset (Inchley et al., 2016; Kraus et al., 2016; Kuntsche, Rossow, Engels & Kuntsche, 2016). Frequency of use and drunkenness are commonly measured by the frequency for any time period of more than one week (Inchley et al., 2016; Koning et al., 2010b; Kraus et al., 2016). Additionally, drunkenness could be measured by the age of being drunk for the first time (Inchley et al., 2016; Kraus et al., 2016). In case some other indicator was used to report alcohol use, where appropriate, it was added to one of the three existing categories and highlighted in the text. When multiple alcohol use-related indicators had been used, all were treated separately. 56 2.2.4 Exposures The primary exposure was parental self-reported attitudes towards children’s alcohol use. The secondary exposure was children’s perceptions of parental attitudes. As researchers use different definitions when referring to alcohol-related attitudes, approval/disapproval and norms were included (Donovan & Molina, 2014; Hayes, Smart, Toumbourou, & Sanson, 2004; Ryan et al., 2010; Sharmin et al., 2017b; Sieving, Maruyama, Williams, & Perry, 2000; Yap et al., 2017). 2.2.5 Data extraction The Cochrane Collaboration’s data collection form for interventions (Effective Practice and Organisation of Care, 2015) was modified to the review’s context. To ensure consistency across reviewers, all authors independently tested the form on two included studies. Disagreements were resolved through discussion. If a study had been published in multiple publications, different sources were considered as one paper if the same results were reported; papers were treated separately if different outcome measures were used. When data differed across publications or important information was missing, the study authors were contacted. The following data were extracted from each article: title, population description, setting, inclusion/exclusion criteria, method(s) of recruitment, study design, data collection method, duration of study, unit of allocation, analysis method, participants’ description, baseline sample size, attrition rate, clusters, definitions of exposure and outcome measures, time-points measured and distribution, results and number of participants included in the analysis. All the following results were extracted when reported on: 1) both parents separately and as a combined measure; 2) different subgroups; and 3) cross-sectionally and longitudinally within a study. 2.2.6 Quality assessment All authors independently assessed the quality of included studies using a modified version of guidelines developed by Hayden, Cote, and Bombardier (2006) and previously used in other similar reviews (Visser et al., 2012; Yap et al., 2017). Six domains were assessed for potential biases (Table 2), and quality items under domains were rated either “yes, +”, “partly, ±”, “no, -” or “unsure, ?” and given a score of 2, 1, 0 or 0, respectively. The bias for each domain was calculated by adding all scores and assessing if the final score was below/equal or above 50% of the maximum, 50% or less indicating a bias. Thereafter the 57 number of biases was counted; the total score for each study ranged between 0 and 6, the latter indicating more biases. All three authors independently assessed the quality of included studies. Any disagreements were resolved by discussion. Table 2. Methodological quality assessment criteria Domain Description Study participation A. The sampling frame and recruitment are described adequately, including the period and place of recruitment B. The inclusion and exclusion criteria are described adequately C. The sample is representative and or/random D. The baseline study sample (i.e. individuals entering the study) is adequately described for relevant key characteristics (at least for age and gender) Study attrition E. The response rate at follow-up is adequate F. Participants lost to follow-up are adequately described for key characteristics G. There are no major differences in key characteristics and outcomes between participants who provided data at time 2 and those who did not Predictor measurement H. The description or definition of the predictor variable is clear I. Continuous variables are reported, or appropriate cut-points (i.e. not data- dependent) were used J. The predictor measurement and method are adequately valid and reliable to limit misclassification bias Outcome measurement K. The description or definition of the outcome variable is clear L. The outcome measurement and method were adequately valid and reliable to limit misclassification bias Confounding measurement M. Confounders are accounted for in the study design (matching for key variables, stratification, or initial assembly of comparable groups) or in the analysis Analysis N. There is sufficient presentation of data to assess the adequacy of the analysis O. The strategy for model building (i.e. inclusion of variables) is appropriate and is based on a conceptual framework or model P. The selected analysis model is adequate for the design of the study Q. There is no selective reporting of results 2.2.7 Meta-analyses In order to perform meta-analysis, effect sizes had to be available from at least two studies. The following rules were applied for the main analysis: 1) combined parental attitudes were preferred over maternal only and maternal attitudes over paternal only; 2) the most commonly reported subgroup across studies was included – if it was not possible to choose the most common subgroup, the one with the highest rate was included; 3) longitudinal results were preferred over cross-sectional – if several follow-ups were reported, the one that 58 was most commonly reported across studies was included; and 4) to maximise comparability of the studies (as different studies select different factors for adjustment and using coefficients facilitates combining data in meta-analysis), correlation coefficients, unadjusted results and direct/total effects were preferred over regression coefficients, adjusted results and indirect effects. A summary statistic was identified/calculated and the weighted average calculated for each study; pooled effect size (OR was calculated using a random-effects model) (Deeks, Higgins, & Altman, 2011; Higgins, Thompson, Deeks, & Altman, 2003). The identified effect sizes were inverted when the attitude measure was scaled from lenient to restrictive. As some studies presented only ORs and P-values, a method suggested by Altman and Bland (2011) was used to obtain CIs from P-values. A value of 0.05λ [λ depends on the value of standardized regression coefficients (β), 1 when positive and 0 when negative] was added to βs to impute correlation coefficients (Peterson & Brown, 2005). If the study did not report an exact P-value and it was not possible to calculate it, the value for a non-significant result was treated as P>0.5 and for a significant result as P≤0.05. Heterogeneity was assessed by calculating the Q-statistic (Higgins & Thompson, 2002), T2-statistic (Borenstein, Hedges, Higgins, & Rothstein, 2009a), and I2-statistic (Higgins et al., 2003). In case of high heterogeneity (I2>75%), a subgroup analysis of studies [minimum four per subgroup (Fu et al., 2011)] was performed taking into account related factors. Originally, participants’ age and gender and study design were planned to be included. Due to low number of studies reporting subgroup information on measures, only study design (cross-sectional, longitudinal) was included, and three additional indicators – sample size (number of participants below and more than 500), location (Europe, USA), alcohol use frequency (lifetime, last year) – were added. Meta-regression between the age at the final assessment (Table 3) and the effect size of the main outcomes was conducted. As the minimum suggested number of studies included in the analysis is 10 (Borenstein, Hedges, Higgins, & Rothstein, 2009b), only studies reporting on alcohol use frequency and drunkenness were included. Age ranges (e.g. 11–13) were transformed to a single (an average) number (e.g. 12), to be included in the analysis. Sensitivity analyses were performed to assess the robustness of the results by excluding studies that highly influenced heterogeneity (Deeks et al., 2011). In addition, studies were omitted one at a time to ensure that the results were not influenced by a single study. It was also assessed how the results were affected by changing the target groups, follow-up times and outcome measures in individual studies when multiple results were presented. If comparable data were available from at least 10 studies (Sterne, Egger, & Moher, 2011), 59 funnel plots were used to assess the publication bias. Comprehensive Meta-Analysis 3.3.0 software was used to conduct meta-analyses (Borenstein, Rothstein, & Cohen, 2005). Table 3. Participants’ age (final assessment) used in meta-regression Study Age at baseline Age at the (final) assessment Aas et al., 1996 13.3 13 (baseline) Brody et al., 2000 10–12 13 (24 months) Colder et al., 2018 12 19 (seven years) Ennett et al., 2001 13.6 (12–14) 15 (12 months) Gerrard et al., 2000 (H) 16 (15/17) 17 (12 months) Gerrard et al., 2000 (L) 16 (15/17) 17 (12 months) Glatz et al., 2012 13.54 (13–14) 16 (24 months) Järvinen & Østergaard, 2009 15 15 (baseline) Kerr et al., 2012 7 13 (72 months) Koning et al., 2010b 12.2 (11–14) 14 (22 months) Margulies et al., 1977 14–18 16 (5–6 months) Mares et al., 2011a (F-O) 15.22 18 (36 months) Mares et al., 2011a (M-Y) 13.36 16 (36 months) Murphy et al., 2016 17 17 (baseline) Özdemir & Koutakis, 2016 12–13 14 (18 months) Özdemir & Koutakis, 2016 12–13 14 (18 months) Peterson et al., 1994 12–13 14 (24 months) Pettersson et al., 2011 13 15 (27 months) Sieving et al., 2000 (C) 12 13 (12 months) Sieving et al., 2000 (I) 12 13 (12 months) Strandberg et al., 2014 (B) 13 15 (30 months) Strandberg et al., 2014 (G) 13 15 (30 months) Tael-Öeren et al., 2019b, (I) 11 13 (30 months) Tael-Öeren et al., 2019b (C) 11 13 (30 months) Van der Vorst et al., 2006 (F-O) 15.22 16 (12 months) Van der Vorst et al., 2006 (M-Y) 13.36 14 (12 months) Yu, 2003 15–18 16 (baseline) H – high self-esteem, L – low self-esteem, F-O – fathers-older children, M-Y – mothers-younger children, C – control, I – intervention, B – boys, G – girls 2.3 Results In total, 12823 articles were identified through database searching (Figure 3). After removing duplicates, 7468 articles were included in the initial screening. Additionally, three articles from other sources (i.e. previously published reviews – Hayes et al., 2004; Ryan et al., 2010; Sharmin et al., 2017b; Yap et al., 2017) were included. Of 7471 articles, 65 articles met the inclusion criteria, and 38 had unclear information in the abstract, being eligible for full-text screening. In total, 29 articles were included in this review, comprising data from 16477 children and 15229 parents. Additional data from 985 children and 790 parents were included from the Effekt trial (see Chapter 4). 60 Figure 3. PRISMA flow diagram of the study selection process. 2.3.1 Study characteristics Characteristics of the 24 included studies represented in 30 articles are shown in Table 4. Twelve studies were longitudinal (Andrews, Hops, Ary, Tildesley, & Harris, 1993; Ary, Tildesley, Hops, & Andrews, 1993; Brody, Ge, Katz, & Arias, 2000; Colder, Shyhalla, & Frndak, 2018; Donovan & Molina, 2011, 2014; Ennett, Bauman, Foshee, Pemberton, & Hicks, 2001; Gerrard et al., 2000; Glatz et al., 2012; Kerr, Capaldi, Pears, & Owen, 2012; Mares, Van der Vorst, Engels, & Lichtwarck-Aschoff, 2011a; Margulies, Kessler, & Kandel, 1977; Peterson, Hawkins, Abbott, & Catalano, 1994; Van der Vorst, Engels, Meeus, & Deković, 2006), six cross-sectional (Aas et al., 1996; Jackson et al., 2012; Järvinen & Østergaard, 2009; Murphy, O’Sullivan, O’Donovan, Hope, & Davoren, 2016; Needle et al., 12823 articles identified through database searching (Medline, n=2074; PsycINFO, n=2205; EMBASE, n=2942; Scopus, n=4521; Web of Science, n=1081) S cr ee n in g In cl u d ed E li g ib il it y Id en ti fi ca ti o n 7471 articles after duplicates removed 7368 articles excluded by titles and abstracts screened 103 full-text articles assessed for eligibility Full-text articles excluded Full text not available (n=1) Full text not in English (n=3) Conference abstract/poster presentation/other (n=10) Relationship between parental attitudes and children’s alcohol use not reported (n=59) Children’s age above 18 (n=1) 29 (30) articles included in the data synthesis 3 articles identified through other sources Tael-Öeren et al., 2019b 61 1986; Yu, 2003), four were RCTs (Koning et al., 2010b; Koning et al., 2013; Koning, Verdurmen, Engels, Van den Eijnden, & Vollebergh, 2012a; Sieving et al., 2000; Strandberg, Bodin, & Romelsjö, 2014; Tael-Öeren et al., 2019b) and two quasi-experimental studies (Özdemir & Koutakis, 2016; Pettersson, Özdemir, & Eriksson, 2011). More than half (14) of the studies were carried out in the United States (Andrews et al., 1993; Ary et al., 1993; Brody et al., 2000; Colder et al., 2018; Donovan & Molina, 2008, 2011, 2014; Ennett et al., 2001; Gerrard et al., 2000; Jackson et al., 2012; Kerr et al., 2012; Margulies et al., 1977; Needle et al., 1986; Peterson et al., 1994; Sieving et al., 2000; Yu, 2003), with the other ten in Europe: four in Sweden (Glatz et al., 2012; Özdemir & Koutakis, 2016; Pettersson et al., 2011; Strandberg et al., 2014), two in the Netherlands (Koning et al., 2010a, 2010b, 2012a, 2013; Mares et al., 2011a; Van der Vorst et al., 2006) and one in each of the following: Denmark (Järvinen & Østergaard, 2009), Estonia (Tael-Öeren et al., 2019b), Ireland (Murphy et al., 2016) and Norway (Aas et al., 1996). The sample size of the included studies varied between 118 (Kerr et al., 2012) and 2599 (Koning et al., 2010a). Seven studies (Aas et al., 1996; Andrews et al., 1993; Brody et al., 2000; Donovan & Molina, 2008, 2011, 2014; Mares et al., 2011a; Murphy et al., 2016; Needle et al., 1986; Van der Vorst et al., 2006) collected data from both parents, two studies only from mothers (Gerrard et al., 2000; Jackson et al., 2012), and the rest did not specify parents by gender (Ary et al., 1993; Colder et al., 2018; Ennett et al., 2001; Glatz et al., 2012; Järvinen & Østergaard, 2009; Kerr et al., 2012; Koning et al., 2010a, 2010b, 2012a, 2013; Margulies et al., 1977; Özdemir & Koutakis, 2016; Peterson et al., 1994; Pettersson et al., 2011; Sieving et al., 2000; Strandberg et al., 2014; Tael-Öeren et al., 2019b; Yu, 2003). 2.3.2 Quality of included studies More than half the studies had low quality on at least three domains out of six, outcome measurement being the most common, due mainly not reporting on the measure’s validity and/or reliability (Table 5, Appendix 3.2 Table S 2). One study (Koning et al., 2013) did not provide any effect sizes due to non-significant results, and four studies (Andrews et al., 1993; Koning et al., 2012a; Needle et al., 1986; Pettersson et al., 2011) presented only those results reaching statistical significance. 62 Table 4. Description of characteristics of the included studies Study Country Design Adolescents’ age at BL No. of participants* included in the analysis Exposure O utcome Results ● Measure Target group Age a Measured Measure Time frame Target group Age a Measured Aas et al., 1996 NOR CS M=13.3 (7 th grade) 348 (B), 257 (G), 605 (O) Attitudes on alcohol use M, F OC BL Alcohol use b LT b B, G, O – BL G-M – r=0.17 (P=≤0.01), G-F – r=0.06 (P>0.05), B-M – r=0.09 (P=≤0.05), B-F – r=0.10 (P=≤0.05), O-M – r=0.08 (P=≤0.05), O-F – r=0.12 (P=≤0.01) 343 (B), 256 (G) Perception of parental attitudes on alcohol use – B, G OA G-M – r=0.18 (P=≤0.01), G-F – r=0.11 (P=≤0.05), B-M – r=0.09 (P>0.05), B-F – r=0.20 (P=≤0.001) 464 (B), 380 (G), 844 (O) Perception of parental attitudes on alcohol use B, G, O OA Alcohol use b LT b B, G, O – B – r=0.22 (P=≤0.001), G – r=0.19 (P=≤0.001), O – r=0.21 (P=≤0.001) Andrews et al., 1993 USA LNG 11–15 (M=13.2) MS – 180 (init .), 288 (maint.); FS – 125 (init.), 129 (maint.) Attitudes on alcohol use M, F AG BL IN, maintenance (≤6 MO) IN, ≤6 MO A – 12 MO IN (M) – no data presented (P>0.05), IN (F) – aOR**=1.20 (95% CI=1.10–1.33); ≤6 MO – no data presented (P>0.05) Perception of parental attitudes on alcohol use A IN (M) – aO R**=1.30 (95% CI=1.14–1.48), IN (F) – aOR**=1.18 (95% CI=1.08– 1.28); ≤6 MO – no data presented (P>0.05) Ary et al., 1993 USA LNG 11–17 173 Attitudes on alcohol use P c OC BL Alcohol use d LT d A – BL, 12 MO BL-BL – r=0.59 (P=≤0.05), BL- 12 MO – r=0.64 (P=≤0.05) P OC/AG e Perception of parental attitudes on alcohol use – OA/AGe BL r=0.40 (P=≤0.001) Brody et al., 2000 USA LNG 10–12 132 Acceptability of alcohol use M, F 14, 15 BL Beer, wine use ≤6 MO A – 24 MO Beer – M (14yo) – r=0.09 (P>0.05), M (15yo) – r=0.08 (P>0.05), F (14yo) – r=0.16 (P>0.05), F (15yo) – r=0.15 (P>0.05); Wine – M (14yo) – r=0.11 (P>0.05), M (15yo) – r=0.10 (P>0.05), F (14yo) – r=0.31 (P=≤0.05), F (15yo) – r=0.28 (P=≤0.05) 63 Study Country Design Adolescents’ age at BL No. of participants* included in the analysis Exposure O utcome Results ● Measure Target group Age a Measured Measure Time frame Target group Age a Measured Colder et al., 2018 USA LNG 12 740 (sipping), 530 (alcohol use) Acceptability of alcohol use P AG BL SP, alcohol use SP (BL), ≤12 MO (7 Y) A – BL, 7Y SP – aO R**=1.54 (95% CI=1.19–1.99), ≤12 MO – aO R**=1.02 (95% CI=1.00– 1.04) Donovan & Molina, 2008 USA CS 8, 10 204 (8yo), 222 (10yo) Approval of sipping M, F SAC BL SP SP A – BL 8yo (M) – OR=1.29 (95% CI=1.1–1.5), 10yo (M) – OR=1.18 (95% CI=1.1–1.3); 8yo (F) – OR=1.02 (95% CI=0.9–1.2), 10yo (F) – OR=1.18 (95% CI=1.03–1.3) Perception of parental approval of sipping A SOA 8yo – OR=1.78 (95% CI=1.3– 2.4), 10yo – OR***=1.84 (95% CI=1.53–2.21) Donovan & Molina, 2011 ● USA LNG 10 393 (MS), 297 (FS) Approval of drinking M, F SAC At child age 10 IN IN A – By age 14 M – OR=0.99 (95% CI=0.81– 1.21), F – OR=1.17 (95% CI=1.01–1.36) 393 Perception of parental approval of drinking A SOA OR=1.12 (95% CI=1.02–1.22) Donovan & Molina, 2014 ● USA LNG 8/10 286 (PS), 286 (MS), 206 (FS) Approval of sipping M, F, P SAC BL SP SP A – By age 12 P – O R=1.20 (95% CI=1.06– 1.37), M – OR=1.15 (95% CI=1.02–1.28), F – OR=1.16 (95% CI=1.03–1.31) 237 Perception of parental approval of sipping A SOA O R=2.37 (95% CI=1.38–4.09) Ennett et al., 2001 USA LNG 12–14 (M=13.6) 195 (init.), 281 (esc.) Disapproval of alcohol use P OC BL IN, escalation IN, LT A – 12 MO IN – aO R=1.39 (95% CI=0.86– 2.24); LT – aO R=1.96 (95% CI=1.39–2.76) 64 Study Country Design Adolescents’ age at BL No. of participants* included in the analysis Exposure O utcome Results ● Measure Target group Age a Measured Measure Time frame Target group Age a Measured Gerrard et al., 2000 USA LNG 15/17 126 (A f ), 125 (A g ) Attitudes on alcohol use M OC BL, 12 MO Alcohol use h ≤3 MO A f , A g – BL, 12 MO BL-BL – rf=0.12 (P>0.05), rg=- 0.07 (P>0.05), BL-12 MO – r f =0.04 (P>0.05), r g =-0.12 (P>0.05), 12 MO-12 MO – r f =0.19 (P=≤0.05), rg=0.02 (P>0.05) Perception of parental attitudes on alcohol use – OA BL-BL – rf=0.38 (P=≤0.05), r g =0.06 (P>0.05), BL-12 MO – r f =0.18 (P=≤0.05), rg=0.03 (P=≤0.05), 12 MO-12 MO – r f =0.24 (P=≤0.05), rg=0.30 (P=≤0.05) Perception of parental attitudes on alcohol use A f , A g OA Alcohol use h ≤3 MO – BL-BL – rf=0.21 (P=≤0.05), r g =0.25 (P=≤0.05), BL-12 MO – r f =0.38 (P=≤0.05), rg=0.20 (P=≤0.05), 12 MO-12 MO – r f =0.32 (P=≤0.05), rg=0.35 (P=≤0.05) Glatz et al., 2012 SWE LNG 15–16 638, 494 (only 24 MO) Attitudes on alcohol use P SAC BL, 24 MO Drunkenness ≤12 MO A – BL, 24 MO BL-BL – β**▲=0.11 (P>0.05), BL-24 MO – β**▲=0.06 (P>0.05), 24 MO-24 MO – r**=0.15 (P=≤0.01) Jackson et al., 2012 USA CS M=9.2 (3 rd grade) 1050 Prosipping beliefs M AG BL SP SP A – BL O R=2.29 (95% CI=1.78–2.94) Järvinen & Østergaard , 2009 DNK CS 15 (9 th grade) 1034 (st. 2 & 4), 1032 ( st . 5) Attitudes on alcohol use (statements 2, 4 &5) P AG BL Binge drinking i ≤1 MO A – BL Statement 2 – OR=1.16 (95% CI=0.89–1.50), statement 4 – O R=1.50 (95% CI=1.17–1.91), statement 5 – OR=1.20 (95% CI=0.94–1.53) Kerr et al., 2012 USA LNG 7 118 Attitudes on alcohol use j P OC 7 & 9 Y Early alcohol use k LT A, P, OS – BL–72MO r=0.28 (P=≤0.001) Koning et al., 2010a NLD CS 11–14 (M=12.16) 2599 (LT), 2122 (IF), 1494 (FR) Acceptability of alcohol use P 12/13 BL Alcohol use LT, WY (IF, FR) A – BL LT – r**=0.24 (P=≤0.001), IF – aOR**=2.56 (95% CI=1.69– 3.85), FR – aOR**=4.00 (95% CI=2.38–6.67) Koning et al., 2010b ●● NLD RCT 11–14 (M=12.16) 2051 Acceptability of alcohol use P 13/14 10 MO Alcohol use WY A – 22 MO r**=0.17 (P=≤0.001) Koning et al., 2012a●● NLD RCT 11–14 (M=12.16) 2381 Acceptability of alcohol use P 12/13 BL Onset WY l , HEWY m A – 34 MO WY – aOR=1.20 (95% CI*** 0.92–1.56); HEWY – no data presented (P>0.05) Koning et al., 2013 ●● NLD RCT 11–14 (M=12.16) 1064 Acceptability of alcohol use P 15 34 MO Alcohol use HEWE m A – 50 MO No data presented (P>0.05) 65 Study Country Design Adolescents’ age at BL No. of participants* included in the analysis Exposure O utcome Results ● Measure Target group Age a Measured Measure Time frame Target group Age a Measured Mares et al., 2011a NLD LNG Younger – M=13.36 Older – M=15.22 428 Acceptability of alcohol use M, F 13 BL Binge drinking i ≤1 MO A – 36 MO M (younger) r=0.16 (P=≤0.001), M (older) r=0.16 (P=≤0.001), F (younger) r=0.17 (P=≤0.001), F (older) r=0.17 (P=≤0.001) Margulies et al., 1977 USA LNG 14–18n 1142 (A), 1199 (G), 735 (B) Attitudes on alcohol use P OC BL Use of distilled spirits LT A, G, B – 5–6 MO A – r=0.09 (P=≤0.01), G – β=- 0.001 (P>0.05), B – β=-0.01 (P>0.05) 1936 Perception of parental attitudes on alcohol use A OA A r=0.05 (P=≤0.05) Murphy et al., 2016 IRL CS Median 17 338 (MS, st. 3 & 6), 266 (FS, st . 3), 267 (FS, st . 6) Attitudes on alcohol use (statements 3 & 6) M, F AG BL Hazardous drinking o ≤12 MO A – BL Statement 3 – M – O R=1.91 (95% CI=1.21–3.00), F – OR=1.34 (95% CI=0.80–2.25); statement 6 – M – OR=3.38 (95% CI=2.12–5.38), F – OR=4.41 (95% CI=2.57–7.58) Needle et al., 1986 USA CS 11–13 196 (beer), 188 (wine), 197 (liquor) Disapproval of alcohol use M, F OC BL Beer, wine, liquor use ≤12 MO A – BL No data presented (P>0.05) MD – 223p, 186 q (beer), 224 p , 187 q (wine), 217 p , 183 q (liquor); FD – 211p, 180 q (beer), 212 p , 181 q (wine), 207 p , 177 q (liquor) Perception of parental disapproval of alcohol use A OA A p , A q MD – r**=0.10 (P>0.05)p, r**=0.09 (P>0.05) q (beer), r**=0.02 (P>0.05) p , r**=0.10 (P>0.05) q (wine), r**=0.06 (P>0.05) p , r**=0.13 (P>0.05) q (liquor); FD – r**=0.18 (P=≤0.05)p, r**=-0.11 (P>0.05)q (beer), r**=0.12 (P>0.05) p , r**=- 0.12 (P>0.05) q (wine), r**=0.15 (P>0.05) p , r**=0.03 (P>0.05) q (liquor) 66 Study Country Design Adolescents’ age at BL No. of participants* included in the analysis Exposure O utcome Results ● Measure Target group Age a Measured Measure Time frame Target group Age a Measured Özdemir & Koutakis, 2016 SWE QE 12–13 (7th grade) A r – 339 (BL), 256 (18 MO) 264 (30 MO); A s – 312 (BL), 268 (18 MO) 242 (30 MO) Attitudes on alcohol use P SAC BL, 18, 30 MO Drunkenness (≤1 MO), onset of monthly drunkenness (BL → 18/30 MO) ≤1 MO, BL → 18/30 MO A r , A s – BL, 18, 30 MO ≤1 MO – BL-BL – r**r=-0.08 (P>0.05), r** s =0.06 (P>0.05), BL-18 MO – r**r=0.09 (P>0.05), r** s =0.01 (P>0.05), BL-30 MO – r** r =0.03 (P>0.05), r** s =0.03 (P>0.05), 18 MO-18 MO – r**r=- 0.05 (P>0.05), r** s =0.11 (P>0.05), 30 MO-30 MO – r** r =0.06 (P>0.05), r** s =0.09 (P>0.05); BL → 18/30 MO – BL- 18 MO – r**r=0.04 (P>0.05), r** s =0.03 (P>0.05), BL-30 MO – r** r =0.08 (P>0.05), r** s =0.01 (P>0.05), 18 MO-18 MO – r**r=- 0.07 (P=≤0.01), r**s=0.12 (P>0.05), 30 MO-30 MO – r** r =0.16 (P=≤0.01), r**s=0.05 (P>0.05) Peterson et al., 1994 USA LNG 12–13 (7th grade) 450 Parent norms about alcohol use (statement 1) (OC); acceptability of alcohol use (statements 2 & 3) (AG) P OC, AG BL Current alcohol use ≤1 MO A – 24 MO Statement 1 – aOR**=1.08 (95% CI=0.78–1.47), statement 2 – aOR**=1.05 (95% CI=0.74– 1.49), statement 3 – aOR**=1.27 (95% CI=0.89–1.82) Pettersson et al., 2011 SWE QE ~13 (7 th grade) 229 (A t ), 280 (A u ) Attitudes on alcohol use P SAC BL Alcohol use, drunkenness LT A t , A u – 27 MO Alcohol use – βt**▲=0.21 (P=≤0.01); βu=no data presented (P>0.05); drunkenness – βt***▲=0.27 (P=≤0.01); βu=no data presented (P>0.05) Sieving et al., 2000 USA RCT 6 th grade 200 (A r ), 213 (A v ) Parent norms about alcohol use P AG w BL Alcohol use ≤12 MO x A r , A v – 12, 24 MO BL-12 MO – rr=0.00 (P>0.01), r v =0.28 (P=≤0.001), BL-24 MO – r r =0.12 (P>0.05), r v =0.19 (P=≤0.05) Strandberg et al., 2014 SWE RCT ~13 (7 th grade) 895 (G), 857 (B) Attitudes on alcohol use P SAC BL Drunkenness LT, ≤1 MO G, B – 30 MO LT – G – aOR**=1.18 (95% CI=0.71–1.96), B – aOR**=1.43 (95% CI=0.50–4.17); ≤1 MO – G – aO R**=1.75 (95% CI=1.11– 2.78), B – aOR**=1.14 (95% CI=0.50–2.56) 67 Study Country Design Adolescents’ age at BL No. of participants* included in the analysis Exposure O utcome Results ● Measure Target group Age a Measured Measure Time frame Target group Age a Measured Tael- Öeren et al., 2019b EST RCT ~11 (5 th grade) A r – 480 (BL), 429 (18 MO), 372 (30 MO), A s – 505 (BL), 455 (18 MO), 415 (30 MO) Attitudes on alcohol use P AG 18 MO IN (LT), alcohol use (≤12 MO), drunkenness (LT) LT, ≤12 MO A r , A s – 30 MO IN – LT – O R**r=1.69 (95% CI=0.85–3.45), OR**s=2.08 (95% CI=1.00–4.35); ≤12 MO – O R** r =2.38 (95%CI=1.28– 4.55), OR** s =2.27 (95% CI=1.23–4.17); LT – O R** r =1.32 (95%CI=0.63– 2.78), OR** s =1.32 (95% CI=0.52–3.33) Perception of parental attitudes on alcohol use A SOA 30 MO IN – LT – O R**r=1.28 (95% CI=0.65–2.50), OR**s=1.92 (95% CI=0.99–3.85); ≤12 MO – O R** r =1.43 (95%CI=0.82– 2.50), OR** s =1.85 (95% CI=1.08–3.23); LT – OR**r=1.79 (95% CI=0.88–3.70), OR**s=1.69 (95% CI=0.86–3.33) Van der Vorst et al., 2006 ●●● NLD LNG Younger – M=13.36 Older – M=15.22 428 (BL), 416 (12 MO) Acceptability of alcohol use M, F 13 BL Alcohol use ≤1 MOy A – BL, 12 MO BL-BL (total effects) – M (younger) β▲=0.30 (P=≤0.05), M (older) β▲=0.19 (P=≤0.05), F (younger) β=-0.01 (P=≤0.05), F (older) β▲=0.08 (P=≤0.05); BL- 12 MO (total effects) – M (younger) β▲=0.23 (P=≤0.05), M (older) β▲=0.14 (P=≤0.05), F (younger) β=-0.01 (P=≤0.05), F (older) β▲=0.07 (P=≤0.05) Yu, 2003 USA CS 15–18 639 (LT), 470 (≤1 MO) Attitudes on alcohol use P OC BL Alcohol use LT, ≤1 MO A – BL LT – r=0.13 (P=≤0.01), ≤1 MO – r=0.08 (P>0.05) 593 Perception of parental attitudes on alcohol use – OA r=0.36 (P=≤0.01) 640 (LT), 470 (≤1 MO) Perception of parental attitudes on alcohol use A OA Alcohol use LT, ≤1 MO – LT – r=0.25 (P=≤0.01), ≤1 MO – r=0.10 (P=≤0.05) *Per target group. **Inverted effect size. ***New CI was calculated due to asymmetry. ●Same study as Donovan & Molina, 2008. ●●Same study as Koning et al., 2010a. ●●●Same study as Mares et al., 2011a. ▲0.05 added to the effect size. aTarget group’s attitudes towards alcohol use at specific age. bAlcohol use score was calculated as a sum of six z-transformed items that assessed the frequency of life-time and last three months use and frequency of drunkenness in the past 6 months. cCombined measure – parent self-report attitudes + perceived parent attitudes. dCombined measure – lifetime use and a monthly rate variable indicating the number of times alcohol was used. eCombination of three measures – first two assess OC/OA, the last AG. fAdolescents with low self-esteem. gAdolescents with high self-esteem. hLatent variable (frequency of drinking + excessive drinking). iFive or more units on one occasion. jCombined measure – the mean of mother and father binary variables (below and above 21) at both waves was calculated to form the construct. kMean of 68 standardized indicators from child self-reports (ages 7–13), parent reports (ages 5–13), and observer report (9 years). lOnly among monthly users. mBoys minimum six, girls minimum five glasses per week/weekend. n Secondary school students. o AUDIT-C. p With older siblings. q Without older siblings. r Intervention group. s Control group. t Programme group. u Comparison group. v Reference group. w Combined measure – measuring beliefs about the acceptability of underage alcohol use, as well as parents’ attitudes and practices related to their own child’s use + perceived parent norms around underage drinking. xAlcohol use (past month and year) + alcohol misuse (episodes of 5 or more drinks in a row over the past 2 weeks, frequency of being “really drunk so you fell down or got sick”, and frequency of alcohol use in the past week). yCombined – alcohol use frequency in the past 4 weeks + intensity of drinking (number of glasses of alcohol the respondents had drunk in the previous week during weekdays and during the weekends in contexts at home and outside the home. A – adolescent, AG – adolescents in general, aOR – adjusted OR, B – boy, BL – baseline, CS – cross-sectional, F – father, FD – fathers’ disapproval, FR – frequent, G – girl, HEWE – heavy weekend use, HEWY – heavy weekly use, IN – initiation, IF – infrequent, LNG – longitudinal, LT – lifetime, M – mother, MD – mothers’ disapproval, MO – month(s), O – offspring, OA – own age, OC – parents’ own child, OS – observer, P – parents, PS – parents’ sample, QE – quasi-experimental, SAC – children the same age as (own) child, SOA – same as own age, SP – sipping, WY – weekly, Y – year(s). Bold – included in the main analyses. 69 Table 5. Methodological quality assessment of included studies* Study Study participation Study attrition Predictor measurement Outcome measurement Confounding measurement Analysis Number of biases A B C D E F G H I J K L M N O P Q Aas et al., 1996 ± - ? + NA NA NA + ? ? ± ± - + ± + + 4 Andrews et al., 1993 ± + - + + ± - ± + ? ± ? + + + + ± 3 Ary et al., 1993 ± + ± ± ? - ? + + ± ± ? ± + + + + 3 Brody et al., 2000 ± + + ± ? - - ± + ± ± ± - + + + + 3 Colder et al., 2018 + + + - + - ? ± ? ? ± ? + + + + + 3 Donovan & Molina, 2008 + + ± + NA NA NA ± + ± ± ? ± + + + + 2 Donovan & Molina, 2011 + + ± + + ± ± ± + ± + ± + + + + + 0 Donovan & Molina, 2014 + + ± + + ± + ± + ± + ? + + + + + 1 Ennett et al., 2001 - - + + + ± ± + + ? + ? + + + + + 2 Gerrard et al., 2000 ± + ? ± ± - ? ± + ? ± ± - + + + + 5 Glatz et al., 2012 + + ? ± - ± ± + ± ? ± ? ± + ± ± + 4 Jackson et al., 2012 ± + - ± NA NA NA + + ± + ? + ± + + ± 2 Järvinen & Østergaard, 2009 + ± + - NA NA NA + + ? + ? - + ± ± - 3 Kerr et al., 2012 ± + - + ? ? ? + + ? ± ± ± + + + ± 3 Koning et al., 2010a + - + + NA NA NA ± + ± + ± + + + + + 0 Koning et al., 2010b + + ± ± + ± ± ± + ± + ? ± + + + + 2 Koning et al., 2012a + + ± + ± - ? ± + ± + ? + + + + ± 2 Koning et al., 2013 ± - ? ± - ± ± ± + ± + ? ± + + + ± 4 Mares et al., 2011a ± + ? ± + ± ± ± + ± ± ? ± + + + + 3 Margulies et al., 1977 ± - ? ± - - ? - ? ? ± ? + ± + + + 4 Murphy et al., 2016 ± ± ± + NA NA NA + + ? + + + + ± + + 0 Needle et al., 1986 ± + + + NA NA NA - ? ? ± + ± + + + ± 2 Özdemir & Koutakis, 2016 - - ? - + - ± ± + ? + ? - + + + + 4 Peterson et al., 1994 ± - - ± - ± - + + ? + ? + + + + + 3 Pettersson et al., 2011 ± ± ? - ± ± - + + ? + ? ± ± + + ± 4 Sieving et al., 2000 ± - ? ± ? - ? ± ? ? ± ? ± + + + + 5 Strandberg et al., 2014 ± + ± ± + + - ± + ? + ? + + + + + 2 Tael-Öeren et al., 2019b + + ? + ± + - + + ? + ? + + + + + 2 Van der Vorst et al., 2006 ± + ? + + - ? ± + ± ± ? ± + + + ± 3 Yu, 2003 ± ± ± + NA NA NA + + ? + ? + + + + + 1 *The description of the symbols (A–Q) and the calculation is presented in section 2.2.6; “+” – yes, 2 points, “±” – partly, 1 point “-” – no, 0 points “?” – unsure, 0 points, NA – not applicable 70 2.3.3 Meta-analyses findings Parental attitudes and children’s alcohol use initiation Of 29 included articles, 13 reported on parental attitudes (Aas et al., 1996; Andrews et al., 1993; Ary et al., 1993; Gerrard et al., 2000; Glatz et al., 2012; Järvinen & Østergaard, 2009; Kerr et al., 2012; Margulies et al., 1977; Murphy et al., 2016; Özdemir & Koutakis, 2016; Pettersson et al., 2011; Strandberg et al., 2014; Yu, 2003), five on parental (dis)approval (Donovan & Molina, 2008, 2011, 2014; Ennett et al., 2001; Needle et al., 1986), nine on acceptability (Brody et al., 2000; Colder et al., 2018; Koning et al., 2010a, 2010b, 2012a, 2013; Mares et al., 2011a; Peterson et al., 1994; Van der Vorst et al., 2006), two on parental norms (Peterson et al., 1994; Sieving et al., 2000) and one on beliefs (Jackson et al., 2012) related to children’s alcohol use. Alcohol use initiation was measured in seven papers [two were excluded from the analysis, as they were from the same study (Donovan & Molina, 2008, 2011)] using the following indicators: initiation (Andrews et al., 1993; Donovan & Molina, 2011; Ennett et al., 2001) and sipping/tasting (Colder et al., 2018; Donovan & Molina, 2008, 2014; Jackson et al., 2012). The odds of children initiating alcohol use were 1.45 (95% CI=1.17–1.80) times higher if their parents had less restrictive attitudes (Figure 4), with evidence of high heterogeneity (χ2=25.47, P≤0.001, I2=84.3%). By adding the results from the Effekt trial, the pooled effect size was slightly increased (OR=1.49, 95% CI=1.23– 1.82); χ2=27.63, P≤0.001, I2=78.3%). Figure 4. Forest plot for meta-analysis of parental attitudes towards children’s alcohol use and children’s alcohol use initiation. 71 Parental attitudes and children’s alcohol use frequency Of 29 articles, 18 reported on the relationship between parental attitudes and children’s alcohol use frequency (Aas et al., 1996; Andrews et al., 1993; Ary et al., 1993; Brody et al., 2000; Colder et al., 2018; Ennett et al., 2001; Gerrard et al., 2000; Kerr et al., 2012; Koning et al., 2010a, 2010b, 2012a; Margulies et al., 1977; Needle et al., 1986; Peterson et al., 1994; Pettersson et al., 2011; Sieving et al., 2000; Van der Vorst et al., 2006; Yu, 2003). Alcohol use frequency was measured by assessing lifetime use (Aas et al., 1996; Ary et al., 1993; Ennett et al., 2001; Kerr et al., 2012; Koning et al., 2010a; Margulies et al., 1977; Pettersson et al., 2011; Yu, 2003), use in the past 12 (Colder et al., 2018; Needle et al., 1986; Sieving et al., 2000), six (Andrews et al., 1993; Brody et al., 2000), three (Gerrard et al., 2000), one (Peterson et al., 1994; Van der Vorst et al., 2006; Yu, 2003) month(s) and weekly use (Koning et al., 2010a, 2010b, 2012a). Sixteen associations from 13 papers (Aas et al., 1996; Brody et al., 2000; Colder et al., 2018; Ennett et al., 2001; Gerrard et al., 2000; Kerr et al., 2012; Koning et al., 2010b; Margulies et al., 1977; Peterson et al., 1994; Pettersson et al., 2011; Sieving et al., 2000; Van der Vorst et al., 2006; Yu, 2003) were included in the meta- analysis [five studies were excluded due to reporting no data (Andrews et al., 1993; Needle et al., 1986), being from the same study that was already included (Koning et al., 2010a; Koning et al., 2012a) and reporting a result that was considered as an outlier (Ary et al., 1993)], and the results indicate that children had 1.52 (95% CI=1.24–1.86) times higher odds consuming alcohol if their parents had less restrictive attitudes (Figure 5) with evidence of high heterogeneity (χ2=149.36, P≤0.001, I2=90.0%). The results from the Effekt trial slightly increased the pooled effect size (OR=1.57, 95% CI=1.29–1.91; χ2=162.04, P≤0.001, I2=89.5%). Visual inspection of the funnel plot (Figure 6) showed asymmetry and suggested the presence of publication bias, this was supported by the Egger’s test (t=3.87, P=0.002). One of the studies (Colder et al., 2018) was outside the funnel on top left and it was decided to conduct an analysis without the study to see if it might have had any influence on the publication bias. By removing the study, the result became statistically non-significant. It might be that the frequency rates among participants in this study were more similar than in other studies and these participants represented a different population (19-year-old students included in the analysis were all alcohol drinkers). This could have contributed to the small standard error and have such an effect in the bias analysis. Removing one study from publication bias 72 analysis is not a common practice and the conducted analysis was exploratory. The study was still included in main meta-analyses. The repeated publication bias analysis with the data from the Effekt trial supported previous findings, suggesting the presence of publication bias when the study by Colder and colleagues (2018) was included, and no bias when it was excluded. H – high self-esteem, L – low self-esteem, I – intervention, C – control, F-O – fathers-older children, M-Y – mothers-younger children Figure 5. Forest plot for meta-analysis of parental attitudes towards children’s alcohol use and children’s alcohol use frequency. Parental attitudes and children’s drunkenness Of 29 papers, nine reported on the relationship between parental attitudes and children’s drunkenness (Glatz et al., 2012; Järvinen & Østergaard, 2009; Koning et al., 2012a, 2013; Mares et al., 2011a; Murphy et al., 2016; Özdemir & Koutakis, 2016; Pettersson et al., 2011; Strandberg et al., 2014). Drunkenness was assessed using the following measures: lifetime drunkenness (Pettersson et al., 2011; Strandberg et al., 2014), drunkenness (Glatz et al., 2012) or hazardous drinking (Murphy et al., 2016) in the past 12 months, drunkenness (Özdemir & Koutakis, 2016; Strandberg et al., 2014) or binge drinking in the past month (Järvinen & Østergaard, 2009; Mares et al., 2011a), onset of monthly drunkenness in the past 30 months 73 F-O – fathers-older children, M-Y – mothers-younger children, C – control, I – intervention, H – high self-esteem, L – low self-esteem Figure 6. Funnel plot of standard error by log OR for alcohol use frequency. 74 (Özdemir & Koutakis, 2016), onset of heavy weekly alcohol use (Koning et al., 2012a) and heavy alcohol use during weekends (Koning et al., 2013). Ten associations from seven papers (Glatz et al., 2012; Järvinen & Østergaard, 2009; Mares et al., 2011a; Murphy et al., 2016; Özdemir & Koutakis, 2016; Pettersson et al., 2011; Strandberg et al., 2014) included in the analysis [two studies were excluded due to not presenting any data (Koning et al., 2012a, 2013)] showed that having less restrictive attitudes increased the odds of children reporting having been drunk by 1.58 (95% CI=1.35–1.85) times with low-moderate heterogeneity (χ2=14.65, P=0.10, I2=38.6%) (Figure 7). Adding the results from the Effekt trial made no meaningful change in the outcome (OR=1.56, 95% CI=1.35–1.80; χ2=14.65, P=0.19, I2=26.4%). F-O – fathers-older children, M-Y – mothers-younger children, I – intervention, C – control, B – boys, G - girls Figure 7. Forest plot for meta-analysis of parental attitudes towards children’s alcohol use and children’s drunkenness. Perceived parental attitudes and children’s alcohol use initiation Nine of 29 papers had information on children-reported perceived parental attitudes (Aas et al., 1996; Andrews et al., 1993; Donovan & Molina, 2008, 2011, 2014; Gerrard et al., 2000; Margulies et al., 1977; Needle et al., 1986; Yu, 2003) related to children’s alcohol use, and four addressed the relationship between perceived parental attitudes and children’s alcohol use initiation (Andrews et al., 1993; Donovan & Molina, 2008, 2011, 2014). The data from two papers (Andrews et al., 1993; Donovan & Molina, 2014) were included in the meta- 75 analysis [the other two had the same sample as one of the included studies and were excluded from the analysis (Dononvan & Molina, 2008, 2011)], and the results indicate no evidence of the relationship between perceived attitudes and alcohol use initiation (OR=1.65, 95% CI=0.93–2.94) (Figure 8), with evidence of high heterogeneity (χ2=4.44, P=0.04, I2=77.5%). Including the data from the Effekt trial resulted in a slightly weaker but statistically significant association (OR=1.55, 95% CI=1.14–2.11; χ2=5.50, P=0.14, I2=45.4%). Perceived parental attitudes and children’s alcohol use frequency Six papers reported on the relationship between perceived parental attitudes and children’s alcohol use frequency (Aas et al., 1996; Andrews et al., 1993; Gerrard et al., 2000; Margulies et al., 1977; Needle et al., 1986; Yu, 2003). Based on seven associations from five studies [one study was excluded from the analysis due to missing data (Andrews et al., 1993)], children perceiving their parents’ attitudes as less restrictive was associated with 1.76 (95% CI=1.29–2.40) times higher odds of reporting consuming alcohol (Figure 9). There was evidence of high heterogeneity (χ2=32.42, P≤0.001, I2=81.5%). Including the data from the Effekt trial resulted in a slightly weaker association (OR=1.73, 95% CI=1.34–2.25; χ2=32.84, P≤0.001, I2=75.6%). Figure 8. Forest plot for meta-analysis of children’s perception of parental attitudes towards children’s alcohol use and children’s alcohol use initiation. Parental attitudes and perceived parental attitudes Perceived parental attitudes were reported as outcome measures in four papers (Aas et al., 1996; Ary et al., 1993; Gerrard et al., 2000; Yu, 2003). Six associations from four studies were included in the meta-analysis, and the results indicate that there is a weak positive correlation (r=0.27, P≤0.001) between the measures (Figure 10), with evidence of high heterogeneity (χ2=20.63, P≤0.001, I2=75.8%). 76 H – high self-esteem, L – low self-esteem, WS – with older sibling(s), WOS – without older sibling(s) Figure 9. Forest plot for meta-analysis of children’s perception of parental attitudes towards children’s alcohol use and children’s alcohol use frequency. B-F – boys and fathers, G-M – girls and mothers, H – high self-esteem, L – low self-esteem Figure 10. Forest plot for meta-analysis of children’s perception of parental attitudes towards children’s alcohol use and parental attitudes reported by parents. 2.3.4 Subgroup analyses and meta-regression Planned subgroup analyses did not indicate any statistically significant differences by study design (χ2=0.24, P=0.63), sample size (χ2=0.34, P=0.56), study location (χ2=2.28, P=0.13) and alcohol use frequency (χ2=1.67, P=0.20) among studies that reported on parental attitudes and children’s alcohol use frequency (Figure 11, Figure 12, Figure 13, Figure 14). However, heterogeneity decreased among three subgroups: studies involving fewer than 500 participants (I2=94.6% vs I2=73.1%), studies from Europe (I2=82.8% vs I2=67.2%) and 77 studies reporting on lifetime alcohol use (I2=90.5% vs I2=58.0%). Including the unpublished data from the Effekt trial did support the previous finding of not having any significant differences among moderators and decreased the heterogeneity a little further among the mentioned subgroups. H – high self-esteem, L – low self-esteem, F-O – fathers-older children, M-Y – mothers-younger children, C – control, I – intervention Figure 11. Forest plot for meta-analysis of parental attitudes towards children’s alcohol use and children’s alcohol use frequency by study design. 78 H – high self-esteem, L – low self-esteem, C – control, I – intervention, F-O – fathers-older children, M-Y – mothers-younger children Figure 12. Forest plot for meta-analysis of parental attitudes towards children’s alcohol use and children’s alcohol use frequency by sample size. F-O – fathers-older children, M-Y – mothers-younger children, H – high self-esteem, L – low self-esteem, C – control, I – intervention Figure 13. Forest plot for meta-analysis of parental attitudes towards children’s alcohol use and children’s alcohol use frequency by study location. 79 H – high self-esteem, L – low self-esteem, C – control, I – intervention, F-O – fathers-older children, M-Y – mothers-younger children Figure 14. Forest plot for meta-analysis of parental attitudes towards children’s alcohol use and children’s alcohol use frequency by frequency (lifetime vs last year). The results from the meta-regression analysis indicate that a one-year increase in age relates to a 0.10 reduction in the association (logOR) between parental attitudes and alcohol use frequency (b=-0.10, P=0.02; R2=58%) (Figure 15). A similar effect was not seen for drunkenness (b=0.08, P=0.21, R2=12%) (Figure 16). Including the data from the Effekt trial did not change the slopes of the lines but increased the total between-study variances explained by the models, R2=63% and R2=33%, respectively. Figure 15. Meta-regression of the effect of age on the association between parental attitudes and alcohol use frequency across studies 80 Figure 16. Meta-regression of the effect of age on the association between parental attitudes and drunkenness across studies 2.3.5 Sensitivity analyses Omitting each study by turn did not have an effect on the results. Additionally, changing the target groups, follow-up times and outcome measures within studies did not have an impact on the results. Initial analysis on the relationship between parental attitudes and children’s alcohol use frequency included the results from a study by Ary and colleagues (1993), but the effect size of that study (OR=20.52, 95% CI=10.09–41.73) differed from other results, and it was considered as an outlier. The study was removed from the analysis to assess its impact, but there was no meaningful influence. 2.4 Discussion 2.4.1 Summary of main findings This systematic review and meta-analysis is reportedly the first one, to assess the relationship between parental attitudes and children’s alcohol use in longitudinal and cross-sectional studies that have included parent-child dyads. The pooled estimates indicate that less restrictive parental attitudes increase the odds of children’s alcohol use onset, alcohol use frequency and drunkenness. Perceived parental attitudes had a similar effect on children’s alcohol use, but there was no evidence regarding the relationship between attitudes and alcohol use initiation. Although, by including the results from the Effekt trial, the relationship was present. The relationship between parent-reported and perceived parental attitudes was weak, indicating that children have little awareness of their parents’ attitudes towards their, or 81 children their age, drinking. Including the data from the Effekt trial was in line with the pooled estimates. 2.4.2 Key considerations The weak correlation between parents’ reports and children’s perceptions on attitudes does not mean that children perceive parental attitudes completely different from parents; rather, their perceptions might be more skewed towards lenient attitudes. To some degree, participants might reflect on their own values and expectations (Cohen & Rice, 1997; Ungar et al., 2012), but this can also be the result of ignorance among children if parents have not expressed their attitudes. Also, if the social norm seems to support parents introducing alcohol to children (Jones, Andrews, & Francis, 2017; Raudne, 2012), it may override parents’ own beliefs, even if they reject the norm, because they assume that other parents are following it (Prentice & Miller, 1996). This, in turn, encourages them to engage in the behaviour (Gilligan & Kypri, 2012; Kypri et al., 2007) and makes it more likely for children to perceive their parents as more lenient, while parents considering themselves as still restrictive. Another option is that parents might overestimate the level of restrictiveness regarding their attitudes. For example, when parents were asked about their current drinking habits, around 70% of respondents perceived themselves drinking either at or below the recommended level (Valentine et al., 2010). At the same time, only 30% of parents corresponded to that level when their past week’s heaviest drinking day alcohol amounts were assessed. Human behaviour is influenced by a variety of factors, attitudes and social norms being two of them (Ajzen & Fishbein, 2005), and as this review addresses two inter-related pathways (one’s own attitudes, other’s behaviour), the reasons behind the investigated relationship are likely to be complex. Parents have been shown to have stricter attitudes towards alcohol use when children are younger (Koutakis et al., 2008; Mares, Lichtwarck-Aschoff, Burk, Van der Vorst, & Engels, 2012; Prins, Donovan, & Molina, 2011). As children grow older the need for autonomy increases, there are more external factors [e.g. peer pressure, media (Jernigan, Noel, Landon, Thornton, & Lobstein, 2016; Nesi, Rothenberg, Hussong, & Jackson, 2017)] influencing behaviour, and the strength of the parent-child relationship can be under pressure (Darling, Cumsille, Caldwell, & Dowdy, 2006; McGue, Elkins, Walden, & Iacono, 2005; Steinberg, 2001). Becoming more lenient is one way of granting autonomy and balancing the changes in the relationship (Masche, 2010). This is also supported by the finding from current 82 meta-regression analysis, that the relationship between attitudes and alcohol use frequency was moderated by the children’s age. The positive parent-child relationship is a central part of authoritative parenting, which has shown to have a protective effect against alcohol use (Baumrind, 1991; Čablová et al., 2014). This parenting style incorporates a myriad of factors and the combination of restrictive parental attitudes with other factors (e.g. quality of the relationship, rule-setting, communication) may explain the relationship between attitudes and alcohol use (assuming that parents with restrictive attitudes are more authoritative) (Koning, Van den Eijnden, Verdurmen, Engels, & Vollebergh, 2012b). For example, Raitasalo and Holmila (2017) investigated how Finnish parents’ alcohol use related to their parenting practices and found that frequent and high volume alcohol use were both related to more lenient practices, including less restrictive attitudes towards children’s alcohol use. Glatz and colleagues (2012) and Chramostova (2014) have suggested that it is children’s alcohol use that predicts parental attitudes, rather than the other way around. Huver and colleagues (2007) have made similar conclusions regarding adolescents’ smoking and related parental factors but added that the attitude affecting behaviour relationship, while being weaker, still existed. When parents are confronted with children’s alcohol use, it is possible that they change their attitudes towards the behaviour, potentially to reduce cognitive dissonance (Chramostova, 2014). The findings from a qualitative study among 12–17-year- old Dutch students showed that the students who had already initiated alcohol use perceived their parents more permissive than the students who were not consuming alcohol (Janssen, Mathijssen, Van Bon-Martens, Van Oers, & Garretsen, 2014). Therefore, taking parental knowledge on children’s alcohol use into account in future studies may offer valuable insight, as becoming less restrictive towards a child’s alcohol use could be a coping mechanism in response to becoming aware of their actual alcohol use (Glatz et al., 2012). This contradicts Koning and colleagues’ (2010a) view that suggests attitudes’ superiority over behaviour. To gain a clearer understanding of causal pathways and the combined influence of indicators, both children and parents, and their attitudes and behaviour should be assessed in future research. 2.4.3 Comparability with previous studies Comparing the findings with results from previous reviews (Ryan et al., 2010; Sharmin et al., 2017b; Yap et al., 2017) provides an indication that the source reporting on attitudes might have an effect on the outcome. For example, two reviews (Ryan et al., 2010; Yap et al., 2017) 83 presented contradictory findings on the relationship between parental attitudes and alcohol use initiation, but removing the studies with parent-reported attitudes is likely to support the result stating that there is no evidence for a relationship between perceived attitudes and behaviour. Sharmin and colleagues (2017b) focused on risky drinking, which would be approximately equivalent to drunkenness in this review, and found that perceived parental approval of alcohol use was related to higher risky drinking, but there was no evidence for the relationship between perceived disapproval and lower risky drinking. The current review did not assess the relationship between perceived parental attitudes and drunkenness but, based on parents’ own reports, there is a clear relationship between the two. 2.4.4 Limitations and strengths This paper has several limitations. First, the search criteria were softened during the screening, as there were abstracts with unclear information. Despite this limitation, the final list of included papers provides a comprehensive overview of the topic. Secondly, all models except one showed high between-study heterogeneity, limiting the generalisability of the findings. The results from the sensitivity analyses showed that excluding and grouping the studies did not change the direction and significance of the relationships. High heterogeneity could be due to variation in follow-up times, outcome and predictor measures, and confounding factors (e.g. parenting variables, psychosocial variables, demographics). Thirdly, some studies excluded non-significant results, which might have resulted in an increased pooled effect size. However, this is somewhat unlikely, as the missing results were related to different outcomes and including one or two studies would not have a significant influence on the pooled effect size. Fourthly, definitions of measures varied among studies (e.g. from single questions to complex constructs regarding attitudes, and from weekly use to lifetime regarding alcohol use). Despite these limitations, this paper has the following strengths. Parental attitudes were reported by parents themselves, as there can exist a mismatch between actual and perceived attitudes (Aas et al., 1996; Smith et al., 1999; Van der Vorst et al., 2005; Williams et al., 2003). Both longitudinal and cross-sectional studies were included, creating a larger evidence base and increasing the analysis power. There can be a considerable time-lag, as the change in one’s attitude might not be followed by the change in other’s behaviour in a short time- frame (Sutton, 2004), but attitudes are not stable and might change over time. Including cross-sectional studies adds an opportunity to investigate the indicators measured at the same 84 time. As the results indicate, regarding alcohol use frequency, the pooled effect size is higher for cross-sectional studies, but there is no statistically significant difference between cross- sectional studies and studies longitudinal in nature. 2.5 Conclusions This paper provides novel insights when assessing the relationship between parental attitudes and children’s alcohol use by focusing on parent-child dyads, including a broader set of studies and assessing the effect on various moderating factors on the relationship. The results show that children’s awareness of their parents’ attitudes can be different from actual attitudes, and there is an association supporting that parents may delay alcohol use onset, but also influence children’s behaviour afterwards, although the effect may attenuate as the age increases. Alcohol use does not occur in a closed system, and therefore is not influenced by only one factor. Greater emphasis should be placed on assessing the individual, combined and bidirectional influence of parental factors to understand which indicators have a stronger effect alone and combined, and what kind of effects the indicators have on each other. Focusing on parents’ role in alcohol use prevention and reduction is crucial, as parents’ role stays important throughout childhood. 85 Chapter 3. Adaptation and delivery of a universal parent-oriented alcohol use prevention programme in Estonia 86 3.1 Development of the parent-oriented prevention programme At the end of 1990s Sweden faced a problem of alcohol use increase among adolescents (Hvitfeldt, Andersson, & Hibell, 2004) and to tackle the issue, the Swedish National Institute of Public Health offered a grant to develop a universal prevention programme. Developmental psychologists from the University of Örebro developed an alcohol prevention programme named ÖPP, which was carried out in 1999–2001. In 2012, the programme was renewed and renamed to Effekt. Effekt is based mainly on findings from two studies conducted by the research team in the University of Örebro, which showed that restrictive parental attitudes and children’s involvement in extracurricular activities were related to lower alcohol use among children (Koutakis, 2011). The “10 to 18” (“10 till 18” in Swedish) study was a longitudinal study carried out among 10–18-year-old students and their parents in a middle-sized town Köping in Sweden (Stattin, 2002). The second study was carried out among children and parents in three localities (Hallsberg, Kumla, Örebro) in Örebro County, in Sweden who participated in ÖPP in 1999–2001 (Koutakis, 2011). Additionally, earlier findings presented by other researchers supported the findings of these two studies and strengthened the rationale behind the programme’s objective to postpone and reduce alcohol use among youth by maintaining parents’ restrictive attitudes towards adolescents’ alcohol use and by increasing youth’s involvement in extracurricular activities (Koutakis, 2011). To achieve the objective parents received newsletters and attended six 20–30-minute school meetings twice a year, when their children were in grades 7– 9 (ages 13–15). 3.2 Previous evaluation of the programme As of 2011, the programme had shown promising results in tackling alcohol use among adolescents (Koutakis et al., 2008), and the adaptation of Effekt had been started in several countries (the Netherlands, Norway3, Iceland, Belarus, Finland, Russia and Slovenia4) (Koutakis, 2011). The initial evaluation (a quasi-experimental design with matched controls) carried out in 1999‒2001 by ÖPP developers among 900 students and their parents in Sweden showed a lower rate for adolescents’ last month drunkenness (13% vs 27%) and stricter attitudes among 3 The possibility to adapt the programme to Norwegian conditions was investigated in 2010–2012 (Endresen, 2013), but the delivery was never carried out (based on the personal communication with the Norwegian representative). 4 Piloting started in 2014/2015. 87 parents (M=3.81 vs M=3.46) at the final, 30 months follow-up in the intervention group (Koutakis et al., 2008). Regarding children’s involvement in organised extracurricular activities, the participation rate decreased in both groups, and there were no differences between groups. Additional analysis on the mediating effects of parental attitudes showed that the more restrictive attitudes were, the less the drunkenness among intervention group students increased and the fewer adolescents started to get drunk on a monthly basis (Özdemir & Koutakis, 2016). The analysis on the moderating effects of demographic factors, alcohol use by peers and parents and the quality of parent-child relationship on parental attitudes and youth drunkenness showed that there were no effects in the intervention group, but the immigration status and the quality of parent-child relationship moderated parental attitudes over time in the control group – immigrant background and lower relationship quality between mother and the child resulted in becoming less restrictive over time. Results from a separate cRCT in 2007‒2010 among 1752 13–16-year-old Swedish students and their parents appeared to be inconsistent with those reported from the first study by showing that there was no evidence on the intervention effect on delaying drunkenness (60% vs 59%) or reducing last month drunkenness (18% vs 19%) among youth by the end of the programme (30 months follow-up), despite the more restrictive attitudes among intervention group parents (M=3.78 vs 3.56, 89% vs 82%) (Bodin & Strandberg, 2011; Strandberg & Bodin, 2011). These findings raised a discussion between the two research groups in Sweden. The ÖPP developers pointed out that these results should be re-examined, because: 1) the fidelity was poor – the number of meetings that took place was lower than expected (M=4.7); 2) dichotomization of the alcohol use measures lowered the variance; 3) mediating effects of parental attitudes were not investigated (Özdemir & Stattin, 2012). After re-analysing the data, they concluded that ÖPP had a direct and indirect effect (mediated by parental attitudes) on youth drunkenness. Bodin (2012) addressed the concerns and pointed out that reaching the maximum number of meetings on a broader scale delivery is difficult and the average number of 4.7 meetings reflects more the regular than optimal situation. The issue of dichotomising the data is considered a valid option, especially when the data is highly skewed (e.g. when prevalence rates are small). Bodin also emphasised that the number of participants was high enough to detect small to moderate effect sizes and re-analysing the data did not have enough basis for rejecting the null hypothesis. Regarding the mediating effects, Bodin agreed on the importance of mediation analysis, but she also pointed out the potential influence of national media campaigns that may have increased restrictiveness among parents. 88 ÖPP was also carried out in Dutch schools (Koning et al., 2009). A cRCT with almost 3000 13-year-old students and their parents was started in 2006, and the participants were divided into four groups (i.e. parent intervention – ÖPP, student intervention – Prevention of Alcohol use in Students, parent + student intervention, control). The parent intervention was adapted, and the number of meetings was reduced from six to two, taking place at the beginning of the school years in 2006 and 2007. Compared to the two trials in Sweden, the trial period was longer (50 months), continuing until participants reached the legal age, 16, to consume alcohol, and more assessments were conducted (baseline – T0; 10 months – T1; 22 months – T2; 34 months – T3; and 50 months – T4) (Koning et al., 2009, 2010b, 2013). Over the years, the combined intervention showed the best results in reducing students drinking. The incidence of heavy weekly use, weekly use and frequency of monthly use in the combined intervention group was lower at T1 compared to the control group, and the latter two showed similar effects at T2 (Koning et al., 2009), T3 (Koning, Van den Eijnden, Verdurmen, Engels, & Vollebergh, 2011) and T4 (only heavy weekly use) (Koning et al., 2013). A separate analysis conducted to assess the growth of weekly drinking up to T3 showed that the increase was less steep among baseline drinkers and non-drinkers in the combined intervention group (Koning, Lugtig, & Vollebrgh, 2014). The effects on weekly and/or heavy weekly use incidence were moderated by students’ education level and externalising behaviour at T2, (Verdurmen, Koning, Vollebergh, Van den Eijnden, & Engels, 2014) and by self-control and perceived rules on alcohol at T3 (Koning et al., 2012a). The analysis of the mediating effects of adolescents’ self-control, alcohol rules and parental attitudes (both perceived by children and reported by parents) at T1 on weekly alcohol use initiation at T2 showed that the combined intervention had an effect on drinking behaviour being mediated by all aforementioned indicators except children’s perception of parental attitudes (Koning et al., 2010b). Koning, Maric, MacKinnon, and Vollebergh (2015) found that while the combined intervention had an effect on perceived rules about alcohol at T1, this had an effect on adolescents’ self-control at T2, which in turn had an effect on the amount of weekly alcohol use at T3. The effect of the combined intervention on the prevalence of heavy drinking at T4 was mediated by adolescents’ self-control and perceived rules about alcohol at T3 and on the amount of weekly alcohol use at T4 by adolescents’ self-control and alcohol initiation at T3 (Koning et al., 2013). 89 3.3 Bringing Effekt to Estonia Estonia is a small north-eastern European country with 1.3 million inhabitants, which is known for its innovative approaches in the digital sphere (Hammersley, 2017) and high alcohol use rates among adults (World Health Organization, 2018) and adolescents (Kraus et al., 2016). Total alcohol per capita (15+) consumption in 2016 was 11.6 litres of pure alcohol, being almost two litres higher than the average of the WHO European region. The prevalence of HED5 – 42% (63% among men and 24% among women) – is the fourth/fifth highest (same as Czechia) in the region, after Lithuania, Luxembourg and Latvia. Although the burden of alcohol-related harm has lingered in the country for decades, alcohol is perceived as a part of the culture and everyday life, thereby supporting the pro-alcohol norm in the society (Raudne, 2012). This is also reflected among youth – Estonian adolescents initiate alcohol use at an early age, 12 on average and typically the first episode of drunkenness takes place a year later (Aasvee & Rahno, 2015); by the age of 15–16 more than 85% of students have consumed alcohol (Kraus et al., 2016). Despite the severity of the problem, there is no nationally applied and coordinated holistic and systematic approach to effectively prevent and reduce alcohol use among adolescents (Ministry of Interior, 2019). Talking about alcohol, tobacco and illegal drugs has been in the school curriculum as a part of the health education since 1996 (Estonian lower and middle education national curriculum, 1996), but it is impracticable to assess its direct impact on alcohol use. There has been a myriad of activities carried out over the years, including short- term programmes and one-off activities targeting children, such as including ex-users, introducing substances, using scare tactics. Only in recent years, the substance use issue has been addressed more vigorously and several guidelines and evidence-based prevention programmes have been developed and carried out (Kull, Saat, Kiive, & Põiklik, 2015; National Institute for Health Development, 2018a; Streimann & Pertel, 2016; Streimann, Selart, & Trummal, 2019). As of February 2019, the cross-sectoral prevention concept is under development to provide a science-based foundation when making decisions and allocating funds in the field of prevention (Ministry of Interior, 2019). But none of it existed in the early 2010s. From here on, the chapter will follow a chronological order of the programme’s adaptation and delivery. 5 Consumption of 60+ grams of pure alcohol on at least one occasion in the past 30 days 90 In 2011, after several schools approached NIHD and asked for activities to tackle the alcohol use issue, it was decided to develop a systematic and sustainable solution to prevent and reduce alcohol use among adolescents. NIHD had the opportunity to use state funding to address the problem mentioned above by delivering a prevention programme, of which the author (M.TÖ) was appointed to lead. The initial project team included two additional people from NIHD, who had a background in public health, psychology and social policy. The selection process of existing alcohol prevention programmes was initiated in the autumn of 2011, and five criteria were applied: 1) low long-term costs, 2) administratively easy to deliver, 3) time-efficient (especially from the target group’s perspective), 4) promising results on reducing and delaying alcohol use, 5) parents included in the target group (Foxcroft & Tsertsvadze, 2011a). Out of different programmes considered including Unplugged (Faggiano et al., 2002), Project Northland (Williams, Perry, Farbakhsh, & Veblen-Mortenson, 1999) and Effekt (Koutakis et al., 2008), the latter met the criteria most closely. 3.4 Preparations to deliver the programme The data from the 2009/2010 Health Behaviour of School-aged Children (HBSC) study showed that approximately 40% of 11-year-old students, 75% of 13-year-old students and 92% of 15-year-old students had initiated alcohol use (Aasvee & Minossenko, 2011). Drunkenness prevalence among the same age groups was 7%, 33% and 63%, respectively. The findings showed that Estonian students initiate alcohol use at the age of 12–13 (~6–7th grade), which meant that to delay it, the target group of the intervention should be below that age group. While the programme delivered in Sweden targeted parents whose children were 13–15-year- old, the project team decided to lower the age to 11–13-year-old in Estonia, as alcohol use onset in Estonia took place among younger children than in Sweden. In addition, the focus of the programme was set only on parental attitudes, and not on children’s involvement in extracurricular activities. This was due to the lack of a nationally integrated financial support system for extracurricular activities and the variation in availability, diversity and quality of the activities by regions, but also due to the reported findings by Koutakis and colleagues (2008), that indicated reduced participation in extracurricular activities and no evidence of difference between the intervention and control groups. The format (two meetings and newsletters per school year) and length (three years) of the programme were left unchanged. 91 The project team received all the programme related materials (i.e. Powerpoint presentations, templates for meeting summaries and newsletters, programme manual, questionnaires for parents and children) from the programme’s developer in February 2012, beyond which the materials were translated from Swedish to Estonian. At the same time, a two-day training event was carried out in English at NIHD by the programme’s developer, Nikolaus Koutakis, to train potential facilitators who would deliver the programme to parents. Out of 22 participants (included all three people from the project team), eight became facilitators. Their background varied from pedagogy, social policy and psychology to health promotion and public health. In May 2012, the project team carried out a pilot to test the programme’s content among parents. As the content of the meetings was very similar (some variation in background colours and pictures used in the presentations), the project team decided to pilot only the content of the first meeting. Three facilitators piloted the presentation with two parent groups in two schools and carried out the meeting as expected in the programme. The majority of the participants agreed on the necessity of the approach, as youths’ alcohol use was perceived as a serious problem. The feedback on the content was very positive, with some minor ideas to change the outlook of the slides. The most valuable input was received regarding the existing topic-related myths that might raise questions among parents (e.g. it is better to offer the alcohol at home than letting children do it secretly with their friends; letting children sip strong alcoholic beverage would keep them off from trying out alcoholic beverages in the future; the forbidden fruit tastes the sweetest, so parents should not hide alcohol from children). Any other materials were not piloted among parents. The content of the meeting summaries was kept the same, but as the content of the newsletters was heavily focused on extracurricular activities, the project team decided not to use the existing templates and write new ones on an ongoing basis. At the same time the piloting took place, the project team was actively engaged with recruiting schools. It was important to include schools that were willing to participate in the programme for three years, thus instead of approaching all schools with grades 5–7 (N=449), the project team decided to contact the schools in the Network of Health Promoting Schools (NHPS) (Stewart Burgher, Barnekow Rasmussen, & Rivett, 1999). According to the data from the Estonian HBSC study, belonging to the network did not predict lower alcohol use prevalence, but it was hypothesised that the network schools would be more likely to 92 participate in the programme, and support and motivate parents for the whole period. The detailed recruitment process is described in Chapter 4 . In total, 34 schools received the intervention. The schools varied by:  school and class size (the smallest school had 55 students, six of them in the fifth grade, while the biggest school had 901 students, 74 of them in the fifth grade)  settlement type (eight schools were from the capital – Tallinn, 14 from urban areas, with >1000 inhabitants, and 12 from rural areas, with <1000 inhabitants  location (14 out of 15 counties were represented by at least one school) To increase facilitators’ level of preparedness, all facilitators went through a base (demo) training in August 2012, to rehearse the presentation in an environment similar to actual meeting and answer difficult questions that may arise during the meetings; other facilitators took the role of parents who would put the presenter into challenging situations. As a result of this event, two things were decided: 1) in order to ensure that the content presented at the meetings stays similar irrespective of the facilitator, a manual (pre-written script of the presentation’s content) will be written for each meeting, 2) demo training will take place before each meeting wave (in total six waves). The manuals and demo presentations were not required by the programme’s developer, but the project team found the approach supportive for the facilitators and it also allowed aligning facilitators’ standpoints with the programme’s vision, if different. Summarising the adaptations as of September 2012, when the programme was initiated, following changes had been introduced: 1) the start of the programme was moved from seventh to fifth grade, 2) the focus on extracurricular activities was excluded, 3) supportive demo training with the facilitators was included, 4) it was decided to renew the content of the newsletters, 5) the name of the programme was “Estonianised” to Efekt. A matched-pair cRCT among parents and children was carried out in parallel with the programme in 2012–2015 to assess the programme’s effectiveness in preventing and reducing alcohol use among children (see Chapter 4). In addition, individual interviews on the programme’s perceived effectiveness and delivery process were carried out with teachers and parents in spring 2015 (see Chapter 6), and a focus group on similar topics was carried out with the programme’s facilitators in winter 2015 (see Chapter 5). An overview of all the assessments that were carried out during the programme is presented in Appendix 4.1. 93 3.5 Meetings I–II The programme was formally initiated among fifth grades in 34 schools (60 classes) in September 2012. At the beginning of the school year, all teachers in the intervention schools received information on the programme via e-mail, which was also the main communication line with teachers during the whole programme. To reduce prejudice towards the programme, the teachers gave a more general description of the programme (e.g. focus on parental support on child’s development) in the first meeting invitation, and the full description was presented at the meeting. Thereafter, the information sent in the invitation included the topics discussed in the upcoming meetings. It was expected that fewer parents would attend the meeting if there would be a separate meeting related to the programme. So, the programme’s meeting was a part of the general school meeting, which gives an overview of past and future class and school activities and enables face-to-face communication between parents and the teacher but also between parents themselves. The part of the general meeting that focused on the programme was carried out by a facilitator who made a PowerPoint presentation and encouraged parents to take part in the discussion. The length of the meetings varied across schools, mostly due to parents’ willingness to participate in the discussion. The formal part (presentation) was about 20–25 minutes long, and it was typically followed by a guided discussion, lasting up to 30 minutes. In total, the length of the meeting varied between 30 to 60 minutes. During the first meeting, a comprehensive overview of programme’s objectives, format (i.e. length, frequency, distribution of materials), and Swedish origin (including programme’s results) was introduced to the parents. Although existing studies showed that children who see their parents consuming alcohol are more likely to initiate and consume alcohol themselves (Ryan et al., 2010), the team decided that parents will be made aware of the fact, but no particular emphasis would be put on reducing alcohol use among parents. In addition, statistics on alcohol use in Estonia, alcohol use consequences among youth and parents’ role in prevention were covered at the meeting. The latter was a cross-cutting theme in all the meetings (thus, this topic will not be covered separately, when describing the content of the following meetings). The aim of talking about parents’ role in prevention was to encourage parents to invest in the trusting and supportive parent-child relationship and to be a role model 94 for children, especially when they might perceive their importance decreasing and friends’ importance increasing. The theme conveyed three main messages:  do not offer alcohol to your child  express clearly your restrictive attitudes towards adolescents’ alcohol use  talk to your child (general communication, including alcohol) At the end of the meeting, parents were encouraged to establish similar rules by making verbal agreements, which aimed to increase collaboration between the parents and contribute to shaping the norms to support children’s development in a safe environment. Facilitators wrote down all agreements and included them in the meeting summaries, which additionally included the number of parents who participated and summarised the topics discussed at the meetings. The summaries were sent to all parents by e-mail (via the teacher), irrespective of their participation at the meetings; although, the prerequisite was that the meeting took place. Parents were encouraged to notify their children about the agreements, so children would know that the same rules are (ideally) applied at all homes. To assess the attendance6 by classes and schools, the facilitators counted the number of participants. In total, 536 participants (47.1%) from 56 classes (91.8%) and 32 schools (94.1%) attended the first meeting. The first meeting wave ended with all the teachers receiving a two-sided newsletter (on paper), which was forwarded to parents (via children). The newsletter covered the topics discussed at the meeting in more depth and included additional information on the reasons why adolescents drink alcohol. After the meetings were over, a new feature was introduced – a reflection event, where facilitators shared their experiences and (verbal) feedback from parents and teachers. The facilitators raised initial concerns regarding the repetitive content of the presentations, and it was suggested to add a new element to the second presentation to facilitate the discussion with parents and increase their active involvement. While all facilitators were satisfied with their performance at the meetings, several schools contacted the project leader to express their dissatisfaction with one of the facilitators. As a result, the contract with that facilitator was not extended. 6 Attendance and participation in the context of Effekt programme are used interchangeably in this thesis; both refer to attending the meetings. The word engagement has been used when refering to being active at the meetings, e.g. taking part of a discussion. 95 As the workload was intensive for the facilitators, the project team organised a new training event in the spring of 2013, which resulted in three new facilitators, whose background covered health promotion, social work and business development. Altogether, these ten facilitators (seven from the previous training and three from the current; one male, nine females, mean age 31.9 years, range 25–39) formed the programme’s core team and stayed together until the end of the programme. Facilitators had the following tasks throughout the programme: taking part in the demo training and reflection events, training parents, and giving feedback to the project leader regarding participation rates and parents’ agreements. Before the second demo training took place, the project team met to discuss the possibility to add an interactive task to engage parents. As empirical findings from several research studies have shown that various risk factors influence children's alcohol use, the project team decided to increase parents’ awareness of environmental factors. Parents were presented with four factors – alcohol availability, peers’ influence, media, parental attitudes – and then asked to rank the factors based on their impact on children’s alcohol use; after that, a discussion was meant to take place. The exercise aimed to show parents that even if they cannot control all factors, they do have control over their attitudes. Additionally, as several participants had raised the question “How to explain own reasons for drinking alcohol?”, the team added one slide with possible explanations to the presentation. Expert input from psychologists was used when developing the reasons to share with children. To assess the delivery and perceived usefulness of materials, and to get an immediate reflection on parents’ thoughts on the meetings, the project team developed a short feedback form (Appendix 4.2), which was meant to be distributed to parents at the end of the meeting. Collecting feedback was a form of process evaluation and was meant to work as a systematic “internal audit” of the programme. Programme’s activities informed the development of the questions and the participants were asked about: 1) the attendance of the previous meeting and receiving the summary of that meeting, 2) receiving and reading newsletters, 3) the helpfulness of the newsletters and meetings in talking about alcohol with the child, 4) the meeting’s content planned to share with family members, 5) the facilitators’ credibility and performance. After the project meeting and demo training event took place, the teachers were contacted to arrange the dates of the spring meetings. As mentioned earlier, the content of the second presentation largely repeated the content presented in the first meeting, with the exception of a 96 task to rank environmental factors and possible options on how to explain own alcohol use to children. Parents received the ranking task on a two-page leaflet that they could keep; the other side of the leaflet presented questions to use when talking about alcohol with children (same questions as presented during the meeting). The leaflet was introduced to enhance the learning experience (parents ranked the factors on paper) and to increase the likelihood of parents making use of what was discussed at the meeting. At the end of the meeting, all previously made agreements between parents were reviewed and adjusted, if necessary (this approach was used at the end of all subsequent meetings). Thereafter, the participants received a one-page feedback form (on paper) and were asked to answer the questions in the classroom. Majority of the participants answered the questions in five minutes or less. The summary of the results is presented in Chapter 3.7. In total, 342 participants (30.1%) from 47 classes (77.0%) and 29 schools (85.2%) attended the second meeting. Right after the meeting, parents received the second newsletter, which covered the topics discussed at the meeting and included additional information on energy drinks (Appendix 4.3). The latter topic was introduced due to participants’ concern in increasing rates in children’s energy drinks use nationally. After the meetings were over, a reflection event with the facilitators took place. The verbal and written feedback from the parents, teachers, and facilitators implied strong reluctance among participants to attend due to the repetitive nature of the information. Thus, it was decided to introduce additional topics to the programme, while keeping the main three messages and the format unchanged. By the end of the second meeting wave, three additional adaptations had been made to the programme: 1) an exercise to rank environmental factors that influence children’s alcohol use, 2) distributing programme-related leaflets, 3) introduction of systematic feedback collection. 3.6 Meetings III–VI As the project team decided to incorporate new topics to meetings’ content starting with the third meeting, the input from parenting and alcohol-related research and additional experts (e.g. family therapists, psychologists, educational scientists, teachers and public health experts) informed the development of the new content. As the time to develop new content for the meeting was rather short, it was decided to focus on one meeting at a time. 97 The third meeting focused more closely on the myths related to parental alcohol supply (alcohol availability at home), which were found to be common among Estonian parents (Raudne, 2012). The parents were presented with four statements they had the chance to agree or disagree in the following discussion: 1) by offering a bit of alcohol to the child, it will reduce their interest in it; 2) by offering alcohol to the child, they assume that the parents approve alcohol use; 3) children who are allowed to drink alcohol at home do it outside the home as well; 4) children can drink a bit of alcohol during New Year’s Eve. Additionally, to emphasise the reality of possible negative alcohol-related consequences, which tend to increase in summer months, parents were presented with national statistics on alcohol-related accidents (e.g. drunk driving, drowning) and violations (e.g. minors buying alcohol). The myths and statistics were also presented on a leaflet that parents could take with them. A separate focus was put on alternative activities (e.g. extracurricular activities) that could improve children’s communication skills, provide new experiences, reduce stress, increase cohesion and self-confidence – providing similar outcomes as to why children tend to drink alcohol (based on the analysis of 2009/2010 national data from the HBSC study). In total, 453 participants (39.8%) from 56 classes (91.8%) and 33 schools (97.1%) attended the third meeting. As the content of the meeting was improved, the original templates for meeting summaries were adjusted accordingly. The third newsletter, sent out after the third meeting wave, addressed additional topics on cannabis and snus use, and the inevitability of alcohol use. As the direct target group of the programme included parents, the project team decided to incorporate parenting-related topics that would support parents in preventing children’s alcohol use. Recent work by Raudne (2012) on Estonian parents’ parenting styles regarding children’s alcohol use gave input to address the topic more in-depth. Raudne implied that while several participants found the authoritative parenting style the most supportive regarding alcohol use, participants’ skills were lacking and application of other styles (e.g. permissive, authoritarian) was more common due to predefined beliefs. With the help of external experts, the project team developed content related to parenting styles, which was based on the typology by Baumrind (1967) and Maccoby and Martin (1983), for the fourth meeting. The participants were given a brief overview of all parenting styles and a more in- depth overview of the authoritative parenting style, with examples of how to apply it at home. To practice applying authoritative parenting-related skills, the parents were presented with four realistic and common alcohol-related scenarios (Appendix 4.4), that were developed in 98 collaboration with psychologists. The parents had to choose one scenario out of the four to resolve and were then divided into two groups, one acting as parents and the other as children. The participants were asked to reflect on their feelings and describe how they would resolve the situation. In parallel, the facilitators incorporated into the discussion a universal four-step problem-solving model, which was developed in collaboration with the psychologists and incorporated some parts (i.e. active listening, I-messages) from the Parent Effectiveness Training developed by Thomas Gordon (Gordon, 2008; see Appendix 4.4). The four steps included staying calm in the situation, applying active listening, sharing information and making agreements. The information on both the parenting styles and problem-solving approach was distributed to parents on a leaflet. In addition, everyone who wished, could take a brochure “Learning is work – help your child” (National Institute for Health Development, 2011a). Due to a common view among parents of intervention school children not drinking alcohol, one additional topic was added to the fourth meeting – baseline results from the programme’s trial, which confirmed the national results presented to parents earlier. At the end of the meeting, the participants filled the feedback, which had been slightly changed – two questions on the usefulness of the activity and materials were added (these questions were kept until the end of the programme). In total, 292 participants (25.7%) from 47 classes (77.0%) and 30 schools (88.2%) attended the fourth meeting. The fourth newsletter sent to parents after the meeting included additional information on e-cigarette use and children’s self-esteem and assertiveness. The content of the fifth meeting focused on two main topics – the media’s role in children’s alcohol use and alcohol’s influence on the developing brain. The role of media was covered due to two reasons: 1) the active topic-related debate in the society, 2) children’s high exposure to alcohol-related media (especially on the internet). At first, parents were shown a well-known advertisement for Estonian beer7, and it was explained that while parents can assess the content of advertisements critically and separate fiction from real life, children’s reasoning skills are still developing. Then, the facilitators analysed the advertisement with parents and introduced parents a variety of questions (also provided in a leaflet) they could use when talking about alcohol advertisements at home (Appendix 4.5). The aim of the questions was for the parents to understand how children (and their friends) perceive the content of alcohol advertisements, explain their views and attitudes, and support children’s critical thinking in general. Talking about critical thinking gave input on addressing the 7 https://www.youtube.com/watch?v=a1JXxNLTBI0 99 relationship between alcohol and the developing brain, as one part of the brain that alcohol influences is prefrontal cortex, which is related to several essential life skills, including critical processing of information (Galinsky, 2010); the list of the skills was presented on the other side of the leaflet with advertisement questions. In total, 323 participants (28.4%) from 52 classes (85.2%) and 31 schools (91.2%) attended the fifth meeting. The fifth newsletter, sent after the fifth meeting, focused on the topics covered at the meeting and additionally addressed parent-child communication when talking about alcohol-related risks. The main topic of the final meeting was peer pressure, as the research literature implies that as children get older, the influence of peers starts to increase. With the help of psychologists, a roleplay was developed to talk about group norms and possible strategies to handle peer pressure. Before conducting the roleplay, parents were familiarised with two stories from a fictional 15-year-old girl’s life (Appendix 4.6). The first one described a situation that took place at her birthday party, where her mother agreed to offer her some alcohol. The second story was about a party where her classmates pressured her to drink alcohol. Regarding the first story, parents discussed possible reasons why the mother had offered alcohol and possible alternatives. Regarding the second story, parents were invited to play the parts of the teenagers and to solve the situation without the girl drinking alcohol. After the roleplay, parents were reminded of the agreements and encouraged to keep them after the programme had finished. By the final meeting, 220 agreements were made between parents – 112 were related to substance use (alcohol, tobacco products, drugs, energy drinks) and 108 to different daily issues (see Appendix 4.7). The feedback form had been adjusted, and a short block of questions related to the agreements had been added. All participants who wished could take home brochures “Support the development of your child’s self-esteem” (Saat, 2011) and “Be assertive” (National Institute for Health Development, 2011b). In total, 159 participants (14.0%) from 31 classes (50.8%) and 20 schools (58.8%) attended the final, sixth meeting. The final newsletter sent to parents included practical information on how to plan the 9th grade graduation party, and how to act when parents see other children acting delinquently. By the end of the sixth meeting, the average number of meetings taken place across 34 intervention schools was 4.78 (SD=1.30). Out of 60 classes, 41 (69%) received at least five meetings, 17 (28%) received 3–4 meetings, one received two and one did not receive any meetings. 100 The adaptations made to the programme after the second meeting wave included: 1) addressing parental alcohol supply related myths, 2) providing information on alcohol-related negative consequences, on alternative activities to alcohol use, on parenting styles, with a special focus on authoritative style, and on alcohol’s influence on the developing brain 3) providing practical examples on how to solve a problematic situation, how to analyse alcohol advertisement, and how to handle peer pressure, 4) improved feedback form and meeting summaries, 5) distributing leaflets and brochures. Figure 17 describes the logic behind the programme’s expected impact on children’s alcohol use. While the main focus of the programme is to target parents’ attitudes that are expected to, directly and indirectly, have an impact on children’s alcohol use, other possible influential pathways support achieving the outcome, e.g. increasing parents’ knowledge and skills that relate to general and alcohol-specific authoritative parenting. Additionally, communication and collaboration between parents can have an influence on alcohol-related norms at home that, in turn, can reduce parental alcohol supply and thus lead to delayed and reduced alcohol use among children. Although the programme does not focus on parents’ own alcohol use, it is possible that established alcohol-related norms and improved parenting skills might have an impact on it. The content and expected outcomes of the programme described in this chapter are summarised in a logic model in Figure 18. The expected outcomes of the programme are divided into three categories – immediate, short- and long-term impact. While the immediate impact relates more to increasing awareness and reshaping common misconceptions among parents, short-term impact focuses on authoritative parenting style. Long-term impact relates to the aim of the programme – delaying and reducing alcohol use among children by maintaining parents' restrictive attitudes. Ideally, the programme would be a part of a systematic and sustainable approach that contributes to alcohol use prevention among children. 101 Parents’ intervention Figure 17. The expected programme’s impact pathways to delayed and reduced alcohol use among children Parents’ attitudes towards children’s alcohol use Parents’ substance use and parenting related knowledge Alcohol use related norms at home Communication and collaboration between parents General and alcohol-specific authoritative parenting skills - Establishing rules - Monitoring - Solving problems - Communication - Teaching refusal skills to children - Tackling environmental use related factors, e.g. media, availability Parental alcohol supply Parents’ alcohol use Children’s perception of parents’ attitudes Children’s alcohol use related attitudes Children’s alcohol use Delayed initiation among children Reduced use among children The dashed line represents a factor or a pathway that was not specifically targeted by the programme’s activities. 102 Figure 18. The logic model of the Effekt programme and expected outcomes 103 3.7 The results from the internal audit – feedback from the parents In general, participants, the majority of whom were women, were willing to provide feedback – the answer rate was above 85% in all meetings (Table 6). Table 6. Parents who participated in the meetings and gave feedback Meeting I Meeting II Meeting III Meeting IV Meeting V Meeting VI Parents 536 342 453 292 323 159 Feedback - 349 (102.0%)* 389 (85.9%) 280 (95.9%) 284 (87.9%) 167 (105.0%)* Sex (%) Female - 81.4 85.9 85.0 90.1 86.2 Male - 17.8 13.9 14.3 9.9 13.2 *As teachers gave feedback in some classes, the number is higher compared to the participants. The participants were asked about previous meeting attendance – the rate varied between 60– 80% over the years, typically being lower at the spring meetings (Table 7). Although it was assumed that all parents received the summaries of the meetings, only 60% of parents attending the second meeting said that they had received the summary of the previous meeting. Over the three years, the rate of receiving the summary increased, reaching 87% at the final meeting. In addition to summaries, parents received newsletters – approximately three-quarters of participants reported receiving the material, and more than 90% reading it. At the end of the second and third meeting parents were asked if the previous meeting and/or newsletter(s) had been useful when talking to the child about alcohol – almost 80% agreed that they had. In the following two meetings, the question asked included only the usefulness of newsletters and rates were significantly lower, around 25%. This could indicate that high rates reported earlier were related to meetings, and most parents did not find newsletters very useful. When asked about the meetings’ content that was planned to share with family members, around 50% of feedback providers answered the question (the only exception was the sixth meeting when almost two-thirds of participants replied). Alcohol-related topics (e.g. harmfulness, consequences) and parents’ role in prevention (e.g. not offering alcohol, own use, expressing restrictive attitudes, talking about alcohol – how and what) were the common topics on a rolling basis that were planned to share with others (mostly children and spouses). Five additional topics that were mentioned more frequently but related to specific meetings were: parenting styles and problem-solving approach (Meeting IV), alcohol advertisements 104 and alcohol’s influence on the brain (Meeting V), and peer pressure (Meeting VI). While several parents planned to talk about everything they could recall about the topics, only a few specifically pointed out that they had no plans to share the content of the meetings (although not answering the question at all might have implied the same). Parents who already had similar viewpoints as presented in the meetings felt more confident presenting them to family members. Table 7. Parents’ feedback on attending the meetings and receiving materials, meetings II–VI Meeting II (%) Meeting III (%) Meeting IV (%) Meeting V d (%) Meeting VI e (%) Attended previous meeting* 59.3 a 65.7 e 82.1 i 66.4 k 83.2 m Received previous meeting’s summary 59.6 a 65.7 e 80.0 i 71.3 k 87.4 m Received newsletter(s) 79.9 b 73.8 f 76.8 i 75.7 l - Read newsletter(s)** 94.6 c 95.5 g - - - Previous meeting/ newsletter(s) was/were useful*** 77.2 d 77.8 h - - - Newsletter(s) was/were helpful**** - - 25.1 j 24.2 j - *Only if it took place **Only if the newsletter(s) was/were received ***Only if the previous meeting was attended and/or the newsletter(s) was/were received ****Only if the newsletter(s) was/were received; an=337, bn=349, cn=279, dn=307, en=321, fn=389, gn=287, hn=316, in=280, jn=215, kn=268, ln=284, mn=167 Participants were asked to assess the facilitators based on four criteria – credibility, engagement, intelligibility, and capability of answering questions. The general impression of the facilitators was very positive throughout the programme – more than 90% of the participants found the facilitators credible and intelligible, over 85% found the facilitators enthusiastic and capable of answering the questions (Table 8). Table 8. Parents’ feedback on facilitators, meetings II–VI Meeting II (%; n=349) Meeting III (%; n=389) Meeting IV (%; n=280) Meeting V (%; n=284) Meeting VI (%; n=167) Credibility 93.1 92.3 93.9 92.6 95.8 Enthusiasm 85.7 94.1 96.1 94.7 98.2 Intelligibility 94.3 96.9 95.0 94.4 97.6 Answered questions 87.4 91.8 93.6 92.3 91.0 Additional materials were distributed during the fourth and fifth meeting. The participants found it useful both times, 261 out of 280 participants (93.2%) in the fourth meeting and 258 out of 284 participants in the fifth meeting (90.8%). Analysing a case (fourth meeting) and 105 alcohol advertisement (fifth meeting) also received positive feedback from participants – 97.1% and 94.0% of the parents found it useful, respectively. At the end of the sixth meeting, participants (n=167) were asked to evaluate the effectiveness of the agreements made between parents. The awareness of agreements was high – 92.2% and 94.6% of the participants agreed the aim and the content of the agreements were clear and understandable, respectively. While 94.0% of the participants had tried to follow the agreements themselves, 67.1% found the agreements lacking the effect, because other parents were not following them. Despite that, 92.8% of participants hoped that parents would continue making agreements in the future. 3.8 Discussion In 2012, a three-year adaptation and delivery process of a universal parent-oriented prevention programme Effekt was initiated in Estonia by NIHD. The programme aimed to prevent and reduce alcohol use among 11–13-year-old adolescents, by targeting their parents’ attitudes towards children’s alcohol use. The programme was revised and included new content on parenting and risk behaviours, included newsletters (and leaflets) with new content and more communication with parents (e.g. use of active learning methods). New topics on parenting aimed to expand parents’ knowledge of authoritative parenting and give a better understanding of the interconnectedness of topics and how the topics can support parental attitudes. To prepare new topics and materials for upcoming meetings, additional meetings with facilitators and external experts were held in between the meeting waves. In addition, to support the facilitators, demo trainings and reflection events were carried out twice a year. Lastly, the project team systematically collected feedback from parents as part of an internal programme audit. When a programme is modified, it is essential to assess the intervention fidelity, as the adaptation can result in a lower fidelity, which in turn can have an impact on the programme’s outcomes (Dane & Schneider, 1998; Durlak & DuPre, 2008; Elliott & Mihalic, 2004). This issue becomes particularly acute when the delivery of the programme enters the real-world conditions and out of the researchers reach (Dusenbury et al., 2003). Applying the criteria presented by Dane and Schneider (1998) (see Chapter 1.7) to assess the fidelity of the current programme suggests that the level of fidelity could be considered as above moderate. 106 First, despite the adaptation, the programme was delivered as intended (implying high adherence) and the core components from the original programme developed in Sweden were kept unchanged, i.e. main messages focusing on parental attitudes, alcohol supply and communication, the format of having six meetings and parents receiving six newsletters, and making agreements between parents. Second, the meeting attendance was systematically recorded, giving an objective way to assess the delivery of the programme. While the average number of meetings that took place was close to the maximum, it varied greatly across classes. Despite that, more than two-thirds of the classes received at least five meetings, indicating above-average dosage. Additionally, parents were invited to provide feedback on the meetings and delivery of the materials, making it possible to assess the programme’s content and related activities on an ongoing basis. For example, more than 70% of the parents had received the newsletters. Dusenbury and colleagues (2003) suggest that there should be no issue with the delivery of the programme when it is done by paid personnel and led by researchers. While the criteria were met, the low attendance rates imply that there are more factors that affect the delivery of the programme. Third, to ensure consistency and quality of the content presented at the meetings, a training manual was developed for each meeting, and regular and intensive training sessions with the facilitators preceded all six meetings. Regarding the adaptation, it was systematic and evidence-based, and the delivery followed the format developed in Sweden. Several experts in related fields were involved to ensure that the new content covered topical issues and was age-appropriate from the child’s perspective. Fourth, the level of responsiveness was high among parents who attended the meetings, as they were asked to give feedback regarding the programme’s content and delivery. Based on the feedback, parents had positive attitudes towards the facilitators, found additional materials and active learning methods based tasks useful, and approximately half of the participants showed interest to share the topics discussed in the meetings with their family members. Fifth, the programme consisted of distinctive components that can be distinguished from each other, i.e. meetings (including distinctive topics), agreements, newsletters, summaries. While parents were asked to give feedback on all the components, the current assessment was not rigorous enough to give a strong indication of the effectiveness of each component. 107 Despite the suggested above moderate fidelity, the adaptation and delivery had some limitations. First, as the programme was modified on an ongoing basis and the attention was on one meeting at a time, parents and facilitators were not aware of the topics presented in the subsequent meetings. This kind of approach was found to be most appropriate, as changing the content was not initially planned, and it was possible to be more user-centred. With more time, the adaptation process could have been more formal and fuller. Second, the first contact with the teachers was via e-mail and may have lacked an “individual touch” (i.e. meeting face-to-face), which might have led to an increase in resistance towards participation. The confirmation of participation came from the school principals and thus, it was not known how much the teachers were involved in the decision process. 3.9 Conclusions The Effekt programme, initially developed in Sweden, was the first universal parent-oriented alcohol use prevention programme delivered in Estonia. Covering topics on risk behaviour and parenting aimed to support the change in parental attitudes and therefore delay adolescents’ alcohol use onset and reduce alcohol use among current users. The programme was based on the findings from the research literature, and it covered relevant age-appropriate topics. The programme was adapted to the local context, and new topics were introduced due to its repetitive content. The adaptation process was systematic, user-centred and included experts in the field. The feedback collected from the parents showed that in general, parents were satisfied with the content and activities of the programme, although parents’ participation stayed relatively low throughout the programme. In parallel with the programme, two studies were conducted. Chapter 4 provides a detailed description of the cRCT that involved both students and parents from intervention and control schools, and Chapter 5 and Chapter 6 give an overview of the qualitative study that addressed the views of programme’s facilitators and intervention schools’ parents and teachers on the programme and its delivery. 108 Chapter 4. Results from a universal parent-oriented alcohol use prevention programme Effekt in Estonia 109 This chapter has been published in the Drug and Alcohol Dependence journal. Chapters 4.2.3.4 and 4.3.4.1 were not part of the published article and have been included afterwards. 4.1 Introduction Alcohol use is high in Europe – 66% of the population aged 15 and older have consumed alcohol in the past 12 months, and 17% are heavy episodic drinkers (World Health Organization, 2014). Use is typically initiated in adolescence, when attitudes and behaviours develop, and is often associated with increased autonomy and proving oneself to others (Brown et al., 2008; Schulenberg & Maggs, 2002). According to the 2013/14 HBSC study (Inchley et al., 2016), 27% of 15-year-old students have ever consumed alcohol, and 8% have been drunk by the age of 13. Alcohol use prevalence among adolescents in Estonia is considerably higher than in most other European countries. Nearly one in two (49%) 15-year- olds have consumed alcohol, and 19% have been drunk by the age of 13 (Aasvee & Rahno, 2015). Initiation at an early age is related to several negative outcomes, e.g. development of health problems, injuries, early sexual behaviour and delinquent behaviour (Newbury-Birch et al., 2009). In addition, alcohol has a serious negative impact on brain development (Bava & Tapert, 2010; Brown, Tapert, Granholm, & Delis, 2000) which continues up until the mid- twenties (Giedd et al., 1999; Mills et al., 2014). Primary socialization theory (Oetting & Donnermeyer, 1998) postulates that parents are one of the main sources for children when learning norms, values and behaviours. Children tend to imitate their parents to receive recognition and to be perceived more like adults (Kohlberg, 1984). Kindergarten children who role play adults are more likely to buy alcohol and cigarettes if their parents drink alcohol or smoke (Dalton et al., 2005). This suggests that children who see drinking and smoking at home might be more prone to trying it out themselves. However, not only witnessing parents’ drinking influences drinking behaviour (Rossow et al., 2016; Yap et al., 2017); several other parental factors are related, including the provision of alcohol, attitudes, the quality of parent-child relationship, parenting style, monitoring, support and involvement (Čablová et al., 2014; Kaynak et al., 2014; Sharmin et al., 2017a; Yap et al., 2017). Targeting parents and related factors in programmes to prevent and reduce adolescents’ alcohol use has shown positive lasting results (Bo et al., 2018; Foxcroft & Tsertsvadze, 2011a; Smit et al., 2008), in comparison to student-oriented programmes that in general have not shown evidence of their effectiveness (Foxcroft & Tsertsvadze, 2011c; Jones et al., 2007). Favourable aspects covered in effective parent- 110 focused interventions include rule-setting, monitoring and parent-child communication (Kuntsche & Kuntsche, 2016). This article focuses on the parent-oriented programme Effekt (formerly known as ÖPP), which was developed in Sweden at the end of 1990s (Koutakis, 2011; Koutakis et al., 2008). The main objective is to delay and reduce adolescents’ alcohol use by maintaining parental restrictive attitudes towards adolescents’ alcohol use over time. The programme has so far been evaluated in Sweden (Bodin & Strandberg, 2011; Koutakis et al., 2008; Özdemir & Koutakis, 2016; Strandberg & Bodin, 2011) and the Netherlands (Koning et al., 2009; Koning et al., 2010b, 2011, 2013; Koning et al., 2012a; Verdurmen, Koning, Vollebergh, Van den Eijnden, & Engels, 2014), resulting in equivocal findings on adolescents’ alcohol use. The developers of the programme found it effective in reducing the frequency of drunkenness (d=0.35) (Koutakis et al., 2008) and onset of monthly drunkenness, mediated by parental attitudes (Özdemir & Koutakis, 2016). However, in a much larger evaluation of the programme in a different Swedish sample, no evidence on the intervention effect was found on delaying use (OR=0.99, 95% CI=0.61‒1.60) or reducing drunkenness (OR=1.07, 95% CI=0.79‒1.44) (Bodin & Strandberg, 2011). In an evaluation in the Netherlands, where the number of meetings was reduced from the original six to two, only a combined intervention targeting both parents and students directly had a positive effect on delaying heavy weekly alcohol use (Koning et al., 2009, 2010b, 2011, 2013; Verdurmen et al., 2014). Interventions targeting parents and students separately showed no conclusive evidence on the intervention effect on adolescents’ alcohol use. The idea to carry out an alcohol prevention programme in Estonia emerged in 2011 after several schools approached NIHD to request a systematic and sustainable solution to prevent and reduce adolescents’ alcohol use. A systematic review (Foxcroft & Tsertsvadze, 2011a) published the same year indicated that family-based prevention programmes had shown promising results. If effective, a parent-focused intervention could be potentially a lower cost intervention than a combined parent and adolescent focused intervention. Therefore, it was important to identify if a parent only intervention would be effective before rolling out a programme across the entire country. Criteria applied when selecting the programme to deliver were: 1) low long-term costs, 2) administratively easy to deliver, 3) time-efficient (from the perspective of parents and teachers), 4) promising results on reducing and delaying alcohol use. Effekt programme met the criteria most closely and was initiated in 2012. 111 Throughout the delivery process, the content of the programme was adjusted by extending the topics on alcohol use and parenting and by increasing the interaction between parents and facilitators. The aim of the present study was to evaluate the effectiveness of the Effekt programme, modified for the cultural context in Estonia. Specifically, it was examined whether allocation to the intervention had an effect on adolescents’ alcohol use, alcohol supply by parents, and parental attitudes. 4.2 Methods A matched-pair cRCT was conducted among adolescents and their parents in 2012‒2015. The trial was approved by the Tallinn Medical Research Ethics Committee (KK2710, 19.04.12). 4.2.1 Recruitment, allocation and participants In May 2012 all Estonian speaking schools in the NHPS that had at least seven grades received an electronic invitation to participate in the trial (schools for children with special needs were excluded). Out of 138 schools that met the criteria, 68 (49.3%) agreed to participate. All parallel classes in fifth grade were included. To allocate schools to groups, pairs and triplets were compiled, based on schools’ and classes’ (5th grades) size and spatial proximity (Figure 19). An online program “Research Randomizer“ (Urbaniak & Plous, 2013) was used to randomly allocate school(s) from each pair/triplet to intervention or control group. Immediately after randomisation two control schools withdrew. Due to this change, out of 66 remaining schools, 34 received the intervention and 32 schools were control schools. All parents received a consent form to confirm adolescents’ (n=2246) participation in the trial. Out of 2246 parents (one parent per household), 35.5% did not give their consent and 18.4% did not send the form back (Figure 20). The baseline assessment (T1) was carried out in September‒October 2012, the first follow-up at 18 months (T2) and the second follow- up at 30 months (T3). Students completed self-report questionnaires during one school lesson. Each student received a sealed envelope with a prepaid envelope and a parent’s questionnaire inside to take home. Unique sequence numbers were used to link parent’s and adolescent’s data. 112 Figure 19. Description of the allocation and randomisation process in the trial. 113 Figure 20. Consolidated Standards of Reporting Trials 2010 flow diagram. 114 The final number of students and parents participating at T1 was 985 and 790, respectively (43.9% and 35.2% of the whole sample) (Table 9). All participants who completed questionnaires at T1 were invited to participate at T3, irrespective of their participation at T2. Table 9. Adolescents’ and parents’ participation rates at the intervention and control schools at T1, T2 and T3 Intervention (n=480) Control (n=505) P-value Students T1 480 (100.0%) 505 (100.0%) 0.10 T2 429 (89.4%) 455 (90.1%) 0.71 T3 372 (77.5%) 415 (82.2%) 0.07 Parents T1 384 (80.0%) 406 (80.4%) 0.88 T2 252 (52.5%) 295 (58.4%) 0.06 T3 202 (42.1%) 267 (52.9%) ≤0.001 4.2.2 Intervention The universal parent-oriented alcohol prevention programme targeted parents, whose children were 11–13 years old (grades 5–7). Six meetings, two meetings per year (autumn and spring) were held at all schools by qualified facilitators, who underwent intensive training throughout the programme. The mean number of meetings in the 34 intervention schools (60 classes) was 4.78 (SD=1.30). Out of 60 classes, 22 received all six meetings, 19 received five, seven received four, 10 received three, one received two and one did not receive any meetings. The first meeting wave had the highest parents’ attendance throughout the programme – 47.1% (Table 10). The participation rate dropped during the second meeting wave in spring 2013 – less than a third of the parents attended the meetings. A similar trend occurred on the second and third year – higher attendance in the autumn and lower attendance in the spring. The participation rate reached its lowest point in the final meeting – 159 parents (14.0%) from 31 classes (51.0%) in 20 schools (59.0%). Table 10. Number of participants (schools, classes and parents) and parents’ participation rates in the meetings Participating schools (n=34; %) Participating classes (n=61; %) Participating parents (n=1138; %) Meeting I 32 (94.1%) 56 (91.8%) 536 (47.1%) Meeting II 29 (85.2%) 47 (77.0%) 342 (30.1%) Meeting III 33 (97.1%) 56 (91.8%) 453 (39.8%) Meeting IV 30 (88.2%) 47 (77.0%) 292 (25.7%) Meeting V 31 (91.2%) 52 (85.2%) 323 (28.4%) Meeting VI 20 (58.8%) 31 (50.8%) 159 (14.0%) 115 After each meeting, teachers received a summary by e-mail and forwarded it to all parents in the class, irrespective of their participation in the meeting (prerequisite was that the meeting had taken place). Twice a year parents also received two-page newsletters. The objective of the meetings and newsletters was to increase parents’ knowledge and awareness of children- related alcohol topics and parenting skills (Table 11). Three main messages were repeated in all the meetings and newsletters: 1) talk to your child (general communication, including alcohol); 2) do not offer alcohol to your child; 3) express clearly your restrictive attitudes towards children’s alcohol use. In addition, parents were encouraged to make agreements with other parents in the class to support children’s development; agreements were included in the meetings’ summaries. Table 11. Topics covered in the meetings and/or newsletters Meeting and newsletter Meeting only Newsletter only Autumn 2012 Introduction of the programme Parents’ role in prevention (main messages) Alcohol use consequences among youth Parental alcohol supply Statistics on alcohol use Programme’s results in Sweden Reasons why adolescents drink Spring 2013 Parents’ role in prevention Alcohol use consequences among youth Communication (e.g. how and what to talk about alcohol) Introductory exercise “Which factors influence adolescents’ alcohol use the most, and what can parents do about that?” Programme’s results in Sweden Energy drinks Autumn 2013 Parents’ role in prevention Importance of recreational activities Introductory exercise “Do you agree with the following statements on alcohol supply?” Statistics on alcohol-related accidents Programme’s results in Sweden Cannabis and alcohol Snus Is adolescents’ alcohol use inevitable? Spring 2014 Parents’ role in prevention Parenting styles Initial results from the Estonian study Demonstration and practice of solving a problematic situation Reasons why adolescents drink Child’s self-esteem and assertiveness E-cigarettes Autumn 2014 Parents’ role in prevention Alcohol advertisements Alcohol and the brain Analysis of an alcohol advertisement How to talk to the child about alcohol-related risks? Spring 2015 Parents’ role in prevention Peer pressure and group norms How to handle peer pressure – roleplay with parents How to act, when the parent sees another child behaving delinquently? How to plan the 9 th grade graduation party? 116 The feedback from parents and facilitators after the first two meetings implied that the repetition of the content – as done in the original programme – created reluctance among parents to participate in the following meetings. The main messages and making agreements were kept the same as in the original programme, but the content was modified after the second meeting (e.g. additional topics, more emphasis on discussion, roleplay). Involving a team of experts (e.g. family therapists, psychologists, educational scientists, teachers and public health experts) ensured the topics covered in the programme were age-appropriate. 4.2.3 Measures Identical questionnaires were used for all adolescents. Parents’ questionnaires at intervention schools had minor differences (i.e. additional questions related to the programme) from those in control schools. All sociodemographic and socioeconomic characteristics are described in Table 15. 4.2.3.1 Primary outcome The primary outcome was adolescents’ alcohol use initiation indicator – “Have you ever tried an alcoholic beverage (more than a sip)? Yes/no”. While several studies have considered alcohol initiation as having even a sip of alcohol, the project team decided to follow the practice of the international HBSC Study (Aasvee & Minossenko, 2011), by using one sip as a cut-off point. In that way, children who accidentally have drunk alcohol (e.g. orange juice vs alcoholic drink) would be excluded. Alcohol use can be initiated only once and having more than one sip might refer to a conscious choice to drink alcohol at that specific occasion, i.e. to initiate use. One of the aims of the programme was to delay the first encounter, especially the one where parents offer a bit (sips) of alcohol to children. 4.2.3.2 Secondary outcomes Secondary outcomes included: 1) adolescents’ past year alcohol use – “How frequently have you consumed the following alcoholic beverages (beer, wine, strong alcohol, light alcoholic beverages and cocktails) in the last 12 months? Never/seldom/every month/week/day”. This measure was dichotomised (irrespective of beverage type) into have not consumed versus have consumed alcohol in the past year; 2) adolescents’ lifetime drunkenness – “Have you ever consumed so much alcohol that you got drunk? No, never/ yes, once/2‒3 times/4‒10 times/more than 10 times” (Currie et al., 2012). This measure was dichotomised into never 117 versus at least once. Both measures were dichotomised due to very low rates of monthly and more frequent alcohol use and being drunk more than once. 4.2.3.3 Intermediate outcomes The main intermediate outcome was parental attitudes towards adolescents’ alcohol use, which focused on the child’s age: “At what age do you feel adolescents could try an alcoholic drink for the first time (at least one sip)?”. The item was dichotomised into below 18 (lenient) versus 18 and over (restrictive). Previously developed measure by Koutakis and colleagues (2008) was used as an additional measure to assess parental attitudes. The parents were asked to choose one of the following four options that describe their attitudes the best: “A child my son or daughter’s age is way too young to try alcoholic beverages. I think it is obvious that adolescents under 18 years are not allowed to drink alcoholic beverages”; “For me, it is unacceptable that my child would drink alcoholic beverages outside the home. Although it has happened that my child has tasted wine or beer, when we have drunk it (during dinner)”; “A child, my son or daughter’s age, is mature enough to be responsible for his/her own actions. If he/she wants to drink alcoholic beverages, then he/she will do it, despite what parents think or say. It has happened that I have offered my child an alcoholic beverage, so it would not seem so exciting to him/her. This way I have control over what my child drinks and therefore it reduces the risk that my child would obtain illegal alcohol or drugs”; “It is natural for a child my son or daughter’s age to be curious about trying alcoholic beverages. I believe that my child drinks in a responsible way”. The measure was dichotomised into restrictive (the first statement) and lenient (the latter three statements that all indicated alcohol use by children). Additional intermediate outcomes included: 1) parental alcohol supply – “From where have you usually acquired alcoholic beverages?”. Two options – parent(s) gave to try, and parent(s) allow(s) alcohol use – were combined and dichotomised (yes/no); 2) adolescents’ perception of parental attitudes – “How do your parents feel about adolescents your age consuming alcohol? Bad/neutral/tolerant/I do not know”. This measure was dichotomised into at least one parent has restrictive (“bad”) attitudes versus neither parent has restrictive attitudes. 118 4.2.3.4 Programme-related measures At the final assessment, parents in the intervention group were asked: 1) how many programme-related meetings they or their family members attended (with an option up to five meetings), 2) how many other meetings that addressed alcohol prevention topic they attended (parents in the control group were asked the same question), 3) how many programme-related newsletters they received (with an option up to five newsletters), and 4) to rate the usefulness of the programme’s meetings and newsletters for themselves and other parents in the class on a 5-point scale, where 1 meant “Not useful at all“ and 5 “Very useful“. All parents irrespective of the intervention condition were asked if they found it important (“Yes/No”) to raise parents’ awareness of adolescents’ alcohol use prevention. Those who chose “Yes” were asked an additional question – “When should the aforementioned awareness-raising targeting parents be started? When children are younger than 11 years/between 11–13-years/14–15-years/16–17-years”. 4.2.4 Sample size and power analysis As part of the study design, the emphasis was placed on identifying schools with a low likelihood of dropping out; therefore, schools from the NHPS were included. The project team estimated that compared to non-NHPS schools, the NHPS schools are more likely to participate in the programme for three years and also to support and motivate parents. However, this was a somewhat limited pool of schools, and the evaluation was dependent on schools opting into the trial, which reduced control over the sample size. Therefore, a formal sample size calculation was not undertaken. A similar approach was reported by Streimann and colleagues (2017) as it is difficult to include a large number of schools in trials in Estonia. 4.2.5 Data analysis Statistical analysis was carried out using Stata version 14.2 for Windows (StataCorp, 2015). Pearson’s Chi-squared test and Fisher’s exact test were used to assess the relationship between baseline characteristics and non-participation at T2 and T3, and between intervention condition and programme-related measures. As the sample consisted of matched pairs, Pearson correlation analysis was performed to assess the need to take into account the design in the following analysis. The correlation between members of pairs regarding alcohol 119 use initiation at T1 was very weak; therefore the pairs were broken, and the analysis performed was unmatched (Diehr, Martin, Koepsell, & Cheadle, 1995; Donner, Taljaard, & Klar, 2007). Two-level logistic regression was performed to account for school-level clustering when estimating how intervention condition predicted adolescents’ alcohol use, parental alcohol supply and parental attitudes at T2 and T3, and how parental attitudes (both self-reported and perceived) predicted adolescents’ alcohol use at T2 and T3. The alcohol use initiation and lifetime drunkenness models included only adolescents who had not initiated specific behaviour at T1. All models were adjusted to account for background characteristics at the exact follow-up and random effect for school. In addition, all models except alcohol use initiation and lifetime drunkenness were adjusted to account for the baseline outcome measure. The number needed to treat was calculated as the inverse of the risk difference (Cook & Sackett, 1995) if the intervention condition predicted a statistically significant change in outcomes. Two-level logistic regression analysis was also carried out to assess the dose-response relationship between the number of meetings and outcome measures at the intervention schools. Due to the low number of participants at the final assessment, it was decided not to use the programme-related measures in the multilevel modelling analyses. As previous studies have shown that parental attitudes predict children’s alcohol use (Ryan et al., 2010; Yap et al., 2017), two-level logistic regression was conducted to assess if parental attitudes at T2 and perception of parental attitudes at T3 predict children’s alcohol use at T3. Different time-points (i.e. T2 for self-reported attitudes, T3 for perceived attitudes) were used in the analysis, to take into account the time lag aspect (Sutton, 2004), as the change in parental attitudes might not instantly have an effect on children’s behaviour. Additional analysis was carried out to assess if parental attitude related measures might act as mediators in the relationship between the intervention condition and the primary outcome. Thereafter multilevel generalized structural equation modelling (StataCorp, 2013) was performed with indicators that predicted alcohol use initiation at T3. Bayes Factor (West, 2016) was computed for the primary outcome using an online calculator (Dienes, n.d.). Half-normal distribution, with mode set to 0 (indicating no effect), one-tailed and standard deviation equal to the expected effect size [(OR=0.71, 95% CI=0.54‒0.94; obtained from a meta-analysis by Smit and colleagues (2008)] was used for prediction. Additionally, Phi coefficient was calculated to assess the relationship between parental attitude related measures. 120 “Logical” imputation was used on alcohol use initiation and lifetime drunkenness to treat inconsistencies and replace missing values based on longitudinal data (Table 12). This approach is suggested to give a more accurate view of alcohol use behaviour, as the longer the time between the behaviour and reporting about it, the less accurate it might be (Spoth, Shin, Guyll, Redmond, & Azevedo, 2006). Table 12. Logical imputation of missing values and inconsistencies of adolescents’ alcohol use initiation and drunkennessa Alcohol use initiation Lifetime drunkenness Inconsistencies Alcohol use initiation Lifetime drunkenness Logical imputation N N N N –/0/1 → 0/0/1 –/0/0 → 0/0/0 0/–/0 → 0/0/0 –/1/– → –/1/1 0/1/– → 0/1/1 1/–/– → 1/1/1 1/–/1 → 1/1/1 1/1/– → 1/1/1 –/0/– → 0/0/– –/–/0 → 0/0/0 2 1 20 1 31 21 16 49 0 0 1 1 37 1 20 7 7 8 3 1 1/0/0 → 0/0/0 1/1/0 → 1/1/1 0/1/0 → 0/1/– 1/0/1 → 1/–/– 1/0/– → 1/–/– 1/–/0 → 1/–/– 28 20 37 30 12 6 2 2 8 2 0 1 a0 – no, 1 – yes This approach was not used with past year alcohol use and parental alcohol supply, as the answers from one wave were not logically dependent on the previous one(s). Table 13 presents the number of students who had not reported alcohol use (i.e. initiation) but indicated alcohol use in the past year and/or alcohol use as parent(s) had provided alcohol. Table 14 shows the distribution of missing data among variables; no information regarding students’ gender is presented as the gender of all students was known from the baseline study. The missing data on outcomes were handled under four scenarios (Bodin & Strandberg, 2011): 1) completers only, 2) missing data treated as negative (no) – the best-case scenario, 3) missing data treated as positive (yes) – worst-case scenario, 4) multiple imputation. The latter was performed via fully conditional specification for multilevel data under missing at random assumption in Blimp 1.0 (Enders, Keller, & Levy, 2017; Keller & Enders, 2017). To reduce the sampling variability, 100 datasets were created, an imputed data set was created after every 1000th computational cycle, and 1000 iterations were performed before saving the first set. The seed value was set at 90291. Additional options incorporated in the imputation were: 1) the Gibbs option – used when some clusters might have few or no cases, 2) common 121 residual variance for all clusters, 3) cluster means as additional predictors. All variables with missing data were included in the multiple imputation process. Table 13. Inconsistencies between alcohol use initiationa and past year useb and parental alcohol supplyc at T1, T2 and T3 Past year use Past year use and parental alcohol supply Parental alcohol supply T1 41 29 68 T2 8 5 9 T3 4 4 6 aAdolescents reported no alcohol use initiation. bAdolescents reported alcohol use in the past year. cAdolescents reported alcohol supply by parents. 4.3 Results 4.3.1 Baseline characteristics The students sample (n=985) consisted of 51.1% of girls at T1 (Table 15). Most of the participants were 11-years old (88.5%) and lived in urban areas (86.5%). The parents' sample (n=790) consisted of 90.9% of females at T1. Almost half (44.9%) of the participating parents had higher education (i.e. a degree from the university) at T1. 4.3.2 Attrition analysis At T2 884 students (89.7% of baseline) and 547 parents (55.5% of baseline) completed study measures (Table 9). At T3, the rates were 79.9% and 47.6%, respectively. Non-participants in the intervention group differed (P≤0.05) from completers at both follow-ups by family structure, alcohol use initiation, past year use and lifetime drunkenness and at T3 by alcohol supply by parents (Table 16). Non-participants in the control group differed from completers at T2 by alcohol use initiation, lifetime drunkenness and parents’ perception of family wealth, and at T3 by living area and alcohol use initiation. 122 Table 14. Missing values among primary, secondary and intermediate outcomes and background characteristics at T1, T2 and T3 Intervention (n=480) Control (n=505) P-value Students Age, n (%) T1 T2 T3 Family income, n (%) T1 T2 T3 Nuclear family, n (%) T1 T2 T3 1 (0.2) 51 (10.6) 109 (22.7) 6 (1.3) 58 (12.1) 113 (23.5) 7 (1.5) 52 (10.8) 90 (17.8) 2 (0.4) 50 (9.9) 91 (18.0) 5 (1.0) 53 (10.5) 93 (18.4) 2 (0.4) 52 (10.5) 109 (22.7) 0.52 a 0.71 0.07 0.47 a 0.43 0.05 0.08 a 0.86 0.06 Alcohol use initiation, n (%) T1 T2 T3 Alcohol use in the past 12 months, n (%) T1 T2 T3 Alcohol supply by parents, n (%) T1 T2 T3 Lifetime drunkenness, n (%) T1 T2 T3 Perception of parents’attitudes, n (%) T1 T2 T3 2 (0.4) 41 (8.5) 94 (19.6) 5 (1.0) 58 (12.1) 138 (28.8) 2 (0.4) 24 (5.0) 60 (12.5) 0 (0.0) 41 (8.5) 125 (26.0) 24 (5.0) 61 (12.7) 115 (24.0) 1 (0.2) 46 (9.1) 79 (15.6) 1 (0.2) 65 (12.9) 114 (22.6) 1 (0.2) 21 (4.2) 45 (8.9) 1 (0.2) 41 (8.1) 107 (21.2) 19 (3.8) 755 (10.9) 92 (18.2) 0.48 a 0.75 0.10 0.10 a 0.71 0.03 0.48 a 0.53 0.07 0.51 a 0.81 0.07 0.34 0.38 0.03 Parents Gender, n (%) T1 T2 T3 Education, n (%) T1 T2 T3 Family income, n (%) T1 T2 T3 Parents’ attitudes towards adolescents’ alcohol use, n (%) T1 T2 T3 100 (20.8) 232 (48.3) 282 (58.8) 101 (21.0) 232 (48.3) 282 (58.8) 101 (21.0) 229 (47.7) 279 (58.1) 128 (26.7) 243 (50.6) 292 (60.8) 104 (20.6) 211 (41.8) 241 (47.7) 104 (20.6) 137 (27.1) 242 (47.9) 100 (19.8) 212 (42.0) 239 (47.3) 127 (25.2) 232 (45.9) 257 (50.9) 0.93 0.04 <0.01 0.86 0.04 <0.01 0.63 0.07 <0.01 0.59 0.14 <0.01 Chi-square/Fisher ’s one-tailed exacta test was performed with each variable to compare intervention and control schools at T1, T2 and T3. 123 Table 15. Baseline characteristics of the sample School level Intervention (n=34) Control (n=32) Area a , n (%) Tallinn Urban Rural 8 (23.5) 14 (41.2) 12 (35.3) 4 (12.5) 15 (46.9) 13 (40.6) School size, mean (SD) 411.79 (265.54) 423.91(401.56) Cluster size, mean (SD) 14.12 (8.29) 15.78 (15.77) Individual level Intervention (n=480) Control (n=505) Students Girls, n (%) 243 (50.6) 260 (51.5) Age, n (%) 10yo 11yo 12yo Missing Area, n (%) Capital Urban Rural 24 (5.0) 430 (89.6) 25 (5.2) 1 (0.2) 135 (28.1) 269 (56.0) 76 (15.8) 29 (5.7) 442 (87.5) 32 (6.3) 2 (0.4) 137 (27.1) 311 (61.6) 57 (11.3) Family income Good Average Bad Missing Nuclear family, n (%) Nuclear Non-nuclear Missing 318 (66.3) 147 (30.6) 9 (1.9) 6 (1.3) 319 (66.5) 154 (32.1) 7 (1.5) 345 (68.3) 148 (29.3) 7 (1.4) 5 (1.0) 346 (68.5) 157 (31.1) 2 (0.4) Parents Participation, n (%) Gender Female Male Missing Family income, n (%) Good Average Bad Missing Parents’ education, n (%) At least one has higher education No parent has higher education Missing 384 (80.0) 349 (72.7) 31 (6.5) 100 (20.8) 90 (18.8) 256 (53.3) 33 (6.9) 101 (21.0) 154 (32.1) 225 (46.9) 101 (21.0) 406 (80.4) 369 (70.1) 32 (6.3) 14 (20.6) 101 (20.0) 266 (52.7) 38 (7.5) 100 (19.8) 201 (39.8) 200 (39.6) 104 (20.6) aCapital – Tallinn, urban – 1000+ inhabitants, rural – less than 1000 inhabitants. SD=standard deviation 124 Table 16. Baseline rates of primary, secondary and intermediate outcomes and background characteristics among follow-up completers and non-participants by intervention condition T2 T3 Completers, n (%) Non- participants, n (%) P-value Completers, n (%) Non- participants, n (%) P-value Students Girls, n (%) Intervention 221 (51.5) 22 (43.1) 0.26 196 (52.7) 47 (43.5) 0.09 Control 238 (52.3) 22 (44.0) 0.27 221 (53.3) 39 (43.3) 0.09 Area, n (%) Capital Intervention 119 (27.7) 16 (31.4) 0.59 99 (26.6) 36 (33.3) 0.17 Control 125 (27.5) 12 (24.0) 0.60 106 (25.5) 31 (34.4) 0.09 Urban Intervention 244 (56.9) 25 (49.0) 0.29 217 (58.3) 52 (48.2) 0.06 Control 275 (60.4) 36 (72.0) 0.11 256 (61.7) 55 (61.1) 0.92 Rural Intervention 66 (15.4) 10 (19.6) 0.44 56 (15.1) 20 (18.5) 0.39 Control 55 (12.1) 2 (4.0) 0.10 53 (12.8) 4 (4.4) 0.03 a Family income, n (%) Good Intervention 288 (67.1) 30 (58.8) 0.24 245 (65.9) 73 (67.6) 0.74 Control 315 (69.2) 30 (60.0) 0.18 283 (68.2) 62 (68.9) 0.90 Average Intervention 126 (29.4) 21 (41.2) 0.08 116 (31.2) 31 (28.7) 0.62 Control 130 (28.6) 18 (36.0) 0.27 123 (29.6) 25 (27.8) 0.73 Bad Intervention 9 (2.1) 0 (0.0) 0.61 a 6 (1.6) 3 (2.8) 0.43 a Control 5 (1.1) 2 (4.0) 0.15 a 4 (1.0) 3 (3.3) 0.11 a Nuclear family, n (%) Intervention 292 (68.1) 27 (52.9) 0.04 a 258 (69.4) 61 (56.5) 0.03 a Control 315 (69.2) 31 (62.0) 0.46 a 291 (70.1) 55 (61.1) 0.18 a Alcohol use initiation, n (%) Intervention 105 (24.5) 21 (41.2) 0.04 a 82 (22.0) 44 (40.7) <0.001 a Control 133 (29.2) 21 (42.0) 0.01 a 117 (28.2) 37 (41.1) 0.04 a Alcohol use in the past year, n (%) Intervention 87 (20.3) 17 (33.3) 0.05 a 72 (19.4) 32 (29.6) 0.03 a Control 126 (27.7) 17 (34.0) 0.47 a 113 (27.2) 30 (33.3) 0.38 a Lifetime drunkenness, n (%) Intervention 10 (2.33) 6 (11.8) <0.01 a 7 (1.9) 9 (8.33) <0.01 a Control 13 (2.86) 6 (12.00) <0.01 a 13 (3.1) 6 (6.67) 0.28 a Supply by parents, n (%) Intervention 117 (27.3) 19 (37.3) 0.32 a 95 (25.5) 41 (38.0) 0.01 a Control 129 (28.4) 15 (30.0) 0.10 a 117 (28.2) 27 (30.0) 0.83 a Perception of parents’ restrictive attitudes, n (%) Intervention 309 (72.0) 33 (64.7) 0.47 a 271 (72.9) 71 (65.7) 0.32 a Control 333 (73.2) 35 (70.0) 0.77 a 304 (73.3) 64 (71.1) 0.62 a 125 T2 T3 Completers, n (%) Non- participants, n (%) P-value Completers, n (%) Non- participants, n (%) P-value Parents Female, n (%) Intervention 318 (74.13) 31 (60.8) 0.07 a 267 (71.8) 82 (75.9) 0.43 a Control 331 (72.8) 38 (76.0) 0.51 a 305 (73.5) 64 (71.1) 0.24 a Family income, n (%) Good Intervention 81 (18.9) 9 (17.7) 1.00 a 71 (19.1) 19 (17.6) 0.73 Control 96 (21.1) 5 (10.0) 0.06 a 88 (21.2) 13 (14.4) 0.15 Average Intervention 235 (54.8) 21 (41.2) 0.07 197 (53.0) 59 (54.6) 0.76 Control 240 (52.8) 26 (52.0) 0.92 220 (53.0) 46 (51.1) 0.74 Bad Intervention 31 (7.2) 2 (3.9) 0.56 a 26 (7.0) 7 (6.5) 1.00 a Control 30 (6.6) 8 (16.0) 0.04 a 28 (6.8) 10 (11.1) 0.16 At least one parent has higher education, n (%) Intervention 140 (32.6) 14 (27.5) 0.08 121 (32.5) 33 (30.6) 0.87 Control 186 (40.9) 15 (30.0) 0.30 165 (39.8) 36 (40.0) 0.36 Parents’ restrictive attitudes b , n (%) Intervention 107 (24.9) 7 (13.7) 0.06 a 92 (24.7) 22 (20.4) 0.47 Control 88 (19.3) 7 (14.0) 0.67 a 84 (20.2) 11 (12.2) 0.21 Chi-square/Fisher ’s exacta test was performed to compare completers and non-participants. Statistically significant differences (P=<.05) are presented in bold. bParents’ attitudes towards adolescents’ alcohol use were measured by asking parents the age, when adolescents could try alcoholic drink for the first time (at least one sip), age below 18 implies lenient attitude and age at 18 or over implies restrictive attitude. SD=standard deviation. 4.3.3 Primary outcome Around 30% of students had initiated alcohol use at the baseline (Table 17), and over time the proportion of new initiators increased (Figure 21), but there was no conclusive evidence on the intervention effect on alcohol use initiation (T2 – OR=1.21, 95% CI=0.81‒1.81; T3 – OR=0.87, 95% CI=0.59‒1.29) (Table 18). Adjusting the model for background characteristics did not change the results (Table 19). The Bayes Factor at T2 was 1.11 and at T3 0.88, the former indicating anecdotal evidence for the experimental hypothesis and the latter for the null hypothesis (Beard, Dienes, Muirhead, & West, 2016). While the results at two separate time points are in a different direction, the results close to one are in general considered of little evidence. 126 Table 17. Baseline rates of primary, secondary and intermediate outcomes by intervention condition Intervention (n=480) Control (n=505) Students Alcohol use initiation, n (%) Yes No Missing Lifetime drunkenness, n (%) Yes No Missing Alcohol use in the past 12 months, n (%) Yes No Missing Alcohol supply by parents, n (%) Yes No Missing Perception of parental attitudes towards adolescents’ alcohol use, n (%) At least one with restrictive attitude None with restrictive attitude Missing 126 (26.3) 352 (73.3) 2 (0.4) 16 (3.3) 464 (96.7) 0 (0.0) 104 (21.7) 371 (77.3) 5 (1.0) 136 (28.3) 342 (71.3) 2 (0.4) 342 (71.3) 114 (23.8) 24 (5.0) 154 (30.5) 350 (69.3) 1 (0.2) 19 (3.8) 485 (96.0) 1 (0.2) 143 (28.3) 361 (71.5) 1 (0.2) 144 (28.5) 360 (71.3) 1 (0.2) 368 (72.9) 118 (23.4) 19 (3.8) Parents Attitudes towards adolescents’ alcohol usea, n (%) Lenient Restrictive Missing 238 (49.6) 114 (23.8) 128 (26.7) 283 (56.0) 95 (18.8) 127 (25.2) aParental attitudes towards adolescents’ alcohol use were measured by asking parents the age, when adolescents could try an alcoholic drink for the first time (at least one sip), age below 18 implies lenient attitudes and age at 18 or over implies restrictive attitudes. 127 *Statistically significant difference (p≤0.05) within groups between T1 and T2; **Statistically significant difference (p≤0.05) within groups between T1 and T3, and T2 and T3; Nintervention=352/333/280, Ncontrol=350/ 331/297 Figure 21. Alcohol use initiation rates at baseline (T1), first (T2) and second (T3) follow-ups at the intervention and control schools among adolescents who had not initiated alcohol use at T1. 4.3.4 Secondary and intermediate outcomes There were no statistically significant differences between groups regarding their past year alcohol use, lifetime drunkenness, parental alcohol supply and perception of parental attitudes (Table 18, Table 19, Figure 22, Figure 23, Figure 24, Figure 25). *Statistically significant difference (p≤0.05) within groups between T1 and T2; **Statistically significant difference (p≤0.05) within groups between T1 and T3, and T2 and T3; Nintervention=475/422/342, Ncontrol=504/440/391 Figure 22. Alcohol use in the past 12 months at baseline (T1), first (T2) and second (T3) follow-ups among adolescents at the intervention and control schools. 30.9* 42.1** 0 26.6* 44.1** 0 10 20 30 40 50 60 Baseline First follow-up Second follow-up % Intervention Control 21.9 35.6* 43.6** 28.4 37.3* 47.8** 0 10 20 30 40 50 Baseline First follow-up Second follow-up % Intervention Control 128 Table 18. Unadjusted two-level logistic regression models on the effect of intervention condition on all outcomes at T2 and T3 T2 T3 N ICC OR 95% CI P NNT N ICC OR 95% CI P NNT Alcohol use initiation a Completers 664 0.04 1.21 0.81–1.81 0.36 – 577 0.03 0.87 0.59–1.29 0.49 – Best case 705 0.04 1.23 0.83–1.84 0.30 – 705 0.04 0.82 0.56–1.19 0.30 – Worst case 702 0.04 1.17 0.80–1.71 0.41 – 702 0.02 1.02 0.73–1.42 0.92 – MI 702 0.04 1.20 0.80–1.81 0.37 – 702 0.04 0.88 0.59–1.31 0.52 – Lifetime drunkenness b Completers 871 0.11 0.97 0.51–1.84 0.92 – 721 0.05 1.10 0.69–1.75 0.69 – Best case 950 0.11 0.93 0.49–1.77 0.84 – 950 0.07 1.00 0.63–1.61 0.99 – Worst case 949 0.07 1.00 0.63–1.58 1.00 – 949 0.01 1.25 0.94–1.66 0.13 – MI 949 0.13 0.94 0.50–1.79 0.85 – 949 0.08 1.06 0.66–1.72 0.80 – Past year use Completers 862 0.06 0.94 0.64–1.39 0.77 – 733 0.06 0.80 0.54–1.20 0.29 – Best case 985 0.05 0.96 0.67–1.37 0.81 – 985 0.06 0.74 0.51–1.08 0.12 – Worst case 985 0.06 0.92 0.65–1.32 0.67 – 985 0.02 0.98 0.72–1.33 0.90 – MI 985 0.05 0.96 0.67–1.39 0.84 – 985 0.07 0.86 0.57–1.29 0.46 – Alcohol supply Completers 940 0.06 1.02 0.70–1.49 0.92 – 880 0.01 0.76 0.55–1.03 0.07 – Best case 985 0.05 1.01 0.70–1.45 0.97 – 985 0.01 0.73 0.54–0.98 0.04 17 Worst case 985 0.06 1.05 0.72–1.52 0.80 – 985 0.02 0.86 0.64–1.16 0.33 – MI 985 0.06 1.02 0.70–1.49 0.91 – 985 0.01 0.72 0.53–1.00 0.05 16 Parental attitudes Completers 510 0.05 2.05 1.32–3.17 <0.01 6 436 0 1.92 1.31–2.83 <0.01 6 Best case 985 0.03 1.72 1.22–2.42 <0.01 9 985 0 1.97 1.47–2.64 <0.01 8 Worst case 985 0.01 1.50 1.08–2.09 0.02 15 985 0.02 1.20 0.84–1.73 0.31 – MI 985 0.11 1.81 1.09–3.02 0.02 7 985 0.09 2.03 1.22–3.38 <0.01 6 Perception of attitudes Completers 869 0 1.25 0.86–1.82 0.24 – 778 0 1.11 0.77–1.59 0.59 – Best case 985 0 1.27 0.88–1.85 0.20 – 985 0 1.20 0.84–1.70 0.31 – Worst case 985 0.01 1.05 0.77–1.43 0.77 – 985 0.02 0.87 0.64–1.17 0.34 – MI 985 0 1.24 0.84–1.82 0.28 – 985 0 1.05 0.72–1.55 0.79 – aAmong adolescents who had not initiated alcohol use at T1. bAmong adolescents who had not been drunk at T1. MI=multiple imputation. 129 Table 19. Adjusted two-level logistic regression models on the effect of intervention condition on all outcomes at T2 and T3 T2 T3 N ICC OR 95% CI P NNT N ICC OR 95% CI P NNT Alcohol use initiation a, b Completers 633 0.03 1.21 0.81–1.81 0.34 – 541 0.02 0.83 0.55–1.24 0.36 – Best case 637 0.03 1.24 0.83–1.85 0.30 – 580 0.02 0.78 0.53–1.16 0.22 – Worst case 635 0.03 1.21 0.81–1.81 0.34 – 578 0.01 0.90 0.62–1.31 0.58 – MI 702 0.04 1.19 0.79–1.79 0.42 – 702 0.02 0.87 0.59–1.30 0.50 – Lifetime drunkenness b, c Completers 822 0.05 1.13 0.62–2.07 0.69 – 697 0.06 1.00 0.60–1.65 0.99 – Best case 848 0.05 1.09 0.60–1.97 0.79 – 758 0.05 0.97 0.59–1.57 0.89 – Worst case 847 0.02 1.09 0.67–1.77 0.72 – 757 0.04 1.15 0.79–1.66 0.47 – MI 949 0.07 1.04 0.57–1.91 0.90 – 949 0.06 1.05 0.65–1.70 0.85 – Past year use d Completers 826 0.05 1.10 0.74–1.65 0.64 – 722 0.05 0.96 0.63–1.45 0.84 – Best case 870 0.04 1.10 0.76–1.59 0.62 – 778 0.05 0.93 0.63–1.37 0.71 – Worst case 870 0.04 1.06 0.72–1.55 0.77 – 778 0.05 0.95 0.64–1.40 0.79 – MI 985 0.05 1.11 0.75–1.66 0.59 – 985 0.07 0.96 0.63–1.48 0.86 – Alcohol supply d Completers 868 0.05 1.04 0.70–1.55 0.85 – 769 0 0.78 0.57–1.06 0.11 – Best case 870 0.05 1.06 0.71–1.58 0.77 – 778 0 0.77 0.57–1.04 0.09 – Worst case 870 0.05 1.05 0.71–1.56 0.80 – 778 0 0.82 0.60–1.11 0.20 – MI e 985 0.05 1.04 0.71–1.53 0.83 – 985 0 0.74 0.54–1.00 0.05 16 Parental attitudes f Completers 463 0.02 2.04 1.27–3.28 <0.01 6 402 0 1.93 1.22–3.05 <0.01 6 Best case 528 0.04 1.86 1.17–2.97 0.01 6 455 0 1.79 1.19–2.71 <0.01 7 Worst case g 530 0.04 2.23 1.38–3.61 <0.01 6 455 0 1.84 1.19–2.83 <0.01 7 MI e 985 0.14 1.83 1.00–3.36 0.05 7 985 0.11 2.04 1.12–3.71 0.02 6 Perception of attitudes d Completers 820 0 1.21 0.80–1.83 0.36 – 737 0 1.07 0.73–1.58 0.73 – Best case 870 0 1.27 0.86–1.89 0.23 – 778 0 1.17 0.80–1.71 0.41 – Worst case 870 0 1.15 0.79–1.68 0.45 – 778 0 1.09 0.76–1.58 0.63 – MI e 985 0 1.25 0.83–1.87 0.28 – 985 0 1.08 0.73–1.60 0.71 – aAmong adolescents who had not initiated alcohol use at T1.bAdjusted for student’s gender, age, family structure, perception of family wealth, area of living at the exact follow-up. cAmong adolescents who had not been drunk at T1. dAdjusted for student’s gender, age, family structure, perception of family wealth and area of living at the exact follow-up, and baseline measurement of the outcome. ePerception of family wealth is excluded from the T3 model as the model did not converge. fAdjusted for student’s gender, age, family structure, perception of family wealth, area of living, parent’s gender, education and perception of family wealth at the exact follow-up, and baseline measurement of the outcome. gParents’ perception of family wealth is excluded from the T2 model as the model did not converge. MI=multiple imputation. 130 *Statistically significant difference (p≤0.05) within groups between T1 and T2; **Statistically significant difference (p≤0.05) within groups between T1 and T3, and T2 and T3; Nintervention=464/425/341, Ncontrol=485/446/380 Figure 23. Lifetime drunkenness rates at baseline (T1), first (T2) and second (T3) follow-ups at the intervention and control schools among adolescents who had not been drunk at T1. *Statistically significant difference (p≤0.05) within groups between T1 and T3; Nintervention=478/456/420, Ncontrol=504/484/460 Figure 24. Parental alcohol supply reported by adolescents at baseline (T1), first (T2) and second (T3) follow-ups at the intervention and control schools. 7.5* 19.1** 0 7.4* 17.1** 0 4 8 12 16 20 Baseline First follow-up Second follow-up % Intervention Control 28.5 32.2 29.1 28.6 30.8 34.8* 0 5 10 15 20 25 30 35 40 Baseline First follow-up Second follow-up % Intervention Control 131 *Statistically significant difference (p≤0.05) within groups between T1 and T2; **Statistically significant difference (p≤0.05) within groups between T1 and T3; Nintervention=456/419/365, Ncontrol=486/450/413 Figure 25. Perception of restrictive parental attitudes towards adolescents’ alcohol use among adolescents at baseline (T1), first (T2) and second (T3) follow-ups at the intervention and control schools. The proportion of parents with restrictive attitudes towards alcohol use increased over time in both groups (Figure 26), and the intervention condition predicted restrictive attitudes at T2 (OR=2.05, 95% CI=1.32‒3.17) and T3 (OR=1.92, 95% CI=1.31‒2.83) (Table 18). *Statistically significant difference (p≤0.05) within groups between T1 and T2; **Statistically significant difference (p≤0.05) within groups between T1 and T3; Nintervention=352/237/188, Ncontrol=378/213/159 Figure 26. Restrictive parental attitudes towards adolescents’ alcohol use among parents at baseline (T1), first (T2) and second (T3) follow-ups at the intervention and control schools. 75.0 86.4* 81.9** 75.7 83.6* 80.4 0 20 40 60 80 100 Baseline First follow-up Second follow-up % Intervention Control 32.4 48.5* 50.5** 25.1 31.9 34.7** 0 10 20 30 40 50 60 Baseline First follow-up Second follow-up % Intervention Control 132 4.3.4.1 The relationship between (perceived) parental attitudes and children’s alcohol use The results from the two-level logistic regression analysis showed that parental restrictive attitudes at T2 predicted lower alcohol use initiation rates at T3, but only among control group students (OR=0.48, 95% CI=0.23–1.00 vs OR=0.59, 95% CI=0.29–1.18). Parental attitudes at T2 and children’s drunkenness at T3 seemed to be unrelated in both groups. Restrictive attitudes at T2 predicted lower past year alcohol use at T3 in both the intervention (OR=0.42, 95% CI=0.22–0.78) and control (OR=0.44, 95% CI=0.24–0.81) groups. Regarding perceived parental attitudes and children’s alcohol use outcomes at T3, no evidence was found for a relationship between attitudes and alcohol use initiation and drunkenness but perceived restrictive attitudes predicted lower past year use among control group students (OR=0.54, 95% CI=0.31–0.93 vs OR=0.70, 95% CI=0.40–1.22). To see if parental self-reported and perceived attitudes mediated the relationship between intervention condition and children’s alcohol use initiation, the relationship between attitudes and alcohol use initiation was assessed. The results show that only the self-reported attitudes predicted alcohol use initiation at T3 (Table 20). Table 20. Unadjusted and adjusted two-level logistic regression models on the effect of attitude related measures on adolescents’ alcohol use initiation at T3 among students who had not initiated alcohol at T1 Unadjusted Adjusted N ICC OR 95% CI P N ICC OR 95% CI P Parental attitudes a,b 314 0.04 0.55 0.33–0.91 0.02 294 0 0.42 0.24–0.72 <0.01 Perception of parental attitudes c,d 542 0.05 0.64 0.40–1.02 0.06 536 0.02 0.70 0.43–1.14 0.15 aAdjusted to student’s gender, age, family structure, area of living at T3, and parent’s gender, education, perception of family wealth at T2. bMeasured at T2. cAdjusted to student’s gender, age, family structure, area of living at T3. dMeasured at T3. The results of the mediation analysis indicate that intervention condition did not have a direct effect on adolescents’ alcohol use initiation at T3 (β=0.04, SE=0.28, P=0.87), but did have an effect on parental attitudes at T2 (β=0.78, SE=0.24, P≤0.01) and the latter had an effect on adolescents’ alcohol use at T3 (β=-0.60, SE=0.26, P≤0.05). Alcohol use initiation may also have been indirectly affected by parental attitudes, but this was not significant at the 5% level (β=-0.47, SE=0.25, P=0.06); the total effect was not statistically significant (β=-0.43, SE=0.35, P=0.22). 133 4.3.4.2 Sensitivity analyses of parental attitude measures While the results based on the original attitude measure (Koutakis et al., 2008) where parents had to choose between four options that describe their attitude the best, showed that more than 70% of parents at baseline had restrictive attitudes towards adolescents’ alcohol use, the results based on the new measure, which asked the parents the age when children could try alcohol for the first time, were more than twice as low (Figure 27). This suggests that the measure used in previous research to measure parental attitudes might overestimate the restrictiveness of attitudes to some degree. The relationship between these two attitude variables was statistically significant at baseline (χ2=69.68, P<0.001, φ=0.31), at the first follow-up (χ2=70.17, P<0.001, φ=0.38) and at the second follow-up (χ2=94.46, P<0.001, φ=0.47). The values of the phi coefficients indicate a moderate positive correlation between the two measures. Attitude 1 refers to the measure used by Koutakis et al., 2008; Attitude 2 refers to the new measure introduced in the current trial. *Statistically significant difference (p≤0.05) within groups between T1 and T2; **Statistically significant difference (p≤0.05) within groups between T1 and T3; Nintervention=374/198/169, Ncontrol=399/450/413 Figure 27. Restrictive parental attitudes towards adolescents’ alcohol use among parents at baseline (T1), first (T2) and second (T3) follow-ups at the intervention and control schools. Separate multilevel analyses were performed with the original attitude measure to see if the intervention had an effect on the maintenance of restrictive attitudes among parents over time (Table 21). There were significant differences between the intervention and the control schools at the first and the second follow-up, and the obtained odds of having restrictive attitude were higher compared to the results with the new measure presented in Table 18 and Table 19. 32.4 48.5* 50.5** 74.6 84.9* 80.7 25.1 31.9 34.7** 72.7 68.0 63.5** 0 20 40 60 80 100 Baseline First follow-up Second follow-up % Intervention Control Attitude 1 Attitude 2 134 Table 21. Unadjusted and adjusted two-level logistic regression models on the effect of intervention condition on the original parental attitudes measure at the first and second follow-ups First follow-up Second follow-up N ICC OR 95% CI P N ICC OR 95% CI P Parental attitudes a Completers 542 0 2.63 1.72–4.02 <0.001 463 0 2.40 1.56–3.70 <0.001 Best case 985 0 2.63 1.76–3.92 <0.001 985 0 2.77 1.86–4.12 <0.001 Worst case 985 0.01 1.22 0.92–1.62 0.16 985 0.03 0.96 0.69–1.34 0.81 MI 985 0.06 2.18 1.33–3.58 <0.01 985 0.08 2.14 1.26–3.63 <0.01 Parental attitudes b Completers 517 0 3.31 1.97–5.56 <0.001 443 0 2.66 1.55–4.53 <0.001 Best case 528 0 3.03 1.84–4.99 <0.001 455 0 2.72 1.62–4.58 <0.001 Worst case 528 0 3.30 1.99–5.46 <0.001 455 0 2.45 1.47–4.09 <0.01 MI 985 0.08 2.70 1.50–4.87 <0.01 985 0.11 2.47 1.29–4.73 <0.01 aUnadjusted bAdjusted to student’s gender, age, family structure, perception of family wealth, area of living, parent’s gender, education, perception of family wealth, baseline measurement of the outcome. MI=multiple imputation 135 4.3.5 Dose-response relationship The number of meetings was not significantly associated with outcome measures at T2 and T3 (Table 22). Table 22. Unadjusted two-level logistic regression models on the effect of the number of meetings on primary, secondary and intermediate outcomes at T2 and T3 at the intervention schools T2 T3 N ICC OR 95% CI P N ICC OR 95% CI P Alcohol use initiation a Completers 333 0.02 0.94 0.61–1.47 0.80 280 0.02 1.04 0.77–1.40 0.82 Best case 354 0.02 0.95 0.61–1.48 0.84 354 0.04 1.06 0.79–1.41 0.70 Worst case 352 0.01 0.98 0.64–1.48 0.91 352 0 1.01 0.79–1.28 0.96 MI 352 0.19 0.96 0.61–1.50 0.86 348 0.01 1.00 0.70–1.41 0.98 Lifetime drunkenness c Completers 425 0.14 1.94 0.71–5.35 0.20 341 0.03 1.14 0.79–1.64 0.48 Best case 464 0.14 1.85 0.68–5.00 0.23 464 0.07 1.16 0.80–1.69 0.79 Worst case 464 0.08 1.64 0.86–3.13 0.13 464 0 1.01 0.82–1.26 0.91 MI 464 0.16 1.91 0.71–5.16 0.20 464 0.04 1.15 0.80–1.66 0.45 Past year use Completers 422 0.03 1.18 0.78–1.80 0.44 342 0 0.96 0.75–1.23 0.76 Best case 480 0.01 1.08 0.73–1.61 0.69 480 0.01 0.94 0.74–1.18 0.58 Worst case 480 0.04 1.29 0.86–1.94 0.22 480 0 1.00 0.81–1.24 0.97 MI 480 0.03 1.19 0.78–1.82 0.42 480 0 0.98 0.76–1.26 0.87 Alcohol supply Completers 456 0.06 1.13 0.71–1.78 0.61 420 0 0.79 0.62–1.01 0.06 Best case 480 0.05 1.10 0.70–1.73 0.67 480 0.01 0.82 0.65–1.04 0.10 Worst case 480 0.06 1.16 0.74–1.81 0.51 480 0 0.81 0.65–1.00 0.05 MI 480 0.07 1.14 0.72–1.81 0.59 480 0 0.81 0.63–1.04 0.10 Parental attitudes Completers 237 0.06 0.92 0.52–1.60 0.76 188 0 0.93 0.65–1.31 0.67 Best case 480 0.06 0.87 0.53–1.41 0.56 480 0.01 0.97 0.74–1.29 0.86 Worst case 480 0 0.95 0.63–1.43 0.80 480 0.01 0.94 0.73–1.23 0.67 MI 480 0.10 0.92 0.52–1.61 0.76 480 0.07 0.86 0.60–1.23 0.40 Perception of attitudes Completers 419 0.05 0.94 0.52–1.71 0.85 365 0.05 1.06 0.76–1.49 0.73 Best case 480 0.04 0.98 0.55–1.75 0.95 480 0.05 1.08 0.78–1.50 0.64 Worst case 480 0.03 0.79 0.50–1.25 0.32 480 0.02 0.98 0.77–1.24 0.87 MI 480 0.07 0.99 0.54–1.80 0.97 480 0.06 1.02 0.72–1.44 0.90 aAmong adolescents who had not initiated alcohol use at T1. bFour cases (zero meetings) were excluded as the model did not converge. cAmong adolescents who had not been drunk at T1. MI=multiple imputation. 136 Table 23. Adjusted two-level logistic regression models on the effect of the number of meetings on primary, secondary and intermediate outcomes at T2 and T3 at the intervention schools T2 T3 N ICC OR 95% CI P N ICC OR 95% CI P Alcohol use initiation a, b Completers 317 0.01 0.98 0.62–1.54 0.92 260 0.03 0.99 0.72–1.36 0.94 Best case 319 0.02 1.00 0.63–1.58 0.99 284 0.04 0.99 0.72–1.36 0.95 Worst case 317 0.01 0.98 0.62–1.54 0.92 283 0.01 1.00 0.74–1.35 1.00 MI e 352 0.02 0.95 0.60–1.51 0.84 352 0.01 1.03 0.77–1.38 0.83 Lifetime drunkenness b,c Completers 400 0.10 1.64 0.59–4.60 0.34 326 0.01 1.09 0.74–1.61 0.65 Best case 412 0.10 1.57 0.57–4.35 0.38 359 0.07 1.13 0.76–1.67 0.55 Worst case 412 0.04 1.75 0.75–4.05 0.19 359 0 0.99 0.74–1.33 0.96 MI f 464 0.01 1.81 0.93–3.55 0.08 464 0.01 1.14 0.79–1.62 0.49 Past year use d Completers 400 0 0.95 0.62–1.45 0.81 335 0 0.93 0.71–1.22 0.59 Best case 421 0 0.95 0.63–1.43 0.82 366 0 0.89 0.68–1.15 0.37 Worst case 421 0 0.96 0.63–1.46 0.86 366 0 0.94 0.73–1.21 0.62 MI f 480 0 0.95 0.62–1.46 0.80 480 0 0.94 0.73–1.22 0.65 Alcohol supply d Completers 419 0.06 1.08 0.66–1.74 0.77 359 0 0.84 0.65–1.09 0.19 Best case 421 0.06 1.10 0.67–1.78 0.71 366 0 0.84 0.64–1.08 0.18 Worst case 421 0.06 1.09 0.67–1.77 0.73 366 0 0.87 0.67–1.13 0.28 MI 480 0.06 1.06 0.66–1.70 0.82 480 0 0.82 0.64–1.04 0.11 Parental attitudes g Completers 211 0.07 0.67 0.32–1.42 0.29 172 0 0.88 0.57–1.36 0.57 Best case 242 0.05 0.95 0.51–1.79 0.88 195 0 0.89 0.60–1.32 0.56 Worst case 242 0.11 0.90 0.46–1.75 0.75 195 0 0.84 0.56–1.26 0.41 MI h 480 0.15 0.83 0.42–1.62 0.58 480 0.08 0.77 0.50–1.17 0.21 Perception of attitudes d Completers 391 0.05 1.07 0.53–2.16 0.85 344 0 1.08 0.79–1.49 0.63 Best case 421 0.03 1.00 0.53–1.91 1.00 366 0.03 1.07 0.76–1.51 0.68 Worst case 421 0.05 1.07 0.58–1.99 0.82 366 0.01 1.07 0.77–1.49 0.68 MI 480 0.08 1.08 0.54–2.17 0.83 480 0.07 1.01 0.69–1.46 0.97 aAmong adolescents who had not initiated alcohol use at T1.bAdjusted for student’s gender, age, family structure, perception of family wealth, area of living at the exact follow-up. cAmong adolescents who had not been drunk at T1. dAdjusted for student’s gender, age, family structure, perception of family wealth and area of living at the exact follow-up, and baseline measurement of the outcome. ePerception of family wealth is excluded from the T2 model as the model did not converge. fPerception of family wealth is excluded from the T2 model as the model did not converge. gAdjusted for student’s gender, age, family structure, perception of family wealth, area of living, parent’s gender, education and perception of family wealth at the exact follow-up, and baseline measurement of the outcome. hChildren’s perception of family wealth is excluded from the T2 model as the model did not converge. MI=multiple imputation 4.3.6 Programme-related measures Of 202 intervention group parents who participated in the study at the second follow-up that took place before the final meeting, 165 (81.7%) had attended at least one Effekt meeting, and 144 (71.3%) specified the number of meetings they attended. Of these 144 parents, 40 (27.8%) had attended one meeting, 46 (31.9%) two meetings (in two cases this was the 137 maximum number of meetings that took place at the school), 26 (18.1%) three meetings, 12 (8.3%) four meetings (in three cases this was the maximum number of meetings that took place at the school) and 20 (13.9%) all five meetings. Twenty four parents out of 202 reported that some other family member had attended at least one meeting, and out of 19 who specified the number of meetings, 15 said that other family member had attended the meeting once, three said the family member had attended two meetings, and in one case the other family member had attended four meetings. When asked about reading programme-related newsletters, 169 participants (83.7%) had read at least one newsletter, while 21 participants (10.4%) had not received any of the five newsletters sent out to parents; 12 parents did not answer the question. Out of 134 parents who specified the number of newsletters read, 34 (25.4%) had read one newsletter, 41 (30.6%) two newsletters, 27 (20.2%) three newsletters and 32 (23.9%) four newsletters. Parents’ perception of the usefulness of the meetings and the newsletters was relatively high, but higher for other parents in the class compared to themselves, M=3.8 and M=3.5 respectively. Out of 462 parents who shared their view on the importance of raising parents’ awareness of children’s alcohol use prevention, 406 (87.9%) found it important. There was a statistically significant difference between the intervention and the control group, as more parents in the intervention group (93.0% vs 83.9%; χ2=8.88, P<0.01) found raising parents’ awareness important. More than half of the parents (52.4%) who supported the approach, found that it should be started when children are 11–13-years-old. Almost 40% of parents thought it should be started earlier, and around 9% of parents found it more useful when children are between the ages of 14 and 17. Compared to the control group, there were slightly fewer parents in the intervention group than in the control group who supported the idea to raise awareness when children are older, 5.0% vs 11.5% (χ2=7.67, P<0.10) respectively. When asked about meetings or trainings that covered alcohol use prevention topic and took place around the same time the programme was carried out, nine parents (4.5%) in the intervention group and 74 parents (27.7%) in the control group said that they had attended such events. 138 4.4 Discussion 4.4.1 Summary of main findings The Effekt programme was the first universal parent-oriented alcohol prevention programme to be carried out in Estonia. Assessing the programme in the North-Eastern-European cultural context gives valuable input on its adaptability to other countries with high rates of adolescents’ alcohol use. It was expected that the programme would help to delay and reduce adolescents’ alcohol use by maintaining parental restrictive attitudes over time. However, while the intervention appeared to increase parents’ restrictive attitudes towards adolescents’ alcohol use, there was no conclusive evidence on the intervention effect on adolescents’ behaviour. Similar results regarding parental attitudes have been presented in all previous studies (Koning et al., 2010b; Koutakis et al., 2008; Özdemir & Koutakis, 2016; Strandberg & Bodin, 2011; Verdurmen et al., 2014), in addition, two, rather than six meetings have been suggested to be enough to see a change in attitudes (Strandberg & Bodin, 2011; Koning et al., 2009). Equivocal results have been presented on the programme’s effect on adolescents’ alcohol use (see Section 4.1). 4.4.2 Comparability with previous studies The Estonian findings are not directly comparable with other studies that have evaluated the Effekt programme as the content of the programme was modified. It was important to adjust the programme to the current situation in Estonia and to ensure that parents would attend the meetings, without changing the main messages and the format. Similarly, there were changes incorporated in the Dutch version by reducing the number of meetings from six to two (Koning et al., 2009). Additionally, the emphasis of the programme was mainly on prevention and therefore it was started among fifth graders (~11-years-old), while the Dutch version of the programme was started among adolescents with mean age of 12.6 (Koning et al., 2009) and the Swedish version among seventh graders (~13-years-old) (Koutakis et al., 2008). This resulted in choosing different outcome measures that addressed more of the lifetime alcohol use aspect (initiation of use and drunkenness) and supply, while previous studies focused more on heavy drinking and frequency of use. One study investigated the effect of the programme on lifetime drunkenness and concluded that there was no significant effect (Bodin & Strandberg, 2011). 139 As Koutakis and colleagues (2008) original attitude measure includes statements regarding restrictive attitudes that do not exclude each other (i.e. “a child my son or daughter’s age” and “adolescents under 18 years”), it was expected that parents may find it difficult to give a clear answer and generalise their attitude by choosing one or the other option. This could also result in overestimating restrictive parental attitudes, as parents who agree only with the statement that a 13-year-old should not drink alcohol are still considered restrictive even if they approve use in later age but before turning 18. Thus, it was decided to use a measure that asked the participants the age when children should try alcohol for the first time. While the findings were similar, the strength of the effect of the intervention condition on parental attitudes was smaller with the new measure. Besides the Effekt programme, there are other parent-oriented programmes that have addressed parental attitudes to reduce children’s alcohol use in Europe. For example, the programme “Strong and Clear” (“Stark och Klar” in Swedish) that was evaluated in six schools in Värmland, in Sweden, targeted the parents of 13–16-year-old students and focused on maintaining parents’ restrictive attitudes and postponing and reducing alcohol use among children (Pettersson et al., 2011). The programme was considered effective in maintaining parental restrictive attitudes, but also in delaying alcohol use onset and episodes of drunkenness. The evaluation of the Norwegian programme “Youth & Alcohol” (“Unge & Rus”) was carried out in 40 schools in Oslo and Akershus, in Norway, among 13–16-year-old students and their parents (Strøm et al., 2015). The programme aimed to delay and reduce alcohol use among children and while it showed positive effects in increasing students’ alcohol-related knowledge, there was no evidence on the intervention effect on alcohol use nor parents’ attitudes (Adolfsen et al., 2017; Strøm et al., 2015). Adolfsen and colleagues (2017) pointed out that not achieving a change in parents’ attitudes could have been due to originally high levels of restrictiveness. 4.4.3 Key considerations Although parental attitudes were influenced by the programme in Estonia, it is not clear why this did not translate into an effect on adolescents’ alcohol use. Reasons for this could include the programme starting too late – around 30% of the participants had already initiated alcohol use at T1; thus, they might have influenced other classmates. At the same time, evidence suggests that despite the increasing influence of peers, the role of parents does not decrease over time (Wood, Mitchell, Read, & Brand, 2004). Furthermore, due to low participation rate 140 in the meetings, many parents did not have direct contact with other parents and facilitators and thereby lacked behavioural practice (Michie, Atkins, & West, 2014), although all parents received the summaries and newsletters, irrespective of their participation. Another reason could be having parents as the main target group, as combining student- and parent-oriented programmes have shown more promising results (Newton et al., 2017; Van Ryzin et al., 2016). To predict future behaviour, attitudes have to be stable over time (Glasman & Albarracin, 2006), nevertheless it has been shown that when adolescents mature, parents become more lenient towards adolescents’ alcohol use (Glatz et al., 2012; Kelly et al., 2011; Özdemir & Koutakis, 2016; Prins et al., 2011; Zehe & Colder, 2014). However, the results from the current trial show that parents in the control group did not become more lenient over time (although the trend was present when a different attitude measure was used) but compared to the intervention group still had a significantly lower prevalence of restrictive attitudes. Another aspect that may explain why the programme showed no evidence in delaying and reducing children’s alcohol use is the that the parental attitudes predicted only lower past year use and not initiation (except in the control group) nor drunkenness. Also, if attitudes play a part in changing behaviour, there might be other factors involved (e.g. behavioural intentions, perceived behavioural control, subjective norms) (Ajzen & Fishbein, 2005). Despite having restrictive attitudes, parents might offer alcohol to children when they perceive that other parents are applying the same practice or when they reflect on their own childhood experiences (Jones, 2016). Additionally, adolescents’ alcohol use is a multifaceted behaviour, influenced by several factors in addition to the role of family, e.g. personal characteristics, environmental, social and cultural factors (Koning et al., 2009; Maggs & Staff, 2017; Velleman, Templeton, & Copello, 2005). Finally, the main focus of the programme was on alcohol, while several researchers (Bo et al., 2018; Foxcroft & Tsertsvadze, 2011a; Kuntsche & Kuntsche, 2016; Robertson et al., 2003; Stockings et al., 2016) have suggested that increasing awareness of substance use is very common, but prevention should be universal in its content and focus more on reducing risk factors and enhancing protective factors. 4.4.4 Limitations and strengths This study has some limitations. A limitation was using non-random sampling, as only schools from the NHPS, who were willing to participate, were included. Thus, the participants may not be representative of students and parents in Estonia but taking into consideration that these schools were motivated to participate, the results should rather 141 overestimate the outcome than the opposite. As the evaluation was dependent on schools opting into the trial, it reduced control over the sample size. Just before the trial started in autumn 2012, almost half of the network schools were willing to participate, resulting in ~2200 students. The high drop-out rate at the beginning of the trial (of all the adolescents and parents, only 44% and 35% participated at T1, respectively) was not expected, and at that point, it was not possible to include any new schools. Low participation rates among adolescents were mainly due to parents not giving consent for their children to participate in the study or not sending the form back at all. Instead of using the traditional active consent (parents’ signature required to confirm/refuse participation), an alternative (passive) consent approach (signature needed only to deny the participation) could have resulted in higher participation rates (Frissell et al., 2004). Parents might also have disliked the approach of using unique numbers that are linked to participants’ names. Adolescents whose home conditions (e.g. high parental alcohol use, violence) could have put them more at risk might have been excluded from the study. Also, it is possible that children whose parents attended meetings did not participate in the trial and vice versa. Although an a priori sample size calculation was not planned given the pragmatic nature of this trial, the original aim was to reduce the difference between HBSC average and Estonian rates in alcohol use initiation by 50%, which amounts to a 12% absolute difference (Currie et al., 2012). A retrospective calculation shows that assuming an ICC of 0.02 (based on the baseline model of alcohol use initiation) and α=0.05, a sample size of 985 adolescents (the achieved sample) in 33 clusters per arm gives 79% power to detect a 12% reduction in initiation rate from control to intervention condition. The Bayes factor calculations that incorporate a predicted value from a previously published meta-analysis on family-based RCTs (Smit et al., 2008), indicate that there was relatively low uncertainty about the programme not being effective, suggesting that study power was unlikely to be a limiting factor. In addition to low survey participation rates, similar trend of low participation were observed in the meetings and reported by parents at the final assessment. This can influence the dose-response relationship outcome as the dose reflects the number of meetings that took place (e.g. five meetings took place, but the participation rate was low). Another limitation was using adolescents’ self-reported alcohol use. While self-reports on initiation age have shown that children’s perception of the age at first use in longitudinal assessment tends to increase (Engels, Knibbe, & Drop, 1997), results from several studies on validity and reliability confirm that students can be trusted to accurately report alcohol use (Donovan et 142 al., 2004; Hibell et al., 2012; Molinaro, Siciliano, Curzio, Denoth, & Mariani, 2012; Wagenaar, Komro, McGovern, Williams, & Perry, 1993). Finally, parents’ own alcohol use was not measured, and this can have a significant influence on adolescents’ alcohol use (Rossow et al., 2016; Yap et al., 2017). Despite the limitations, the study had several strengths. First, follow-up rates among students were high, more than 80% at both follow-ups. Second, follow-up times were long, meaning that the time lag (Sutton, 2004) between parental attitudes and adolescents’ behaviour was sufficient to see any change in the behaviour. Third, using different approaches to take into account missing data showed that similar results to complete case analysis were obtained. Fourth, two measures were used to assess parental attitudes. The new measure was considered more robust by asking for the specific age when parents perceived children could try an alcoholic drink for the first time. Fifth, additional programme-related questions were asked from the parents at the final assessment, which showed that parents find it important to raise awareness on alcohol use prevention, but on the other hand, find such an approach more useful for others. 4.5 Conclusions The Effekt programme has received a rating of “Likely to be beneficial” in the registry of evidence-based prevention programmes, Xchange (European Monitoring Centre for Drugs and Drug Addiction, 2017) and a rating of “Promising intervention” in the Blueprints for Healthy Youth Development database (Blueprints for Healthy Youth Development, 2018). At the same time, the findings are contradictory, and this article in combination with others (Bodin & Strandberg, 2011; Koning et al., 2009) provides evidence that targeting parental attitudes might not be sufficient to delay and/or reduce adolescents’ alcohol use. It is important to understand how the programme works in different countries and cultural contexts, but also to allow the programme to be adjusted to the local situation. It has been suggested that parent-oriented programmes may be effective in preventing and reducing adolescents’ alcohol use, but this may depend on various factors, such as adolescents’ age, parents’ characteristics and intensity of the programme (Kuntsche & Kuntsche, 2016). Future research should focus on combining parent and adolescent programmes, starting the programme earlier, addressing more general protective factors, such as life skills and less alcohol-related awareness. Ensuring high participation rates is another crucial part of universal prevention programmes because reaching only the people who have the necessary 143 skills and knowledge is not enough to see a change. Additional attention should be directed to the qualitative assessment of the interventions to obtain a better understanding of potential barriers (e.g. low participation rates), but also components that work. 144 Chapter 5. Preventing children’s alcohol use by targeting parents: A qualitative study of facilitators delivering the Effekt programme 145 5.1 Introduction Transitions take place across the life-course and during these periods humans tend to be more vulnerable and more receptive to changes in behaviour (Catalano & Hawkins, 1996; Jindal- Snape, Cantali, MacGillivray, & Hannah, 2019). Physical, emotional and physiological changes occurring when children are transitioning from childhood to adolescence and from one school level to another may result in different risk behaviours (Catalano & Hawkins, 1996; Burdzovic Andreas & Jackson, 2015). The most prevalent risk behaviour among children is alcohol use – 27% of 15-year-old students are estimated to initiate alcohol use at age 13 or earlier, and 22% report having been drunk at least twice by the age of 15 (Inchley et al., 2016). Although the initiation rate has dropped significantly over time (46% in 2002), situations in countries differ. For example, Estonia is one of the few countries in Europe where initiation rates among girls did not change between 2002–2014 and the reduction over time among boys was below 10%, whereas the average across Europe was 20% (Inchley et al., 2018). While the context of use varies (Mair, Lipperman-Kreda, Gruenewald, Bersamin, & Grube, 2015), studies have shown that home and family have an essential role regarding children’s alcohol use (Ryan et al., 2010; Yap et al., 2017). A recent review shows an association between having parents with lenient attitudes towards children’s alcohol use and children’s alcohol use initiation, frequency and drunkenness (Tael-Öeren, Naughton, & Sutton, 2019a). Multiple programmes have been developed to target the home/family setting when tackling alcohol use among children. Although several reviews have indicated that these programmes have a small and lasting effect, especially when focusing on the aspect of psychosocial development and developing skills (Bo et al., 2018; Stockings et al., 2016; Van Ryzin et al., 2016), the latest findings by Gilligan and colleagues (2019) contradict this. The study included findings from 46 papers, where the assessment of universal (i.e. targeting general populations), selective (i.e. targeting at-risk populations) and indicated (i.e. targeting users) prevention programmes had been reported. The pooled estimates showed that there were no differences between groups at all prevention levels regarding alcohol preventing and reducing alcohol use (Gilligan et al., 2019). These findings emphasise the importance of a thorough evaluation to understand the reasons behind successes and failures. The evaluation of substance use prevention programmes is typically carried out quantitatively, but researchers have emphasised that collecting and analysing qualitative data 146 can contribute to enriching the findings (Allen et al., 2008; Hopson & Steiker, 2010; Steckler, McLeroy, Goodman, Bird, & McCormick, 1992; Strandberg, 2014). Qualitative data can add valuable insights into participants’ views, perceptions and experiences, and broaden the understanding of successes and barriers. In 2012–2015, a universal parent-based alcohol prevention programme Effekt was carried out and evaluated in Estonia by NIHD. The programme was initially developed in Sweden, and previous studies showed mixed findings on its effectiveness (Bodin & Strandberg, 2011; Koning et al., 2011; Koutakis et al., 2008). The results from the cRCT in Estonia implied that while the intervention was effective in maintaining and increasing parental restrictive attitudes, there was no conclusive evidence on its effect on children’s alcohol use (Tael- Öeren, Naughton, & Sutton, 2019b). In addition to the trial, a focus group with the programme’s facilitators (findings presented in the current chapter) and individual interviews with parents and teachers who participated in the programme (see Chapter 6) was carried out at the end of the programme. This chapter aims to get a better understanding of 1) facilitators’ views on the adaptation and delivery of the Effekt programme, 2) facilitators’ perception on the programme’s impact and attractiveness, and participation barriers. This is known to be the first time when Effekt programme has been evaluated qualitatively. 5.2 Methods 5.2.1 Epistemology and ontology The previous chapter where the findings from a quantitative study were presented is an example of positivist research, where the focus is on causal effect and generalisation of the results, and which offers more objective perspective, as it is relatively independent of the researcher (Crotty, 1998; Johnson & Onwuegbuzie, 2004). This chapter relies on the interpretivist paradigm, where understanding and interpretation of the phenomenon are of interest, and where the findings rely more on the researcher’s perception of the topic (Johnson & Onwuegbuzie, 2004; Yanow, 2014). If the content and delivery of the intervention rely on theory and evidence-based practice, then ideally, it should work. But, from an ontological perspective, social norms, people’s attitudes, experiences and interpretations of the world vary; thus, it provides a basis to imply that there are different versions of social realities (Guba, 1990). This leads to a situation where intervention’s effectiveness may exist in one case but vanish in another. From an 147 epistemological perspective, to get a deeper understanding of the underlying mechanisms, a qualitative view is an essential part of the research, offering a complementary input to quantitative findings (Johnson & Onwuegbuzie, 2004). 5.2.2 Study design This qualitative study, in addition to a cRCT (see Chapter 4), is part of a universal parent- oriented alcohol use prevention programme carried out in Estonia in 2012–2015 among 5–7th grade students and parents (see Chapter 3). A focus group study with a semi-structured interview schedule was carried out with the programme’s facilitators. This interview method was chosen as the participants were part of a pre-existing group with good dynamic and it was expected that the communication is more open-ended and facilitating the interaction between the group members brings out collective experience nuances that might be overlooked when conducting individual interviews (Finch, Lewis, & Turley, 2014). The study was approved by the Tallinn Medical Research Ethics Committee (KK 932, 12.02.15). 5.2.3 Study sample and recruitment The study sample consisted of ten facilitators who carried out the programme related activities. The project leader (M.TÖ) who was also one of the facilitators was deliberately left out from the study to reduce response bias among other participants. M.TÖ contacted all facilitators via e-mail and invited them to participate in the study, and eight out of nine agreed. The eight facilitators (seven women, one man; M=32.3 years, range 26–39) who participated in the focus group varied in terms of their background and training experience. Three facilitators had more than five years of training experience (equivalent to high experience), and five facilitators had less than five years of training experience (equivalent to low experience). Facilitators’ background covered public health, psychology, social policy, pedagogy, health promotion and social work. 5.2.4 Data collection and analysis The focus group was conducted in Estonian at NIHD, in February 2015 (before the final, sixth meeting with parents) and it included eight out of the ten facilitators. The focus group was moderated by a researcher who had vast experience (>20 years) in qualitative research (including the topic of alcohol) and who was not related to the programme. The focus group lasted approximately two hours, was audio-recorded and later transcribed by M.TÖ. All 148 facilitators gave their verbal consent to be interviewed and were paid an hourly rate of 10€ for their participation. The semi-structured interview schedule was developed by M.TÖ and the researcher who moderated the focus group, and it addressed the following topics: 1) reflection of own training experience, 2) views on the adaptation and delivery process of the programme, 3) perception of the programme’s impact, strengths and weaknesses, 4) suggestions regarding how the programme could be improved (e.g. format, content, length, communication with participants). Data analysis was conducted in two stages. At stage 1, the researcher who moderated the focus group documented moderator notes, analysed and summarised the data shortly after it took place. At stage 2, M.TÖ applied thematic analysis (Braun & Clarke, 2006) to the transcript to identify patterns in the data. The thematic analysis consists of six phases that enable the researcher to move back and forth between the phases when needed to capture the more coherent and in-depth interpretation of the data (Braun & Clarke, 2006). Phase 1 involved getting familiar with the data, i.e. reading and re-reading the transcript to get an initial sense of the data and noting down thoughts and ideas. In Phase 2, the initial coding framework was generated, and the data were organised. Although the interview schedule gave ideas on possible themes and codes, inductive and open coding was used, i.e. all text was coded with no codes predetermined. Next, the codes (n=24) were sorted into potential themes and sub-themes (Phase 3 – searching for themes). Phase 4 focused on reviewing the preliminary themes in two stages – reviewing code extracts, and then the data as a whole – to ensure that the codes collated under themes and sub-themes conveyed the meaning and reflected the data set. Phase 5 consisted of the final refinement of the themes – five codes were discarded as there was an overlap with other codes, resulting in 19 codes under three themes. All the work, as mentioned above, was done on paper, and no qualitative software program was used. Phase 5 was followed by a comparison of findings obtained by both researchers who had reached similar results, with minimal discrepancies that were resolved. The analysis concluded with a written report describing the themes that are supported by illustrative quotes, that is the final stage of the thematic analysis. As it would be possible to identify participants based on their age, sex and background, only the level of training experience has been used as an identifier next to the quotes. 149 5.3 Results The data were categorised into three main themes: 1) The perceived value of the programme, 2) Encountered challenges, 3) Long-term continuation of the programme. 5.3.1 The perceived value of the programme Parents’ perspective All facilitators said that they felt that the programme would lead to positive outcomes, but one key concern was about the proportion of parents and ultimately children it would reach due to low participation rates. Several topics addressed parents and created discussion, including alcohol’s influence on the brain development, ranking factors that have an impact on children’s alcohol use, using a practical problem-solving approach and talking about expressing emotions, using analogues and reflecting on own childhood experience. Taking part in the discussion showed that parents thought through the topics, and this helped to reinforce the programme’s messages. „I found it very gratifying, especially in the last year when there were new parents in the class, who had not heard anything about previous meetings, and we started from the main things, e.g. why you cannot offer alcohol to children, and it was a good opportunity to let other parents answer it, and they rolled their eyes and said that how they (new parents) cannot understand why it is not allowed. This was so awesome, at least somebody had received the message.” (Facilitator #4, low) The facilitators believed that the parents who attended meetings were more likely to retain restrictive attitudes towards children’s alcohol use, although it was pointed out that attitudes may not directly influence behaviour, and people’s enthusiasm may wane in the home environment. Personal value While there were several fears and challenges facilitators had to overcome, self-development was identified as a key factor in overcoming these. Positive aspects of self-development included increased knowledge on topics, improved communication skills, adapting someone 150 else’s material and presenting it as their own, and effectively coping in unfamiliar environments. While there were challenging situations with parents, the facilitators perceived the general tone from parents to be positive, and they valued the interaction highly. “It was such a nice experience. Sometimes you came back and had a tear in your eye, in a positive way, because the classes and parents were so nice, and the community was nice, and children brought you apples.” (Facilitator #5, low) Carrying out the meetings and talking about the topics made the facilitators feel that they were contributing to the greater good and changing the norm. 5.3.2 Encountered challenges Fears before the start of the meetings Depending on facilitators’ previous training experience, the fears varied. Those with minimal or no training experience, reported doubting their skills and knowledge, especially when confronted with questions from parents. Most facilitators had not carried out activities related to parent education, and as parents did not attend the meetings to hear about alcohol use prevention, but to listen about their children’s school life, facilitators were feeling anxious. In addition, as the facilitators had no information on parents’ professional background, they indicated that they feared the possibility that someone in the audience knew more about some specific topic, e.g. a doctor. When parents had different views on the topics, facilitators said they quickly had to adjust and adequately explain their point of view. The first meeting helped to alleviate the fears in most cases, but there were some exceptions where facilitators had the opposite feelings: “… I did not have any fears, but during my first meeting … I was shocked. … You have someone who knows something particular about the topic, and he had excellent knowledge, and I felt that my knowledge was a bit [weak] and I should have known more… it was such a shocking experience for me..” (Facilitator #1, high) Facilitators with more training experience found that resistance from parents would be normal in a situation where the target group is previously not made aware of the meeting’s content, and it should not frighten the facilitators. But as other facilitators were not used to educating people who did not willingly attend meetings, the view above was not shared by all. 151 Content related challenges Before it was decided to change the content of meetings, facilitators felt unsure about repeating the same content for each session, as expected according to the original programme. On the other hand, changing the content and including information on additional substances and parenting created another kind of challenge – facilitators felt that they could not manage the vast amount of information provided to them and some topics were perceived to be more challenging to grasp than others, e.g. alcohol and the developing brain. The facilitators had to go through the topics thoroughly and be prepared to answer any questions parents had, e.g. if somebody asks about tobacco use. It was important not just to be a messenger who delivers information, but to convey the content as you would have developed the presentation yourself. While the content was the same for all facilitators, the knowledge, attitudes and skills of facilitators varied. Thus, it was found challenging to address everyone’s questions/needs without changing the content. Additionally, the programme’s gradual modification increased uncertainty about the topics covered in the subsequent meetings. Organisational challenges Each meeting was preceded by rehearsal presentations, where all facilitators had to carry out a demo meeting and go through difficult imaginary situations (e.g. hostile parents, difficult questions), which were meant to encourage facilitators and make them more confident. Although it did not always have this effect: “I felt the most insecure … I think it was the second presentation when I was the first one to make the demo, and then all this critique towards the presentation and me. And I could have coped with the feeling in front of parents, in an unfamiliar environment, but here, I did not have the courage, or I could not do it, and I shut up, and it made me so insecure. So, I made only one presentation during the season. But the next demo meetings have been better and more positive, all training has gone well, and I have felt more confident.” (Facilitator #3, low) When meeting waves started, there was a high possibility that several classes had their meetings on the same dates, meaning it was very time demanding for the facilitators to cover all classes. Also, while the participating schools were scattered over the country, the facilitators were located in the capital. This required long drives and waiting at schools, and facilitators who did not work in NIHD had to leave their main job earlier or rearrange the whole day: 152 “…there were these annoying situations, where you drive for two hours, then you wait for two hours, talk for 30 minutes and drive back for two hours. You have spent a full working day to spend 30-45 minutes to pass on your message.” (Facilitator #2, medium) There were unexpected issues that sometimes happened before or during the meetings, and it required creative approach to find a quick solution, e.g. technology failure, training time was shortened to 10 minutes, not a single parent wanted to stay or the opposite, the class was full of parents, and everyone had something to say. Parent and teacher-related challenges Facilitators pointed out that although the programme aimed to prevent children’s alcohol use, many parents did not differentiate prevention from harm reduction, and found that the alcohol topic should be addressed when children are older, and problems are present. Facilitators implied that it was very challenging to involve parents who may have had difficulties with disciplining their children or whose children had already drunk alcohol. Getting parents to attend the meetings was a recurrent problem, which was not solely related to the programme, as it is a general issue in the Estonian school network. Facilitators assumed that it might have been due to parents’ perception about the importance of their role (i.e. perceived to decrease over the years). Regarding the alcohol topic, facilitators indicated that it was possible that parents considered their knowledge as good enough and did not perceive a need for new information. Low participation rates and parents’ indifference towards the programme required the facilitators to keep themselves motivated and keep a positive attitude towards achieving the programme’s objective. One of the biggest challenges was related to teachers’ involvement in the programme. The decision to participate in the programme usually came from the highest level in school (i.e. school principal), and facilitators perceived that in many cases, teachers were not included in the process. This led to situations where teachers did not express any interest in the programme (e.g. not participating in the meetings, not finding the programme useful as there were no problems with children, being afraid to talk about the topic), which made it more difficult to interact with parents, as the teacher was the link between facilitators and parents. On the contrary, some teachers were very eager to answer questions and lead the discussions started by the facilitators, thereby decreasing parents’ motivation to be engaged in the 153 discussion. Facilitators recommended that to overcome the issue, facilitators had to have good group management skills. 5.3.3 Long-term continuation of the programme The facilitators emphasised that the whole concept of scaling up the programme assumes that it is effective. Facilitators gave ideas on how to improve the programme, including addressing weaknesses (Table 24), renewing the content, increasing facilitators’ responsibilities, how to reach the target group and new potential facilitators. Table 24. Perceived strengths and weaknesses related to the programme (P) and its delivery (D) Perceived strengths Perceived weaknesses  Interesting topics (P)  Thorough preparation (P, D)  Same content for all (D)  Self-development (D)  Contributing to the greater good (D)  Strong project leader (P)  Weak engagement of teachers (D)  Low participation rate (P, D)  Same content for all (P, D)  The repetitiveness of topics (P, D)  Too few people at the management level (D)  Participants unsupportive attitudes towards the topic (P, D)  Vast amount of information to cover (D)  Intensive workload for short periods (D)  Challenging target group (D)  Uncertainty of the topics covered in the subsequent meetings (P, D) Ideas on improvement As the teachers were left out from the planning in the current programme, the future version should put a stronger focus on the collaboration between facilitators and teachers, e.g. holding a special training event for teachers, more effective communication during the programme. Teachers’ engagement was perceived as a critical factor to overcome most of the challenges related to the delivery process. Facilitators recommended putting particular focus on teachers who join the programme in the middle of the process. 154 “Teachers had often received some abstract, some kind of order from above to participate and there were teachers who said that they did not care, and order is one thing, but the way they comply with it another – how effectively do they forward the information, motivate parents to participate, understand the programme’s objectives.” (Facilitator #5, low) The fact that the programme’s content would not have to be developed from zero, at least initially, and to have the same order of content across all meetings, was found supportive, but it was emphasised that there was the need to renew or restructure the content. For example, the first meeting should focus more on the nature of prevention, and the repetitiveness of topics should be reduced. “When we started, we did not know where we “land” at the end of the final meeting. We wrote each “chapter”, without knowing what happens after that. ... Now, when we look at all six meetings, then seemingly, there is a coherent story.” (Facilitator #6, high) Additionally, facilitators suggested continuing to send newsletters to parents after the formal part, i.e. meetings, of the programme is finished. A thorough ongoing training process (e.g. demo presentations, training manuals) with the facilitators was found to be a crucial part of the programme, as it gave the facilitators necessary skills and knowledge to carry out the meetings. At the same time, facilitators did not have a clear idea of how to carry it out if the number of facilitators increases significantly. Regarding the process itself, it was suggested to finalise the presentation before the practice starts; otherwise, wrong elements would get entrenched. In the current programme, facilitators did not have their “own” classes. This was due to the short time frame when most of the meetings took place and classes were divided based on facilitators’ availability. Some facilitators thought that the classes in the future programme should have the same facilitator from the start to end, with the possibility to rotate facilitators if the need arises. The facilitators suggested making changes to the programme’s organisation, by involving a few extra people in the management, to ease the high workload of the project leader, and increase facilitators’ responsibilities, e.g. sending meeting summaries. Also, having a strong project leader with good organising skills was found essential to maintain as it helps to reach the programme’s objective. 155 Reaching the target group It was expected that the direction of the Institute-schools relationship would be different in the future, and instead of the Institute looking for schools, the schools would notify the Institute of their wish to participate in the programme. This could mean that the schools would be more motivated to engage the parents and organise the meetings. Although, several facilitators were sceptical in seeing an increase in interest from the schools. They pointed out, that if this situation would actually occur, there would be a much more serious issue regarding meeting the expected high demand, as it would require a significantly higher number of facilitators and thereby a network. It was pointed out that the issue of low participation rates has to be addressed; otherwise, the impact of the programme would be limited. Parents’ interest in the programme could be increased by using a broader approach in raising parents’ awareness in parenting, and by involving parents who already have had a positive experience with the programme. By hearing about the programme from other parents and not from specialists may help create broader interest. At the same time, it was understood that the participation issue is difficult to solve as it may be common to other participation aspects in schools and education and related to the school-teacher-parent dynamics. The participants emphasised that the way to reach parents is dependent on the school grade when the programme would start. For example, it was suggested that the likelihood of involving more parents in the programme would be higher in earlier grades, but it would also need a different approach, e.g. focusing on the parent-child relationship and not on alcohol, as parents’ attitudes towards talking about the latter topic might be less tolerant. In general, the facilitators agreed that the start of the programme should be around fourth-fifth grade (age of 10–11). Potential new facilitators Based on facilitators’ current experience, all were ready to recommend others to become a facilitator in the programme, if the other person has training experience and some knowledge on the topic. At the same time, sharing their own full experience was not found to be reasonable, as it might be perceived as frightening. 156 “We cannot “overload” them with our own experiences right at the start . It might scare them. ... Yes, we can provide guidelines on how to handle stressful situations, but not share our jolly and fun negative experiences.” (Facilitator #7, high) When the facilitators were asked to rate their recommendation of becoming a facilitator on a 10-point scale, the average score was 8.6 points, with everybody giving at least 7 points. Teachers, psychologists, social pedagogues, youth centres workers, health promotion and youth work specialists were seen as potential facilitators. It was suggested to carry out a test to assess training skills and get a better understanding of a person’s view of the topic. The facilitators recommended providing supervision to new facilitators and emphasising that programme’s delivery is a process that should be taken step by step, as a vast amount of information needs to be covered. 5.4 Discussion 5.4.1 Summary of main findings and key considerations This study explored the views and perceptions of eight facilitators who carried out training- related activities as part of a parent-oriented alcohol use prevention programme in Estonia. The feedback collected at the end of the programme showed that facilitators believed the programme to have a positive outcome, although they were unsure whether the impact would only reach parents, by improving their skills and knowledge, rather than their children’s behaviour. Evaluations of the programme’s impact in Estonia (Tael-Öeren et al., 2019b), Sweden (Bodin and Strandberg, 2011) and the Netherlands (Koning et al., 2010b) confirm this perception, as all showed increases in parents’ restrictive attitudes in intervention groups compared to control groups, but no evidence of change in adolescents’ alcohol use (the only exception was the evaluation carried out by Koutakis and colleagues (2008), who additionally found the programme to be effective in reducing children’s alcohol use). Although there is a clear link between parental attitudes and children’s alcohol use (Tael-Öeren et al., 2019a), the evaluation findings suggest that other factors might be involved that suppress the attitudinal influence or that the direction of causality is reversed where attitudes are altered in accordance with adolescent alcohol use, in line with cognitive dissonance theory (Festinger, 1957). The facilitators perceived the programme’s impact to be strongly related to participation rates in the meetings. It is a broader problem that is prevalent in Estonian schools, but similar 157 challenges related to low participation rates in prevention programmes have been reported in other studies as well (Al-Halabi Diaz et al., 2006; Dusenbury, 2000; McKay et al., 2018; Pettersson, Linden-Boström, & Eriksson, 2009; Spoth & Redmond, 2000). Baker, Arnold, and Meagher (2011), who carried out a prevention programme to address children’s conduct problems, found that while the attendance at the first session was high (84%), it dropped to 41% at the penultimate session, followed by a 50% increase at the last session. The researchers point out that a higher rate at the final session was due to researchers’ active parental engagement, and this approach should always be used when parents miss a session. While some of the Effekt’s facilitators tried to plan the next meeting time with parents, this approach did not seem to work, and the participation stayed low. This might have been due to short contact time between the facilitators and parents who only met during the meetings twice a year. Whittaker and Cowley (2012) emphasise the importance of facilitators’ skills and characteristics, as their experience, skilful managing of the group and having sound knowledge on the topic contribute to higher attendance rates. Another issue mentioned by the facilitators and suggested by researchers is parents’ perception of programme-related aspects, e.g. how relevant is the topic for parents and do they perceive attending the meetings as beneficial (Al-Halabi Diaz et al., 2006; Pettersson et al., 2009; Spoth & Redmond, 2000). While parents might perceive the general situation regarding alcohol use among children as serious (Bogenschneider, Wu, Raffaelli, & Tsay, 1998), they tend to underestimate the exposure to their own children (Bogenschneider et al., 1998; Berge, Sundell, Öjehagen, Höglund, & Håkansson, 2015), which may lead to decreased need for preventive measures from parents’ perspective. Also, the schools in Estonia are obliged to organise only one annual meeting with parents (The Basic Schools and Upper Secondary Schools Act, 2010), which usually takes place at the beginning of the school year. Many schools have an individual developmental conversation with the child and his/her parents in the second part of the school year. As the participation rates dropped further in spring, this might have been due to the low need to get parents together, because teachers had already met them individually. In Estonia, where the study took place, the discussion of alcohol has been stigmatised for a long time, and social norms have been more supportive of introducing alcohol to children (Raudne, 2012). Thus, approaching the target group with a sensitive topic would be expected to be a challenge on its own. Current participants did not focus on that aspect but listed other factors, such as the limited experience of training parents, covering lots of material, dealing with unexpected 158 issues, and teachers’ and parents’ indifference towards the programme. Therefore, supporting the facilitators, keeping them motivated and helping to alleviate encountered difficulties is crucial from the programme’s perspective, as they have a direct relationship with the target group (Dobbie et al., 2017; Dusenbury, 2000; United Nations Office on Drugs and Crime, 2009). This could mean providing the facilitators with additional theoretical background materials, conducting training to improve specific skills (e.g. group management), offering mentoring and self-reflection throughout the programme, and rehearsing training. Another aspect is to have a rigorous pre-selection process for facilitators to ensure that all have essential knowledge and skills. An additional target group who received less attention but were perceived as important to address as the primary target group were teachers, who had the role of intermediaries. Facilitators implied that some of the teachers expressed less supportive attitudes towards the programme, which might be due to minimal involvement before the meetings started, but also due to the perception of there not being a problem. Teachers are the ones who typically organise the meetings and have a direct relationship with parents. Thus, their attitude towards the prevention programme could have influenced parents’ willingness to attend the meetings. A study conducted with facilitators and teachers in the United States regarding Life Skills Training showed that teachers who were more involved with the programme were also more enthusiastic toward the programme in general (Hahn, Noland, Rayens, & Myers, 2002). Another aspect is teachers’ perception of the time they need to allocate for the programme. West and colleagues (2008) found that teachers who perceived the prevention programme related activities as additional workload were more resistant toward it. Despite keeping teachers’ tasks regarding the Effekt programme to a minimum, there was no consistent contact with the teachers before the programme besides an introductory e-mail, so they could have perceived the programme as creating additional workload. This implies the need to have a separate intervention for teachers before the programme starts, to give a proper introduction to the programme, but also provide theoretical background, as teachers’ attitudes towards alcohol may differ. All facilitators supported the idea of continuing the programme in the long term if it shows effectiveness, but it was suggested to renew and restructure some of the content, increase the collaboration between facilitators and teachers, address the crucial problem with participation rates, and include additional facilitators and managerial people to share the workload. 159 5.4.2 Limitations and strengths One of the limitations of the study is related to the reliability of the facilitators’ feedback – although the tenth facilitator (M.TÖ – the project leader) was not included in the study, participants were made aware that she would analyse the data and possibly would understand who said what. Thus, not making the study anonymous in the analysis process might have reduced facilitators’ eagerness to reflect their negative feelings fully and made the participants more consensual in their answers. At the same time, the interviewer who conducted the focus group was not related to the programme and acted as a neutral intermediary. Based on the findings, it did not appear as though participants were inhibited in giving honest views, although this remains a risk. Another limitation is M.TÖ’s involvement in the programme. To be more neutral and contribute to a more objective interpretation of the data, the analysis was carried out by two people (M.TÖ and E.K), who both reached similar conclusions. Additionally, the findings may not apply to other similar programmes or if other individuals were to deliver this programme elsewhere. But as the replicability of the findings was not the aim of this study, it is not a major concern. The main strength of this study is including almost all people who were involved with the delivery of the programme (eight out of 10 facilitators); thus, nearly all views were represented. Another strength is that it complements the findings summarised from the interviews with parents and teachers in Chapter 6, offering a more holistic assessment of the programme. 5.5 Conclusions Parent-oriented alcohol use prevention programmes have become more common, but the reasons behind their successes and failures in the delivery process and their impact on the outcome are seldom assessed. The findings from this focus group show that the facilitators of the Effekt programme perceived the intervention to be successful in changing parental attitudes, but the effect on children’s alcohol use was questionable due to a small number of parents attending the meetings. The facilitators supported the idea of continuing the programme in the long term after some modifications, e.g. increasing the participation rates, more emphasis on the collaboration with teachers, renewing the content. The current findings provide possible explanations for why there was no difference between intervention and control groups regarding delaying and reducing children’s alcohol use. Also, the findings can 160 help other researchers to improve the content and activities of substance use prevention programmes in the future and thereby increase the likelihood of achieving planned outcomes. 161 Chapter 6. Preventing adolescents’ alcohol use by targeting parents: A qualitative study of parents’ and teachers’ views on the Effekt programme 162 6.1 Introduction Countries all over the world are tackling the issue of underage youth consuming alcohol and facing a variety of short and long-term negative consequences (World Health Organization, 2018). Therefore a myriad of activities has been carried out by targeting risk and protective factors on individual, family, school, peer and societal level (Donovan, 2004; Foxcroft & Tsertsvadze, 2011a, 2011b, 2011c; Hawkins et al., 1992; Stocking et al., 2016). Research findings support a recommendation to combine different target-group dependent measures when targeting children’s alcohol use. Suggested measures include, for example, developing skills (e.g. social/life/parenting), applying behavioural techniques, incorporating the whole family in the programme, using trained professionals, holding recurrent structured meetings, and making activities appealing, e.g. active engagement, interactive methods (Onrust, Otten, Lammers, & Smit, 2016; United Nations Office on Drugs and Crime, 2018). Despite evidence-based recommendations, non-effective or even harmful measures are still frequently used, focusing on one-off activities, fear arousal (e.g. adverse outcomes), awareness-raising, non-interactive methods (e.g. lecturing), and using ex-users and police personnel as lecturers (United Nations Office on Drugs and Crime, 2018). Providing information may inform the target group, but using it as the only option is not found to be effective in changing behaviour (Bo et al., 2018; Foxcroft & Tsertsvadze, 2011a; Kuntsche & Kuntsche, 2016; Public Health England, 2015; Robertson et al., 2003; Stocking et al., 2016). Family-based programmes have gained more attention in the previous 20–25 years and findings from studies have supported their wider delivery, indicating a small and lasting effect on reducing children’s alcohol use (Stocking et al., 2016). This knowledge was recently refuted by Gilligan and colleagues (2019), who stated that the programmes aiming to prevent or reduce alcohol use by targeting the family showed no intervention effect on reducing alcohol use and the amount of alcohol consumed. This suggests that the question of whether and how family-based approaches might be effective requires deeper exploration The reasons behind the effectiveness of the programmes may not always be straightforward, as different factors can impact the outcome, e.g. how the programme is delivered, current social norms, target group’s perceptions regarding the topic and the programme (Henriksen, 2012; Javakhishvili, Javakhishvili, Miovksý, Razmadze, & Kandelaki, 2014; Valentine et al., 2010; West et al., 2008). Moore and colleagues (2015) have proposed a guidance framework for process evaluation, which “aims to understand the functioning of an intervention, by 163 examining implementation, mechanisms of impact, and contextual factors” and which updates the previous version developed by Craig and colleagues (2008). A thorough evaluation can provide a better understanding of mechanisms behind the effectiveness, but also the ineffectiveness of interventions. For example, the Strengthening Families Programme has been carried out in various countries, and while it has shown effectiveness in reducing children’s alcohol use in the United States (Spoth, Redmond, & Shin, 2001), a similar effect has not been reported in European countries (Baldus et al., 2016; Foxcroft, Callen, Davies, & Okulicz-Kozaryn, 2017; Skärstrand, Sundell, & Andréasson, 2013). Several studies have been conducted to assess the cultural adaptation of the programme (Allen et al., 2008; Roulette, Hill, Diversi, & Overath, 2016; Skärstrand, Larsson, & Andréasson, 2008), but in order to get a better understanding of possible reasons behind the programme’s ineffectiveness which will also complement the outcome findings, the evaluation should continue throughout the programme (Furlong, Leckey, & McGilloway, 2017). By conducting a process evaluation of the drug use prevention programme Unplugged in Brazil, it was discovered that while all lessons were delivered in more than 90% of the classes, the full curriculum was delivered in only 57% of the classes (Medeiros, Cruz, Schneider, Sanudo, & Sanchez, 2016). Also, while children’s and teachers’ satisfaction with the programme was high, teachers implied that they do not have enough time to plan and carry out all programme-related activities. Thus, in the future, more emphasis should be put on supporting the teachers. This study complements the randomised controlled trial described in the previous chapter and the focus group study with the programme’s facilitators in Chapter 3, by adding the perspectives of parents and teachers who participated in the Effekt programme. The objective of this study is to better understand parents’ and teachers’ experiences, but also views and attitudes on the content and delivery of the Effekt programme. As children’s alcohol use did not differentiate between the groups at the end of the programme, but parental attitudes did, participants’ reflections on their experience can give valuable insights to assess possible reasons behind programme’s strengths and barriers. 164 6.2 Methods 6.2.1 Epistemology and ontology See chapter 5.2.1. 6.2.1 Study design This qualitative study, in addition to a cluster-randomised controlled trial, is part of a universal parent-oriented alcohol use prevention programme carried out in Estonia in 2012– 2015 among 5–7th grade students and parents (see Chapter 3). Semi-structured face-to-face individual interviews were conducted with 7th grade teachers and parents at the end of the programme. This research method was preferred over focus groups for several reasons. First, the focus was not on the collective perspective, but to get a more in-depth reflection of individual behaviour, attitudes and experiences (Yeo et al., 2014). Second, participants are expected to be more open to express their personal views on sensitive topics during one-on- one interviews without being afraid of being judged by other group members (Gaskell, 2000). Although the focus group might enable an exploration on consensus, participants’ varying attitudes and experiences regarding children’s alcohol use might create conflict situations in a group setting. Third, as the interest was to include parents from different schools, the distance between locations would have made it challenging to organise the groups. The study was approved by the Tallinn Medical Research Ethics Committee (KK 932, 12.02.15). 6.2.3 Study sample and recruitment All 7th grade classes that participated in the intervention programme were divided into groups based on: 1) the attendance rate in meetings – medium/high (more than half of the meetings had an attendance rate of 25% or more) and low (half of the meetings had an attendance rate below 25%), 2) the type of settlement – small (≤1000 inhabitants), medium (1001–9999) and big (≥10000). The former was chosen, as low rates can reflect on participants’ opposing attitudes and the latter, as the group dynamics and perception of the topic may differ in different sized settlements, therefore ensuring higher variability in answers. Twelve classes (two per criteria) were randomly selected by settlement type and participation rate (Figure 28). 165 Figure 28. Description of the allocation process in the qualitative study. All selected teachers were sent an e-mail invitation to participate in the study. In case the teacher refused to participate, a new class (i.e. teacher) with the same attributes (i.e. attendance rate and settlement size) was selected. For the interviews with parents, each teacher was asked to recruit two parents to participate in the study, while considering that one parent should have participated in most of the meetings or have a somewhat positive attitude towards the programme and the other parent with opposite attitude or low attendance rate. Two teachers were not successful with recruiting parents, so substitutes were selected from different classes with the same attributes. In total, 12 teachers (11 women, one man) from 11 schools and 24 parents (all women) from 13 schools agreed to participate in the study. 6.2.4 Data collection and analysis All interviews were conducted in Estonian by four interviewers who were not related to the intervention programme in February–April 2015. The interviews took place in school facilities (e.g. in a classroom or a teacher’s cabinet), with one exception, where the interview was held in a local cafe. All participants gave their verbal consent to participate in the study and received a 30€ gift card after finishing the interview; the interviewees were not notified about the gift card before the interview. The interviews varied between 27–63 minutes (M=42min) among teachers and 23–54 minutes (M=39min) among parents, were audio- recorded and later transcribed by M.TÖ. Also, the interviewers wrote a summary after each interview (no personal details were recorded, keeping the interviewees anonymous). Due to one interviewer accidentally deleting audio files of six interviews, prior written summaries were used in the data analysis. 166 The semi-structured interview schedule was developed by M.TÖ and E.K and consisted of two parts: 1) general views on adolescents’ alcohol consumption and its prevention, 2) programme-related matters, i.e. attitudes towards the programme (including the delivery process, participation’s barriers and attractiveness), perception of the programme’s impact, strengths and weaknesses, and suggestions regarding how the programme could be improved. Thematic analysis was applied to both parents’ and teachers’ interviews. First, the initial thematic framework was created by M.TÖ, and it was based on the teachers’ interview schedule (initial themes) (see Chapter 5.2.4), using a semi-inductive approach. Then, to validate the framework’s content and applicability and assess coding consistency, M.TÖ and E.K used the framework and independently coded one teacher’s interview using the qualitative software program NVivo 12 (QSR International Pty Ltd., 2018). The results were compared and discussed to improve coding coherence. Although there were some differences in researchers’ interpretations of the text, the process confirmed that their context-related understanding was similar. The differences occurred due to the interpretation of codes’ wording accuracy, codes’ logical location in the text, and interpretation of subcodes, which was related to the careful reading of long paragraphs. After the coding framework was refined, both researchers independently coded an additional teacher’s interview. Next, a similar dual coding process was undertaken with one parent’s interview. The framework was refined several times and eventually resulted in 35 codes under six themes. The preliminary findings were discussed between researchers with a consensus reached. To give direct examples of participants’ opinions, illustrative quotes have been added in the text. The source of the quote (i.e. parent/teacher), size of the area where the school is located (i.e. big/medium/small) and the general participation rate at the class (i.e. high/low) are presented after the quotes. 6.3 Results The data were categorised into six main themes: 1) Perceiving and tackling the problem 2) Experience of the first contact 3) Low participation rates and meeting the objective 4) Views on the content and delivery 5) Perceived value of the programme 6) Long-term continuation of the programme 167 Before focusing on the programme, participants were asked to reflect on their views of (children’s) alcohol use issue in general and discuss possible options to support parents in preventing and reducing it, if this was felt necessary. 6.3.1 Perceiving and tackling the problem Perception of the general situation In general, participants perceived alcohol use in society as a problem, e.g. too much alcohol is consumed, many people are addicted, drunk driving is prevalent (especially among younger adults), minors drink alcohol, and drunk people are seen in public spaces. One of the parents shared her view that the situation regarding alcohol use is unchangeable because people have consumed alcohol in the past, and they keep doing it in the future. “But maybe we will get there one day. I think that it may change, maybe with the next generation. The society cannot change itself that fast. /…/ I do not believe that alcohol use will drastically decrease in the next 20 years. I honestly do not believe in that. /…/ I can give an easy example of why it never stops; alcohol is prohibited for minors, everybody knows it, there is a law, everything is fine. But it is also prohibited (by the law) to cross the road when the red light is on… everybody does not get hit by a car.” (Parent #1, big-low) “I had this situation; I think it was in a bus station, as young people usually meet there. So, I was waiting for a bus, and there were underage girls who were talking about a party that took place the previous day. They talked about how they drank alcohol and were caught by the police, and the officer said that how could he detain such beautiful girls and that they should leave. If already the police have that kind of attitude of not doing anything and not preventing it, then how do we move forward in a small city like this. Everybody has to contribute.” (Parent #2, middle-high) „How to root it out, I do not know; it seems practically impossible. People perceive it as normal, in every level of society, and people drink anyway, the end of the year school trip, it is normal.“ (Teacher #1, middle-high) “We know that the birthday parties at the end of secondary school include alcohol. It is a public secret that we are aware of.” (Teacher #2, small-high) 168 “For example, in Austria, I was told that children learn to drink wine from early on and thus they are much less threatened because they can value it, and wine is not very strong, it is different drinking culture. /…/ In Estonia, we do not have this kind of light alcoholic beverages drinking culture; it is still developing. I find this necessary, as people here drink much more strong alcoholic beverages, we have no wine culture, and then children choose that (strong alcoholic beverages) too.” (Parent #3, big-high) While some saw alcohol use as an increasing problem, others found the current situation improved compared to the Soviet era. “It is definitely a problem, but a decreasing one. There were times when people consumed alcohol senselessly. Adults who are in their second part of life, who have Soviet experience, they have problems. People who come from a Soviet society where vodka was used as a means of payment, for example, in a car repair shop.” (Parent #4, big-low) The two main perceived factors influencing alcohol use were alcohol availability and advertisements. For example, alcohol was seen as easily available, as minors get easily to buy it (both off- and on-license) or find an adult who would buy them alcohol. Also, it was highlighted that “alcopops” that have a similar amount of alcohol as beer, have a mild lemonade-like taste and a “tempting outlook” are getting more popular among children. Although the situation was perceived problematic, most participants pointed out that these problems did not exist among their family and friends. It was suggested that people who had more direct experience with the issue also perceived it as more acute. “Among my family and friends, it is not a problem. Well, we and our friends, we do not consume too much alcohol. We do drink at parties and events. But what others have said and what you hear from the radio, it seems to be a problem.” (Parent #5, middle-high) Participants agreed that alcohol use initiation takes place at an early age, but it does not apply to everyone (i.e. to their children). 169 “Kindergarten teachers have talked about it for a long time. You do not have that many playthrough situations at school, but at kindergartens, they play the weekend on Monday mornings and clink the glasses. This means that there has been a party on the weekend.” (Parent #6, big-low) “When the child is already a teenager, 14 and so, maybe s/he does not come to parents to ask vodka or some other drink. S/he tries it out on her/his own, maybe even finds a way to buy it from somewhere. But children below 10, they ask, “Why do you drink” and “I want to try”, and maybe they have an even bigger interest in this.” (Parent #7, middle-low) The same was found regarding heavy drinking, e.g. problematic among some, but not among all. Main perceived reasons for early-onset and use in general included:  parental role modelling at home  parents’ permissive attitudes or children’s lack of knowledge on parents’ attitudes  the prevalent concept of “the forbidden fruit tastes the sweetest” in the society  parents’ minimal interference in children’s lives (e.g. trusting children to be home alone that can result in having parties)  high availability of alcohol (e.g. at home)  being a part of a group where somebody offers alcohol Some parents did find that problems with alcohol use were less serious among children than adults, as parents’ attitudes have become more restrictive compared to the parents’ childhood. Also, young people were perceived to be more aware and self-confident to say no if someone offers them alcohol; something that was not common decades ago. The alcohol use problem among children, but not among young adults, was perceived as less serious, but not non- existent in smaller/rural places as everyone knows each other. At the same time, as the incomes are lower in rural areas, children have fewer opportunities to engage in recreational activities, which has an impact on children’s alcohol use. Also, it was highlighted, that if the child has alcohol use-related problems, then most probably the scope of the problem is much more extensive and relates to other risk behaviours as well. The situation at home and school Participants agreed that children are aware of alcohol and the reasons for not using it and have had exposure to alcohol, e.g. television, public places, family get-togethers and family 170 members’ alcohol use. Two main occasions were mentioned in connection with parents’ alcohol use in front of the children: adults’ birthday parties and New Year’s Eve. However, most parents who reported on their alcohol use said that they try not to do it in front of the children. “At the child’s birthday party – of course, it (alcohol use) is not okay. But for example, on New Year’s Eve, when we have friends visiting, we do not change these things because children are present.” (Parent #6, big-low) Parents and teachers indicated that in general, children had not expressed interest in alcohol, and there had been only single occasions where they wanted to sip an alcoholic beverage or had been interested in alcohol-free cider/beer. Some parents had offered alcohol to their children or knew that the children had tried it. On the other hand, as children are maturing, the interest in alcohol was perceived to become more frequent. “I have not had this situation, where my child asked to try alcohol. I know that he has asked about the reasons why I drink or why we have to drink. /…/ Of course, you do not always have to drink alcohol, but maybe when you have grown up that way, that during birthdays there is a bottle on the table, then…” (Parent #7, middle-low) “We have not offered alcohol to her, during parties, and I do think that we should not do it. But maybe in two years, we have to re-evaluate the situation and change the approach. Right now, she is going through a period where the development is very fast, and things and attitudes change.” (Parent #8, small-high) “I know that she has tried alcohol, and she has done it at home, I have offered. I know that there are opinions that it is not right, but if she wants to try it, then I think it is better when I know it, rather than hearing from someone else that she was offered alcohol or asked it from a stranger.” (Parent #9, small-high) Although discussing alcohol use was perceived as important from the health and law perspective (latter more), the majority of parents found the children to be too young to talk about it. It was also suggested, that talking about it might increase children’s interest in trying it out. Despite that, some parents had had conversations with their children. 171 “I remember when my son asked me if fathers actually act like this – he saw something from the TV where father came home drunk and started to beat his child, and the child hid himself under the bed. My son asked if this kind of thing is possible and I remember that then a discussion developed out from this topic – what alcohol does and it takes your mind “away”, so you are capable of doing this kind of thing to your child.” (Parent #10, big-low) Teachers’ experiences with children’s alcohol use were much more comprehensive, as, over the years, they had taught hundreds of students. Mostly, teachers had examples of problems regarding older students, or if the children with problems were younger, they were not from current classes. "We had this one incident when fifth-graders took vodka and went to have a party in the forest. Somebody saw it and got called names." (Teacher #3, middle-low) Parental attitudes towards children’s alcohol use Parents’ reports on their attitudes towards children’s alcohol use were divided. Parents with restrictive attitudes pointed out that children should turn 18 before drinking alcohol. They also felt that the use of alcohol-free alternatives (e.g. alcohol-free beer/cider) and drinking alcohol in front of the children were not acceptable. Parents said that after all, everything comes down to children’s attitudes, as children make the final decision, taking input from the surrounding environment, such as friends. Parents and teachers with less restrictive attitudes found that children get more curious, and it is okay when 15–16-year-olds try alcohol and keeping the child away from alcohol only increases resistance. They also highlighted the importance of being a role model by consuming alcohol moderately and raising children’s awareness of the topic. “As I drink wine, my child knows from early on which glasses belong to specific beverages. /…/ Just recently I told her to bring me a wine glass, and as she saw that it was red wine, she brought the right glass. Then I thought that she did not bring me just any glass but looked for the right one. This means that I have taught her something.” (Parent #3, big-high) 172 “If it is the New Year’s Eve, a glass of champagne, or a birthday party. It should be coded to the child that alcohol exists, but you use it moderately and reasonably. I do not support zero-tolerance, because we already know that the more forbidden a thing is, the more interesting it is.” (Parent #6, big-low) "I come from another generation, and I do think that at some point you have to talk about how to drink, and how to drink culturally and what to drink with what. For example, setting the table is important. This could be taught at the handicraft class, where to put the glasses and what to drink with what. /.../ I think that they try alcohol anyway, so why not to do it culturally.“ (Teacher #1, middle-low) When asked about other parents’ attitudes towards children’s alcohol use, parents perceived others more permissive, e.g. due to their own childhood experience and current alcohol use practice. Participants gave examples of parents at the meetings who said that they would not stop their children from trying alcohol. They suggested that these parents prefer their children to drink at home rather than at a party. “In those families where parents drink alcohol, the attitude is different, and they let their children try alcohol, they think it is normal.” (Parent #11, middle-low) Based on parents’ feedback from individual meetings with teachers, teachers perceived that parents might not have the motivation to talk about the alcohol topic at home, as children are too young, and parents do not think that their children would consume alcohol. On the other hand, discussion on the topic was found relevant, if problems already existed. The former was confirmed by most parents. One of the teachers suggested to explain and talk about the issue as it is (i.e. not making alcohol to look “utterly evil”) and not prohibit and let children try it if interest exists. This way, the parent can be present, and children would not look out for alternative options on how and where to drink alcohol. Teachers implied that although the general attitude among parents seemed to be rather restrictive, it was somewhat dependent on children’s age and individual differences. It was also suggested that some parents might not understand the seriousness of the problem (e.g. how many children consume alcohol) or if they do, they do not want to talk about it. 173 The collaborative approach to support parents Participants agreed that the home has the crucial role in children’s alcohol use and its prevention – how parents behave (being a role model), what kind of attitudes they have – but it was also said that all stakes should not be set there; school environment, friends and society all have their influence. It was perceived that the approach has to be holistic and address similar principles at home, school and society. One of the teachers indicated that if the home’s influence is negative, it is difficult for the school to achieve impact. “… environment, community, school and family, having a united front that can help and support each other.” (Parent #12, small-high) “Society has a big influence, what are the prevalent attitudes. For example, when smoking was prohibited in public indoor spaces, and there were separate rooms, then younger people took it over, and smoking is not popular anymore.” (Parent #13, big- low) “Unfortunately, you cannot do it alone. Talking to three mothers does not help that much. We need the shop assistant to be active, the janitor. /…/ Nobody should look away.” (Teacher #3, middle-low) “I am really sad that we do not have that kind of communes like abroad. Whenever a problem arises, everybody starts to work on it – the combined effort is bigger, and someone might know more than you do.” (Parent #2, middle-high) Including school in prevention was perceived as helpful by parents, as it repeats the message delivered at home and children may not always listen to their parents. Teachers had similar views; only the perspective of the support was different – parents repeat the same message as the school does. Teachers said that the school itself could not do much besides educating and monitoring (e.g. surveys) the children, and initiating the dialogue that conveys the message, thereby supporting the parents. Another view expressed by teachers was that in a way, parents move the responsibility from themselves to school and assume that the school has to do all the talking and explaining, and not only regarding the discussion about alcohol. One of the parents suggested that if parents are to be involved in alcohol use prevention, the tradition of educational meetings needs to be created at an early age, focusing on other topics at first. Participants found that raising awareness is more important among those parents who 174 already are at the risk-group, assuming that highly educated parents have more knowledge on the topic. Regarding supporting parents, some participants pointed out that support itself is very welcome, but the parents’ attitude and experience play an essential role in the outcome. For example, parents who are more open-minded and ready to learn new skills and information probably value the support more. “Somebody should support parents, but if you overdo it, then it creates defiance. Why do you teach me? I have heard it enough. But where to draw the line, how much is enough and too little, who knows.” (Parent #13, big-low) Several teachers and parents suggested that from the state’s perspective, the law should be better enforced, alcohol availability should be decreased, the tax increased, and advertisements banned. Also, the media could portray more positive examples/role models. “I think that the most important thing is to ban alcohol advertisements. It seems the worst. Well, the child does not go to bed at 9 PM, or the TV would not be turned off at that time, and you can see it on the internet, everywhere – how having a party is cool and how alcohol is a part of this.” (Parent #9, small-high) Some parents expressed their concern that limiting or restricting alcohol availability would not help to tackle the issue, and the problem may be the result of excessive restrictions. 6.3.2 Experience of the first contact Prevalent lack of awareness In interviews with teachers, teachers being left out of the decision-making process was a dominant issue. Teachers implied that the first contact was made with the school administration and often the initial information about the programme was received by an e- mail from the project leader just before the first meeting took place. Lack of awareness appeared to create the feeling of not being part of the programme, being last in the “school chain” and thus, it was more difficult to understand why the programme was needed and how it was carried out. The first meeting was reported to give some insights, but several teachers indicated they were already negatively pre-tuned. This “practice” was not prevalent at all schools, and teachers, who were made aware of the programme by the school administration, said they saw it as a good opportunity to involve the home and support the school curriculum 175 as the topic of alcohol was planned to be covered in health education classes around the same time. “… I am quite sure that the administration got the overview of the programme, but class teachers were not aware of it. /…/ … the information came through different channels and somehow it was decided.” (Teacher #4, big-low) Teachers perceived themselves as the connecting link (i.e. mediator) between the facilitators and parents – they distributed materials and held meetings. Their readiness to contribute was perceived to be related to their workload, level of awareness, understanding of the objective and own perception of the role. Some teachers did not feel as “equal partners”, had no sense of ownership, and, perhaps consequently, were more reluctant to be part of the whole process, e.g. not attending the meetings and reading provided materials superficially or not at all. “I perceived it as an obligation, not a bad thing, but I felt it was an obligation that has to be done. I did not feel devoted.” (Teacher #5, middle-high) In interviews with parents, similar issues related to lack of awareness were found. It was apparent that teachers’ ignorance transferred to parents, and they did not appear to understand why the programme had to start that early (fifth grade), as children were not yet consuming alcohol. In some schools, parents did not report having any knowledge of the programme before the first meeting, and they were notified either just before the meeting or during it. I went to the class meeting and was faced with the fact that before the class meeting we will have a lecture. And then I hoped that it would not take place the next time.” (Parent #14, big-low) Parents indicated they had no clear idea on what to expect, and while some were more open- minded and found it reasonable to participate and listen (as it concerns children and helps parents to improve their parenting skills), others had more negative views. Motivation to participate When asked about their motivation to attend the subsequent meetings, two main reasons were highlighted: an opportunity to broaden their horizon, and sense of duty. The former seemed to be related to learning new skills and insights which can be useful at home (e.g. how to cope with a teenager) and understanding if their practice had been “right”. Parents who attended the meetings out of a sense of duty pointed out that if a promise had been made, it should be 176 adhered to. Parents also said they found that attending meetings would be a great example of responsible behaviour for children. Teachers perceived the same motivational factors for parents and implied that the teacher’s own negative or indifferent attitude towards the programme might have negatively influenced parents, including those who were willing to attend the meetings. 6.3.3 Low participation rates and meeting the objective Declining participation rates Both parents and teachers found that parents’ attitudes towards the programme were reflected in meeting participation rates. While most of the interviewed parents (or their spouses) had attended more than half of the meetings, the general participation rate gradually decreased over time, being lower in spring than in autumn. This trend of declining participation was considered to be a common problem after elementary school and not particularly specific to the programme. Although there were some cases where spring meetings were held only because of the programme, and the attendance was even lower than in the class meeting. Parents expressed their concern that prior experience with a teacher who gives the same talk every meeting, does not increase parents’ expectations to hear anything new, and they decide not to continue attending the meetings. In some classes, parents “learned to be late” or deliberately delayed the formal part of the class meeting to avoid the part where the programme’s content was covered. Several participants implied that parents who should have attended the meetings (e.g. had a drinking problem, were from a problematic family), did not do it, and people who did not have any problems or did not express a specific need for the content were mostly the ones who attended. “One parent who has problems with alcohol use was horrified when she was called and introduced to the alcohol use prevention programme. She has not attended any of the meetings.” (Teacher #6, small-high) Perception of the programme’s objective Previous findings show that some participants’ perception of the programme’s objective was different from the actual main objective – to delay alcohol use initiation. In general, three 177 different types of objectives were identified from the parent and teacher interviews: 1) prevention: “The objective is that we would behave correctly at home, direct our children away from alcohol, and increase our awareness – what alcohol does to the child, especially alcohol use at an early age. In a nutshell, the objective is to delay alcohol use onset as far as possible” (Parent #4, big-low) “I understand that the objective was the same as in Sweden. On the one hand, to reduce children’s alcohol use or raise awareness through parents and on the other hand, to delay or avoid alcohol use initiation. And I understand that indirectly the influence is wider, not just alcohol, but other things too. To develop a restrictive attitude towards alcohol use through parents, this has been my impression. (Teacher #7, big-low) 2) harm reduction – addressing the problem (and problematic families): “I believe that we have problematic children and children from problematic families, where children start to experiment with alcohol at the age of 11 or 12, and maybe this (programme) is important and useful for them.” (Parent #8, small-high) “Well, what has happened is that this class is a bit too good for this programme. I am not saying that we do not have any problems in rural schools, we do! But this class, we have these families, and honestly, we do not have any problems with smoking and alcohol use.” (Teacher #2, small-high) 3) education – paying attention, discussing, and making parents think about alcohol use related topics (e.g. consequences, harmfulness): “To educate parents and how to cope with problems, drugs etc. I remember the last one – what to do and how to behave when your child comes home while being drunk.” (Parent #10, big-low) One of the parents emphasised the aspect of increasing collaboration: “…the hope was to connect parents, via agreements, parents and school and I do not know, maybe the whole society.” (Parent #15, big-high) 178 Reasons behind low participation rates Teachers and parents proposed different reasons for low participation rates, emphasising that the reasons were speculative, due to minimal or lacking communication between parents present in most classes. For example, parents might not have acknowledged the programme’s usefulness to themselves, e.g. own knowledge was found to be good enough, or if needed, the information could be found on the web. It was also possible that at some point, parents might have felt that they had heard enough and did not need any additional information. Parents suggested that the initial positive reaction might have changed due to the repetitive content and long pauses between meetings. On the other hand, if a parent’s first experience was negative, it was perceived difficult to reverse it. Participants sensed that some parents might have felt that their parenting skills were questioned, especially if they also had older children, making the situation uncomfortable and hostile. Few of the parents pointed out that despite their attending, they did not plan to act on the provided information, as they already had enough knowledge and skills. Having liberal views on children’s alcohol use (e.g. children should try alcohol at home, “I am not raising a teetotaller”) might have made some parents reluctant to participate. It was also proposed that it is not possible to convince parents with liberal views, but more likely to influence those parents who have not “made up their mind”. One of the teachers pointed out that it is difficult for parents to relate if they do not have any personal exposure. “For example, offering alcohol at home, finding that it is better to let the child drink at home. I think that you cannot persuade them. /…/ These parents who had not made up their mind, maybe they were persuaded. But I will say that these who have made up their mind… /…/ …the same happened to me, and it has always been like this, and it will stay like this.” (Parent #16, big-high) “I do not care what others have agreed on, because I drink wine when I want to, and I do not hide, this is absurd. I will not change my habits before and after this programme.“ (Parent #14, big-low) Several parents who perceived the programme as a harm reduction intervention did not express great interest in attending the meetings if they reported not having any alcohol use- related problems at home. In addition, it was suggested that people are prone to think that bad things would not happen to them and thus are less interested in that kind of programmes. 179 High number of different training sessions at school was perceived as a factor that might have created indifference towards the topic. This is especially acute in a situation where the alcohol topic is widely covered in the media. “The topic is relevant; it has to be dealt with, but… I do not know, I mean, if everyone talks about it all the time, then at one point it is just too much, and it turns into background noise, and nobody would not listen to it anymore.” (Parent #17, middle- high) In general, schools had a maximum of two meetings per year and it was found reasonable to assume that the proposed time did not suit everyone (e.g. shift workers, people not able to find a babysitter for younger children). One of the teachers indicated that parents might not like discussing delicate issues in front of others and may prefer a more individual approach, i.e. a developmental conversation between the teacher and parent that usually takes place once a year. 6.3.4 Views on the content and delivery Content Participants’ views on the content covered three aspects: meetings, agreements and materials. Regarding meetings, the majority of participants who had positive attitudes towards the programme also expressed positive attitudes towards the covered topics, as otherwise, these topics would not have been addressed. However, some participants found the repetition of some topics annoying. Participants perceived the basic principles consistent, examples practical and visual, remembered most of the topics when presented with a list of topics, and found them interesting and informative. One of the teachers who had a rather neutral attitude towards the programme did not agree with others and found that there could have been more practical and tailored examples because parents, children and classes are not the same everywhere. Parents who had a negative attitude towards the programme mostly found the topics useful, if they were relevant and included new information, but it was the format they disliked. 180 “In general, all of the topics are normal, but I do not know, the presentation form, it makes this…/…/ The lecturer acted like she was the teacher and we were the children, like we were half-witted, and said that we should not drink alcohol with our children and so on. Well, I do not know; this format does not work very well.” (Parent #14, big-low) Regarding the topics, the most memorable included the importance of parental attitudes, not offering alcohol to children, communication with the child, alcohol’s influence on the brain, taking a problem-solving approach, and alcohol advertisement analysis. It was easier to remember topics when they addressed personal needs, e.g. energy drinks use was an acute problem. While some participants would have wanted to hear about alcohol’s influence on children every single meeting, others found it less useful, due to common knowledge. Participants found that when the topic-related knowledge was scarce, the first meeting made parents aware of the alcohol use problem among children and “brought the problem closer to home”. Despite remembering the topics covered in the meetings, most of the participants could not recall anything specific that was talked about at the first meeting. Agreements that were made between parents were perceived by many participants being important, giving a “theoretical basis” and helping to create a unified front, assuming everyone would adhere to them. But making agreements requires communication between parents, and as the interaction between parents in most classes was irregular, the feasibility of the agreements was highly questionable. In addition, low participation rates influenced agreements’ effectiveness – people who did not attend the meetings did not have their say and reading about the agreements at home was not the same as making them in the classroom. Some parents also said that the parents were obliged to make agreements, and this created reluctance among participants. Based on children’s feedback to teachers, agreements were not always adhered at home and teachers suggested that parents’ initial enthusiasm waned at home. Many participants remembered an agreement that was related to notifying another parent when his/her child has been seen drinking alcohol or consuming any other substance. Although the purpose was found noble, people were sceptical about other parent’s reaction in terms of whether it would be perceived as help or as “whistle-blowing” and how it would affect the children of “whistle-blowers” at school afterwards. It was pointed out that in a real situation, parents would not think about the agreement or how to act on it. 181 “… it is difficult when you go and say, I saw your child. It is always difficult. It is like saying that I saw your husband with another; it is welcomed, but I do not know how it will be received.” (Parent #18, big-high) “… and then some parents said that they could not do it because their children will be bullied.” (Parent #5, middle-high) As most parents and teachers had not had any alcohol use-related issues with the children, it was found difficult to assess the effectiveness of agreements related to alcohol use. And some agreements were perceived to be too rigid (e.g. children do not consume alcohol before turning 18) or not relevant, e.g. agreement on limiting children’s energy drink use when only some consume it. On the other hand, agreements were found more useful, also outside the classroom, when there were an acute personal need and an actual change in the outcome: “I had more interest when people talked about excessive computer use. I remember there was an agreement that children would not be behind the computer after 9 pm. This was, well I thought that I should start applying this more often.“ (Parent #14, big-low) “We agreed that the class parties would end at 19:00, and it works. /…/ …children do not beg anymore, and they follow the agreement.” (Teacher #8, small- low) “We, parents, agreed, that on school days there is a certain time when children have to be home, and we followed it and said it to children, that we had our meeting, four families, and we decided on the curfew time for school days. /…/ And it worked, and children were okay with it and were home at the right time.” (Parent #13, big-low) Participants agreed that it would be reasonable to introduce the approach in earlier grades, so parents could develop a habit of communicating with each other and making agreements; otherwise, it is easy to forget the agreements, especially without any exposure. It was emphasised that agreements have to be easily applicable and to be lasting, parents need to have intensive and consistent collaboration. One of the parents suggested that the true value of that approach will be seen when actual problems arise. When asked about receiving any additional information in between the meetings, at first many parents did not remember receiving anything, but after the interviewers specified the format, they did recall receiving newsletters and meetings’ summaries, although the latter 182 was recalled less frequently. While half of the participants found materials generally useful, most of the teachers and parents found that the main target group for materials would be parents who are interested in the topics, have less knowledge, have problems or do not attend the meetings. The latter was dependant on the reasons for not attending – parents suggested that if people were not interested in the meetings, then most probably they did not read the summaries too. Participants’ views on the format were divided, while some preferred “to hold it in their hand”, others would have liked to receive the materials by email: “When you receive it and read it right away, then the information gets somewhere. But when you put it on a cupboard, even for a moment, it probably will stay there. So, maybe it could arrive by e-mail. I am more of a person who sits behind the computer.” (Parent #19, big-high) The newsletters were found informative and supportive, although some participants found them too general and theoretical, loaded with a lot of information, that was not new per se, and rather summarised the information already covered at the meetings. Some participants, who had received meeting summaries, found them useful to recall agreements made between parents. Teachers indicated that parents would probably not get this kind of summarised information from anywhere else (e.g. on the web). There were parents who, for example, created a special folder for newsletters at home, to use if necessary and at some homes, newsletters were used to engage other family members (e.g. spouse, children) in the discussion. Delivery Regarding the delivery, parents supported the approach of facilitators rotating, as it was interesting to hear different perspectives; although, it required the new facilitators to take some additional time to “blend in”. In general, facilitators were perceived to be too young, and this created scepticism among some parents. These parents felt it was challenging to take facilitators seriously when they did not have any experience in raising children, and when their knowledge was purely theoretical. 183 “There were several meetings when someone asked the facilitators: “Do you have children?” I think they did not have any. Maybe I am wrong, but at least the facilitator we had did not have any children. And then there were these comments: “First have children and then come and teach us.” I do not think that the age, but then again if you have not struggled with a teenager, you do not know how things are.” (Parent #17, middle-high) “I am a parent, and I can talk about my own experience, so from that perspective, I presented both sides. But this experience I have, personal and teaching related, they are quite extensive. So, these talks (by the facilitators) sound a bit naïve.” (Teacher #9, big-low) For some of the parents, the initial suspicion faded after attending the meeting, and it was acknowledged that the facilitators managed to overcome the challenge of not being parents. One parent suggested that as the final target group is children, young people should be the facilitators as their point of view is different from parents’ view. Participants pointed out the importance of facilitators being able to answer questions, know the topic, lead discussions, give examples, be friendly and charismatic, and blend in. Most agreed that these prerequisites were filled. “When the facilitator is interesting when it is interesting to hear her/him, then it is very positive, you feel engaged and remember more.” (Parent #2, middle-high) “You could see that they (facilitators) were interested in the topic; they found that they are doing the right thing. I find it very important. The person who trains believes in the content.” (Parent #1, big-low) Despite the positive experience, some participants expressed their dissatisfaction with the facilitators. This was due to poor interaction with parents (e.g. teacher-students approach), strong reliance on slides, coming across as moralising, and presenting the content “dryly”. Some parents perceived that in case there were disagreements between parents and facilitators, facilitators sounded too “strong” when implying that their views were the right ones. 184 6.3.5 The perceived value of the programme Teachers’ perspective Most of the teachers found that despite their existing knowledge on discussed topics, there were new valuable aspects (e.g. information on the hidden content of alcohol advertisements, brain development, statistics) covered in the meetings. In addition to new information, teachers acquired new methods which were seen as useful in the classroom to support the school curriculum (e.g. using a problem-solving approach, roleplay). Involving parents was perceived very important, as, without home’s support, the school cannot do much about the alcohol topic. This also gave teachers a new perspective, as the alcohol topic was usually covered only with children. “… what I always say to children, your behaviour is important, you cannot change another person, but you can always improve yourself.” (Teacher #5, middle-high) All teachers were asked to imagine a situation where their colleague would ask their advice on joining the programme and give a score from 1–10 on the likelihood of recommending. Out of 12 teachers, eight gave a score of six or higher. “In general, I would recommend, because it is an important topic. If the format and content stayed the same, I would decrease the score a bit; I would say seven.” (Teacher #7, big-low) “10 points. Hoping that the parents in another school would be readier to attend.” (Teacher #8, small-low) Teachers who gave lower scores (five or less) indicated that teachers should be more involved in the programme, but that would require additional time from them. One teacher firmly stated that she would not recommend participating at all if the programme’s format and content would be kept the same. “If I would recommend participating, to understand the process, the involvement should be bigger. Maybe this way it would be possible to involve and direct parents more effectively.” (Teacher #4, big-low) 185 Parents’ perspective Parents and teachers who found the programme beneficial for parents, in general, indicated why with the following reasons:  it broadened horizons and gave a theoretical foundation, as it gave a systematic overview of the alcohol use problem and proposed ideas and practical examples on how to address the topic at home. At the same time, this gave parents input to think about their influence on children, e.g. limiting alcohol use in front of children, not offering alcohol, in general becoming better in parenting  regular meetings (unlike one-off activities) with good content and clear objective (i.e. prevention and harm reduction) created an opportunity to discuss, share worries and ask questions  information was brought together and made available to all parents, assuming that most parents would not search it for themselves  newsletters were good to rely on, e.g. in a situation when asked about the topics covered at the meetings or when in an alcohol use related situation  collaboration between the school and home and also between parents was facilitated, e.g. making agreements  parents received confirmation that they had done the right thing, and there were others with negative attitudes towards children’s alcohol use, encouraging parents to stay persistent Parents gave examples of good practices, e.g. their own alcohol use was reduced, attitude towards children’s alcohol use was more restrictive, they paid more attention to child’s activities, increased communication with the child and other parents, agreements were successfully used outside the class environment, information covered at the meetings and newsletters was shared with other family members. “We managed to agree on limiting children’s computer time. Mutual agreements, there is no point for the parent to play Don Quixote alone, to tilt at windmills. We created a unified front with other parents, so, after nine PM, no child was allowed to stay behind the computer.” (Parent #12, small-high) “You have to be persistent, not offer alcohol yourself. /…/ This is a big thing that I have done, offered champagne “Want to try? Well, come and try.” /…/ I am not doing it anymore, not anymore.” (Parent #11, middle-low) 186 Both parents and teachers agreed that the programme’s usefulness was related to participation rates at the meetings, and parents’ attitudes and interest towards the topics. Teachers suggested that the programme had an impact on parents who attended the meetings, who were actively engaged in discussions and interested in their child’s activities in general. At the same time, it was pointed out that the programme gives the input to act, but the final step has to be taken alone, at home. One of the teachers perceived the programme as a “small droplet in the sea”, as changes take time, but at least the process was initiated. It was also suggested that the programme’s usefulness is dependent on the quality of the parent-child relationship and becomes clear when the first problems arise. Those parents who did not find the programme beneficial indicated that parents who had enough knowledge, no problems and had their views on the topic did not gain much from the programme: “Parents do need support, but it depends on their age. For example, when the person became a parent at the age of 16, then, of course, his/her worldview needs to be shaped. But for example, when a 40-year-old person becomes parent, this is something different, this person has already a developed worldview, that would most likely not be changed by some programme.” (Parent #4, big-low) The same question that was asked of teachers regarding recommending participation in the programme was asked of parents. Out of 22 parents, 17 gave a score of 6 or higher, indicating a higher likelihood of recommending another parent to participate in the programme. Two parents did not specify the score but said that they would agree to recommend. It was suggested that an essential aspect is other parent’s interest in participation; although if the parent is already asking, then it probably indicates at least some interest. “Five. /…/ I would say that it is useful, but there is a lot of repetition. It is useful to attend the first two meetings.” (Parent #5, middle-high) “Yes, I would recommend if the parent has time to participate. I would not vilify it; I would not say that it was bad. /…/ 9–10, I would say. There were professionals, and the meetings were well prepared.” (Parent #19, big-high) 187 “If the parent is interested, then, of course, it is reasonable to participate, and maybe if there are problems with the child. /…/ Eight. /…/ You understand that you know these things and if there is no specific need and you are not that interested, if I would have a specific problem, of course, I would say ten.” (Parent #6, big-low) One of the parents who had a negative attitude towards the programme and did not find the programme useful pointed out that before recommending anything, she should know more about the other parent’s situation, e.g. are there any problems, what kind of knowledge the parent has on the topic. If there were indications of problems, she would recommend participating in the programme (giving a score of 6–7) but without any specific need, she would not give her recommendation (a score of 1). Other parents who gave lower scores expressed the same belief, that if there are no problems they would rather not recommend participating in the programme. None of these parents perceived the programme’s objective to prevent children’s alcohol use, but as to reduce harm and educate parents. Children’s perspective Many participants indicated that it was difficult to assess the programme’s impact on children. Both teachers and parents perceived that most children had a rather negative attitude towards alcohol and had not initiated alcohol use. Therefore, it was difficult to see any impact or change. “My child is so good; there could not have been any change. As I have not had any problems with my children, I would rather say that I have not seen any changes. /…/ This whole class, regarding these topics, alcohol is a no, and it is interesting that there is no risk group. Is it because of the programme or are they essentially that good? I do not know.” (Parent #12, small-high) Participants perceived that in order to make the programme beneficial for children, parents had to share the information and recommended tips (e.g. how to be more assertive) that were discussed at the meetings and newsletters with children, and that this must be consistent, because it is hard to change behaviour if the exposure is low. On the other hand, there was doubt regarding this approach, because it was perceived more likely that the child would accept a drink when offered alcohol by peers who were drinking, than think of what their parents had said at home. One teacher suggested that the influence could be greater if children would be the direct target group, e.g. training during health education classes. 188 6.3.6 Long-term continuation of the programme Participants indicated that in general, the programme was useful, and they would support its upscaling after weaknesses were addressed and values kept (Table 25). However, they indicated that encouraging parents to attend the meetings is a serious issue to tackle, as there is no point in having meetings if only a few of them participate. Parents’ and teachers’ suggestions regarding the continuation of the programme addressed four issues: target group, length, format and content, and facilitators. Table 25. Perceived values and weaknesses of the programme among parents (P) and teachers (T) Values Weaknesses  A systematic, holistic and consistent approach (P, T)  Useful and relevant topics, evidence-based information, practical examples (P, T)  Supportive, helpful and well- prepared newsletters (P, T)  External professional facilitators – had a thorough knowledge of the topics and involved parents in discussions (P, T)  Facilitators’ rotation (P, T)  Making agreements (creating a unified front) (P)  Supports existing attitudes (P)  Diverse presentation of materials (e.g. videos, pictures) (T)  Involving parents (T)  Face to face interaction (not just sending materials) (T)  Facilitated teacher-parent collaboration (T)  Supportive programme’s management – reminders, notifications, flexibility (T)  Delivered by the National Institute for Health Development (T)  The objective and the process were not well described (P, T)  Incomplete notification at the start (P, T)  Low participation rates (P, T)  Too few practical examples, which were rather general (P, T)  Examples and agreements difficult to implement in practice (P, T)  Too many slides and too much emphasis on the slides (P, T)  Too long (P, T)  Not involving the children (P, T)  Took place at the end of the workday – people were tired (P, T)  Those in need were not participating (P, T)  Facilitators did not have real-life experience in parenting (P, T)  Unsolicited format (P)  Repetitive content (P)  A minimal amount of new information (P)  The material was not tailored to individuals (P)  Not all parents received newsletters and meetings summaries (P)  Meetings were longer than initially planned (P)  Facilitators’ rotation (P)  Too infrequent (twice a year), people forgot the topics (T)  Teachers were not involved enough (T) 189 Target group Teachers expressed their concern that being involved with the programme takes extra time for them, but their involvement is crucial, and they should be included from an early stage, suggesting a separate intervention for teachers. Initially, the programme’s target group has been parents, but it was also recommended to include children, to increase the programme’s effectiveness. Most of the parents found 4–5th grade the right time to start with the programme. Parents suggested that while children might be perceived to be too young to talk about alcohol and not yet interested in the topic, they slowly begin to take examples from their friends. Thus, the issue has to be addressed, and younger children’s parents might be more interested in participating. “At first it seemed a bit funny, does my child, at that age, need to know all these things. But when you think about it more thoroughly, children at that age want to try and know everything.” (Parent #20, middle-high) Some participants suggested starting the programme in kindergarten, as that was perceived as the best time to establish rules and create the foundation: “Fundamental principles and personality characteristics will develop at an early age. After that, we are dealing with consequences. When you think about it, a five-year-old has received the fundamental principles and developed “spine”. It is challenging to change it afterwards. The foundation has to be strong.” (Teacher #5, middle-high) Teachers had similar views as parents – to start in 4–5th grade, but agreed that from a parent’s perspective, starting in 6–7th grade would be more reasonable. Those parents who favoured starting the programme in 6–7th grade, pointed out that these grades are more acceptable because at this point, experimentation with alcohol start to happen and problems become real. Length Participants’ views on the programme’s duration varied greatly. While some favoured more frequent meetings over three years (e.g. 3–4 times a year, twice in a semester), others preferred to meet less often (e.g. once a year). It was indicated that regular meetings would make it easier to remember agreements made between parents, strengthen the bond between the facilitator and participants, and the programme would be more effective if the meetings would be more concentrated in time. Some parents suggested that if the alcohol topic would 190 be addressed regularly, then maybe at some point those parents who have problems would start to attend. It was also expressed that a higher frequency might annoy parents and getting parents to attend meetings more frequently could be a great challenge. Participants who favoured meeting less often implied that this way there would be no need to make the additional meeting after the general one in the autumn. Some parents preferred to have programme meetings as separate events because the objective of the class meeting was different. “Maybe it was too frequent, or not, but still, getting parents to school twice a year, maybe one would be enough. It felt like we forcefully got them here, to continue with the project.” (Teacher #1, middle-low) As class teachers change and classes get restructured at many schools at the beginning of secondary school, participants suggested taking these factors into account when considering the perfect time to start the programme. Format and content Regarding the content, participants agreed to keep the discussed themes, but to offer more practical examples that would take into account the local situation and cover the following topics in more detail: group norms, parenting styles, energy drinks, brain development, problem solving approach, importance of parents’ role, tobacco products, alcohol advertisements, and communication with the teenager. New topics that were suggested for inclusion were illegal drugs, digital addiction, sexual health, laws (related to substance use), school’s inner life (related to the main topics covered at the meetings), and alcohol-free drinks. “For example, does the parent know, what happens when the child breaks the law, and how does it depend on the child’s age. I do not know does the parent know. I am not sure this topic was covered in this programme, that when can children legally drink alcohol, what do you lose in life when you break the law. For example, American visa and so on, it will leave a mark.” (Teacher #3, middle-low) Participants agreed that when parents are already making an effort to attend meetings, it would be reasonable to cover different relevant topics, not just alcohol. Due to low participation rates, parents suggested that the most important topics should be moved to the 191 first meetings, which would also create the theoretical framework and show parents the clear impact of the programme, so they would understand direct influences, e.g. the difference new knowledge and skills make, and what they get from this. Teachers pointed out that frequent repetition of the topic might reduce its usefulness, as parents get tired of it, and addressing the alcohol topic might increase the child’s interest, and thus, backfire. Parents expressed their concern that the content was not tailored and thus, the future approach should be more individual, involving some pre-assessment or pre-selection. Different suggestions were made about ways in which the meetings’ format and content should be changed. For example, using fewer slides, making it less formal, and the use of more illustrative materials, e.g. pictures, videos. It was also suggested that facilitators put more emphasis on discussion, use a more creative approach (e.g. hands-on activities for participants) and give extreme examples such as scare tactics, send newsletters by e-mail, and at the end of the meeting briefly introduce the topics to be covered at the next meeting. If children would be included as the target group, then it was suggested to use games and to give real-life examples when talking about alcohol’s influence and carry out active learning classes, e.g. movie night with parents, followed by a discussion. Facilitators Participants preferred external facilitators, as they were perceived to have a different point of view, more knowledge and new ideas, to be more neutral and to have more time to prepare the meetings. Parents emphasised the importance of facilitators’ knowledge, charisma and open-mindedness, ability to talk interestingly and connect with the audience. It was suggested to include different professionals (e.g. police officers, doctors, social workers) who would cover the topics, but also people who have had personal experience, e.g. former addicts, people who had “hit rock bottom”. While some parents preferred facilitators with more experience in parenting, others found it more important to comply with the above-mentioned requirements; in both ways, parents felt that they would anyway have the final say regarding their child. “Even if we would have a doctor, if he would sound boring, I would not go and listen.” (Parent #10, big-low) 192 On the other hand, including someone from the school or local area was perceived as beneficial, as the person’s knowledge of the local situation would be better. This way, the person could give a thorough overview of what is happening at the school, and there would be more flexibility in coordinating the meetings. Possible perceived options would be to combine the internal (e.g. health education teacher, psychologist, social pedagogue due to their existing knowledge on the topic) and external facilitators. If an external facilitator conveyed the same message as the teachers, it would give a stronger confirmation to parents. In general, teachers found that they could manage the training themselves, but as it would require additional work and they already have enough tasks, they were not interested in being future facilitators. They also expressed the concern that they would more likely drift away from the topic, and as they have a personal relationship with parents, parents would be less eager to ask questions. In addition, if children were added as a target group, participants recommended having younger facilitators who would interact with them more easily. 6.4 Discussion 6.4.1 Summary of main findings and key considerations This study explored the attitudes and experiences of Estonian teachers and parents regarding a parent-oriented alcohol use prevention programme they were a part of, but also their views on the alcohol use issue and its prevention among children in general. While the majority of participants perceived alcohol use in the society as a problem, their own experiences were portrayed more moderate, such as no problems among family and friends and their own alcohol use at a low level. Similarly, participants acknowledged the issue of early alcohol use onset, but considered this a problem for other families and classes, as their children mostly had not expressed interest in drinking alcohol. A qualitative study among Australian parents presents a similar perspective of parents perceiving own behaviour at an acceptable level, but others’ behaviour as more problematic (Jones, Andrews, & Berry, 2016). While alcohol supply by other parents was considered as bad parenting, providing alcohol themselves was considered as teaching children to drink responsibly. The perception of having no or low exposure from alcohol might have been one of the reasons why it was difficult for several participants to understand the necessity behind the prevention programme. Distancing oneself from the problem may suggest the presence of 193 comparative optimism, where the person can objectively assess the situation regarding other people but give an unrealistically positive view of own perspective (Dillard, Midboe, & Klein, 2009; Shepperd, Carroll, Grace, & Terry, 2002). For example, when parents of 8–12th graders were asked about the alcohol use of their children and children’s friends, parents found it more likely for children’s friends to consume alcohol (Bogenschneider et al., 1998). The results from the Effekt programme’s trial show that already at the age of 11, 25% of children had consumed alcohol in the past 12 months, and by the age of 13, the rate had increased to 46% (see Chapter 4). This indicates how both parents and teachers had a distorted view of the real situation. Although, this kind of discrepancy is not unique, as several researchers have shown how parents tend to underestimate their children’s alcohol use (Berge et al., 2015; Williams et al., 2003). However, not relating to the problem might create a situation where parents and teachers are not aware of the existing problems, which may lead to the worsening of the issue for the child (e.g. frequent alcohol use, being involved in other risk behaviours). Another reason behind distancing oneself from the problem may be related to the stigma around alcohol-related issues (Raudne, 2012; Room, 2005), and the topic might be too intimate to discuss in front of other parents, as talking about the issue could be perceived as publicly confirming of having a problem. This, in combination with the comparative optimism, may reduce people’s need for external help, both in preventing and reducing the problem and have an impact on attitudes towards such approaches. On the other hand, if alcohol use is seen as a part of growing up, this could have a similar impact. Davies (2015) shows that teachers tend to perceive alcohol use among adolescents as inevitable, and some suggest introducing safe drinking levels as a coping mechanism. Providing alcohol to adolescents has been suggested by parents, too as a way to minimise future harm (Valentine et al., 2010). When parents and teachers were asked to assess the perceived value of the current programme, they considered it more as an educational harm reduction intervention and found that people who needed the programme the most did not attend the meetings. Baker and colleagues (2011) point out that participants might find it more difficult to perceive the expected benefit when the focus of the programme is to prevent, instead of reducing harm. Rosenman, Goates, and Hill (2012) show that high-risk families are less likely to participate in universal prevention programmes, but the same implies to well-functioning families, as they may not find the need to participate. If the parents perceive their relationship with the 194 child of being of good quality and do not see alcohol as an acute issue, it is more likely they would not participate in a programme that targets alcohol use (Cohen & Rice, 1995). The issue of low participation rates throughout the programme raised by the programme’s facilitators (see Chapter 5) was also pointed out by teachers and parents as one of the main issues possibly limiting the effectiveness of the programme. The reason behind low rates appears to be a combination of different factors. For example, parental educational level is positively related to meeting attendance (Al-Halabi Diaz et al., 2006; Pettersson et al., 2009; Spoth & Redmond, 2000). Regarding practical issues, lack of time and duration of the meeting have been listed as contributing factors to lower attendance rates (Pettersson et al., 2009; Spoth & Redmond, 2000). Also, if parents are used to attending school meetings, then participation in a prevention programme is more likely (Al-Halabi Diaz et al., 2006). Having practical issues as barriers have been commonly reported as one of the main reasons behind non-participation (Mendez, Carpenter, LaForett, & Cohen, 2009; Pettersson et al., 2009; Spoth & Redmond, 2000). When parents were asked about the main barriers of not attending the intervention workshops, the work schedule conflict was reported as the most common reason (Mendez et al., 2009). Although the findings are inconclusive (Baker et al., 2011; Gross, Julion, & Fogg, 2001; Pettersson et al., 2009), being a single parent or raising the child(ren) alone might be one of the reasons why it is difficult for the parent to attend the meetings. Approximately one-third of the children who participated in the programme’s trial were from non-nuclear families (Tael-Öeren et al., 2019a). Parents as a target group are heterogeneous and vary in terms of the level of knowledge, experience, perceived relevance and open-mindedness towards the topics. Parents who are interested in obtaining new information and improving their skills are found to be more open to participating in programmes (McCurdy & Taro, 2001). On the other hand, parents who see their knowledge and skills being good enough might feel that their role as a parent is being questioned, thus feeling reluctant to participate. This indicates that some level of tailoring should be incorporated into the programme. For example, when high-school students were asked about their intention to use alcohol-related services due to friend’s or own alcohol use, the findings were dependent on the level of alcohol use (D’Amico, McCarthy, Metrik, & Brown, 2004). Students who were considered as heavy drinkers were more likely to be reluctant in using any services and students who did not use alcohol were more open in using services, if necessary. 195 Among other factors mentioned above, participants’ attitudes towards children’s alcohol use might have influenced parents’ readiness to participate. Similarly to perceiving alcohol use as an issue among others, parents’ own attitudes towards children’s alcohol use were perceived as more restrictive. Thus, the programme might be seen as more relevant to other parents. The findings from a study undertaken by Valentine and colleagues (2010) implies that parents’ attitudes might not be in direct accordance with their behaviour. Parents may not consider offering alcohol to children as alcohol use per se, as it takes place in a supervised situation and offering small amounts of alcohol to children at home is seen as a good practice to prevent children’s alcohol use with friends. Based on the findings from a study among Australian parents, it is also possible that parents themselves do not consider allowing children to sip alcohol as alcohol use (Gilligan & Kypri, 2012). But by parents allowing children to drink in their presence from time to time, and at the same time emphasising the harmfulness of alcohol sends children a mixed message and can create confusion (Bourdeau et al., 2012; Glowacki, 2016). Also, initiating alcohol use might not always be due to child’s interest, but due to their parents wish to introduce it, as they reflect on their own childhood experience or use it as a potential future harm reduction mechanism (Friese, Grube, Moore, & Jennings, 2012; Valentine et al., 2010). Despite the low participation, the programme was perceived to have a positive impact on those parents who attended the meetings and were interested in the topics and the long-term continuation of the programme after some modifications was supported by the majority of the participants. While the current programme increased the number and breadth of topics, participants pointed out that when parents are already making an effort to attend the meetings, other age-relevant issues should be addressed, e.g. digital addiction, sexual health. The variety of proposed topics shows that there is a need for additional knowledge, but it is questionable as to whether it should be addressed while talking about alcohol. Other more urgent matters might be more of interest and providing fundamental knowledge and examples that are applicable in different situations would make the programmes more valuable for parents (Valentine et al., 2010). Increasing the coverage of topics seems logical from the participants’ perspective, as it is unrealistic to imagine that each topic could have a separate programme (which appears to be rather common practice). Suggestions to expand the coverage of topics have been reported by participants in other programmes as well (West et al., 2008). This issue could be addressed by combining the 196 topics and finding overlapping content, for example, focusing more on life and social skills, and/or quality of the parent-child relationship, which are universally protective for several risk behaviours. In addition to covering more topics, participants recommended incorporating scare tactics when talking about alcohol use. It is a common belief in Estonian schools that overdramatising the consequences, giving shocking examples and involving former addicts would keep children away from alcohol (National Institute for Health Development, 2018b), despite the ineffectiveness and harmfulness of these approaches (United Nations Office on Drug and Crime, 2018). A similar attitude is shared among the general population, where 79% of 18–74-year-old Estonians were confident that by demonstrating alcohol use related negative consequences to children, alcohol consumption would reduce (Orro, Martens, Lepane, Josing, & Reiman, 2017). Another content-related aspect that was recommended to change was the one-size-fits-all approach, which did not take into account the heterogeneity of the group. While it would be possible to carry out pre-assessment and divide parents into groups, it would be organisationally very challenging to involve more than 1–2 facilitators per class. One option to have more facilitators would be to include pedagogues and specialists from participating schools. While external facilitators may have a different point of view, more knowledge and new ideas, and be more neutral, local ones have a better overview of the current situation. Irrespective of having an internal or external facilitator, what matters is the breadth of facilitator’s knowledge, communication skills and experience (Whittaker & Cowley, 2012). Another aspect to consider regarding facilitators is how to solve situations where parents undermine the facilitator’s experience due to not having children. While there were only a few occasions where the facilitators were asked about their experience as parents, the interviews with parents revealed that the issue was more prevalent. Having children seems to increase parents’ trust in facilitators’ knowledge, and it might be easier to relate to another parent than to someone who is purely theoretical, despite summarising the experiences of thousands of parents. A frequent idea proposed by both teachers and parents was to widen the target group of the sessions and include the children. For example, combining the Effekt programme with a student-oriented programme in the Netherlands showed promising results in reducing children’s alcohol use in short- and long-term, while separate interventions had no effect on alcohol use behaviour (Koning et al., 2009; Koning et al., 2011). The current research 197 supports combining parent- and student-oriented interventions to prevent and reduce children’s alcohol use; although it has been stated that further research on the programmes’ delivery (in addition to assessing the outcomes) is needed (Foxcroft & Tsertsvadze, 2011b; Newton et al., 2017). One interesting development that took place during the interviews was the change in participants’ understanding of the time when the programme should start. While several participants did not see a reason why the alcohol topic should be discussed before children begin to drink alcohol, then by the end of the interview most parents and teachers suggested starting the programme when children are 10–11-years-old. A thorough discussion might have given participants another perspective to consider and thus, justify covering the topic before the perceived need occurs. Lack of a problem makes it more difficult for the target group to grasp the expected benefit, so, introducing a broader benefit (e.g. improved parent- child communication) helps to increase the programme’s relevance. This shows how crucial it is to provide a strong first impression and well-explained rationale behind the programme (Whittaker & Cowley, 2012). 6.4.2 Limitations and strengths One of the limitations of this study relates to the low number of parents with less positive attitudes towards the programme. Initially, the intention was to include an equal number of parents with varying attitudes towards the programme to increase the breadth of opinions and the depth of the analysis. When using the recommendation score as an indicator of the attitude, four teachers and five parents had somewhat less positive attitudes towards the programme. Despite the small number, other participants pointed out different programme- related shortcomings as well, giving an input to create a comprehensive list of weaknesses to address. Another limitation was that all parents who participated were mothers. There were significantly fewer fathers attending the meetings (see Chapter 3.7), and in general, fathers tend to participate in parenting programmes less often (Spoth & Redmond, 2000; Wells, Sarkadi, & Salari, 2016). Including fathers in the study would have possibly provided an additional perspective on the programme. Lastly, participants’ views on the programme’s content indicate low acceptability of some components, such as agreements between parents, and approaches, such as repeating the content, which in turn highlights the value of thorough pre-testing of the content and format. 198 The main strength of this study is offering a more in-depth understanding of parents’ and teachers’ attitudes and experiences regarding the Effekt programme, as the programme has not been qualitatively assessed before (in Sweden or the Netherlands). Also, this study is reportedly the first one in Estonia to offer a better understanding of the reasons why parents do not attend school meetings. Another strength of this study is including participants (almost 20% of all teachers who participated in the programme) from different schools and areas (e.g. urban, rural) while taking into account the participation rates (i.e. low/high), thus increasing the variation of people’s experiences. An additional strength is involving interviewers who were not related to the programme. The interviewers emphasised this aspect to all participants to encourage them to give honest feedback, including negative opinions. Lastly, the current findings complement the findings summarised from the focus group with facilitators (see Chapter 5), offering a more holistic assessment of the programme. 6.5 Conclusions While the Effekt programme has been evaluated in several experimental studies, there have been no qualitative studies conducted to better understand participants’ views, attitudes and experiences regarding the programme’s content and delivery. The current study carried out among teachers and parents in Estonia is the first of its kind to assess that. Alcohol use, in general, was seen as a problem, but one that affects others. This point of view was also reflected in participants’ understanding of the programme’s objective, which was more often seen as a harm reduction intervention. Participants found the programme to provide new knowledge and skills that are applicable at home and school when necessary and increase the collaboration between the parents, but also between parents and teachers. Also, it was suggested that parents who attended the meetings and were interested in the discussed topics were more likely to be positively influenced by the programme. Despite that, the programme’s effectiveness was questionable as the attendance rates were low throughout the programme. The upscaling of the programme after thorough adjustments was supported by most of the teachers and parents. These findings offer possible insights regarding the target group and the programme in general when developing programmes that include parents. More emphasis should be put to the engagement process of all involved parties (both directly and indirectly), as lack of 199 understanding of the necessity may significantly reduce participants’ motivation to be engaged in programme-related activities. 200 Chapter 7. Discussion 201 7.1 Overview of the main findings This dissertation focuses on children’s alcohol use and the role of parental attitudes in preventing it and on the assessment of a parent-oriented alcohol use prevention programme, which was carried out in Estonia. Chapter 1 gives a comprehensive overview of the severity of alcohol use problem in today’s society and possible solutions to tackle it among children while focusing on parents and taking into account insights from previous research. Alcohol-related culture, attitudes and behaviour differ by countries, and therefore, it is important to adapt and test interventions before scaling up. Chapter 2 aims to investigate the relationship between parental attitudes and children’s alcohol use. This is the first systematic review and meta-analysis that included parent-child dyads, as previously no distinction has been made between perceived and self-reported parental attitudes. A meta-analysis of 30 studies showed that less restrictive parental attitudes were related to children’s alcohol use initiation, frequency of use, and drunkenness. Perceived lenient attitudes were related to children’s alcohol use initiation and frequency. The correlation between perceived and self-reported parental attitudes was weak, implying that children are not fully aware of their parents’ attitudes towards alcohol use. A comprehensive moderator analysis identified one factor that affected the relationship between parental attitudes and children’s alcohol use frequency – child’s age. It was recommended to use more holistic approach in future studies, on the one hand exploring the mutual and combined influences of parental factors on children’s alcohol use, and on the other hand exploring the influence of children’s alcohol use on parental factors. Chapters 3, 4, 5 and 6 seek to address the issue raised in the introduction, by describing the adaptation, delivery and evaluation of a parent-oriented programme in Estonia. Chapter 3 aims to describe the adaptation and delivery process of the parent-oriented programme that aimed to influence children’s alcohol use by targeting parental attitudes. The content of the three-year programme was conveyed through meetings and materials (i.e. newsletters, summaries, leaflets), that addressed alcohol-related topics. After the first two meetings, the content was modified (without changing the main messages and the format of the programme), and new topics on parenting and other avenues of children’s alcohol use 202 (e.g. impact on the brain development) were introduced, as participants showed reluctance towards the programme content’s repetitive nature. Several other modifications were made to the programme, to support its delivery, e.g. introducing active learning based methods, demo trainings by facilitators before each meeting wave and collecting feedback from the participants. The systematic collection of attendance data shows that participation rates of the meetings were less than 50% throughout the programme and reduced over the years, dropping to 14% at the final meeting. The feedback collected from parents at the end of the meetings showed participants’ satisfaction with the programme related activities and the facilitators. Chapter 4 assessed the programme’s effectiveness in delaying alcohol use initiation and reducing alcohol use among children. A matched-pair cluster RCT was carried out among 66 schools (34 intervention, 32 control) among 985 fifth grade students and their parents (n=790) in 2012–2015. The findings showed that there was no difference between the intervention and control group students regarding alcohol use initiation, drunkenness, past year use and parental alcohol supply. While there was no evidence of a change in children’s perceptions regarding their parents’ attitudes, parents in intervention schools had approximately twice the odds of having restrictive attitudes at the end of the programme. The results from this trial are in line with similar studies conducted in the Netherlands and Sweden, which show that there is a gap between parental attitudes and children’s behaviour. Additionally, the data collected from parents at the second follow-up showed that only 17% had attended all meetings that had taken place at their schools. While most of the parents found it important to carry out preventive activities that raise parents’ awareness of children’s alcohol use, it was suggested that these activities would be more beneficial for other parents. The importance of qualitative assessment was emphasised, as feedback from teachers and parents can give helpful insights on possible barriers. Implications for future interventions included combining parent and child interventions, turning the focus from alcohol to social and life skills, and shifting the start of the programme to earlier grades. Chapter 5 introduced the programme’s facilitators’ views on the adaptation and delivery of the programme, and their perception of the programme’s effectiveness, strengths and weaknesses. The qualitative findings from the focus group showed that the facilitators perceived the programme to be effective in changing parental attitudes, but its effect on children’s alcohol use was questionable. Despite various challenges, the facilitators had to 203 face, such as organisational, content-, parent- and teacher-related challenges, they welcomed the idea to scale up the programme if the data from the trial supports it. The facilitators suggested introducing modifications, such as renewing the content and increasing collaboration with the teachers. This study offers an interesting perspective on the programme’s evaluation, as the facilitators were involved in both the adaptation and delivery of the programme. Chapter 6 aims to assess the programme from the qualitative perspective of teachers and the programme’s main target group, parents. Participants shared their views on the content, delivery and perceived effectiveness of the programme but also their attitudes related to children’s alcohol use. The findings indicate that the issue of low participation rates was seen as one of the main barriers in achieving the expected outcome. Participants provided various reasons that might have influenced this, e.g. participants finding the topic irrelevant as the children were perceived not to have had an experience with alcohol, low engagement of teachers, the general practice of not attending school meetings. While the programme aimed to prevent children’s alcohol use, many participants perceived it as a harm reduction programme targeting high-risk families. On the other hand, parents who had positive attitudes towards the programme found it useful as it increased their awareness and provided practical skills. Similarly to the facilitators, both parents and teachers supported upscaling the programme after some modifications have been introduced, e.g. widening the scope of topics, involving children, tailoring the content. In combination with the findings from the focus group study with the facilitators, this research provides potential explanations for why the programme was ineffective in influencing children’s alcohol use and adds two additional perspectives from the groups of people who were involved with the programme. 7.2 Findings in the context of the literature Prior to the systematic review and meta-analysis presented in this thesis, three reviews had assessed the relationship between parental attitudes and children’s alcohol use (Ryan et al., 2010; Sharmin et al., 2017b; Yap et al., 2017). While all three showed the presence of a relationship at some level, the findings, in general, were mixed. One of the reasons behind this might be that the perceived and self-reported parental attitudes were combined in the analysis. The current review focused on both self-reported and perceived parental attitudes and included studies only with parent-child dyads. All main findings had some overlap with previous reviews; however, different definitions of levels of alcohol use were used. For 204 example, while this review considered initiation as drinking alcohol for the first time, other reviews combined all alcohol use that took place before the age of 15 as alcohol use initiation (Ryan et al., 2010; Yap et al., 2017). The findings from the meta-regression suggest that as the children get older, the relationship between parental attitudes and children’s alcohol use frequency becomes weaker. Previous studies have shown that parents tend to become more lenient and less controlling over time (Koutakis et al., 2008; Mares et al., 2012; Piko & Balázs, 2012; Prins et al., 2011). The findings from studies that have confirmed the relationship between parental factors and children’s alcohol use have been heavily drawn on when developing parent-oriented alcohol use programmes. While there is a large number of such programmes that have been developed and carried out over the past few decades, only a minority of the programmes have been delivered and evaluated in the Northern- and Eastern Europe. Effekt is one such programme. While the original version of the Effekt programme that was developed in Sweden was found to have changed parental attitudes and reduce alcohol use among children, subsequent evaluations found similar effects on parental restrictive attitudes, but there was no evidence of the effect translating into children’s alcohol use (Bodin & Strandberg, 2011; Koning et al., 2010b; Strandberg & Bodin, 2011; Tael-Öeren et al., 2019b). The majority of the facilitators, parents and teachers of the Effekt programme, when evaluated in Estonia as reported in this thesis, reinforced this finding, suggesting it is more likely that the programme has an effect on parents and not children. This could mean that the change in parents’ attitudes was not big enough to have an effect on children’s alcohol use, but also that there may be other factors that affected the outcome. For example, Koning and colleagues (2010b) showed that when the programme was combined with a student-oriented programme, it showed effectiveness. Findings from different reviews indicate the same and show that these programmes can have a positive lasting effect on children’s alcohol use (Newton et al., 2017; Van Ryzin et al., 2016). Involving children was also suggested by the teachers and parents who took part in the Effekt programme. At the same time, an alcohol use prevention programme that was carried out among children and their parents in Norway and addressed parental attitudes, showed no significant differences between intervention and control group students regarding their alcohol use (Strøm et al., 2015). Another prevention programme from Sweden that targeted 205 only parents maintained parental restrictive attitudes and decreased alcohol use (Pettersson et al., 2011). While there are alcohol use prevention programmes that are home-based and do not involve direct face-to-face contact with the parents (Dickinson, Hayes, Jackson, Ennett, & Lawson, 2014; Mares et al., 2011b), the majority of the parent-oriented programmes are meant to be delivered by teachers or external facilitators. This, in turn, assumes that parents turn up to attend the programme-related activities, such as meetings. One of the major problems in the Effekt programme in Estonia was related to attendance, as the number of participants at the meetings stayed low on a regular basis. While it is not possible to compare the rate of attendance with other Effekt studies, as no data were collected on the number of participants at each meeting in other evaluations, there are other programmes that have faced the same problem (McKay et al., 2018; Pettersson et al., 2009; Skärstrand et al., 2014). The reasons behind low attendance suggested by the Effekt programme’s participants broadly align with the ones presented in other studies, e.g. perception of need is underrated, sensitive topic, low motivation among teachers, practical issues, the habit of attending school meetings (Al- Halabi Diaz et al., 2006; Hahn et al., 2002; Jones et al., 2016; Mendez et al., 2009; Pettersson et al., 2009; Raudne, 2012; Spoth & Redmond, 2000). Many participants perceived alcohol use in society as a problem, affected by its availability and exposure by the media. A similar conclusion was reached by the Croatian parents who took part of the Project Northland programme and who highlighted that alcohol has its place in the Croatian culture which makes the prevention work more challenging (West et al., 2008). It is important to take into account the local context when bringing a programme from one country or setting to another. Delivering Effekt in Estonia required adaptations, and while there is a risk in decreasing the programme’s delivery fidelity (Botvin, 2004), the core components of the programme were not changed and the new content and activities introduced were evidence-based. Stigler, Perry, Komro, Cudeck, and Williams (2006) conducted a component analysis to distinguish between the effectiveness of five different components in Project Northland that involved the school, home and community. The findings indicated that only three out of the five components were effective in lowering the tendency to drink alcohol and increasing self- efficacy: classroom curriculum that addressed parent-child communication and self-efficacy, carrying out alcohol-free activities (but only for the youth who planned the activities) and 206 involving parents. Other researchers have identified the effective elements of parent-based programmes as well and concluded that the combination of general and alcohol-specific parenting factors, skills development, increasing active parental involvement, having frequent sessions that in total would not last too long, using digital and face-to-face interactions and including youth as a target group help to reduce alcohol use among children (Allen et al., 2016; Bo et al., 2018; Cairns et al., 2014; Kuntsche & Kuntsche, 2016; Petrie et al., 2007; Smit et al., 2008; Van Ryzin et al., 2016). Although the Effekt programme included many of these elements, there was no conclusive evidence on the intervention effect on children’s alcohol use. 7.3 Limitations The research presented in this thesis has several limitations which are related to the adaptation and delivery of the prevention programme. The input from parents, teachers and facilitators implies that the piloting undertaken before the first meeting was most likely too superficial. As the parents from two schools who attended the pilot meetings had only minor suggestions regarding the content, it is possible that the facilitators did not manage to convey the whole nature of the programme, including its repetitive content. Also, the short piloting period did not enable the project team to gain a more profound knowledge of the programme’s potential acceptability, target group and possible meeting related barriers that may occur. The assumption was that the target group lacks alcohol-specific awareness. While the focus of piloting was on parents’ views on the programme’s content, teachers as a separate target group were not paid enough attention. The interaction with the teachers was too brief, and the project team did not have proper knowledge of teachers’ thoughts on the programme. Limited piloting resulted in modifying the content of the programme in the middle of the delivery process, which in turn increased uncertainty among facilitators and participants, as the new content was developed on an ongoing basis, not all at once. For future work, more time should be planned before starting this kind of programme. While modifying the programme’s content reduces its comparability with the programmes conducted in the Netherlands (although they reduced the number of meetings) (Koning et al., 2009) and Sweden (Koutakis et al., 2008; Bodin & Strandberg, 2011), the core of the programme was kept unchanged. Additionally, the new content did not only address 207 increasing knowledge but also improving skills, an approach suggested to be more effective than just providing information (Stockings et al., 2016). Taking into account the findings from the qualitative study, it is reasonable to believe that the change in the content was worthwhile and contributed to the outcomes (at least parent-related ones). Another limitation relates to low participation rates and some parents deliberately avoiding attending the class meetings or part of the meetings where the programme was covered. While several studies on parent-based programmes have reported a similar problem with low attendance rates (McKay et al., 2018; Pettersson et al., 2009; Spoth & Redmond, 2000; Whittaker & Cowley, 2010), it seems to be a more general problem at the schools as the children get older. Additionally, parents’ participation rates in the trial were also poor, as less than half of the parents filled the questionnaires. The answers from parents who were willing to take part in the trial are reasonably similar to the trends in the society at large, but despite that, the results may not be fully generalisable. While the importance of a thorough process evaluation that should accompany the delivery of an intervention has been emphasised by several researchers (Craig et al., 2008; European Monitoring Centre for Drugs and Drug Addiction, 2012; Moore et al., 2015), no specific framework informed the evaluation of the planning, adaptation or delivery of the current programme. For example, the aim of the pilot study was to assess the acceptability of the content among the target group, but ideally, the approach should be broader to understand the possible facilitators and barriers of the delivery (Moore et al., 2015). Also, the impact of external factors that could have directly and indirectly influenced the programme was not investigated. Although the quantitative and qualitative measures developed by the project team give a reasonable overview of the whole process, adhering to an existing scientific framework would have provided a more in-depth understanding of the programme and the causal mechanisms behind it. Similarly, following the guidelines of high-quality prevention (Brotherhood & Sumnall, 2011) could have made the adaptation process more systematic. A further minor limitation is not describing the programme content in terms of behaviour change techniques (BCTs; Michie et al., 2013). BCTs refer to the most irreducible and discrete techniques used to affect behaviour change and represent the potential “active ingredients” of an intervention. When the Effekt programme was adapted, this taxonomy was not yet in wide use, and so elements of the programme were not constructed using the BCT “language”. Furthermore, although the content of the programme was fixed, there was some 208 flexibility in how the content was delivered by facilitators, and it would be difficult to reliably assess which BCTs were delivered. For future intervention development in this area, it will be important to include frameworks such as the BCT taxonomy to better describe and assess what and how programme content is delivered. While it is common for a PhD student to be a part of the research project, it is less common to be responsible for the whole project and carry out a large-scale evaluation. One could argue that the involvement of the thesis author in the programme delivery and evaluation raises the issue of experimenter bias, which suggests that the outcomes of the research might be influenced by her expectations (Rosenthal & Rubin, 1978). Similarly, Moore and colleagues (2015) suggest that evaluation should be carried out independently to keep the credibility of the evaluation. To reduce the bias, all stages of the programme from piloting to analysing the quantitative and qualitative data should involve independent researchers who do not take part in the delivery of the programme, increasing the likelihood of presenting the results from a neutral point of view. 7.4 Strengths The systematic review and meta-analysis is known to be the first one to focus on child-parent dyads when investigating the possible link between self-reported parental attitudes and children’s alcohol use. A thorough search was conducted across several databases to identify studies that reported results from both experimental and non-experimental studies. The quality of all the included studies was assessed by using previously developed guidelines. Meta-analyses were conducted across a range of alcohol use measures, and additional analyses provided insight on possible factors that might moderate the relationship between attitudes and alcohol use. As the results from the analyses indicated a relationship between parental attitudes and children’s alcohol use, it gives support to the idea of addressing parental attitudes as one of the parental factors when preventing and reducing children’s alcohol use. One of the strengths related to the prevention programme that was the first substance use prevention programme in Estonia targeting parents, is an in-depth description of its adaptation, delivery and assessment. The programme was supported by rigorous evaluation – a matched-pair cRCT with two follow-up assessments that was carried out among children and parents, and qualitative studies with the programme’s participants and facilitators at the 209 end of the programme. Also, feedback was collected from the parents to assess the delivery of the programme on a rolling basis, and the number of participants attending the meetings was counted to get an objective overview of the attendance rates. The findings from the studies have provided input in the development of other preventive measures and strategies on the local level. The adaptation of the programme was systematic, informed by the theory and empirical findings from the research literature and aimed to meet the needs presented by the target group (e.g. through feedback). Increased attention on the facilitators (e.g. demo trainings, providing materials to increase knowledge) supported the programme delivery by ensuring the quality of it was similar across the schools. Another strength relates to the qualitative studies that provided different perspectives on the programme and the topic of alcohol in general from the programme’s facilitators, teachers and parents. This programme has not been evaluated qualitatively in other countries where it has been delivered. The in-depth findings from the qualitative study gave valuable insights regarding the Effekt programme as a whole but also enabled a critical assessment of different components separately. For example, while the participants found that agreements between parents were useful in principle, they were perceived as difficult to implement in practice, as the attendance rates at the meetings were low and the communication between parents minimal. Additionally, the findings from the study raised issues that go beyond the current programme, such as generally low attendance rates at the meetings, limited knowledge of the concept of prevention. In general, qualitative assessment of substance use prevention programmes seems to be carried out infrequently (especially in the form of peer-reviewed publications), and presenting the work to the public is a small step towards making the practice more regular. 7.5 Practical implications Future research could look into the possibility of a reverse relationship of children’s alcohol use having an impact on parental attitudes and investigate how parental factors influence each other and what is the combined effect of the factors on children’s alcohol use. Although parental attitudes are related to children’s alcohol use, targeting primarily the attitudes in a prevention programme does not seem to be effective, at least in Estonia where 210 alcohol use rates among children are high, meeting attendance low and communication between parents minimal. One option to increase the likelihood of a programme’s effectiveness is to combine student- and family-based programmes (Foxcroft & Tsertsvadze, 2011b) and additionally widen the scope of the topics included as many programmes are substance-specific, and the underlying risk factors of different substance use behaviours frequently overlap (Hawkins et al., 1992). The number of programmes targeting the health of children is high, and while the programmes cover a wide array of topics, they often share similar concepts (Peters et al., 2009). Thus, combining the basics of the programmes would offer a more comprehensive way to tackle substance use in general and make it more appealing for parents to take part of the programmes, as it is unreasonable to expect that parents can take part of all the programmes and planned activities. Several reviews (e.g. Kuntshce & Kuntsche, 2016; Van Ryzin et al., 2016) have reported on effective programme components, and this knowledge can be incorporated into improving existing programmes and their delivery. The latest findings from Gilligan and colleagues (2019) indicate that the evidence on the effectiveness of family-based programmes is inconclusive. Replication of these prevention programmes can be complicated by the local context (e.g. alcohol use rates, social norms), but also by the situation on different parenting-related factors. For example, if the overall parent-child relationship quality is low, it is somewhat unrealistic to expect that the child listens to the parent regarding alcohol. Thus, thorough piloting and a proper assessment of the situation and target group (including the difference in their experience, practice, knowledge, expectations and needs) by an existing scientific framework should be conducted before the programme is delivered, as it may have a considerable effect on the participation rates and expected outcomes. Additionally, in-depth process evaluation should accompany the planning and delivery of the programme. Providing a more integrated approach also increases the chance to get long-term funding for the implementation. There are some alcohol use prevention programmes that have been effective in delaying and reducing children’s alcohol use, but this is only a part of the bigger picture. As one of the teachers in the qualitative study pointed out, alcohol use prevention cannot be put on the shoulders of one person or institution. The global strategy to reduce the harmful use of alcohol outlines ten areas of policies and interventions, which mutually reinforce each other and create a holistic approach to reduce alcohol use (World Health Organization, 2010). 211 Increasing awareness is one of the ten areas, although when assessing the effectiveness and cost-effectiveness of various policies and measures which aim to reduce alcohol-related harm, traditional awareness-raising approach alone is generally considered ineffective, as it does not lead to a lasting behaviour change (Anderson, Chisholm, & Fuhr, 2009; Faggiano, Minozzi, Versino, & Buscemi, 2014). The most promising results in reducing alcohol-related harm were shown by environmental approaches, e.g. reducing availability and exposure in media, increasing the price, although these measures are often of a regulatory kind, meaning that there has to be political will to change the laws. Increasing alcohol-specific awareness can support other policy measures, as it might help to steer the pro-alcohol social norm towards opposite direction, but the main focus in prevention should be on the development of universal life and social skills, as they can have a positive impact across different health and risk behaviours. While the researchers might have a good overview of the evidence, the knowledge does not always transfer to teachers, whose primary focus at schools is not on alcohol use prevention. Thus, finding a quick and effective solution for a complex problem without external help is difficult, and the use of ineffective measures is widely prevalent. Teachers’ interest in prevention in Estonia has started to increase (this, for example, led to the delivery of the Effekt programme), but the knowledge on evidence-based approaches still seems to be scarce. A focus group study among prevention field practitioners in Estonia supported the findings of the current study, showing that there is no single understanding of prevention concepts among practitioners, and activities that are carried out are often ineffective (Abel- Ollo & Streimann, 2017). Thus, to promote more extensive use of evidence-based approaches and better allocation of resources, more attention should be put on building a strong foundation in prevention among practitioners, as their attitudes can shape the attitudes of the participants. Last but not least, one of the main concerns in parent-oriented programmes relates to low attendance. To tackle this issue, two aspects should be considered – how to get parents initially to be involved in the programme and how to keep them engaged throughout the programme. Axford, Lehtonen, Kaoukji, Tobin, and Berry (2012) recommend paying attention to the following aspects when engaging participants in parenting programmes: having a transparent and well-thought recruitment process that is part of the programme, incentivising both recruitment and retention, using creative solutions to engage parents, 212 establishing good relationships between facilitators and participants, making the programme easily accessible and addressing the target group’s needs (paying special attention to high- risk families), having realistic expectations on attendance and addressing the barriers, engaging all related parties (also people who are indirectly related) from early on and facilitating collaboration and clear communication. Addressing the needs of the target group was also recommended by the parents and teachers who participated in the qualitative study – many of them perceived the programme as aiming to reduce harm rather than preventing initiation and not providing a possible direct benefit. While several parent-based programmes are meeting-based, having regular meetings may not be the optimal way to keep parents engaged and other options should be explored. For example, the number of digital interventions available has been increasing (Serbanati, Ricci, Mercurio, & Vasilateanu, 2011) and digital interventions targeting adolescents, young adults and parents have shown to be effective in reducing alcohol use (Voolma, 2017; Wurdak, Kuntsche, & Wolstein, 2016). The use of digital solutions, such as e-mails and text messages, may help to maintain participants’ regular engagement in the programme (Alkhaldi et al., 2016). Addressing the participation issue, in general, is crucial, because carrying out an intervention typically needs a considerable amount of financial resources. Furthermore, low attendance makes it more likely for the programme to be less cost-effective. The findings from the qualitative study with parents and teachers in Chapter 6 showed that parents as a group are heterogeneous, having different experiences, knowledge and needs. Also, participants indicated that parents are more likely interested in improving their knowledge and skills when their children are younger, and their engagement with school tends to drop after elementary school. By using these insights, an option would be to approach parents when children start first grade, and parents are more open to external support. Instead of providing the same content to all parents at the same time, parents could take part in an assessment of parenting skills and based on the results receive recommendations to specific modules (such as the combination of a digital and face to face approach). The modules would create and/or strengthen the foundation of parenting skills in earlier grades and gradually introduce age-appropriate topics. This approach could be developed into a tailored whole-school approach, rather than a programme targeting one specific behaviour, to target individual needs. This, in turn, could increase parents’ interest in 213 engaging with the approach, and also improve the collaboration between the parents and the school (Helgøy & Homme, 2017). 7.6 Conclusions This thesis investigated the relationship between self-reported parental attitudes towards children’s alcohol use and their children’s alcohol use and assessed a universal parent- oriented alcohol use prevention programme, which was carried out in Estonia in 2012–2015. The findings presented show that less restrictive parental attitudes are related to higher alcohol use among children, but targeting parental attitudes in a prevention programme showed no evidence of the programme’s effectiveness in delaying and reducing alcohol use among children. The qualitative input provided by the programme’s facilitators, teachers and parents showed that the participants found the programme to impact only parents and not children, whose behaviour change was the main aim of the programme. The main perceived barrier of the programme related to participation rates, which stayed low throughout the programme. While the participants supported the programme’s long-term continuation with some modifications, the final decision by the National Institute for the Health Development was not to continue the programme due to the results from the trial study and the general issue of low attendance at the meetings. The current research gives a good example of how bringing a programme from one cultural context to another might result in considerable adaptation and that targeting parental attitudes to change children’s alcohol use might not always work. 214 References Aas, H., Jakobsen, R., & Anderssen, N. (1996). Predicting 13-year-olds’ drinking using parents’ self-reported alcohol use and restrictiveness compared with offspring’s perception. Scandinavian Journal of Psychology, 37, 113–120. doi:10.1111/j.1467-9450.1996.tb00644.x Aasvee, K., & Minossenko, A. (2011). Eesti kooliõpilaste tervisekäitumise uuring. 2009/2010. õppeaasta. Tabelid [Health Behaviour in School-aged Children survey in Estonia. 2009/2010. Tables]. Tallinn, Estonia: National Institute for Health Development. Aasvee, K., & Rahno, J. (2015). Eesti kooliõpilaste tervisekäitumise uuring 2013/2014. õppeaasta. Tabelid [Health Behaviour in School-aged Children survey in Estonia. 2013/2014. Tables]. Tallinn, Estonia: National Institute for Health Development. Abar, C., Jackson, M., Colby, M., & Barnett, P. (2015). Parent-child discrepancies in reports of parental monitoring and their relationship to adolescent alcohol-related behaviors. Journal of Youth and Adolescence, 44, 1688. doi:10.1007/s10964-014-0143-6 Abel-Ollo, K., & Streimann, K. (2017). Ennetuse õppekava rakendamine Euroopas. Eesti fookusgruppide kokkuvõte [Implementing a prevention training curriculum in Europe. Summary of Estonian focus groups]. Tallinn, Estonia: National Institute for Health Development. Abernathy, K., Judson Chandler, J., & Woodward, J. J. (2010). Alcohol and the prefrontal cortex. International Review of Neurobiology, 91, 289–320. doi:10.1016/S00747742(10) 91009-X Adolfsen, F., Strøm, H. K., Martinussen, M., Handegård, B. H., Natvig, H., Eisemann, M., & Koposov, R. (2017). Parent participation in alcohol prevention: Evaluation of an alcohol prevention programme. Nordic Studies on Alcohol and Drugs, 34, 456–470. doi:10.1177/14 55072517732276 Ajzen, I., & Fishbein, M. (2005). The influence of attitudes on behaviour. In Albarracin, D., Johnson, B. T., & Zanna, M. P. (Eds.), The handbook of attitudes (pp. 173–221). Mahwah, NJ: Lawrence Erlbaum Associates, Inc. Al-Halabi Diaz, S., Secades-Villa, R., Perez, J. M. E., Fernandez-Hermida, J. R., Garcia- Rodriguez, O., & Crespo, J. L. C. (2006). Family predictors of parent participation in an 215 adolescent drug abuse prevention program. Drug and Alcohol Review, 25, 327–331. doi:10.1080/09595230600741149 Alkhaldi, G., Hamilton, F. L., Lau, R., Webster, R., Michie, S., & Murray, E. (2016) The effectiveness of prompts to promote engagement with digital interventions: A systematic eeview. Journal of Medical Internet Research, 18, e6. doi:10.2196/jmir.4790 Allen, D., Coombes, L., & Foxcroft, D. R. (2008). Preventing alcohol and drug misuse in young people: Adaptation and testing of the Strengthening Families Programme 10–14 (SFP10–14) for use in the United Kingdom. Oxford, UK: Oxford Brookes University. Allen, M. L., Garcia-Huidobro, D., Porta, C., Curran, D., Patel, R., Miller, J., & Borowsky, I. (2016). Effective parenting interventions to reduce youth substance use: A systematic review. Pediatrics, 138, e20154425. doi:10.1542/peds.2015-4425 Altman, D. G., & Bland, J. M. (2011). How to obtain the confidence interval from a P value. BMJ, 343, 1–2. doi:10.1136/bmj.d2090 Anderson, P., Chisholm, D., & Fuhr, D. C. (2009). Effectiveness and cost-effectiveness of policies and programmes to reduce the harm caused by alcohol. The Lancet, 373, 2234–2246. doi:10.1016/S0140-6736(09)60744-3 Anderson, K. G., Grunwald, I., Bekman, N., Brown, S. A., & Grant, A. (2011). To drink or not to drink: Motives and expectancies for use and nonuse in adolescence. Addictive Behaviors, 36, 972–979. doi:10.1016/j.addbeh.2011.05.009 Andrews, J. A., Hops, H., Ary, D. V, Tildesley, E., & Harris, J. (1993). Parental influence on early adolescent substance use: Specific and non-specific effects. The Journal of Early Adolescence, 13, 285–310. doi:10.1177/0272431693013003004 Arain, M., Haque, M., Johal, L., Mathur, P., Nel, W., Rais., A., … Sharma, S. (2013). Maturation of the adolescent brain. Neuropsychiatric Disease and Treatment, 9, 449–461. doi:10.2147/NDT.S39776 Ary, D. V, Tildesley, E., Hops, H., & Andrews, J. (1993). The influence of parent, sibling, and peer modeling and attitudes on adolescent use of alcohol. International Journal of the Addictions, 28, 853–880. doi:10.3109/10826089309039661 216 Axford, N., Lehtonen, M., Kaoukji, D., Tobin, K., & Berry, V. (2012). Engaging parents in parenting programs: Lessons from research and practice. Children and Youth Services Review, 34, 2061–2071. doi:10.1016/j.childyouth.2012.06.011 Babor, T. F., & Robaina, K. (2013). Public health, academic medicine, and the alcohol industry’s corporate social responsibility activities. American Journal of Public Health, 103, 206–214. doi:10.2105/AJPH.2012.300847 Baker, C. N., Arnold, D. H., & Meagher, S. (2011). Enrollment and attendance in a parent training prevention program for conduct problems. Prevention Science, 12, 126–138. doi:10.1007/s11121-010-0187-0 Baldus, C., Thomsen, M., Sack, P. M., Bröning, S., Arnaud, N., Daubmann, A., & Thomasius, R. (2016). Evaluation of a German version of the Strengthening Families Programme 10–14: A randomised controlled trial. The European Journal of Public Health, 26, 953–959. doi:10.1093/eurpub/ckw082 Bandura, A. (1977). Social learning theory. Englewood Cliffs, NJ: Prentice Hall. Basic Schools and Upper Secondary Schools Act. (2010, June 6). With amendments as of 13.03.2019. RT I 2010, 41, 240. Baumrind, D. (1991). The influence of parenting style on adolescent competence and substance use. Journal of Early Adolescence, 11, 56–95. doi:10.1177/0272431691111004 Baumrind, D. (1967). Child care practices anteceding three patterns of preschool behavior. Genetic Psychology Monographs, 75, 43–88. Bava, S., & Tapert, S. F. (2010). Adolescent brain development and the risk for alcohol and other drug problems. Neuropsychology Review, 20, 398–413. doi:10.1007/s11065-010-9146- 6 Baxter, L. A., Bylund, C. L., Imes, R., & Routsong, T. (2009). Parent-child perceptions of parental behavioral control through rule-setting for risky health choices during adolescence. Journal of Family Communication, 9, 251–271. doi:10.1080/15267430903255920 Beard, E., Dienes, Z., Muirhead, C., & West, R. (2016). Using Bayes factors for testing hypotheses about intervention effectiveness in addictions research. Addiction, 111, 2230– 2247. doi:10.1111/add.13501 217 Bellis, M. A., Morleo, M., Hughes, K., Downing, J., Wood, S., Smallthwaite, L., & Cook, P. A. (2010). A cross-sectional survey of compliance with national guidance for alcohol consumption by children: Measuring risk factors, protective factors and social norms for excessive and unsupervised drinking. BMC Public Health, 10, 547. doi:10.1186/1471-2458- 10-547 Bellis, M. A., Quigg, Z., Hughes, K., Ashton, K., Ferris, J., & Winstock, A. (2015). Harms from other people’s drinking: An international survey of their occurrence, impacts on feeling safe and legislation relating to their control. BMJ Open, 5, e010112. doi:10.1136/bmjopen-20 15-010112 Belloc, N. B. (1973). Relationship of health practices and mortality. Preventive Medicine, 2, 67–81. doi:10.1016/0091-7435(73)90009-1 Belloc, N. B., & Breslow, L. (1972). Relationship of physical health status and health practices. Preventive Medicine, 1, 409–421. doi:10.1016/0091-7435(72)90014-X Berge, J., Sundell, K., Öjehagen, A., Höglund, P., & Håkansson, A. (2015). Parental awareness of substance use among adolescents in a junior high school sample. Journal of Drug Issues, 45, 263–278. doi:10.1177/0022042615580989 Botvin, G.J. (2004). Advancing prevention science and practice: Challenges, critical issues, and future directions. Prevention Science, 5, 69–72. Blueprints for Healthy Youth Development. (2019a). Providing a registry of experimentally proven programs. Retrieved September 2, 2019, from https://www.blueprintsprograms.org Blueprints for Healthy Youth Development. (2019b). Guiding Good Choices. Retrieved September 2, 2019, from https://www.blueprintsprograms.org/programs/guiding-good-choice s/ Blueprints for Healthy Youth Development. (2018). Effekt. Retrieved May 2, 2018, from https://www.blueprintsprograms.org/factsheet/effekt Bo, A., Hai, A. H., & Jaccard, J. (2018). Parent-based interventions on adolescent alcohol use outcomes: A systematic review and meta-analysis. Drug and Alcohol Dependence, 191, 98– 109. doi:10.1016/j.drugalcdep.2018.05.031 218 Bobrowski, K. J., Pisarska, A., Ostaszewski, K., & Borucka, A. (2014). Effectiveness of alcohol prevention program for pre-adolescents. Psychiatria Polska, 48, 527–539. Bodin, M. C. (2012). Does the Örebro Prevention Programme prevent youth drinking? A response to Özdemir & Stattin. Addiction, 107, 1706–1708. doi:10.1111/j.1360-0443.2012.0 3959.x Bodin, M. C., & Strandberg, A. K. (2011). The Örebro prevention programme revisited: A cluster-randomized effectiveness trial of programme effects on youth drinking. Addiction, 106, 2134–2143. doi:10.1111/j.1360-0443.2011.03540.x Bogenschneider, K., Wu, M-Y., Raffaelli, M., & Tsay, J. C. (1998). "Other teens drink, but not my kid": Does parental awareness of adolescent alcohol use protect adolescents from risky consequences? Journal of Marriage and Family, 60, 356–373. doi:10.2307/353854 Bonomo, Y., Norman, A., Biondo, S., Bruno, R., Daglish, M., Dawe, S., ... Castle, D. (2019). The Australian drug harms ranking study. Journal of Psychopharmacology, 33, 759–768. doi:10.1177/0269881119841569 Borenstein, M., Hedges, L., Higgins, J. P. T., & Rothstein, H. (2009a). Identifying and quantifying heterogeneity. In Borenstein, M., Hedges, L., Higgins, J. P. T., & Rothstein, H. (Eds.), Introduction to meta-analysis (pp. 107–125). Chichester, UK: John Wiley & Sons Ltd. Borenstein, M., Hedges, L.V., Higgins, J. P. T, & Rothstein H. R. (2009b). Meta-regression. In Borenstein, M., Hedges, L., Higgins, J. P. T., & Rothstein, H. (Eds.), Introduction to meta- analysis (pp. 187–203). Chichester, UK: John Wiley & Sons Ltd. Borenstein, M., Rothstein, D., & Cohen, J. (2005). Comprehensive Meta Analysis: A computer program for research synthesis [Computer software]. Englewood, NJ: Biostat. Bourdeau, B., Miller, B., Vanya, M., Duke, M., & Ames, G. (2012). Defining alcohol- specific rules among parents of older adolescents: Moving beyond no tolerance. Journal of Family Communication, 12, 111–128. doi:10.1080/15267431.2011.561140 Brain, K., Parker, H., & Carnwath, T. (2000). Drinking with design: Young drinkers as psychoactive consumers. Drugs: Education, Prevention and Policy, 7, 5–20. doi:10.1080/d ep.7.1.5.20 219 Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3, 77–101. doi:10.1191/1478088706qp063oa Brody, G. H., Ge, X., Katz, J., & Arias, I. (2000). A longitudinal analysis of internalization of parental alcohol-use norms and adolescent alcohol use. Applied Developmental Science, 4, 71–79.doi:10.1207/S1532480XADS0402_2 Brotherhood, A., & Sumnall, H. (2011). European drug prevention quality standards. A manual for prevention professionals. Lisbon, Portugal: European Monitoring Centre for Drugs and Drug Addiction. Brown, S. A., McGue, M., Maggs, J. L., Schulenberg, J. E., Hingson, R., Swartzwelder, S., … Murphy, S. (2008). A developmental perspective on alcohol and youths 16 to 20 years of age. Pediatrics, 121, S290–S310. doi:10.1542/peds.2007-2243D Brown, S. A., Tapert, S. F., Granholm, E., & Delis, D. C. (2000). Neurocognitive functioning of adolescents: Effects of protracted alcohol use. Alcoholism: Clinical and Experimental Research, 24, 164–171. doi:10.1111/j.1530-0277.2000.tb04586.x Burdzovic Andreas, J., & Jackson, K. M. (2015). Adolescent alcohol use before and after the high school transition. Alcoholism: Clinical and Experimental Research, 39, 1034–1041. doi:10.1111/acer.12730 Čablová, L., Pazderková, K., & Miovský, M. (2014). Parenting styles and alcohol use among children and adolescents: A systematic review. Drugs: Education, Prevention and Policy, 21, 1–13. doi:10.3109/09687637.2013.817536 Cairns, G., Purves, R., & McKell, J. (2014). Combining school and family alcohol education: A systematic review of the evidence. Health Education, 114, 451–472. doi:10.1108/HE-12- 2013-0066 Carver, H., Elliott, L., Kennedy, C., & Hanley, J. (2017). Parent-child connectedness and communication in relation to alcohol, tobacco and drug use in adolescence: An integrative review of the literature. Drugs: Education, Prevention & Policy, 24, 119–133. doi:10.1080/0 9687637.2016.1221060 Casey, B. J., Jones, R. M., & Hare, T. A. (2008). The adolescent brain. Annals of the New York Academy of Sciences, 1124, 111–126. doi:10.1196/annals.1440.010 220 Catalano, R. F., & Hawkins, J. D. (1996). The social developmental model: A theory of antisocial behaviour. In: Hawkins, J.D. (Ed.), Delinquency and crime: Current theories (pp. 149–197). Cambridge, UK: Cambridge University Press. Cherpitel, C. J. (2014). Focus on: The burden of alcohol use – trauma and emergency outcomes. Alcohol Research: Current Reviews, 35, 150–154. Chramostova, H. (2014). Changes in parental attitude towards adolescent alcohol consumption as a consequence of adolescent alcohol consumption (Master’s thesis, University of Utrecht, Utrecht, the Netherlands). Retrieved from https://dspace.library. uu.nl/handle/1874/296028 Clark, A. (2013). Whatever next? Predictive brains, situated agents, and the future of cognitive science. Behavioural and Brain Sciences, 36, 181–253. doi:10.1017/S01405 25X12000477. Cleveland, M. J., Lanza, S. T., Ray, A. E., Turrisi, R., & Mallett, K. A. (2012). Transitions in first-year college student drinking behaviors: Does pre-college drinking moderate the effects of parent- and peer-based intervention components? Psychology of Addictive Behaviors, 26, 440–450. doi:10.1037/a0026130 Cohen, D. A., & Rice, J. (1997). Parenting styles, adolescent substance use, and academic achievement. Journal of Drug Education, 27, 199–211. doi:10.2190/QPQQ-6Q1G-UF7D- 5UTJ Cohen, D. A., & Rice, J. (1995). A parent-targeted intervention for adolescent substance use prevention: Lessons learned. Evaluation Review, 19, 159–180. doi:10.1177/0193841X950190 0203 Committee on National Alcohol Policy and Action. (2014). Action plan on youth drinking and on heavy episodic drinking (binge drinking) (2014–2016). Retrieved from https://ec.europa.eu/health//sites/health/files/alcohol/docs/2014_2016_actionplan_youthdrinki ng_en.pdf Colder, C. R., Shyhalla, K., & Frndak, S. E. (2018). Early alcohol use with parental permission: Psychosocial characteristics and drinking in late adolescence. Addictive Behaviors, 76, 82–87. doi:10.1016/j.addbeh.2017.07.030 221 Collins, W. A., & Laursen, B. (2004). Parent-adolescent relationships and influences. In Lerner, R. M., R & Steinberg, L. (Eds.), Handbook of adolescent psychology (2nd ed., pp. 331–361). Hoboken, NJ: John Wiley and Sons, Inc. Cook, R. J., & Sackett, D. L. (1995). The number needed to treat: A clinically useful measure of treatment effect. BMJ, 310, 452–454. doi:10.1136/bmj.310.6977.452 Coombes, L., Allen, D., Marsh, M., & Foxcroft, D. R. (2009). The Strengthening Families Programme (SFP) 10–14 and substance misuse in Barnsley: The perspectives of facilitators and families. Child Abuse Review, 18, 41–59. doi:10.1002/car.1055 Cooper, M. L. (1994). Motivations for alcohol use among adolescents: Development and validation of a four-factor model. Psychological Assessment, 6,117–128. doi:10.1037/1040- 3590.6.2.117 Cooper, M. L., Frone, M. R., Russell, M., & Mudar, P. (1995). Drinking to regulate positive and negative emotions: A motivational model of alcohol use. Journal of Personality and Social Psychology, 69, 990–1005. doi:10.1037/0022-3514.69.5.990 Cox, W. M., & Klinger, E. (1988). A motivational model of alcohol use. Journal of Abnormal Psychology, 97, 168–180. doi:10.1037/0021-843X.97.2.168 Craig, P., Dieppe, P., Macintyre, S., Michie, S, Nazareth, I., & Petticrew, M. (2008). Developing and evaluating complex interventions: The new Medical Research Council guidance. BMJ, 337, a1655. doi:10.1136/bmj.a1655 Crews, F., He, J., & Hodge, C. (2007). Adolescent cortical development: A critical period of vulnerability for addiction. Pharmacology, Biochemistry and Behavior, 86, 189–199. doi:10.1016/j.pbb.2006.12.001 Cronin, C. (1997). Reasons for drinking versus outcome expectancies in the prediction of college student drinking. Substance Use & Misuse, 32, 1287–1311. doi:10.3109/1082608970 9039379 Crotty, M. (1998). Positivism: the march of science. In: Crotty, M. (Ed), The foundations of social research: Meaning and perspective in the research process (pp. 18–41). London, UK: Sage. 222 Cuijpers, P. (2003). Three decades of drug prevention research. Drugs: Education, Prevention and Policy, 10, 7–20. doi:10.1080/0968763021000018900 Currie, C., Zanotti, C., Morgan, A., Currie, D., de Looze, M., Roberts, C., … Barnekow, V. (2012). Social determinants of health and well-being among young people. Health Behaviour in School-aged Children (HBSC) study: International report from the 2009/2010 survey. Copenhagen, Denmark: WHO Regional Office for Europe. Dalton, M. A., Bernhardt, A. M., Gibson, J. J., Sargent, J. D., Beach, M. L., Adachi-Mejia, A. M., … Heatherton, T. F. (2005). Use of cigarettes and alcohol by preschoolers while role- playing as adults: “Honey, have some smokes.” Archives of Pediatrics and Adolescent Medicine, 159, 854–859. doi:10.1001/archpedi.159.9.854 D’Amico, E. J., McCarthy, D. M., Metrik, J., & Brown, S. A. (2004). Alcohol-related services: Prevention, secondary intervention, and treatment preferences of adolescents. Journal of Child & Adolescent Substance Abuse, 14, 61–80. doi:10.1300/J029v14n02_04 Dane, A. V., & Schneider, B. H. (1998). Program integrity in primary and early secondary prevention: Are implementation effects out of control? Clinical Psychology Review, 18, 23– 24. doi:10.1016/S0272-7358(97)00043-3 Darling, N., Cumsille, P., Caldwell, L. L., & Dowdy, B. (2006). Predictors of adolescents’ disclosure to parents and perceived parental knowledge: Between- and within-person differences. Journal of Youth and Adolescence, 35, 667–678. doi:10.1007/s10964-006-9058- 1 Davies, E. L. (2015). A think aloud study to explore the feasibility and acceptability of a digital tool for the prevention of alcohol related harm in adolescents. Retrieved from https://alcoholchange.org.uk/publication/a-think-aloud-study-to-explore-the-feasibility-and- acceptability-of-a-digital-intervention-for-the-prevention-of-alcohol-related-harm-in- adolescents Deeks, J. J., Higgins, J. P. T., & Altman, D. G. (2011). Chapter 9: Analysing data and undertaking meta-analyses. In Higgins, J. P. T., & Green, S. (Eds.), Cochrane handbook for systematic reviews of interventions. Version 5.1.0 (updated March 2011). Retrieved from https://handbook-5-1.cochrane.org/chapter_9/9_analysing_data_and_undertaking_meta_ana lyses.htm 223 DeMarsh, J., & Kumpfer, K. L. (1986). Family-oriented interventions for the prevention of chemical dependency in children and adolescents. Prevention, 18, 117–151. doi:10.1300/J2 74v18n01_07 Dickinson, D. M., Hayes, K. A., Jackson, C., Ennett, S. T., & Lawson, C. (2014). Promoting an alcohol-free childhood: A novel home-based parenting program. American Journal of Health Education, 45, 119–128. doi:10.1080/19325037.2013.875963 Diehr, P., Martin, D. C., Koepsell, T., & Cheadle, A. (1995). Breaking the matches in a paired t‐test for community interventions when the number of pairs is small. Statistics in Medicine, 14, 1491–1504. doi:10.1002/sim.4780141309 Dienes, Z. (n.d). Calculate your Bayes factor. Retrieved January 24, 2018, from http://www.lifesci.sussex.ac.uk/home/Zoltan_Dienes/inference/bayes_factor.swf Dillard, A. J., Midboe, A. M., & Klein, W. M. P. (2009). The dark side of optimism: Unrealistic optimism about problems with alcohol predicts subsequent negative event experiences. Personality and Social Psychology Bulletin, 35, 1540–1550. doi:10.1177/01461 67209343124 Dobbie, F., Purves, R., McKell, J., Dougall, N., White, J., Campbell, R., … Bauld, L. (2017). A process evaluation of the implementation of ASSIST in Scotland: Final report . Stirling, UK: University of Stirling. Donner, A., Taljaard, M., & Klar, N. (2007). The merits of breaking the matches: A cautionary tale. Statistics in Medicine, 26, 2036–2051. doi:10.1002/sim.2662 Donovan, J. E. (2007). Really underage drinkers: The epidemiology of children’s alcohol use in the United States. Prevention Science, 8, 192–205. doi:10.1007/s11121-007-0072-7 Donovan, J. E. (2004). Adolescent alcohol initiation: A review of psychosocial risk factors. Journal of Adolescent Health, 35, 529–e7. doi:10.1016/j.jadohealth.2004.02.003 Donovan, J. E., Leech, S. L., Zucker, R. A., Loveland-Cherry, C. J., Jester, J. M., Fitzgerald, H. E., … Looman, W. S. (2004). Really underage drinkers: Alcohol use among elementary students. Alcoholism: Clinical and Experimental Research, 28, 341–349. doi:10.1097/01.AL C.0000113922.77569.4E 224 Donovan, J. E., & Molina, B. S. G. (2014). Antecedent predictors of children’s initiation of dipping/tasting alcohol. Alcoholism: Clinical and Experimental Research, 38, 2488–2495. doi:10.1111/acer.12517 Donovan, J. E., & Molina, B. S. G. (2011). Childhood risk factors for early onset drinking. Journal of Studies on Alcohol and Drugs, 72, 741–751. doi:10.15288/jsad.2011.72.741 Donovan, J. E., & Molina, B. S. G. (2008). Children’s introduction to alcohol use: Sips and tastes. Alcoholism: Clinical and Experimental Research, 32, 108–119. doi:10.1111/j.1530- 0277.2007.00565.x Doumas, D. M., Turrisi, R., Ray, A. E., Esp, S. M., & Curtis-Schaeffer, A. K. (2013). A randomized trial evaluating a Parent Based Intervention to reduce college drinking. Journal of Substance Abuse Treatment, 45, 31–37. doi:10.1016/j.jsat.2012.12.008 Durlak, J. A., & DuPre, E. P. (2008). Implementation matters: A review of research on the influence of implementation on program outcomes and the factors affecting implementation. American Journal of Community Psychology, 41, 327–350. doi:10.1007/s10464-008-9165-0 Dusenbury, L. (2000). Family-based drug abuse prevention programs: A review. The Journal of Primary Prevention, 20, 337–352. doi:10.1023/A:1021366721649 Dusenbury, L., Brannigan, R., Falco, M., & Hansen, W. B. (2003). A review of research on fidelity of implementation: Implications for drug abuse prevention in school settings. Health Education Research, 18, 237–256. doi:10.1093/her/18.2.237 Eadie, D., MacAskill, S., Brooks, O., Heim, D., Forsyth, A., & Punch, S. (2010). Pre-teens learning about alcohol. Drinking and family contexts. York, UK: Joseph Rowntree Foundation. Eccles, M. P., & Mittman, B. S. (2006). Welcome to Implementation Science. Implementation Science, 1, 1. doi:10.1186/1748-5908-1-1 Effective Practice and Organisation of Care (EPOC). (2015). EPOC resources for review authors. Good practice data extraction form. Retrieved from https://epoc.cochrane.org /resources/epoc-resources-review-authors#protocol 225 Ellickson, P. L., & Hays, R. D. (1991). Antecedents of drinking among young adolescents with different alcohol use histories. Journal of Studies on Alcohol, 52, 398–408. doi:10.15288/jsa.1991.52.398 Elliott, D., & Mihalic, S. (2004). Issues in disseminating and replicating effective prevention programs. Prevention Science, 5, 47–53. doi:10.1023/B:PREV.0000013981.28071.52 Enders, C. K., Keller, B. T., & Levy, R. (2017). A fully conditional specification approach to multilevel imputation of categorical and continuous variables. Psychological Methods, 23, 298–317. doi:10.1037/met0000148 Endresen, E. (2013). Tilrettelegging av Örebro preventionsprogram (ÖPP) for bruk i Norge. Rapport til Helsedirektoratet 2013 [Adaptations of Örebro prevention program (ÖPP) for use in Norway. Report to the Directorate of Health 2013] . Retrieved from http://korus.azure websites.net/wp-content/uploads/2017/01/Opp-rapport_1_.pdf Engels, R. C. M. E., Knibbe, R. A., & Drop, M. J. (1997). Inconsistencies in adolescents’ self-reports of initiation of alcohol and tobacco use. Addictive Behaviors, 22, 613–623. doi:10.1016/S0306-4603(96)00067-6 Ennett, S. T., Bauman, K. E., Foshee, V. A., Pemberton, M., & Hicks, K. A. (2001). Parent- child communication about adolescent tobacco and alcohol use: What do parents say and does it affect youth behavior? Journal of Marriage and Family, 63, 48–62. doi:10.11 11/j.1741-3737.2001.00048.x Ernst, M., Pine, D. S., & Hardin, M. (2006). Triadic model of the neurobiology of motivated behavior in adolescence. Psychological Medicine, 36, 299–312. doi:10.1017/S003329170 5005891 Estonian Lower and Middle Education National Curriculum. (1996, Sept 6). RT I 1996, 65, 1201. European Monitoring Centre for Drugs and Drug Addiction. (2017). Xchange Prevention Registry. EFFEKT/Örebro/PAS – To increase parents’ attitudes, norm-setting and monitoring about alcohol. Retrieved January 24, 2018, from http://www.emcdda.europa. eu/best-practice/xchange/effektörebro_en 226 European Monitoring Centre for Drugs and Drug Addiction. (2012). Guidelines for the evaluation of drug prevention. A manual for programme planners and evaluators (2nd ed). Lisbon, Portugal: European Monitoring Centre for Drugs and Drug Addiction. Faggiano, F., Minozzi, S., Versino, E., & Buscemi, D. (2014). Universal school-based prevention for illicit drug use. Cochrane Database of Systematic Reviews, Issue 12. doi:10.10 02/14651858.CD003020.pub3 Faggiano, F., Siliquini, R., Zunino, B., Siliquini, V., Vigna Taglianti, F., Cuomo, G., … Scatigna, M. (2002). Unplugged: An effective school-based program for the prevention of substance use among adolescents. Retrieved from https://www.eudap.net/PDF/FinalRepo rt1.pdf Festinger, L. (1957). A theory of cognitive dissonance. Stanford, CA: Stanford University Press. Finch, H., Lewis, J., & Turley, C. (2014). Focus groups. In Ritchie, J., Lewis, J., McNaughton Nicholls, C., & Ormston, R. (Eds.), Qualitative research practice. A guide for social science students and researchers (2nd ed., pp. 211–242). London, UK: SAGE Publications Ltd. Foster, J., Bryant, L., & Brown, K. (2017). “Like sugar for adults”: The effect of non- dependent parental drinking on children and families. Retrieved from http://www.ias. org.uk/uploads/pdf/IAS%20reports/rp28102017.pdf Foxcroft, D. R. (2014a). Can prevention classification be improved by considering the function of prevention? Prevention Science, 15, 818–822. doi:10.1007/s11121-013-0435-1 Foxcroft, D. R. (2014b). “Form ever follows function. This is the law”. A prevention taxonomy based on a functional typology. Addiciones, 26, 10–14. Foxcroft, D. R., Callen, H., Davies, E. L., & Okulicz-Kozaryn, K. (2017). Effectiveness of the strengthening families programme 10-14 in Poland: Cluster randomized controlled trial. The European Journal of Public Health, 27, 494–500. doi:10.1093/eurpub/ckw195 Foxcroft, D. R., Ireland, D., Lister-Sharp, D. J., Lowe, G., & Breen, R. (2003). Longer-term primary prevention for alcohol misuse in young people: A systematic review. Addiction, 98, 397–411. doi:10.1046/j.1360-0443.2003.00355.x 227 Foxcroft, D. R., & Tsertsvadze, A. (2011a). Universal family-based prevention programs for alcohol misuse in young people. Cochrane Database of Systematic Reviews, Issue 9. doi:10.1002/14651858.CD009308 Foxcroft, D. R., & Tsertsvadze, A. (2011b). Universal multi-component prevention programs for alcohol misuse in young people. Cochrane Database of Systematic Reviews, Issue 9. doi:10.1002/14651858.CD009307 Foxcroft, D. R., & Tsertsvadze, A. (2011c). Universal school-based prevention programs for alcohol misuse in young people. Cochrane Database of Systematic Reviews, Issue 5. doi:10.1002/14651858.CD009113 Friese, B., Grube, J. W., Moore, R. S., & Jennings, V. K. (2012). Parents' rules about underage drinking: A qualitative study of why parents let teens drink. Journal of Drug Education, 42, 379–391. doi:10.2190/DE.42.4.a. Frissell, K. C., McCarthy, D. M., D’Amico, E. J., Metrik, J., Ellingstad, T. P., & Brown, S. A. (2004). Impact of consent procedures on reported levels of adolescent alcohol use. Psychology of Addictive Behaviors, 18, 307–315. doi:10.1037/0893-164X.18.4.307 Fu, R., Gartlehner, G., Grant, M., Shamliyan, T., Sedrakyan, A., Wilt, T. J., … Trikalinos, T. A. (2011). Conducting quantitative synthesis when comparing medical interventions: AHRQ and the Effective Health Care Program. Journal of Clinical Epidemiology, 64, 1187–1197. doi:10.1016/j.jclinepi.2010.08.010 Furlong, M., Leckey, Y., & McGilloway, S. (2017). An evaluation of the Strengthening Families Programme in Co. Kildare, Ireland. Retrieved from http://sfpcouncilireland.ie/wp- content/uploads/2017/10/Evaluation-of-the-Strengthening-Families-Programme-in-Co.-Kilda re.pdf Galinsky, E. (2010). Mind in the making: The seven essential life skills every child needs. New York, NY: HarperCollins. Garcia-Huidobro, D., Doty, J. L., Davis, L., Borowsky, I. W., & Allen, M. L. (2018). Prevention Science, 19,570–578. doi:10.1007/s11121-017-0853-6 228 Gaskell, G. (2000). Individual and group interviewing. In Bauer, M. W., & Gaskell, G. (Eds.), Qualitative researching with text, image, and sound: A practical handbook (pp. 38– 56). London, UK: SAGE Publications Ltd. Geibprasert, S., Gallucci, M., & Krings, T. (2010). Alcohol-induced changes in the brain as assessed by MRI and CT. European Radiology, 20, 1492–1501. doi: Gerrard, M., Gibbons, F. X., Reis-Bergan, M., & Russell, D. W. (2000). Self-esteem, self- serving cognitions, and health risk behavior. Journal of Personality, 68, 1177–1201. doi:10.1111/1467-6494.00131 Giedd, J. N. (2004). Structural magnetic resonance imaging of the adolescent brain. Annals of the New York Academy of Sciences, 1021, 77–85. doi:10.1196/annals.1308.009 Giedd, J. N., Blumenthal, J., Jeffries, N. O., Castellanos, J. F. X., Liu, H., Zijdenbos, A., … Rapoport, J. L. (1999). Brain development during childhood and adolescence: A longitudinal MRI study. Nature Neuroscience, 2, 861–863. doi:10.1038/13158 Gilligan, C., & Kypri, K. (2012). Parent attitudes, family dynamics and adolescent drinking: Qualitative study of the Australian parenting guidelines for adolescent alcohol use. BMC Public Health, 12, 491. doi:10.1186/1471-2458-12-491 Gilligan, C., Wolfenden, L., Foxcroft, D. R., Williams, A. J., Kingsland, M., Hodder, R.K., ... Wiggers, J. (2019). Family-based prevention programmes for alcohol use in young people. Cochrane Database of Systematic Reviews, Issue 3. doi:10.1002/14651858.CD012287.pub2. Glasman, L., & Albarracin, D. (2006). Forming attitudes that predict future behavior: A meta-analysis of the attitude–behavior relation. Psychological Bulletin, 132, 778–822. doi:0.1037/0033-2909.132.5.778. Glatz, T., Stattin, H., & Kerr, M. (2012). A test of cognitive dissonance theory to explain parents’ reactions to youths’ alcohol intoxication. Family Relations, 61, 629–641. doi:10.11 11/j.1741-3729.2012.00723.x Glowacki, E. M. (2016). Communication about problematic drinking between young adults and their parents: An application of inconsistent nurturing as control theory. Health Communication, 31, 1135–1144. doi:10.1080/10410236.2015.1045578 229 Gordon, T. (2008). Parent Effectiveness Training: The proven program for raising responsible children. New York, NY: Harmony. Graham, M. L., Ward, B., Munro, G., Snow, P., & Ellis, J. (2006). Rural parents, teenagers and alcohol: What are parents thinking? Rural and Remote Health, 6, 3–83. Grant, B. F., & Dawson, D. A. (1997). Age at onset of alcohol use and its association with DSM-IV alcohol abuse and dependence: Results from the National Longitudinal Alcohol Epidemiologic Survey. Journal of Substance Abuse, 9, 103–110. doi:10.1016/S0899- 3289(97)90009-2 Griswold, M. G., Fullman, N., Hawley, C., Arian, N., Zimsen, S. R. M., Tymeson, H. D., … Gakidou, E. (2018). Alcohol use and burden for 195 countries and territories, 1990–2016: A systematic analysis for the Global Burden of Disease Study 2016. The Lancet, 392, 1015– 1035. doi:10.1016/S0140-6736(18)31310-2 Gross, D., Julion, W., & Fogg, L. (2001). What motivates participation and dropout among low‐income urban families of color in a prevention intervention? Family Relations, 50, 246– 254. doi:10.1111/j.1741-3729.2001.00246.x Guba, E. G. (1990). The alternative paradigm dialog. In: Guba, E. G. (Ed.), The paradigm dialog (pp. 17–30). London, UK: Sage. Gupta, R., Koscik, T. R., Bechara, A., & Tranel, D. (2011). Neuropsychologia, 49, 760–766. doi:10.1016/j.neuropsychologia.2010.09.029 Hahn, E. J., Noland, M. P., Rayens, M. K., & Myers, D. M. (2002). Efficacy of training and fidelity of implementation of the Life Skills Training Program. Journal of School Health, 72, 282–287. doi:10.1111/j.1746-1561.2002.tb01333.x Handley, E. D., & Chassin, L. (2013). Alcohol-specific parenting as a mechanism of parental drinking and alcohol use disorder risk on adolescent alcohol use onset. Journal of Studies on Alcohol and Drugs, 74, 684–693. doi:10.15288/jsad.2013.74.684 Hammersley, B. (2017, March 27). Concerned about Brexit? Why not become and e-resident of Estonia. Retrieved from https://www.wired.co.uk/article/estonia-e-resident Hammerton, G., Mahedy, L., Murray, J., Maughan, B., Edwards, A. C., Kendler, K. S., … Heron, J. (2017). Effects of excessive alcohol use on antisocial behaviour across adolescence 230 and early adulthood. Journal of the American Academy of Child and Adolescent Psychiatry, 56, 857–865. doi:10.1016/j.jaac.2017.07.781 Hasking, P., Lyvers, M., & Carlopio, C. (2011). The relationship between coping strategies, alcohol expectancies, drinking motives and drinking behaviour. Addictive Behaviors, 36, 479–487. doi:10.1016/j.addbeh.2011.01.014 Hawkins, J. D., Catalano, R. F., & Miller, J. Y. (1992). Risk and protective factors for alcohol and other drug problems in adolescence and early adulthood: Implications for substance abuse prevention. Psychological Bulletin, 112, 64–105. doi:10.1037/0033-2909.112.1.64 Hayden, J. A., Cote, P., & Bombardier, C. (2006). Evaluation of the quality of prognosis studies in systematic reviews. Annals of Internal Medicine, 144, 427–437. doi:10.7326/0003- 4819-144-6-200603210-00010 Hayes, L., Smart, D., Toumbourou, J. W., & Sanson, A. (2004). Parenting influences on adolescent alcohol use (Research report no. 10). Retrieved from https://aifs.gov.au /sites/default/files/publication-documents/aifsreport10_0.pdf Hearst, M. O., Fulkerson, J. A., Maldonado-Molina, M. M., Perry, C, L., & Komro, K. A. (2007). Who needs liquor stores when parents will do? The importance of social sources of alcohol among young urban teens. Preventive Medicine, 44, 471–476. doi:10.1016/j.ypmed .2007.02.018 Helgøy, I., & Homme, A. (2017). Increasing parental participation at school level: A “citizen to serve” or a “customer to steer”? Nordic Journal of Studies in Educational Policy, 3, 144– 154. doi:10.1080/20020317.2017.1343625 Henriksen, Ø. (2012). “They must be 18 years old ... really”. Alcohol debut discourses at parent meetings within alcohol prevention programmes in school. Nordic Studies on Alcohol and Drugs, 29, 453–466. doi:10.2478/v10199-012-0040-9 Hibell, B., Andersson, B., Bjarnason, T., Kokkevi, A., Morgan, M., & Narusk, A. (1997). The 1995 ESPAD report: Alcohol and other drug use among students in 26 European countries. Stockholm, Sweden: The Swedish Council for Information on Alcohol and Other Drugs. Hibell, B., Guttormsson, U., Ahlström, S., Balakireva, O., Bjarnason, T., Kokkevi, A., & Kraus, L. (2012). The 2011 ESPAD report: Substance use among students in 36 European 231 countries. Stockholm, Sweden: The Swedish Council for Information on Alcohol and Other Drugs. Higgins, J. P. T., & Thompson, H. (2002). Quantifying heterogeneity in a meta-analysis. Statistics in Medicine, 21, 1539–1558. doi:10.1002/sim.1186 Higgins, J. P. T., Thompson, S., Deeks, J., & Altman, D. (2003). Measuring inconsistency in meta-analyses. BMJ, 327, 557–560. doi:10.1136/bmj.327.7414.557 Hill, L. (2017). Alcohol industry. In Quah, S R., & Cockerham, W. (Eds.), International encyclopedia of public health (2nd ed., pp. 62–70). Oxford, UK: Elsevier. Hingson, R., Edwards, E. M., Heeren, T., & Rosenbloom, D. (2009). Age of drinking onset and injuries, motor vehicle crashes, and physical fights after drinking and when not drinking. Alcoholism: Clinical and Experimental Research, 33, 783–790. doi:10.1111/j.1530-0277. 2009.00896.x Hingson, R. W., Heeren, T., Jamanka, A., & Howland, J. (2000). Age of drinking onset and unintentional injury involvement after drinking. JAMA, 284, 1527–1533. doi:10.1001/jama. 284.12.1527 Hingson, R. W., Heeren, T., & Winter, M R. (2006). Age at drinking onset and alcohol dependence: Age at onset, duration, and severity. Archives of Pediatrics & Adolescent Medicine, 160, 739–746. doi:10.1001/archpedi.160.7.739 Hopson, L. M., & Steiker, L. K. H. (2010). Methodology for evaluating an adaptation of evidence-based drug abuse prevention in alternative schools. Children & Schools, 30, 16–27. doi:10.1093/cs/30.2.116 Hurley, E., Dietrich, T., & Rundle-Thiele, S. (2019). A systematic review of parent based programs to prevent or reduce alcohol consumption in adolescents. BMC Public Health, 19, 1451. doi:10.1186/s12889-019-7733-x Huver, R. M. E., Engels, R. C. M. E., Van Breukelen, G., & de Vries, H. (2007). Parenting style and adolescent smoking cognitions and behaviour. Psychology & Health, 22, 575–593. doi:10.1080/14768320600976182 232 Hvitfeldt, T., Andersson, B., & Hibell, B. (2004). Skolelevers drogvanor 2003 [Alcohol and drug use among adolescents]. Retrieved from https://can.se/Publikationer/rapporter/Skolelev ers-drogvanor-2003/ Ichiyama, M. A., Fairlie, A. M., Wood, M. D., Turrisi, R., Francis, D. P., Ray, A. E., & Stanger, L. A. (2009). A randomized trial of a Parent-Based Intervention on drinking behavior among incoming college freshmen. Journal of Studies on Alcohol and Drugs, s16, 67–76. doi:10.15288/jsads.2009.s16.67 Inchley, J., Currie, D., Vieno, A., Torsheim, T., Ferreira-Borges, C., Weber, M. M., … Breda, J. (2018). Adolescent alcohol-related behaviours: Trends and inequalities in the WHO European Region, 2002–2014. Observations from the Health Behaviour in School-aged Children (HBSC) WHO collaborative cross-national study. Copenhagen, Denmark: WHO Regional Office for Europe. Inchley, J., Currie, D., Young, T., Samdal, O., Torsheim, T., Augustson, L., … Barnekow, V. (2016). Growing up unequal: Gender and socioeconomic differences in young people’ s health and well-being. Health Behaviour in School-aged Children (HBSC) study: International report from the 2013/2014 survey. Copenhagen, Denmark: WHO Regional Office for Europe. Jackson, C., Ennett, S. T., Dickinson, D. M., & Bowling, J. M. (2012). Letting children sip: Understanding why parents allow alcohol use by elementary school-aged children. Archives of Pediatrics and Adolescent Medicine, 166, 1053–1057. doi:10.1001/archpediatrics.2012 .1198 Jackson, K. M., Sher, K. J., Cooper, M. L., & Wood, P. K. (2002). Adolescent alcohol and tobacco use: Onset, persistence and trajectories of use across two samples. Addiction, 97, 517–531. doi:10.1046/j.1360-0443.2002.00082.x Janssen, M.M., Mathijssen, J. J. P., Van Bon-Martens, M. J. H., Van Oers, H. A. M., & Garretsen, H. F. L. (2014). A qualitative exploration of attitudes towards alcohol, and the role of parents and peers of two alcohol-attitude-based segments of the adolescent population. Substance Abuse Treatment, Prevention, and Policy, 9, 20. doi:10.1186/1747-597X-9-20 233 Javakhishvili, D. J., Javakhishvili, N., Miovksý, M., Razmadze, M., & Kandelaki, N. (2014). Piloting the comprehensive social influence (“Unplugged”) programme in Georgia: A formative study. Adiktologie, 14, 33–39. Jernigan, D., Noel, J., Landon, J., Thornton, N., & Lobstein T. (2016). Alcohol marketing and youth alcohol consumption: A systematic review of longitudinal studies published since 2008. Addiction, 112, 7–20. doi:10.1111/add.13591 Jindal-Snape, D., Cantali, D., MacGillivray, S., & Hannah, E. (2019). Primary-secondary transitions: A systematic literature review. Retrieved from https://www.gov.scot/publicati ons/primary-secondary-transitions-systematic-literature-review/ Johnson, R. B., & Onwuegbuzie, A.J. (2004). Mixed methods research: A research paradigm whose time has come. Educational Researcher, 33, 14–26. doi:10.3102/0013189X033007014 Jones, S. C. (2016). Parental provision of alcohol: A TPB-framed review of the literature. Health Promotion International, 31, 562–571. doi:10.1093/heapro/dav028 Jones, S. C., Andrews, K., & Berry, N. (2016). Lost in translation: A focus group study of parents’ and adolescents’ interpretations of underage drinking and parental supply. BMC Public Health, 16, 561. doi:10.1186/s12889-016-3218-3 Jones, S. C., Andrews, K., & Francis, K. (2017). Combining social norms and social marketing to address underage drinking: Development and process evaluation of a whole-of- community intervention. PLoS ONE, 12, e0169872. doi:10.1371/journal.pone.0169872 Jones, L., Bates, G., Downing, J., Sumnall, H., & Bellis, M. A. (2010). A review of the effectiveness and cost effectiveness of alcohol and sex and relationship education for all children and young people aged 5–19 years in community settings. Retrieved from https://www.nice.org.uk/guidance/gid-phg0/documents/pshe-evidence-review-community2 Jones, S. C., & Gordon, C. S. (2017). A systematic review of children's alcohol-related knowledge, attitudes and expectancies. Preventive Medicine, 105, 19–31. doi:10.1016/j.yp med.2017.08.005 Jones, L., James, M., Jefferson, T., Lushey, C., Morleo, M., Stokes, E., … Bellis, M. (2007). A review of the effectiveness and cost-effectiveness of interventions delivered in primary and secondary schools to prevent and/or reduce alcohol use by young people under 18 years old. 234 Retrieved from https://www.nice.org.uk/guidance/ph7/evidence/effectiveness-and-cost-effec tivenessreview-369704701 Jones, S. C., Magee, C., & Andrews, K. (2015). “I think other parents might…”: Using a projective technique to explore parental supply of alcohol. Drug and Alcohol Review, 34, 531–539. doi:10.1111/dar.12258. Järvinen, M., & Østergaard, J. (2009). Governing adolescent drinking. Youth & Society, 40, 377–402. doi:10.1177/0044118X08318118 Kaynak, O., Winters, K. C., Cacciola, J., Kirby, K. C., & Arria, A. M. (2014). Providing alcohol for underage youth: What messages should we be sending parents? Journal of Studies on Alcohol and Drugs, 75, 590–605. doi:10.15288/jsad.2014.75.590 Keating, D. P. (1990). Adolescent thinking. In Feldman, S. S., & Elliott, G. R. (Eds.), At the threshold: The developing adolescent (pp. 54–89). Cambridge, MA: Harvard University Press. Keller, B. T., & Enders, C. K. (2017). Blimp user’s manual (Version 1.0). Retrieved from http://www.appliedmissingdata.com/blimpuserguide-4.pdf Kelly, A. B., O’Flaherty, M., Toumbourou, J. W., Connor, J. P., Hemphill, S. A., & Catalano, R. F. (2011). Gender differences in the impact of families on alcohol use: A lagged longitudinal study of early adolescents. Addiction, 106, 1427–1436. doi:10.1111/j.1360- 0443.2011.03435.x Kerr, D. C. R., Capaldi, D. M., Pears, K. C., & Owen, L. D. (2012). Intergenerational influences on early alcohol use: Independence from the problem behavior pathway. Development and Psychopathology, 24, 889–906. doi:10.1017/S0954579412000430 Kohlberg, L. (1984). The psychology of moral development: The nature and validity of moral stages. New York, NY: Harper & Row. Komro, K. A., Perry, C. L., Veblen-Mortenson, S., Farbakhsh, K., Toomey, T. L., Stigler, M. H., ... Williams, C. L. (2008). Outcomes from a randomized controlled trial of a multi- component alcohol use preventive intervention for urban youth: Project Northland Chicago. Addiction, 103, 606–618. doi:10.1111/j.1360-0443.2007.02110.x 235 Koning, I. M., Engels, R. C. M. E., Verdurmen, J. E. E., & Vollebergh, W. A. M. (2010a). Alcohol-specific socialization practices and alcohol use in Dutch early adolescents. Journal of Adolescence, 33, 93–100. doi:10.1016/j.adolescence.2009.05.003 Koning, I. M., Lugtig, P., & Vollebergh, W. A. M. (2014). Differential effects of baseline drinking status: Effects of an alcohol prevention program targeting students and/or parents (PAS) among weekly drinking students. Journal of Substance Abuse Treatment, 46, 522–527. doi:10.1016/j.jsat.2013.11.003 Koning, I. M., Maric, M., MacKinnon, D., & Vollebergh, W. A. M. (2015). Effects of a combined parent-student alcohol prevention program on intermediate factors and adolescents’ drinking behavior: A sequential mediation model. Journal of Consulting and Clinical Psychology, 83, 719–727. doi:10.1037/a0039197 Koning, I. M., Van den Eijnden, R. J. J., Engels, R. C. M. E., Verdurmen, J. E. E., & Vollebergh, W. A. M. (2010b). Why target early adolescents and parents in alcohol prevention? The mediating effects of self-control, rules and attitudes about alcohol use. Addiction, 106, 538–546. doi:10.1111/j.1360-0443.2010.03198.x Koning, I. M., Van den Eijnden, R. J. J. M., Verdurmen, J. E. E., Engels, R. C. M. E., & Vollebergh, W. A. M. (2013). A cluster randomized trial on the effects of a parent and student intervention on alcohol use in adolescents four years after baseline; no evidence of catching-up behavior. Addictive Behaviors, 38, 2032–2039. doi:10.1016/j.addbeh.2012.12. 013 Koning, I. M., Van den Eijnden, R. J. J. M., Verdurmen, J. E. E., Engels, R. C. M. E., & Vollebergh, W. A. M. (2012b). Developmental alcohol-specific parenting profiles in adolescence and their relationships with adolescents’ alcohol use. Journal of Youth and Adolescence, 41, 1502–1511. doi:10.1007/s10964-012-9772-9 Koning, I. M., Van den Eijnden, R. J., Verdurmen, J. E. E., Engels, R. C. M. E., & Vollebergh, W. A. M. (2011). Long-term effects of a parent and student intervention on alcohol use in adolescents: A cluster randomized controlled trial. American Journal of Preventive Medicine, 40, 541–547. doi:10.1016/j.amepre.2010.12.030 Koning, I. M., Verdurmen, J. E. E., Engels, R. C. M. E., Van den Eijnden, R. J. J. M., & Vollebergh, W. A. M. (2012a). Differential impact of a Dutch alcohol prevention program 236 targeting adolescents and parents separately and simultaneously: Low self-control and lenient parenting at baseline predict effectiveness. Prevention Science, 13, 278–287. doi:10.1007 /s11121-011-0267-9 Koning, I. M., Vollebergh, W. A. M., Smit, F., Verdurmen, J. E. E., Van den Eijnden, R. J. J. M., Ter Bogt, T. F. M., … Engels, R. C. M. E. (2009). Preventing heavy alcohol use in adolescents (PAS): Cluster randomized trial of a parent and student intervention offered separately and simultaneously. Addiction, 104, 1669–1678. doi:10.1111/j.1360-0443.2009. 02677.x Koutakis, N. (2011). Preventing underage alcohol drinking through working with parents (Doctoral dissertation, Örebro University, Örebro, Sweden). Retrieved from https://www.diva -portal.org/smash/get/diva2:461182/FULLTEXT02.pdf Koutakis, N., Stattin, H., & Kerr, M. (2008). Reducing youth alcohol drinking through a parent-targeted intervention: The Örebro prevention program. Addiction, 103, 1629–1637. doi:10.1111/j.1360-0443.2008.02326.x Kraus, L., Guttormsson, U., Leifman, H., Arpa, S., Molinaro, S., Monshouwer, K., … Hibell, B. (2016). ESPAD report 2015: Results from the European School Survey Project on Alcohol and Other Drugs. Lisbon, Portugal: European Monitoring Centre on Drugs and Drug Addiction. Kull, M., Saat, H., Kiive, E., & Põiklik, E. (2015). Uimastiennetuse õpetajaraamat põhikoolile [Substance use prevention teacher’s book for secondary schools]. Tallinn, Estonia: National Institute for Health Development. Kumpfer, K. L., Alvarado, R., & Whiteside, H. O. (2003). Family-based interventions for substance use and misuse prevention. Substance Use & Misuse, 38, 1759–1787. doi:10.108 1/JA-120024240 Kumpfer, K. L., Molgaard, V., & Spoth, R. (1996). The Strengthening Families Program for the prevention of delinquency and drug use. In: Peters, R. D., & McMahon R. J. (Eds.), Preventing childhood disorders, substance abuse, and delinquency (pp. 241–267). Thousand Oaks, CA: Sage. 237 Kumpfer, K. L., Xie, J., & O’Driscoll, R. (2012). Effectiveness of a culturally adapted Strengthening Families Program 12–16 Years for high-risk Irish families. Child & Youth Care Forum, 41, 173–195. doi:10.1007/s10566-011-9168-0 Kuntsche, E. (2017). “Do grown-ups become happy when they drink?” Alcohol expectancies among preschoolers. Experimental and Clinical Psychopharmacology, 25, 24–30. doi:10.1 037/pha0000104 Kuntsche, E., Knibbe, R., Engels, R., & Gmel, G. (2007). Drinking motives as mediators of the link between alcohol expectancies and alcohol use among adolescents. Journal of Studies on Alcohol and Drugs, 68, 76–85. doi:10.15288/jsad.2007.68.76 Kuntsche, E., Knibbe, R., Gmel, G., & Engels, R. (2005). Why do young people drink? A review of drinking motives. Clinical Psychology Review, 25, 841–861. doi:10.1016/j.cpr.2 005.06.002 Kuntsche, S., & Kuntsche, E. (2016). Parent-based interventions for preventing or reducing adolescent substance use – A systematic literature review. Clinical Psychology Review, 45, 89–101. doi:10.1016/j.cpr.2016.02.004 Kuntsche, E., Le Mével, L., & Zucker, R. A. (2016). What do preschoolers know about alcohol? Evidence from the electronic Appropriate Beverage Task (eABT). Addictive Behaviors, 61, 47–52. doi:10.1016/j.addbeh.2016.05.004 Kuntsche, E., Rossow, I., Engels, R. C. M. E., & Kuntsche, S. (2016). Is “age at first drink” a useful concept in alcohol research and prevention? We doubt that. Addiction, 111, 957–965. doi:10.1111/add.12980 Kypri, K., Dean, J. I., & Stojanovski, E. (2007). Parent attitudes on the supply of alcohol to minors. Drug and Alcohol Review, 26, 41–47. doi:10.1080/09595230601037018 Lemmens, P., Vandlik, E., & Elekes, Z. (2016). Motives for drinking and non-drinking. In Moskalewicz, J., Room, R., & Thom, B. (Eds.), Comparative monitoring of alcohol epidemiology across the EU. Baseline assessment and suggestions for future action. Synthesis report (pp. 54–72). Warsaw, Poland: PARPA – The State Agency for Prevention of Alcohol Related Problems. 238 Lendrum, A., & Humphrey, N. (2012). The importance of studying the implementation of interventions in school settings. Oxford Review of Education, 38, 635–652. doi:10.1080/03 054985.2012.734800 Liverpool John Moores University. (2019). EDPQS Activities Map. Retrieved August 21, 2019, from http://prevention-standards.eu/countries/ Long, J., & Mongan, D. (2014). Alcohol consumption in Ireland 2013: Analysis of a national alcohol diary survey. Retrieved from https://www.drugsandalcohol.ie/22138/1/HRB018_ NADS_2014_forWeb.pdf Lundborg, P. (2007). Parents’ willingness to provide alcohol and adolescents’ alcohol use – evidence from Swedish data. Vulnerable Children and Youth Studies, 2, 60–70. doi:10.108 0/17450120601130581 MacArthur, G., Caldwell, D. M., Redmore, J., Watkins, S. H., Kipping, R., White, J., ... Campbell, R. (2018). Individual-, family-, and school-level interventions targeting multiple risk behaviours in young people. Cochrane Database of Systematic Reviews, Issue 10. doi:10.1002/14651858.CD009927.pub2 Maccoby, E.E., & Martin, J.A. (1983). Socializations in the context of the family: Parent- child interactions. In: Mussen, P. H. (Ed.), Handbook of child psychology: Vol. 4. Socialization, personality, and social development (pp. 1–102). New York, NY: Wiley. Macdonald, S., Greer, A., Brubacher, J. R., Cherpitel, C., Stockwell, T., & Zeisser, C. (2013). Alcohol consumption and injury. In: Boyle, P., Boffetta, P., Lowenfels, A. B., Burns, H., Brawley, O., Zatonski, W., & Rehm, J. (Eds.), Alcohol: Science, policy and public health (pp. 171–178). Oxford, UK: Oxford University Press. Maggs, J. L., & Staff, J. A. (2018). Parents who allow early adolescents to drink. Journal of Adolescent Health, 68, 245–247. doi:10.1016/j.jadohealth.2017.09.016 Mair, C., Lipperman-Kreda, S., Gruenewald, P. J., Bersamin, M., & Grube, J. W. (2015). Adolescent drinking risks associated with specific drinking contexts. Alcoholism: Clinical and Experimental Research, 39, 1705–1711. doi:10.1111/acer.12806 Mallett, K. A., Ray, A. E., Turrisi, R., Belden, C., Bachrach, R. L., & Larimer, M. E. (2010). Age of drinking onset as a moderator of the efficacy of Parent‐Based, brief motivational, and 239 combined intervention approaches to reduce drinking and consequences among college students. Alcoholism: Clinical and Experimental Research, 34, 1154–1161. doi:10.11 11/j.1530-0277.2010.01192.x Mares, S. H. W., Lichtwarck-Aschoff, A., Burk, W. J., Van der Vorst, H., & Engels, R. C. M. E. (2012). Parental alcohol-specific rules and alcohol use from early adolescence to young adulthood. Journal of Child Psychology and Psychiatry, 53, 798–805. doi:10.1111/j.1469- 7610.2012.02533.x Mares, S. H. W., Van der Vorst, H., Engels, R. C. M. E., & Lichtwarck-Aschoff, A. (2011a). Parental alcohol use, alcohol-related problems, and alcohol-specific attitudes, alcohol- specific communication, and adolescent excessive alcohol use and alcohol-related problems: An indirect path model. Addictive Behaviors, 36, 209–216. doi:10.1016/j.addbeh.2010.10.013 Mares, S. H. W., Van der Vorst, H., Lichtwarck-Aschoff, A., Schulten, I., Verdurmen, J. E. E., Otten, R., & Engels, R. C. M. E. (2011b). Effectiveness of the home-based alcohol prevention program "In control: No alcohol!": Study protocol of a randomized controlled trial. BMC Public Health, 11, 622. doi:10.1186/1471-2458-11-622 Margulies, R. Z., Kessler, R. C., & Kandel, D. B. (1977). A longitudinal study of onset of drinking among high school students. Journal of Studies on Alcohol, 38, 897–912. doi:10.15288/jsa.1977.38.897 Martineau, F., Tyner, E., Lorenc, T., Petticrew, M., & Lock, K. (2013). Population-level interventions to reduce alcohol-related harm: An overview of systematic reviews. Preventive Medicine, 57, 278–296. doi:10.1016/j.ypmed.2013.06.019 Martino, F. P., Miller, P. G., Coomber, K., Hancock, L., & Kypri, K. (2017). Analysis of alcohol industry submissions against marketing regulation. PLoS ONE, 12, e0170366. doi:10.1371/journal.pone.0170366 Masche, J. G. (2010). Explanation of normative declines in parents’ knowledge about their adolescent children. Journal of Adolescence, 33, 271–284. doi:10.1016/j.adolescence.2009.08 .002 McCambridge, J., Mialon, M., & Hawkins, B. (2018). Alcohol industry involvement in policymaking: A systematic review. Addiction, 113, 1571–1584. doi:10.1111/add.14216 240 McCurdy, K., & Daro, D. (2001). Parent involvement in family support programs: An integrated theory. Family Relations, 50, 113–121. doi:10.1111/j.1741-3729.2001.00113.x McGue, M., Elkins, I., Walden, B., & Iacono, W. G. (2005). Perceptions of the parent- adolescent relationship: A longitudinal investigation. Developmental Psychology, 41, 971– 984. doi:10.1037/0012-1649.41.6.971 McKay, M., Agus, A., Cole, J., Doherty, P., Foxcroft, D., Harvey, S., … Sumnall, H. (2018). Steps Towards Alcohol Misuse Prevention Programme (STAMPP): A school-based and community-based cluster randomised controlled trial. BMJ Open, 8, e019722. doi:10.113 6/bmjopen-2017-019722 Medeiros, P. F. P., Cruz, J. I., Schneider, D. R., Sanudo, A., & Sanchez, Z. M. (2016). Process evaluation of the implementation of the Unplugged Program for drug use prevention in Brazilian schools. Substance Abuse Treatment, 11, 2. doi:10.1186/s13011-015-0047-9 Mendez, J. L., Carpenter, J. L., LaForett, D. R., & Cohen, J. S. (2009). Parental engagement and barriers to participation in a community-based preventive intervention. American Journal of Community Psychology, 44, 1–14. doi:10.1007/s10464-009-9252-x Mewton, L., Visontay, R., Chapman, C., Newton, N., Slade, T., Kay-Lambkin, F., & Teesson, M. (2018). Universal prevention of alcohol and drug use: An overview of reviews in an Australian context. Drug and Alcohol Review, 37, S435–S469. doi:10.1111/dar.12694 Meyer, H. C., & Bucci, D. J. (2016). Imbalanced activity in the orbitofrontal cortex and nucleus accumbens impairs behavioral inhibition. Current Biology, 26, P2834–2839. doi:10.1 016/j.cub.2016.08.034 Mialon, M., & McCambridge, J. (2018). Alcohol industry corporate social responsibility initiatives and harmful drinking: A systematic review. The European Journal of Public Health, 28, 664–673. doi:10.1093/eurpub/cky065 Michie, S., Atkins, L., & West, R. (2014). The behaviour change wheel: A guide to designing interventions. London, UK: Silverback Publishing. Michie, S., Richardson, M., Johnston, M., Abraham, C., Francis, J., Hardeman, W., ... Wood, C. E. (2013). The behavior change technique taxonomy (v1) of 93 hierarchically clustered 241 techniques: building an international consensus for the reporting of behavior change interventions. Annals of Behavioral Medicine, 46, 81–95. doi:10.1007/s12160-013-9486-6 Mihalic, S. F., & Elliott, D. S. (2015). Evidence-based programs registry: Blueprints for Healthy Youth Development. Evaluation and Program Planning, 48, 124–131. doi:10.1016/ j.evalprogplan.2014.08.004 Mills, K. L., Lalonde, F., Clasen, L. S., Giedd, J. N., & Blakemore, S. J. (2014). Developmental changes in the structure of the social brain in late childhood and adolescence. Social Cognitive and Affective Neuroscience, 9, 123–131. doi:10.1093/scan/nss113 Ministry of Interior. (2019). Eesti valdkondadeülese ennetuse kontseptsioon. Tööversioon 21.01.2019 [Estonian cross-sector prevention concept. Working version 21.01.2019] . Tallinn, Estonia: Ministry of Interior. Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. (2009). Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. BMJ, 339, b2535–b2535. doi:10.1136/bmj.b2535 Molgaard, V. K., Spoth, R. L., & Redmond, C. (2000). Competency training: The Strengthening Families Program. For Parents and Youth 10–14. OJJDP Juvenile Justice Bulletin, 1–11. Molinaro, S., Siciliano, V., Curzio, O., Denoth, F., & Mariani, F. (2012). Concordance and consistency of answers to the self-delivered ESPAD questionnaire on use of psychoactive substances. International Journal of Methods in Psychiatric Research, 21, 158–168. doi:10.1002/mpr.1353 Monti, P. M., Miranda, Jr., R., Nixon, K., Sher, K. J., Swartzwelder, H. S., Tapert, S. F., … Crews, F. T. (2005). Adolescence: Booze, brains and behaviour. Alcoholism: Clinical and Experimental Research, 29, 207–220. doi:10.1097/01.ALC.0000153551.11000.F3 Moore, G. F., Audrey, S., Barker, M., Bond, L., Bonell., Hardeman, W., … Baird, J. (2015). Process evaluation of complex interventions: Medical Research Council guidance. BMJ, 350, h1258. doi:10.1136/bmj.h1258 242 Morales, M., Varlinskaya, E. I., & Spear, L. P. (2011). Age differences in the expression of acute and chronic tolerance to ethanol in male and female rats. Alcoholism: Clinical and Experimental Research, 35, 1614–1624. doi:10.1111/j.1530-0277.2011.01508.x Moretti, M. M., & Peled, M. (2004). Adolescent-parent attachment: Bonds that support healthy development. Paediatrics and Child Health, 9, 551–555. doi:10.1093/pch/9.8.551 Mrazek, P. J., & Haggerty, R. J. (1994). Reducing the risks for mental disorders: Frontiers for preventive intervention research. Washington, DC: National Academy Press: Institute of Medicine. Murphy, E., O’Sullivan, I., O’Donovan, D., Hope, A., & Davoren, M. P. (2016). The association between parental attitudes and alcohol consumption and adolescent alcohol consumption in Southern Ireland: A cross-sectional study. BMC Public Health, 16, 821. doi:10.1186/s12889-016-3504-0 National Institute for Health Development. (2018a). Substance use prevention. Retrieved from http://terviseinfo.ee/et/truekised/download/1019 National Institute for Health Development. (2018b). Alcohol, tobacco and drugs – what can the school do? Retrieved from https://www.terviseinfo.ee/et/truekised/download/909 National Institute for Health Development. (2011a). Õppimine on töö – aita last [Learning is work – help your child]. Tallinn, Estonia: National Institute for Health Development. National Institute for Health Development. (2011b). Pane ennast maksma [Be assertive] . Tallinn, Estonia: National Institute for Health Development. National Institute on Drug Abuse. (2018). Drugs, brains, and behavior: The science of addiction. Retrieved from https://www.drugabuse.gov/publications/drugs-brains-behavior-sci ence-addiction/preface National Institute on Drug Abuse. (2003). Prevention principles. Preventing drug use among children and adolescents (2nd ed.). Retrieved from https://www.drugabuse.gov/sites/default /files/preventingdruguse_2.pdf Navarro, H. J., Doran, C. M., & Shakeshaft, D. A. (2011). Measuring costs of alcohol harm to others: A review of the literature. Drug and Alcohol Dependence, 114, 87–99. doi:10.1016 /j.drugalcdep.2010.11.009 243 Needle, R., McCubbin, H., Wilson, M., Reineck, R., Lazar, A., & Mederer, H. (1986). Interpersonal influences in adolescent drug use – the role of older siblings, parents, and peers. International Journal of the Addictions, 21, 739–766. doi:10.3109/10826088609027390 Nesi, J., Rothenberg, W. A., Hussong, A. M., & Jackson, K. M. (2017). Friends’ alcohol- related social networking site activity predicts escalations in adolescent drinking: Mediation by peer norms. The Journal of Adolescent Health, 60, 641–647. doi:10.1016/j.jadohea lth.2017.01.009 Newbury-Birch, D., Walker, J., Avery, L., Beyer, F., Brown, N., Jackson, K., … Stewart, S. (2009). Impact of alcohol consumption on young people: A systematic review of published reviews. Retrieved from https://dera.ioe.ac.uk/11355/1/DCSF-RR067.pdf Newton, N. C., Champion, K. E., Slade, T., Chapman, C., Stapinski, L., Koning, I. M., … Teesson, M. (2017). A systematic review of combined student- and parent-based programs to prevent alcohol and other drug use among adolescents. Drug and Alcohol Review, 36, 337– 351. doi:10.1111/dar.12497 Nutt, D. (2010, November 5). David Nutt: I am not a prohibitionist. Retrieved from https://www.theguardian.com/commentisfree/2010/nov/05/alcohol-drug-worse-than-heroin Nutt, D. J., King, L. A., & Phillips, L. A. (2010). Drug harms in the UK: A multicriteria decision analysis. The Lancet, 376, 1558–1565. doi:10.1016/S0140-6736(10)61462-6 Oetting, E. R., & Donnermeyer, J. F. (1998). Primary socialization theory: The etiology of drug use and deviance. I. Substance Use & Misuse, 33, 995–1026. doi:10.3109/10826089 809056252 Oncioiu, S. I., Burkhart, G., Calafat, A., Duch, M., Perman-Howe, P., & Foxcroft, D. R. (2018). Environmental substance use prevention interventions in Europe. Retrieved from http://www.emcdda.europa.eu/system/files/publications/7882/Environmental-substance-use- prevention-Interventions- in-Europe.pdf Onrust, S.A., Otten R., Lammers, J., & Smit, F. (2016). School-based programmes to reduce and prevent substance use in different age groups: What works for whom? Systematic review and meta-regression analysis. Clinical Psychology Review, 44, 45–59. doi:10.1016/j.cpr.201 5.11.002 244 Orro, E., Martens, K., Lepane, L., Josing, M., & Reiman, M. (2017). Alcohol market, consumption and harms in Estonia. Yearbook 2017. Tallinn, Estonia: Estonian Institute of Economic Research. Özdemir, M., & Stattin, H. (2012). Does the Örebro Prevention Programme prevent youth drinking? Addiction, 107, 1705–1706. doi:10.1111/j.1360-0443.2012.03959.x Özdemir, M., & Koutakis, N. (2016). Does promoting parents’ negative attitudes to underage drinking reduce adolescents’ drinking? The mediating process and moderators of the effects of the Orebro Prevention Programme. Addiction, 111, 263–271. doi:10.1111/add.13177 Pascual-Leone, A., Amedi, A., Fregni, F., & Merabet, L. B. (2005). The plastic human brain cortex. Annual Review of Neuroscience, 28, 377–401. doi:10.1146/annurev.neuro.27.07020 3.144216 Pedersen, W., Fjar, E., G., Gray, P., & von Soest, T. (2016). Perceptions of harms associated with tobacco, alcohol, and cannabis among students from the UK and Norway. Contemporary Drug Problems, 43, 47–61. doi:10.1177/0091450916638578 Perry, C. L., Williams, C. L., Forster, J. L., Wolfson, M., Wagenaar, A. C., Finnegan, J. R., ... Anstine, P. S. (1993). Background, conceptualization and design of a community-wide research program on adolescent alcohol use: Project Northland. Health Education Research, 8, 125–136. doi:10.1093/her/8.1.125 Perry, C. L., Williams, C. L., Komro, K. A.,Veblen-Mortenson, S., Forster, J. L., Bernstein- Lachter, R., ... McGovern, P. (2000). Project Northland high school interventions: Community action to reduce adolescent alcohol use. Health Education & Behavior, 27, 29– 49. doi:10.1177/109019810002700105 Perry, C. L., Williams, C. L., Komro, K. A.,Veblen-Mortenson, S., Stigler, M. H., Munson, K. A., ... Forster, J. L. (2002). Project Northland: Long-term outcomes of community action to reduce adolescent alcohol use. Health Education Research: Theory & Practice, 17, 117– 132. doi:10.1093/her/17.1.117 Perry, C. L., Williams, C. L., Veblen-Mortenson, S., Toomey, T. L., Komro, K. A., Anstine, P. S., ... Wolfson, M. (1996). Project Northland: Outcomes of a communitywide alcohol use 245 prevention program during early adolescence. American Journal of Public Health, 86, 956– 965. doi:10.2105/ajph.86.7.956 Peters, L. W., Wiefferink, C. H., Hoekstra, F., Buijs, G. J., Ten Dam, G. T., & Paulussen, T. G. (2009). A review of similarities between domain-specific determinants of four health behaviors among adolescents. Health Education Research, 24, 198–223. doi:10.1093/her/cyn 013 Peterson, R. A., & Brown, S. P. (2005). On the use of beta coefficients in meta-analysis. Journal of Applied Psychology, 90, 175–181. doi:10.1037/0021-9010.90.1.175 Peterson, P. L., Hawkins, D. J., Abbott, R. D., & Catalano, R. F. (1994). Disentangling the effects of parental drinking, family management, and parental alcohol norms on current drinking by black and white adolescents. Journal of Research on Adolescence, 4, 203–227. doi:10.1207/s15327795jra0402_3 Petrie, J., Bunn, F., & Byrne, G. (2007). Parenting programmes for preventing tobacco, alcohol or drugs misuse in children <18: A systematic review. Health Education Research, 22, 177–191. doi:10.1093/her/cyl061 Pettersson, C., Linden-Boström, M., & Eriksson, C. (2009). Reasons for non-participation in a parental program concerning underage drinking: A mixed-method study. BMC Public Health, 9, 478. doi:10.1186/1471-2458-9-478 Pettersson, C., Özdemir, M., & Eriksson, C. (2011). Effects of a parental program for preventing underage drinking – the NGO program strong and clear. BMC Public Health, 11, 251. doi:10.1186/1471-2458-11-251 Piko, B. F., & Balázs, M. Á. (2012). Authoritative parenting style and adolescent smoking and drinking. Addictive Behaviors, 37, 353–356. doi:10.1016/j.addbeh.2011.11.022 Pitkänen, T., Lyyra, A-L., & Pulkkinen, L. (2005). Age of onset of drinking and the use of alcohol in adulthood: A follow-up study from age 8–42 for females and males. Addiction, 100, 652–661. doi:10.1111/j.1360-0443.2005.01053.x Prentice, D. A., & Miller, D. T. (1996). Pluralistic ignorance and the perpetuation of social norms by unwitting actors. Advances in Experimental Social Psychology, 28, 161–209. doi:10.1016/S0065-2601(08)60238-5 246 Prins, J. C., Donovan, J. E., & Molina, B. S. G. (2011). Parent-child divergence in the development of alcohol use norms from middle childhood into middle adolescence. Journal of Studies on Alcohol and Drugs, 72, 438–443. doi:10.15288/jsad.2011.72.438 Public Health Agency. (2017). You, your child and alcohol. Retrieved from https://www.pu blichealth.hscni.net/sites/default/files/You%2C%20your%20child%20and%20alcohol%2005 _17_final.pdf Public Health England. (2015). The international evidence on the prevention of drug and alcohol use. Summary and examples of implementation in England. Retrieved from https://www.gov.uk/government/publications/preventing-drug-and-alcohol-misuse-effective- interventions QSR International Pty Ltd. (2018). NVivo qualitative data analysis software. Version 12 [Computer software]. Doncaster, AU: QSR International Pty Ltd. Raitasalo, K., & Holmila, M. (2017). Practices in alcohol education among Finnish parents: Have there been changes between 2006 and 2012? Drugs: Education, Prevention and Policy, 24, 392–399. doi:10.1080/09687637.2016.1183587 Raudne, R. (2012). Drinking to freedom: Themes of continuity and change in contemporary Estonian alcohol discourse (Doctoral dissertation, Johns Hopkins University, Baltimore, MD). Retrieved from ProQuest Dissertations and Theses database. (UMI No. 3536099) Rehm, J., Gmel Sr., G. E., Gmel, G., Hasan, O. S. M., Imtiaz, S., Popova, S., … Shuper, P. A. (2017). The relationship between different dimensions of alcohol use and the burden of disease – An update. Addiction, 112, 968–1001. doi:10.1111/add.13757 Rehm, J., Mathers, C., Popova, S., Thavorncharoensap, M., Teerawattananon, Y., & Patra, J. (2009). Global burden of disease and injury and economic cost attributable to alcohol use and alcohol-use disorders. The Lancet, 373, 2223–2333. doi:10.1016/S0140-6736(09)60746-7 Reynolds, E. K., MacPherson, L., Matusiewicz, A. K., Schreiber, W. M., & Lejuez, C. W. (2011). Discrepancy between mother and child reports of parental knowledge and the relation to risk behavior engagement. Journal of Clinical Child and Adolescent Psychology, 40, 67– 79. doi:10.1080/15374416.2011.533406 247 Robertson, E. B., David, S. L., & Rao, S. A. (2003). Preventing drug use among children and adolescents: A research-based guide for parents, educators and community leaders. (2nd ed). Retrieved from https://www.drugabuse.gov/sites/default/files/preventingdruguse_2.pdf Room, R. (2005). Stigma, social inequality and alcohol and drug use. Drug and Alcohol Reviews, 24, 143–155. doi:10.1080/09595230500102434 Room, R. (2013). Sociocultural aspects of alcohol consumption. In Boyle, P., Boffetta, P., Lowenfels, A. B., Burns, H., Brawley, O., Zatonski, W., & Rehm, J. (Eds.), Alcohol: Science, policy, and public health (pp. 38–45). Oxford, UK: Oxford University Press. Rosenman, R., Goates, S., & Hill, L. (2012). Participation in universal prevention programs. Applied Economics, 44, 219–228. doi:10.1080/00036846.2010.502111 Rosenthal, R., & Rubin, D. B. (1978). Interpersonal expectancy effects: The first 345 studies. Behavioral and Brain Sciences, 1, 377–386. doi:10.1017/S0140525X00075506 Rossow, I., Keating, P., Felix, L., & McCambridge, J. (2016). Does parental drinking influence children’s drinking? A systematic review of prospective cohort studies. Addiction, 111, 204–217. doi:10.1111/add.13097 Roulette, J. W., Hill, L. G., Diversi, M., & Overath, R. (2016). Cultural adaptations of the Strengthening Families Programme 10–14 in the US Pacific Northwest: A qualitative evaluation. Health Education Journal, 76, 169–181. doi:10.1177/0017896916654726 Rutter, M. (1985). Resilience in the face of adversity: Protective factors and resistance to psychiatric disorder. British Journal of Psychiatry, 147, 598–611. doi:10.1192/bjp.147.6.598 Ryan, S. M., Jorm, A. F., Kelly, C. M., Hart, L. M., Morgan, A. J., & Lubman, D. I. (2011). Parenting strategies for reducing adolescent alcohol use: A Delphi consensus study. BMC Public Health, 11, 13. doi:10.1186/1471-2458-11-13 Ryan, S. M., Jorm, A. F., & Lubman, D. I. (2010). Parenting factors associated with reduced adolescent alcohol use: A systematic review of longitudinal studies. The Australian and New Zealand Journal of Psychiatry, 44, 774–783. doi:10.1080/00048674.2010.501759 Saat, H. (2011). Ole toeks lapse enesehinnangu kujunemisel [Support the development of your child’s self-esteem]. Tallinn, Estonia: National Institute for Health Development. 248 Sanchez-Ramirez, D. C., Franklin, R.C., & Voaklander, D. (2018). Perceptions about alcohol harm and alcohol-control strategies among people with high risk of alcohol consumption in Alberta, Canada and Queensland, Australia. Journal of Preventive Medicine & Public Health, 51, 41–50. doi:10.3961/jpmph.17.112 Schulenberg, J. E., & Maggs, J. L. (2002). A developmental perspective on alcohol use and heavy drinking during adolescence and the transition to young adulthood. Journal of Studies on Alcohol and Drugs, Supplement, s14, 54–70. doi:10.15288/jsas.2002.s14.54 Schulte, M. T., Ramo, D., & Brown, S. A. (2009). Gender differences in factors influencing alcohol use and drinking progression among adolescents. Clinical Psychological Review, 29, 535–547. doi:10.1016/j.cpr.2009.06.003 Serbanati, L.D., Ricci, F.L., Mercurio, G. & Vasilateanu, A. (2011). Steps towards a digital health ecosystem. Journal of Biomedical Informatics, 44, 621–636. doi:10.1016/j.jbi.2011.02. 011 Sharmin, S., Kypri, K., Khanam, M., Wadolowski, M., Bruno, R., Attia, J., … Mattick, R. P. (2017b). Effects of parental alcohol rules on risky drinking and related problems in adolescence: Systematic review and meta-analysis. Drug and Alcohol Dependence, 178, 243– 256. doi:10.1016/j.drugalcdep.2017.05.011 Sharmin, S., Kypri, K., Khanam, M., Wadolowski, M., Bruno, R., & Mattick, R. P. (2017a). Parental supply of alcohol in childhood and risky drinking in adolescence: Systematic review and meta-analysis. International Journal of Environmental Research and Public Health, 14, 287. doi:10.3390/ijerph14030287 Shepperd, J.A., Carroll, P., Grace, J., & Terry, M. (2002). Exploring the causes of comparative optimism. Psychologica Belgica, 42, 65–98. Shield, K. D., Rylett, M., & Rehm, J. (2016). Public health successes and missed opportunities. Trends in alcohol consumption and attributable mortality in the WHO European Region, 1990–2014. Copenhagen, Denmark: WHO Regional Office for Europe. Sieving, R. E., Maruyama, G., Williams, C. L., & Perry, C. L. (2000). Pathways to adolescent alcohol use: Potential mechanisms of parent influence. Journal of Research on Adolescence, 10, 489–514. doi:10.1207/sjra1004_06 249 Silveri, M. M., & Spear, L.P. (1998). Decreased sensitivity to the hypnotic effects of ethanol early in ontogeny. Alcoholism: Clinical and Experimental Research, 22, 670–676. doi:10.111 1/j.1530-0277.1998.tb04310.x Skärstrand, E., Larsson, J., & Andréasson, S. (2008). Cultural adaptation of the Strengthening Families Programme to a Swedish setting. Health Education, 108, 287–300. doi:10.1108/0965428081 0884179 Skärstrand, E., Sundell, K., & Andréasson, S. (2014). Evaluation of a Swedish version of the Strengthening Families Programme. The European Journal of Public Health, 24, 578–584. doi:10.1093/eurpub/ckt146 Smit, E., Verdurmen, J. E. E., Monshouwer, K., & Smit, F. (2008). Family interventions and their effect on adolescent alcohol use in general populations; a meta-analysis of randomized controlled trials. Drug and Alcohol Dependence, 97, 195–206. doi:10.1016/j.drugalcdep.2008 .03.032 Smith, G. T., Miller, T. L., Kroll, L., Simmons, J. R., & Gallen, R. (1999). Children’s perceptions of parental drinking: The eye of the beholder. Journal of Studies on Alcohol, 60, 817–824. doi:10.15288/jsa.1999.60.817 Spear, L. P. (2018). Effects of adolescent alcohol consumption on the brain and behaviour. Nature Reviews. Neuroscience, 19, 197–214. doi:10.1038/nrn.2018.10 Spear, L. P. (2010). The behavioural neuroscience of adolescence. New York, NY: W. W. Norton & Company. Spear, H. J., & Kulbok, P. (2004). Autonomy and adolescence: A concept analysis. Public Health Nursing, 21, 144–152. doi:10.1111/j.0737-1209.2004.021208.x Spoth, R., & Redmond, C. (2000). Research on family engagement in preventive interventions: Toward improved use of scientific findings in primary prevention practice. The Journal of Primary Prevention, 21, 267–284. doi:10.1023/A:1007039421026 Spoth, R.L, Redmond, C., & Shin, C. (2001). Randomised trial of brief family interventions for general populations: Adolescent substance use outcomes 4 years following baseline. 250 Journal of Consulting and Clinical Psychology, 69, 627–642. doi:10.1037//0022-006x.69.4 .627 Spoth, R.L, Redmond, C., & Shin, C. (1998). Direct and indirect latent-variable parenting outcomes of two universal family-focused preventive interventions: Extending a public health-oriented research base. Journal of Consulting and Clinical Psychology, 66, 385–399. doi:10.1037/0022-006x.66.2.385 Spoth, R., Shin, C., Guyll, M., Redmond, C., & Azevedo, K. (2006). Universality of effects: An examination of the comparability of long-term family intervention effects on substance use across risk-related subgroups. Prevention Science, 7, 209–224. doi:10.1007/s11121-006- 0036-3 Stafford, M., Kuh, D. L., Gale, C. R., Mishra, G., & Richards, M. (2016). Parent-child relationships and offspring’s positive mental wellbeing from adolescence to early older age. The Journal of Positive Psychology, 11, 326–337. doi:10.1080/17439760.2015.1081971 StataCorp. (2015). Stata Statistical Software: Release 14 [Computer software]. College Station, TX: StataCorp LP. StataCorp. (2013). Stata structural equation modeling. Reference manual. (Release 13). Retrieved from https://www.stata.com/manuals13/sem.pdf Stattin, H. (2002). “10 till 18” i Köping – en design för ungdomsstudier [“10 to 18” in Köping – a design for youth studies]. In Nygard, B. (Ed.), Ungdomsforskning – skilda världar och värderingar [Youth research – separated worlds and values] (pp. 36–39). Retrieved from https://www.yumpu.com/sv/document/read/18347194/ladda-ner- fas/5 Stautz, K., & Cooper, A. (2013). Impulsivity-related personality traits and adolescent alcohol use: A meta-analytic review. Clinical Psychology Review, 33, 574–592. doi:10.1016/j.cpr .2013.03.003 Steckler, A., McLeroy, K. R., Goodman, R. M., Bird, S. T., & McCormick, L. (1992). Toward integrating qualitative and quantitative methods: An introduction. Health Education Quarterly, 19, 1–8. doi:10.1177/109019819201900101 Steinberg, L. (2010). A dual systems model of adolescent risk-taking. Developmental Psychobiology, 52, 216–224. doi:10.1002/dev.20445 251 Steinberg, L. (2008). A social neuroscience perspective on adolescent risk-taking. Developmental Review, 28, 78–106. doi:10.1016/j.dr.2007.08.002 Steinberg, L. (2001). We know some things: Parent-adolescent relationships in retrospect and prospect. Journal of Research on Adolescence, 11, 1–19. doi:10.1111/1532-7795.00001 Sterne, J. A. C., Egger, M., & Moher, D. (2011). Chapter 10: Addressing reporting biases. In Higgins, J. P. T., & Green, S. (Eds.), Cochrane handbook for systematic reviews of interventions. Version 5.1.0 (updated March 2011). Retrieved from https://handbook-5- 1.cochrane.org/chapter_10/10_addressing_reporting_biases.htm Stewart Burgher, M., Barnekow Rasmussen, V., & Rivett, D. (1999). The European Network of Health Promoting Schools: The alliance of education and health. Retrieved from http://www.euro.who.int/__data/assets/pdf_file/0004/252391/E62361.pdf?ua=1 Stigler, M. H., Perry, C. L., Komro, K. A., Cudeck, R., & Williams, C. L. (2006). Teasing apart a multiple component approach to adolescent alcohol prevention: What worked in Project Northland? Prevention Science, 7, 269–280. doi:10.1007/s11121-006-0040-7 Stockings, E., Hall, W. D., Lynskey, M., Morley, K. I., Reavley, N., Strang, J., … Degenhardt, L. (2016). Prevention, early intervention, harm reduction, and treatment of substance use in young people. The Lancet Psychiatry, 3, 280–296. doi:10.1016/S2215- 0366(16)00002-X Strandberg, A. (2014). Evaluation of a Swedish parental prevention program: Youth drunkenness, alcohol-specific parenting and gender differences. (Doctoral dissertation, Karolinska Institutet, Stockholm, Sweden). Retrieved from https://openarchive.ki.se/xmlui /bitstream/handle/10616/42240/Thesis_Anna_Strandberg.pdf?sequence=1 Strandberg, A. K., & Bodin, M. C. (2011). Alcohol‐specific parenting within a cluster‐ randomized effectiveness trial of a Swedish primary prevention program. Health Education, 111, 92–102. doi:10.1108/09654281111108526 Strandberg, A. K., Bodin, M. C., & Romelsjö, A. (2014). Gender differences in the prediction of parental servings of alcohol to adolescents and youth drunkenness. Substance Use & Misuse, 49, 1857–1866. doi:10.3109/10826084.2014.913628 252 Streimann, K., & Pertel, T. (2016). Uimastiennetusjuhend koolidele [Substance use prevention guidelines for schools]. Tallinn, Estonia: National Institute for health Development. Streimann, K., Selart, A., & Trummal, A. (2019). Effectiveness of a universal, classroom- based preventive intervention (PAX GBG) in Estonia: A cluster-randomized controlled trial. Prevention Science. doi:10.1007/s11121-019-01050-0 Streimann, K., Trummal, A., Klandorf, K., Akkermann, K., Sisask, M., Toros, K., & Selart, A. (2017). Effectiveness of a universal classroom-based preventive intervention (PAX GBG): A research protocol for a matched-pair cluster-randomized controlled trial. Contemporary Clinical Trials Communications, 8, 75–84. doi:10.1016/j.conctc.2017.08.013 Strøm, H. K., Adolfsen, F., Handegård, B. H., Natvig, H., Eisemann, M., Martinussen, M., & Koposov, R. (2015). Preventing alcohol use with a universal school-based intervention: Results from an effectiveness study. BMC Public Health, 15, 337. doi:10.1186/s12889-015- 1704-7 Stueve, A., & O’Donnell, L. N. (2005). Early alcohol initiation and subsequent sexual and alcohol risk behaviors among urban youths. American Journal of Public Health, 95, 887–893. doi:10.2105/AJPH.2003.026567 Substance Abuse and Mental Health Services Administration. (2017). Risk and protective factors and estimates of substance use initiation: Results from the 2016 National Survey on Drug Use and Health. Retrieved from https://www.samhsa.gov/data/sites/default/files/ NSDUH-DR-FFR3-2016/NSDUH-DR-FFR3-2016.htm Sutton, S. (2004). Determinants of health-related behaviours: Theoretical and methodological issues. In Sutton, S., Baum, A., & Johnston, M. (Eds.), The SAGE handbook of health psychology (pp. 94–126). London, UK: SAGE Publications Ltd. Tael-Öeren, M., Naughton, F., & Sutton, S. (2019a). The relationship between parental attitudes and children’s alcohol use: A systematic review and meta-analysis. Addiction, 114, 1527–1546. doi:10.1111/add.14615 Tael-Öeren, M., Naughton, F., & Sutton, S. (2019b). A parent-oriented alcohol prevention programme “Effekt” had no impact on adolescents’ alcohol use: Findings from a cluster- 253 randomised controlled trial in Estonia. Drug and Alcohol Dependence, 194, 279–287. doi:10.1016/j.drugalcdep.2018.10.024 Taylor, K., Babineau, K., Keogan, S., Whelan, E., & Clancy, L. (2016). ESPAD 2015: European Schools Project on Alcohol and Other Drugs in Ireland. Retrieved from http://alcoholireland.ie/download/reports/alcohol_health/children_young_people/Irish-ESPA D-Report-20151.compressed.pdf Taylor, B., Irving, H. M., Kanteres, F., Room, R., Borges, G, Cherpitel, C. J., … Rehm. J. (2010). The more you drink the harder you fall: A systematic review and meta-analysis of how acute alcohol consumption and injury or collision risk increase together. Drug and Alcohol Dependence, 110, 108–116. doi:10.1016/j.drugalcdep.2010.02.011 Testa, M., Hoffman, J. H., Livingston, J. A., & Turrisi, R. (2010). Preventing college women's sexual victimization through Parent Based Intervention: A randomized controlled trial. Prevention Science, 11, 308–318. doi:10.1007/s11121-010-0168-3 Tildesley, E. A. & Andrews, J. A. (2008). The development of children’s intentions to use alcohol: Direct and indirect effects of parent alcohol use and parenting behaviors. Psychology of Addictive Behaviors, 22, 326–339. doi:10.1037/0893-164X.22.3.326 Turrisi, R., Jaccard, J., Taki, R., Dunnam, H., & Grimes, J. (2001). Examination of the short- term efficacy of a parent intervention to reduce college student drinking tendencies. Psychology of Addictive Behaviors, 15, 366–372. doi:10.1037/0893-164x.15.4.366 Turrisi, R., Larimer, M. E., Mallett, K. A., Kilmer, J. R., Ray, A. E., Mastroleo, N. R., ... Montoya, H. (2009). A randomized clinical trial evaluating a combined alcohol intervention for high-risk college students. Journal of Studies on Alcohol and Drugs,70, 555–567. doi:10.15288/jsad.2009.70.555 Ungar, W. J., Boydell, K., Dell, S., Feldman, B. M., Marshall, D., Willan, A., & Wright, J. G. (2012). A parent-child dyad approach to the assessment of health status and health-related quality of life in children with asthma. Pharmacoeconomics, 30, 697–712. doi:10.2165/11597 890-000000000-00000 254 United Nations Office on Drug and Crime. (2018). International standard on drug use prevention. Second updated edition. Vienna, Austria: United Nations Office on Drug and Crime. United Nations Office on Drug and Crime. (2009). Guide to implementing family skills training programmes for drug abuse prevention. Vienna, Austria: United Nations Office on Drug and Crime. Urbaniak, G. C., & Plous, S. (2007). Research Randomizer (Version 3.0) [Computer software]. Retrieved from http://www.randomizer.org/ Valentine, G., Jayne, M., Gould, M., & Keenan, J. (2010). Family life and alcohol consumption: A study of the transmission of drinking practices. Retrieved from https://www.jrf.org.uk/file/40688/download?token=S0x1eDII&filetype=full-report Van den Bulk, B. G., Koolschijn, P. C. M. P., Meens, P. H. F., van Lang, N. D. J., Van der Vee, N. J. A., Rombouts, S. A. R. B., … Crone, E. A. (2013). How stable is activation in the amygdala and prefrontal cortex in adolescence? A study of emotional face processing across three measurements. Developmental Cognitive Neuroscience, 4, 65–76. doi:10.1016/j.dcn .2012.09.005 Van der Vorst, H., Engels, R. C. M. E., Meeus, W., & Deković, M. (2006). The impact of alcohol-specific rules, parental norms about early drinking and parental alcohol use on adolescents’ drinking behavior. Journal of Child Psychology and Psychiatry and Allied Disciplines, 47, 1299–1306. doi:10.1111/j.1469-7610.2006.01680.x Van der Vorst, H., Engels, R. C. M. E., Meeus, W., Deković, M., & Van Leeuwe, J. (2005). The role of alcohol-specific socialization in adolescents’ drinking behaviour. Addiction, 100, 1464–1476. doi:10.1111/j.1360-0443.2005.01193.x Van Ryzin, M. J., Roseth, C. J., Fosco, G. M., Lee, Y. K., & Chen, I. C. (2016). A component-centered meta-analysis of family-based prevention programs for adolescent substance use. Clinical Psychology Review, 45, 72–80. doi:10.1016/j.cpr.2016.03.007 Velleman, R. D. B., Templeton, L. J., & Copello, A. G. (2005). The role of the family in preventing and intervening with substance use and misuse: A comprehensive review of family interventions, with a focus on young people. Drug and Alcohol Review, 24, 93–109. doi:10.1080/09595230500167478 255 Verdurmen, J. E. E., Koning, I. M., Vollebergh, W. A. M., Van den Eijnden, R. J. J. M., & Engels, R. C. M. E. (2014). Risk moderation of a parent and student preventive alcohol intervention by adolescent and family factors: A cluster randomized trial. Preventive Medicine, 60, 88–94. doi:10.1016/j.ypmed.2013.12.027 Visser, L., de Winter, A. F., & Reijneveld, S. A. (2012). The parent-child relationship and adolescent alcohol use: A systematic review of longitudinal studies. BMC Public Health, 12, 1. doi:10.1186/1471-2458-12-886 Voogt, C., Beusink, M., Kleinjan, M., Otten, R., Engels, R., Smit, K., & Kuntsche, E. (2017a). Alcohol-related cognitions in children (aged 2–10) and how they are shaped by parental alcohol use: A systematic review. Drug and Alcohol Dependence, 177, 277–290. doi:10.1016/j.drugalcdep.2017.04.006 Voogt, C., Otten, R., Kleinjan, M., Engels, R., & Kuntsche, E. (2017b). Alcohol-related knowledge and alcohol-related norms in 4- to 6-year-olds – Evidence from the Dutch electronic Appropriate Beverage Task. Alcoholism: Clinical and Experimental Research, 41, 1637–1647. doi:10.1111/acer.13452 Voolma, S-R. (2017). The development and feasibility testing of a digital health intervention for reducing Estonian adolescent and young adult alcohol and tobacco consumption (Doctoral dissertation, University of Cambridge, Cambridge, UK). Retrieved from https://www.repository.cam.ac.uk/bitstream/handle/1810/267955/Voolma-2017-PhD.pdf?seq uence=1&isAllowed=y Wagenaar, A. C., Komro, K. A., McGovern, P., Williams, C. L., & Perry, C. L. (1993). Effects of a saliva test pipeline procedure on adolescent self-reported alcohol use. Addiction, 88, 199–208. doi:10.1111/j.1360-0443.1993.tb00803.x Wallace, D. S., Paulson, R. M., Lord, C. G., & Bond Jr., C. F. (2005). Which behaviors do attitudes predict? Meta-analyzing the effects of social pressure and perceived difficulty. Review of General Psychology, 9, 214–227. doi:10.1037/1089-2680.9.3.214 Wells, M. B., Sarkadi, A., & Salari, R. (2016). Mothers’ and fathers’ attendance in a community-based universally offered parenting program in Sweden. Scandinavian Journal of Public Health, 44, 274–280. doi:10.1177/1403494815618841 256 West, R. (2016). Using Bayesian analysis for hypothesis testing in addiction science. Addiction, 111, 3–4. doi:10.1111/add.13053 West, B., Abatemarco, D., Ohman-Strickland, P. A., Zec, V., Russo, A., & Milic, R. (2008). Project Northland in Croatia: Results and lessons learned. Journal of Drug Education, 38, 55–70. doi:10.2190/DE.38.1.e Whittaker, K. A., & Cowley, S. (2012). An effective programme is not enough: A review of factors associated with poor attendance and engagement with parenting support programmes. Children & Society, 26, 138–149. doi:10.1111/j.1099-0860.2010.00333.x Wickman, M. E., Anderson, N. L. R., & Smith Greenberg, C. (2008). The adolescent perception of invincibility and its influence on teen acceptance of health promotion strategies. Journal of Pediatric Nursing, 23, 460–468. doi:10.1016/j.pedn.2008.02.003 Williams, C. L., Grechanaia, T., Romanova, O., Komro, K. A., Perry, C. L., & Farbakhsh, K. (2001). Russian-American partners for prevention adaptation of a school-based parent-child programme for alcohol use prevention. European Journal of Public Health, 11, 314–321. doi:10.1093/eurpub/11.3.314 Williams, R. J., McDermitt, D. R., Bertrand, L. D., & Davis, R. M. (2003). Parental awareness of adolescent substance use. Addictive Behaviors, 28, 803–809. doi:10.1016/S030 6-4603(01)00275-1 Williams, C. L., Perry, C. L., Farbakhsh, K., & Veblen-Mortenson, S. (1999). Project Northland: Comprehensive alcohol use prevention for young adolescents, their parents, schools peers and communities. Journal of Studies on Alcohol, s13, 112–124. doi:10.1528 8/jsas.1999.s13.112 Windle, M., Spear, L. P., Fuligni, A. J., Angold, A., Brown, J. D., Pine, D., … Dahl, R. E. (2008). Transitions into underage and problem drinking: Summary of developmental processes and mechanisms: Ages 10–15. Pediatrics, 121, S273–S289. doi:10.1542/peds.2 007-2243C Winters, K. C., & Lee, C-Y. S. (2008). Likelihood of developing an alcohol and cannabis use disorder during youth: Association with recent use and age. Drug and Alcohol Dependence, 92, 239–247. doi:10.1016/j.drugalcdep.2007.08.005 257 Wolf, J. P., & Chávez, R. (2015). “Just make sure you can get up and parent the next day”: Understanding the contexts, risks, and rewards of alcohol consumption for parents. Families in Society, 96, 219–228. doi:10.1606/1044-3894.2015.96.28 Wood, M. D., Fairlie, A. M., Fernandez, A. C., Borsari, B., Capone, C, Laforge, R., & Carmona-Barros, R. (2010). Brief motivational and parent interventions for college students: A randomized factorial study. Journal of Consulting and Clinical Psychology, 78, 349–361. doi:10.1037/a0019166 Wood, M. D., Mitchell, R. E., Read, J. P., & Brand, N. H. (2004). Do parents still matter? Parent and peer influences on alcohol involvement among recent high school graduates. Psychology of Addictive Behaviors, 18, 19–30. doi:10.1037/0893-164X.18.1.19 World Health Organization. (2018). Global status report on alcohol and health 2018. Geneva, Switzerland: World Health Organization. World Health Organization. (2014). Global status report on alcohol and health 2014. Geneva, Switzerland: World Health Organization. World Health Organization. (2010). Global strategy to reduce the harmful use of alcohol. Geneva, Switzerland: World Health Organization. World Health Organization. (1999). Health 21 – The health for all policy framework for the WHO European Region. Retrieved from http://www.euro.who.int/__data/assets/pdf_file/001 0/98398/wa540ga199heeng.pdf?ua=1 Wurdak, M., Kuntsche, E., & Wolstein, J. (2016). Effectiveness of an email-based intervention helping parents to enhance alcohol-related parenting skills and reduce their children’s alcohol consumption – A randomised controlled trial. Drugs: Education, Prevention and Policy, 24, 144–151. doi:10.1080/09687637.2016.1201459 Yanow, D. (2014). Thinking interpretively: Philosophical presuppositions and the human sciences. In Yanow, D., & Schwartz-Shea, P. (Eds), Interpretation and method. Empirical research methods and the interpretive turn (2nd ed., pp. 5–26). New York, NY: Routledge. 2nd ed. Yap, M. B. H., Cheong, T. W. K., Zaravinos-Tsakos, F., Lubman, D. I., & Jorm, A. F. (2017). Modifiable parenting factors associated with adolescent alcohol misuse: A systematic 258 review and meta-analysis of longitudinal studies. Addiction, 112, 1142–1162. doi:10.1111/a dd.13785 Yeo, A., Legard, R., Keegan, J., Ward, K., McNaughton Nicholls, C., & Lewis, J. (2014). In- depth interviews. In Ritchie, J., Lewis, J., McNaughton Nicholls, C., & Ormston, R. (Eds.), Qualitative research practice. A guide for social science students and researchers (2nd ed., pp. 177–210). London, UK: SAGE Publications Ltd. Yu, J. (2003). The association between parental alcohol-related behaviors and children’s drinking. Drug and Alcohol Dependence, 69, 253–262. doi:10.1016/S0376-8716(02)00324-1 Zehe, J., & Colder, C. (2014). A latent growth curve analysis of alcohol-use specific parenting and adolescent alcohol use. Addictive Behaviors, 39, 1701–1705. doi:10.1007 /s11103-011-9767-z.Plastid 259 Appendices 260 Appendix 1. Systematic review and meta-analysis on the relationship between parental attitudes and children’s alcohol use published in Addiction 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 Appendix 2. Quantitative evaluation of the Effekt trial published in Drug and Alcohol Dependence 281 282 283 284 285 286 287 288 289 Appendix 3. Supplementary material from the meta-analysis Appendix 3.1 Search strategy Table S 1. Description of search strategy in all included electronic databases Medline ((((parent*).ti,ab OR (mother*).ti,ab OR (father*).ti,ab OR exp PARENTING/ OR exp PARENTS/) AND ((child*).ti,ab OR (underage*).ti,ab OR (youth*).ti,ab OR (daughter*).ti,ab OR exp ADOLESCENT/ OR (adolescen*).ti,ab OR (son).ti,ab OR (sons).ti,ab OR (teen).ti,ab OR (teens).ti,ab OR (teenage*).ti,ab OR (young*).ti,ab)) AND ((alcohol*).ti,ab OR (drunk*).ti,ab OR exp BINGE DRINKING/ OR exp ALCOHOL DRINKING/ OR exp ALCOHOLIC INTOXICATION/ OR (drink*).ti,ab OR ("binge drink*").ti,ab OR (intoxicat*).ti,ab)) AND ((norm*).ti,ab OR ((attitude*).ti,ab OR exp ATTITUDE/ OR (approv*).ti,ab OR (disapprov*).ti,ab)) PsycINFO ((((parent*).ti,ab OR (mother*).ti,ab OR (father*).ti,ab OR exp PARENTING/ OR exp PARENTS/) AND ((child*).ti,ab OR (underage*).ti,ab OR (youth*).ti,ab OR (daughter*).ti,ab OR (adolescen*).ti,ab OR (son).ti,ab OR (sons).ti,ab OR (teen).ti,ab OR (teens).ti,ab OR (teenage*).ti,ab OR (young*).ti,ab)) AND ((alcohol*).ti,ab OR (drunk*).ti,ab OR exp BINGE DRINKING/ OR exp ALCOHOL INTOXICATION/ OR exp UNDERAGE DRINKING/ OR (drink*).ti,ab OR ("binge drink*").ti,ab OR (intoxicat*).ti,ab)) AND ((attitude*).ti,ab OR exp ATTITUDES/ OR (approv*).ti,ab OR (disapprov*).ti,ab OR (norm*).ti,ab) EMBASE ((((parent*).ti,ab OR (mother*).ti,ab OR (father*).ti,ab OR exp PARENT/) AND ((child*).ti,ab OR (underage*).ti,ab OR (youth*).ti,ab OR (daughter*).ti,ab OR (adolescen*).ti,ab OR (son).ti,ab OR (sons).ti,ab OR (teen).ti,ab OR (teens).ti,ab OR (teenage*).ti,ab OR (young*).ti,ab OR exp ADOLESCENT/)) AND ((alcohol*).ti,ab OR (drunk*).ti,ab OR exp BINGE DRINKING/ OR exp DRINKING/ OR exp ALCOHOL INTOXICATION/ OR (drink*).ti,ab OR ("binge drink*").ti,ab OR (intoxicat*).ti,ab OR ("heavy episodic drinking").ti,ab)) AND ((norm*).ti,ab OR ((attitude*).ti,ab OR exp ATTITUDE/ OR (approv*).ti,ab OR (disapprov*).ti,ab)) Scopus (TITLE-ABS-KEY("alcohol*" OR "binge drink*" OR "drunk*" OR "drink*" OR "intoxicat*")) AND (TITLE-ABS-KEY("parent*" OR "mother*" OR "father*")) AND (TITLE-ABS-KEY("teen" OR "teens" OR "teenage*" OR "child*" OR "underage*" OR "youth*" OR "young*" OR "son" OR "sons" OR "daughter*" OR "adolescen*")) AND (TITLE-ABS-KEY("attitude*" OR "approv*" OR "disapprov*" OR "norm*")) Web of Science TS=(alcohol* OR binge drink* OR drunk* OR intoxicat* OR drink*) AND TS=(parent* OR mother* OR father*) AND TS=("teen" OR "teens" OR teenage* OR child* OR underage* OR youth* OR "son" OR "sons" OR daughter* OR adolescen* OR young*) AND TS=(attitude* OR approv* OR disapprov* OR norm*) Refined By: [excluding]: Databases: (KJD OR MEDLINE OR DIIDW OR ZOOREC) 290 Appendix 3.2 Quality assessment Table S 2. Methodological quality assessment of included studies Study Study participation Study attrition Predictor measurement Outcome measurement Confounding measurement Analysis Number of biases Aas et al., 1996 - NA - - - + 4 Andrews et al., 1993 + - - - + + 3 Ary et al., 1993 + - + - - + 3 Brody et al., 2000 + - + - - + 3 Colder et al., 2018 + - - - + + 3 Donovan & Molina, 2008 + NA + - - + 2 Donovan & Molina, 2011 + + + + + + 0 Donovan & Molina, 2014 + + + - + + 1 Ennett et al., 2001 - + + - + + 2 Gerrard et al., 2000 - - - - - + 5 Glatz et al., 2012 + - - - - + 4 Jackson et al., 2012 - NA + - + + 2 Järvinen & Østergaard, 2009 + NA + - - - 3 Kerr et al., 2012 + - + - - + 3 Koning et al., 2010a + NA + + + + 0 Koning et al., 2010b + + + - - + 2 Koning et al., 2012a + - + - + + 2 Koning et al., 2013 - - + - - + 4 Mares et al., 2011a - + + - - + 3 Margulies et al., 1977 - - - - + + 4 Murphy et al., 2016 + NA + + + + 0 Needle et al., 1986 + NA - + - + 2 Özdemir & Koutakis, 2016 - + - - - + 4 Peterson et al., 1994 - - + - + + 3 Pettersson et al., 2011 - - + - - + 4 Sieving et al., 2000 - - - - - + 5 Strandberg et al., 2014 + + - - + + 2 Tael-Öeren et al., 2019b + - + - + + 2 Van der Vorst et al., 2006 + - + - - + 3 Yu, 2003 + NA + - + + 1 291 Appendix 4. Supplementary material from the Effekt programme Appendix 4.1 An overview of the assessments carried out during the programme 292 Appendix 4.2 Feedback form An example of a form used to collect programme-related feedback from parents at the end of the meetings: 1) Did you attend the previous meeting in autumn? ▢ Yes ▢ No 2) Did you receive the meeting’s summary in autumn? ▢ Yes ▢ No 3) Did you receive the programme’s newsletter(s)? ▢ Yes, I received the first one at the beginning of the year ▢ Yes, I receieved the second one in April ▢ No 4) Did you read the newsletter(s)? ▢ Yes, I read the first newsletter ▢ Yes, I read the second newsletter ▢ No 5) Did the newsletter/meeting help you talk to your child about alcohol-related topics? ▢ Yes ▢ No Comments: 6) Which of the topics discussed today are you planning to share with your family members? 7) Please assess our facilitator - was s/he ... ? Yes No I don’t know credible ▢ ▢ ▢ enthusiastic ▢ ▢ ▢ understandable ▢ ▢ ▢ able to answer the questions ▢ ▢ ▢ Comments: 8) What is your sex? ▢ Male ▢ Female Date: 293 Appendix 4.3 Example of a newsletter Example of a newsletter (#2) sent to parents (in Estonian) – “Parent’s attitudes – an important factor in children’s alcohol use prevention” 294 Appendix 4.4 Case study scenarios and a problem-solving model Scenarios presented to parents: 1) The child consumed alcohol at school – the parent is notified about the situation at the school. The teacher explains that one of the students in the class took a bottle of vodka and a pack of juice from home, and then secretly drank alcohol with friends in the school toilet until caught by an older student who notified the teacher; 2) Sixth and seventh-grade students consumed alcohol on a two-day school trip – the teacher heard a noise at night and found drunk minors in one of the rooms. Older students had hidden beer and cider into the luggage compartment in the bus before the bags were checked and quickly moved alcohol away at arrival; 3) The child came home drunk for the first time – it is Saturday evening, and the child arrives home, trying to sneak into his/her room. The parent sees that, and when approaching the child, the parent smells alcohol odour and understands that the child is drunk. This is the first time the parent becomes aware of child’s alcohol use; 4) The child consumed alcohol with friends and was caught by the police – on Friday evening, the parent receives a phone call from the police, being asked to pick up his/her drunk child. The child was supposed to be at a friend’s place. At the police station, the parent hears from the child’s friends that his/her child was the one who had brought alcohol from home. The parent’s attitudes towards alcohol use have always been restrictive, and problems like that have not occurred before. A universal four-step problem-solving model: 1) Stay calm and think the situation through: a. What is it that you wish to know? What is the aim of the discussion with the child – achieve understanding each other so the situation would not repeat; b. Let the primary emotions (e.g. disappointment, anger, hopelessness) wear off and do not express them on the child – scaring, moralising and “interrogating” usually do not work. Think about why you have these feelings; c. Think about your own attitude towards adolescents’ alcohol use – have you clearly expressed it. 295 2) Apply active listening: a. Agree on the time to have a discussion with the child – preferably not on the same day, so both parties can reflect on their thoughts. If the child is drunk, wait until he/she gets sober; b. Express your feelings by using “I” approach (e.g. I was worried, I did not know were you okay, I felt very frightened) and then give the child the chance to explain his/her version of the incident, without fearing your negative reaction; c. The aim is to understand the cause – which circumstances led to this behaviour. It is recommended to use open-ended questions (e.g. Please tell me what happened; How did it affect you? What would you do in the same situation in the future?) and if possible, avoid using “why” questions, as they tend to sound more aggressive. If the child says that he/she enjoyed alcohol use, then do not get angry and tone down your negative emotions; d. Discuss what to do, to avoid the situation recurring (primarily, let the child explore the options). It is important to stay calm and supportive as it encourages the child to be more open; e. This kind of discussion has a positive effect on the parent-child relationship – the child values parents’ rational attitude and dares to talk to them about future problems, the child knows that the parents will listen to him/her. 3) Share information: a. Parents’ role is to be a reliable source of information for the child. Thus, it is recommended to explain why this behaviour is not good for the child – e.g. talking about short-term consequences. Parents can be prepared to talk about the topic, but not scare the child, as it is difficult for them to relate to frightening examples; b. Knowledge on the topic can also help when arguing and justifying own viewpoints; c. If you feel that your knowledge is limited, visit websites and read books that are related to the topic. 296 4) Make agreements: a. The discussion should end with an agreement, that states the behaviour is not acceptable and cannot repeat. Talk about the impact of child behaviour on other people and think of different strategies to apply; b. Depending on the incident, it can be reasonable to interact with other parents and make agreements. Talk to the class teacher, as the teacher might know some additional information that parents are not aware of; c. If the child cannot comply with the agreement(s), then it is important to find out what impedes it. If you cannot find a solution, you can ask advice from the teacher, school psychologist, social pedagogue or from another adult, who the child perceives as an authority. An option is to use an online counselling service; d. Punishing is not an alternative to the agreements, because: i. the child might not approach parents with future problems, ii. studies show that punishing increases behavioural problems, iii. physical punishing might result in the child being violent to others, iv. physical punishing intensifies low self-esteem and might cause anxiety disorders. Whenever an obstacle occurs in the process, the parent can refer back to step two – applying active listening. Active listening, empathy and honest self-expression are crucial factors in the problem-solving process. 297 Appendix 4.5 Alcohol advertisement analysis Questions used in the alcohol advertisement analysis discussion:  What did you notice (e.g. music, colours, emotions, symbols)?  What did you like/dislike about the video?  Is the video realistic?  How much did these people drink?  What kind of information was left out?  Who is the target group?  What kind of thoughts did occur?  Based on this video, what kind of people drink alcohol?  What kind of consequences can this kind of alcohol use have?  How persuasive was the video?  How can children interpret this video? Do children understand the consequences of alcohol use?  How likely can children critically analyse this video on their own? 298 Appendix 4.6 Fictional stories used in the roleplay Story 1 It was Kadri’s 15th birthday party with her family and friends. Adults were drinking alcohol, and as Kadri saw some bottles on the table, she asked her mother’s permission to take a sip. The mother refused, but later the girl decided to try some cider in secret and was caught by her mother. Kadri justified her behaviour by saying that she was almost an adult, and all her friends had tried alcohol, so it was unfair that she was not allowed to try. Although being upset in the beginning, Kadri’s mother was persuaded and allowed Kadri to take a few sips. Story 2 Kadri and her friends went to a party at her classmate’s home, where it was decided to play a game called “The bottle”. The idea of the game is that people sit in a circle and spin a bottle. The person who ends up facing the bottle mouth must drink some alcohol (in this case a shot of strong alcohol). Kadri was not sure about joining the game, as she had not tried strong alcohol before. Her friends were persistent, urging her to join, so she agreed, and soon after she was facing the bottle mouth. She took a sip, but spit it out immediately, as the taste was disgusting. The guy next to her explained that she must drink the whole shot because the rules say so and alcohol is not meant to be wasted. 299 Appendix 4.7 Agreements made between parents Substance related agreements focused on four main topics: 1) zero tolerance on the supply – e.g. “We do not offer alcohol to children”, “Alcoholic beverages are kept out of children’s reach at home”, 2) alcohol-free parties – e.g. “Class parties are alcohol free”, “Children’s birthdays are alcohol free, including for parents), 3) communication with children – e.g. “We talk about alcohol and energy drinks”, 4) informing other parents – e.g. “If we see somebody else’s child consuming alcohol, we inform his/her parents”. The remaining agreements were divided into eight topics: 1) communication with children – e.g. “We talk about boundaries and rules set at home”, 2) communication with parents – e.g. “We increase the communication between parents to discuss important children-related topics”, 3) communication with class teacher – e.g. “Communication between parents and class teacher should be increased – parents notify the teacher if they hear something from the children and the teacher gives more frequent feedback on children’s behaviour at the school”, 4) mobile phone use – e.g. “We explain to children that at the school mobile phone can be used only when necessary. Playing mobile games leads the attention away from the schoolwork”, 5) computer and internet use – e.g. “The computer will be shut down at 9pm during the week”, “We pay more attention to child’s activities in the virtual world”, 6) curfew time – e.g. “We agree that children have to be at home at 9pm during the week and at 10pm during the weekend”, 7) bullying – e.g. “We talk to children about bullying and our zero-tolerance attitude towards it”, 8) children’s free time activities – e.g. “We try to pay more attention to children’s free time activities and support their engagement in positive and healthy activities”.