Parents' views about healthcare professionals having real‐time remote access to their young child's diabetes data: Qualitative study

Abstract Objectives We explored parents' views about healthcare professionals having remote access to their young child's insulin and glucose data during a clinical trial to inform use of data sharing in routine pediatric diabetes care. Research Design and Methods Interviews with 33 parents of 30 children (aged 1–7 years) with type 1 diabetes participating in a randomized trial (KidsAP02) comparing hybrid closed‐loop system use with sensor‐augmented pump therapy. Data were analyzed using a qualitative descriptive approach. Results Parents reported multiple benefits to healthcare professionals being able to remotely access their child's glucose and insulin data during the trial, despite some initial concerns regarding the insights offered into everyday family life. Key benefits included: less work uploading/sharing data; improved consultations; and, better clinical input and support from healthcare professionals between consultations. Parents noted how healthcare professionals' real‐time data access facilitated remote delivery of consultations during the COVID‐19 pandemic, and how these were more suitable for young children than face‐to‐face appointments. Parents endorsed use of real‐time data sharing in routine clinical care, subject to caveats regarding data access, security, and privacy. They also proposed that, if data sharing were used, consultations for closed‐loop system users in routine clinical care could be replaced with needs‐driven, ad‐hoc contact. Conclusions Real‐time data sharing can offer clinical, logistical, and quality‐of‐life benefits and enhance opportunities for remote consultations, which may be more appropriate for young children. Wider rollout would require consideration of ethical and cybersecurity issues and, given the heightened intrusion on families' privacy, a non‐judgmental, collaborative approach by healthcare professionals.


| INTRODUCTION
Type 1 diabetes management is particularly challenging in very young children due to rapid physical and social development, hypoglycemia unawareness, high insulin sensitivity, and unpredictable food intake and activity patterns. 1,2 Diabetes technologies, such as insulin pumps and continuous glucose monitors (CGMs), can help improve glycemic control and reduce severe hypoglycaemia 3 ; hence, pediatric diabetes guidelines recommend their use. 4 Use of diabetes technologies may also alleviate pressures on parents, with many reporting improved sleep, less worry about glucose excursions and hypoglycemia, more flexibility and freedom, and improved familial quality of life. 3 As a result, technology use in this population is growing rapidly. 5,6 The latest CGM technology allows other individuals to access and remotely review users' glucose data, including healthcare professionals (HCPs). Little is known about users' experiences of, and views about, HCPs having remote access to their (or, in the case of parents, their child's) data. Studies suggest that data access may enhance communication by fostering a sense of intimacy and "seamless connectedness" between user and practitioner 7 and, in turn, improve glycemic control. 8 However, while access to glucose/insulin dose data can help HCPs provide better advice and support, the level of insight this allows into private routines and behaviors may make users feel scrutinized or judged. 9,10 Survey studies exploring remote data review and support by HCPs in pediatric populations have reported no notable impact on parent/child psychosocial outcomes, but suggest that it may substantially reduce HCP contact times during and between clinic visits. 11,12 Given the potential for the delivery of routine pediatric diabetes care to be enhanced by HCPs having remote data access, we sought to clarify the above equivocal findings. As part of a broader investigation of parents' experiences of using closed-loop technology (as compared to sensor-augmented pump therapy [SAP]) during a clinical trial (the KidsAP02 study), we explored their views about HCPs having remote access to their child's data during the trial. In doing so, we aimed to help inform use of data sharing with HCPs in routine pediatric diabetes care.

| METHODS
We interviewed parents of young children aged 1-7 years with type 1 diabetes taking part in a 24-month open-label, multi-center, multinational crossover trial (KidsAP02). Children were randomized to 16-week use of a hybrid closed-loop system (CLS) or 16-week use of SAP therapy, before crossing over to the other regimen. 13 In both arms, participants used a CGM sensor (Dexcom G6), which transmit- In both arms, data from the app was automatically streamed to a cloud-based platform (Glooko/Diasend; Göteborg, Sweden) every 5-10 min. These data were accessible to parents and study teams via both the Diasend app and the Diasend website, enabling near realtime sharing of both insulin and glucose data with, and remote review by, HCPs working on the trial. As well as aiding trial evaluation, HCPs could use these data to inform clinical advice given to parents during pre-arranged and unscheduled study contacts; for example, to optimize basal rates (during periods of SAP use) and/or mealtime ratios following transition onto the study equipment and during the trial.
During each study period, HCPs telephoned/emailed parents twice during the first week and again after 1 week to: address any problems or concerns regarding the study devices, review device use, troubleshoot problems and provide additional training as necessary; then monthly, to troubleshoot problems and collect trial-relevant information, for example, adverse events or device deficiencies. Parents could also contact HCPs via a 24-h telephone helpline. Additionally, parents and participants attended two study appointments per study period: one at the start of each period for training purposes, and one at the end for collection of blood samples for HbA1c measurement. When the onset of the COVID-19 pandemic necessitated a transition from face-to-face to remote modes of care delivery, 14 increased emphasis was placed on using shared data to provide participants with support.
See Table 1 for more information about the trial, devices used and remote data access arrangements.
The interview study was conducted by an independent team of qualitative researchers at the University of Edinburgh. Initially, the project researcher (B.K.; a female, PhD-qualified, non-clinical researcher) visited participating sites to gather contextual information and gain broad understanding and awareness of potential country/ site-specific differences. These insights helped inform the development of the interview topic guides and supported subsequent data analysis.
We used semi-structured interviews informed by topic guides to help ensure the discussion remained relevant to the study aims, while offering participants flexibility to raise issues they considered important. The topic guides used open-ended questions and probes and were informed by the site visits, literature reviews and input from clinical co-investigators (via a dedicated qualitative working group comprising representation from all sites), and revised in light of emerging findings (see Table 2 for details of the main topic areas relevant to the reporting in this article). Data collection and analysis took place concurrently, allowing findings from earlier interviews to inform issues explored in subsequent ones.

| Sampling and recruitment
Participants were recruited following randomization to the trial at seven clinical sites in four countries: Austria, Germany, Luxembourg, and the UK. Parents were consented into the interview study when they were consented into the trial. We employed purposive sampling to promote diversity with respect to participating countries and sites as well as the child's age and gender. Recruitment continued until data saturation was reached, that is, when new data ceased to generate new findings.

| Data collection and analysis
Parents were interviewed at the end of the first study period (i.e., after 16 weeks of using CLS or SAP therapy), and at the end of the second study period (i.e., after 16 weeks of using the other regimen). We aimed to interview one parent per child; however, in the The KidsAP02 trial The KidsAP02 trial was conducted at seven clinical centers in four countries: Austria, Germany, Luxembourg, and the UK. Seventy-four children (aged 1-7 years, type 1 diabetes duration ≥6 months, insulin pump use ≥3 months) were randomized to use either a hybrid closed-loop system (intervention) or sensor-augmented pump therapy (control) for 16 weeks. Following a "wash-out" period (1-4 weeks), participants changed to the other regimen for a further 16 weeks. In both study arms, participants used the same component devices (pump, CGM sensor and smartphone). HCPs contacted parents 24 h after starting each study arm to address any immediate problems/concerns regarding the study devices, and again after a further 24-48 h and 1 week to review device use, troubleshoot problems and provide additional training as necessary. For the remainder of each study period, HCPs then telephoned/emailed parents monthly to troubleshoot problems and collect trial-relevant information, such as adverse events or device deficiencies. Parents were also given access to a 24-h telephone helpline for any problems related to the devices or general diabetes management. Additionally, parents and participants attended two study appointments per study period: one at the start of each period for training purposes, and one at the end for collection of blood samples for HbA1c measurement. The CamAPS FX hybrid closed-loop system (CamDiab, Cambridge, UK) The CamAPS FX is a "hybrid" closed-loop system, which combines automated, algorithm-informed delivery of basal (background) insulin with userinitiated mealtime boluses. CamAPS FX comprises the following devices/components: • DANA RS insulin pump (Sooil, Seoul, South Korea), with CGM receiver; • Dexcom G6 factory-calibrated real-time CGM sensor (Dexcom, San Diego, CA, USA), with CGM transmitter; • An unlocked Android smartphone (Galaxy S8, Samsung, South Korea) running Android 8 OS or above, hosting the CamAPS FX app incorporating the Cambridge model predictive control algorithm (CamDiab, Cambridge, UK). The smartphone/app communicates wirelessly with both sensor and insulin pump, subject to being kept within 5-10 m of these devices.

Data-sharing
The CamAPS FX app facilitates automatic data upload to a cloud-based platform (Glooko/Diasend; Göteborg, Sweden), thus enabling data sharing and review by other individuals, including healthcare professionals, parents or other caregivers. Using participant login details, HCPs could view a child's real-time data via the Diasend mobile app (requiring a smartphone) or the Diasend web application (requiring a computer). HCPs had remote access to the following shared data: • "Real-time" and retrospective graphs displaying glucose levels, rate of insulin delivery, meal-time boluses and carbohydrate intake, high/low glucose range, Boost and Ease-off status (for participants using the closed-loop system), and a function which indicates whether the closed-loop was operational (Automode on) or interrupted/switched off (Automode off). [Note: "Automode" remained switched off in the sensor-augmented pump therapy arm of the trial; hence, during this phase, the hybrid closed-loop system was not activated and rates/times of basal insulin delivery were instead pre-set.] • Summary statistics for daily, weekly, monthly or quarterly periods, including: average glucose, estimated HbA1c, time in/below/above target, number and average duration of hypos, total daily dose/bolus/basal insulin, and percentage of time in Automode (for those using the closed-loop system). To reduce bias and enhance rigor, two team members (B.K. and J.L., a highly-experienced non-clinical qualitative researcher) independently analyzed the data using a qualitative descriptive approach, which seeks to provide rich, low-inference descriptions of an event or experience. 15 familiar with the interview content, were available to mediate; however, discrepancies in data interpretation were minimal and resolved through discussion without need for third-party arbitration. We used a qualitative analysis software package, NVivo11 (QSR International, Doncaster, Australia), to facilitate data coding and retrieval. Joint interviews were analyzed using the same approach as that used for single-person interviews.
The trial, including this qualitative sub-study, received approval from all relevant national regulatory authorities and ethics committees in the participating sites/countries. Our reporting is guided by the Standards for Reporting Qualitative Research (SRQR). 17

| RESULTS
We interviewed 33 parents of 30 children at the end of their first study period and 29 parents of 26 children at the end of the second. Four parents were lost to follow-up: one could not be recontacted and three second interviews were not pursued due to staffing challenges caused by the COVID-19 pandemic and consensus that data saturation had already been reached. Sample characteristics and details of the devices used pre-trial are provided in Table 3. Others observed how HCPs no longer needed to spend time during consultations downloading directly from the child's pump or glucose meter. HCPs' ability to "simply push a button and…look how it has been" (017_CLS), as some parents noted, freed up valuable consultation time to review their child's data and determine necessary adjustments: "Before…we had an appointment every three months and we just came, they downloaded the data from the pump, they had 10 minutes on checking the 3-month record, let's say, and how can you adjust any improvement of set-up in 10 minutes?" (015_SAP) Parents also described how HCPs' ability to remotely access their child's data had simplified and facilitated the process of help-seeking outside of pre-arranged appointments, which, before the trial, had, for some, occurred "once to sometimes twice a week, because her levels were so hard to manage" (020_CLS). They observed finding it helpful that HCPs were able to instantly access the child's data on their own computers and "look at the same thing I'm looking at" (002_SAP), as this meant they no longer needed to remember, or gather and trans- Parents also suggested that, when HCPs had offered input and support after having reviewed the child's data, this had been well received because of the sensitive and non-judgmental approach they had employed, which was "never moralising, but more questioning why it was like that and should it be different in future" (010_CLS).
Parents also described welcoming HCPs using a collaborative rather than prescriptive approach in these situations: "They called a few times and were just kind of asking, like: 'Okay, how's it going? We were looking at her blood sugars and we saw that you've been having this problem.
How do you feel like it's been going that way?' Like: 'Should we be looking at that and changing it?' and that. Furthermore, the majority of parents noted (un)anticipated benefits to this change. As well as not having to travel to the hospital (which could be time consuming, require some to take time off work and/or take the child out of school), parents noted quality-of-life benefits to their child by virtue of no longer having to attend disruptive, medicalized appointments where they were not active participants: "With these phone calls…it's better for the patient, because she is less at hospital, so it's not that much in her face. And basically, at the age of five she is not taking part in it anyway, she sits in a corner. You know, she is measured and blood finger pricks and these things, but then she is basically sitting an hour and painting." (017_CLS) Several parents, like this father, also noted having gained more from remote consultations as it had been easier to keep the child entertained in their own home, and this had allowed them to concentrate and have more constructive conversations with HCPs: "The questions that they ask or the questions that you ask…in the phone call and the Zoom call today, they have been answered. Whereas when you've got [child] there, you're kind of keeping half an eye on him and you're not able to give necessarily a full answer." (022_CLS, father) 3.3 | Potential adaptations to routine diabetes care for young children The majority of parents endorsed continued use of data sharing in routine diabetes care, in part because it would allow more consultations to be undertaken remotely, which, following their pandemic experiences, parents now considered a more age-appropriate approach. Parents also noted how, while using the closed-loop system during the trial, they had needed much less clinical input to optimize their child's glycemic control, due to the system automatically regulating basal rates: "It used to be-I would be ringing constantly about, enabled diabetes consultations to be delivered remotely; these being seen as more suitable for young children than face-to-face consultations. Some parents also proposed that remote data access and review by HCPs could allow routine appointments for closed-loop users to be replaced with needs-driven, ad-hoc consultations; this being due, in part, to their needing less support when using a closed-loop system.
In light of their positive experiences, parents recommended use of real-time data sharing in routine clinical care, albeit subject to caveats related to data access, security and privacy.
Some parents reported initially feeling uneasy about HCPs having remote access to their child's data, due to the insights these data could offer into family routines and everyday life. Indeed, in this trial, HCPs were able access insulin as well as glucose data, which potentially provided richer insight into families' lives than CGM data alone.
Given this invasion of privacy, some have argued that closed-loop data sharing should be based on consent and agreement about the amount and type of data to be shared with HCPs. 9 Additionally, ground rules could be agreed, and regularly reviewed, regarding if, when and how frequently HCPs will check these data.
In this study, data review was mostly undertaken by HCPs with whom parents had long-standing relationships and, consistent with others' findings, 18  indeed, others have noted that the optimal frequency for providing remote input requires closer investigation. 12 Finally, some commentators have warned that, despite its advantages, remote data review and support may not meet the needs of all populations. 23 Our findings suggest that, in pediatric populations, remote data access may be particularly beneficial to parents of newly-diagnosed and/or very young children, those feeling more anxious or reluctant to seek advice, and when the child first transitions onto new technologies. However, the benefits and widespread acceptability of this approach reported by all parents in this study indicate value in exploring how remote data review/support might be offered to all those caring for a child with type 1 diabetes who wish to use such provisioning.
Parents highlighted benefits to attending fewer face-to-face consultations, including fewer absences from school or work, which, as others have noted, can be disruptive to both parent and child and reduce productivity and learning. 24 Given the short and long-term health-economic consequences of lost productivity and education, offering real-time data sharing (which facilitates the remote delivery of consultations) in routine care may therefore also deliver wider societal benefits. Mirroring findings by Lawton et al., 25 parents in this study also noted how remote consultations helped lessen the distraction of keeping a young child entertained, thereby enabling more effective conversations with HCPs.
Wider rollout of remote diabetes support based on real-time data sharing would require careful consideration and mitigation of several critical issues. For example, we recommend that HCPs offering input and support by remotely accessing real-time glucose and insulin data employ a sensitive, non-judgmental and collaborative conversational approach, as this was well received by parents in this study.
Furthermore, to promote optimal uptake, technology developers and policymakers should consider potential ethical issues (e.g., balancing increased device complexity with user safety, or the availability of insurance/reimbursement options), 9 and safeguard individuals' cybersecurity with regard to data theft and external device manipulation. 26,27 The parents in our study were mostly in professional occupations and, hence, were not in a state of digital poverty. However, it is important to recognize that individual and socioeconomic disparities exist regarding confidence with, and, importantly, access to, diabetes technologies, reliable internet connectivity and appropriate private space. Care providers therefore need to ensure that individuals most likely to benefit from remote care provision are supported to access this. 28 A key study strength is the inclusion of parents of young children privacy; this will also help identify those most likely to benefit from such provisioning. 29 Finally, research should explore HCPs' experiences of providing remote support and, importantly, their views about how such provisioning might be integrated into routine clinical care. We are immensely grateful to the parents who gave up their time to share their experiences with us. We would also like to thank HCPs at the trial sites for their support with recruitment. For the purpose of Open Access, the author has applied a CC BY public copyright licence to any Author Accepted Manuscript version arising from this submission.