OBJECTIVE: Dietary supplements are commonly consumed but may not be beneficial for everyone. It is known that supplement users have healthy behaviour characteristics but until now concordance between spouses living in the same household has not been investigated and concordance may be an important behavioural determinant. DESIGN: Prospective cohort study, cross-sectional data analysis. SETTING: European Prospective Investigation into Cancer in Norfolk (EPIC-Norfolk) in the UK, recruitment between 1993 and 1998. SUBJECTS: Married (or living as married) participants sharing a household, who attended a health examination and completed a 7 d diet diary were included in the analysis (n 11 060). The age range was 39-79 years. RESULTS: Nearly 75 % of the households in EPIC-Norfolk were concordant in their supplement use, with 46·7 % not using supplements and 27·0 % using supplements. Concordance increased with age; the percentage of concordant couples varied less by other sociodemographic characteristics. Participants who had a spouse who used a supplement were nearly nine times more likely to use a supplement (unadjusted). Depending on participants' sex and type of supplement used, odds ratios for 'supplement use by spouse' in the prediction of participants' supplement use varied between 6·2 and 11·7 adjusted for participants' age, smoking status, BMI, social class, education level and physical activity. CONCLUSIONS: 'Supplement use by spouse' is an independent and the strongest predictor of participants' supplement use. This phenomenon can be useful in the design of studies and health interventions; or when assessing risk of excessive intake from dietary supplements.