OBJECTIVES: To examine patients’ stated preferences for persisting with medications using a discrete choice experiment (DCE) and to explore the relationship with clinical, demographic and psycho-social variables. METHODS: A 4-attribute DCE (mild side-effects, potentially life-threatening but rare side-effects, dose frequency, treatment benefits) with 3-levels identified from literature and expert opinion was developed using a fractional factorial design. Scenarios were folded into nine forced binary choices: Which medicine would you be most likely to continue taking? The survey was translated, piloted and approved for eleven European countries. Target sample was 100≤n≤323 patients prescribed anti-hypertensives per country, recruited by posters in community pharmacies or general practices. Results were analysed in STATA using a random effects logit model. RESULTS: 2856 patients from Austria (n=323), Belgium (n=180), England (n=323), Germany (n=265), Greece (n=289), Hungary (n=323), Netherlands (n=237), Poland (n=323) and Wales (n=323) completed the online questionnaire.   All four attributes influenced persistence with treatment (p<0.01).   Patients were willing to forego chance of improvements in treatment benefits (%) in order to improve other attributes:  -36.10% (95% CI: -41.24 to -32.94) for a very rare risk of life-threatening side-effects; -18.66% (95% CI: -21.51 to -16.67) for once daily dose frequency; -0.74% (95% CI: -0.85 to -0.67) to reduce the risk of mild ADR by 1%. Likelihood ratio tests showed that models controlling for clinical, demographic and psycho-social variables were significantly different from the base-case.  There was limited evidence that self-reported adherence influenced stated preferences to persist.  CONCLUSIONS: Patients were willing to trade potential benefits, harms, and convenience in responding that they would persist with treatment. Clinical, demographic and psycho-social factors influence the extent of the trade-offs between these attributes.  Persistence may therefore be enhanced directly, through selection of medicines meeting preferred levels of attributes, or indirectly through targeting modifiable psycho-social factors that affect trade-off choices.