OBJECTIVES: The aim of this study was to determine patient benefit-risk preference of insulin treatment, and to examine how preference is predicted by health literacy, sociodemographic status and patients own health risk perception.  As a secondary objective we investigated whether elicited preferences were sensitive to the presentation of benefits as either a surrogate or clinical health outcome METHODS: This was a questionnaire based study involving Danish type 2 diabetes patients recruited through a diabetes registry. Laboratory- and clinical data on diabetes management status was obtained from the registry. The questionnaire included a section on diabetes management, health literacy, demographics and a discrete choice experiment (DCE). In the DCE, respondents were asked to choose between two unlabelled insulin treatments. These were defined by improvements in glucose control (HbA or long-term sequelae risk reduction), weight control, hypoglycemic events (severe/minor) and treatment-related heart attack risk. A Baysian efficient design (Ngene v 1.1.1) was used to construct the choice tasks. Logistic regression analysis was used to describe the data relationships. Regression coefficients derived from this model were used to estimate maximum acceptable risk (MAR) of treatment in return of benefit. RESULTS: One thousand and thirty-three patients completed the questionnaire. Overall, the avoidance of a minor increase in heart-attack risk (3 additional people of 1000) was driving choice of treatment across all versions of the DCE.  This was followed by avoiding severe hypoglycemic events. We observe some differences in preference estimates across the DCE versions indicating that patients are sensitive to the presentation of benefit as long-term sequelae risk reduction rather than a surrogate measure of health improvement (HbA). Detailed results on subgroup analysis and predictors of preference are presented at the conference. CONCLUSIONS: Risk-aversion of heart attack, although it is a minor additional risk, influence patient choice of treatment. Final conclusion is presented at the conference.