OBJECTIVES: Preference-evidence for cardiovascular risk and severe versus non-severe hypoglycemic aspects of diabetes treatments is scarce in the literature. This study aims to 1) elicit patient preference for the benefit and risk aspects of diabetes treatment, and 2) identify segments with differences in preference for treatment based on socioeconomic position and health indicators.

METHODS: A self-administered survey including a discrete choice experiment was distributed to insulin users registered in a large-scale health database in the county of Funen, Denmark. Respondents made multiple choices between two hypothetical treatments characterized by five health outcomes: glycemic improvement (A_1c), weight loss, treatment related cardiovascular risk, and frequency of non-severe to severe hypoglycemic events. Development of the discrete choice experiment was based on extensive preparatory work including interviews with patients and health care professionals. Choice- and health registry data, including well-defined clinical measures of metabolic control, were combined. The data were analyzed using a conditional logit model.

RESULTS: Respondents preferred treatments with no severe, or frequent non-severe (eight per month versus none), hypoglycemic events, or no cardiovascular risk over glycemic improvement and weight loss. Respondents (n=787) were elderly (67.1, SD±10.2) and constituted a high risk group for cardiovascular disease. Respondents at very high risk for cardiovascular disease accepted significantly fewer (p<0.001) severe hypoglycemic events in order to avoid further risk than their healthier peers. Respondents valued improvements from moderate-to-good glycemic control over moderate-to-tight glycemic control, preferring an A_1c level of 7.5% (p<0.001). Some indication on preference heterogeneity for tight control (A_1c level of 6.0%) existed for segments based on glycemic control status, hypoglycemic experience and sociodemographic position.

CONCLUSIONS: Avoiding an incremental cardiovascular (heart attack) risk, or severe- or frequent non-severe hypoglycemic events was preferred over efficacy (A_1c improvement). The patient segment with the lowest risk of cardiovascular disease most strongly preferred treatments with no incremental cardiovascular event risk.