Utilities and disutilities for attributes of injectable treatments for type 2 diabetes

Abstract

Although cost-utility models are frequently used to estimate treatment outcomes for type 2 diabetes, utilities are not available for key attributes of injectable treatments. The purpose of this study was to identify the utility or disutility of three injection-related attributes (dose frequency, dose flexibility, injection site reaction) that may influence patient preference. Patients with type 2 diabetes in Scotland completed standard gamble (SG) interviews to assess the utility of hypothetical health states and their own current health state. The EQ-5D, PGWB, IWQOL-Lite, and QIDS were also administered. Construct validity and differences among health states were examined. A total of 151 patients completed interviews. Of the three injection-related attributes, dose frequency was the only attribute with a statistically significant impact on utility (in a multiple regression model, p = 0.01). Weekly injections were associated with an average added utility of 0.023 in comparison to everyday injections. Flexible dosing and injection site reactions resulted in somewhat smaller utility shifts that were in the expected directions (+0.006 and -0.011, respectively). SG utility of current health (mean = 0.897) demonstrated construct validity through statistically significant correlations with patient-reported outcome measures. The three injection attributes were associated with small utility shifts in the expected directions. Dose frequency appears to be the most important of the three attributes from the patients' perspective. The vignette-based SG approach was feasible and useful for assessing added utility or disutility of injection-related attributes associated with treatments for type 2 diabetes.