A discrete-choice experiment to quantify patient preferences for frequency of glucagon-like peptide-1 receptor agonist injections in the treatment of type 2 diabetes

Abstract

OBJECTIVE: Understanding patients' preferences for attributes of injectable antihyperglycemic regimens may improve patient satisfaction and medication adherence. Our objective was to quantify the preferences of patients with type 2 diabetes mellitus (T2DM) for reducing the frequency of glucagon-like peptide-1 receptor agonist injections from once daily to once weekly. METHODS: A total of 643 respondents with a self-reported physician diagnosis of type 2 diabetes completed a web-based discrete-choice experiment survey. The sample included four prespecified subgroups: currently using exenatide once weekly (n = 150), liraglutide once daily (n = 153), insulin (but not exenatide once weekly or liraglutide) (n = 156), and no injectable treatment (n = 184). Device attributes included type of injection device, needle size and pain, injection frequency, refrigeration, and injection-site reactions. Random-parameters logit was used to estimate the relative impact of device attributes on treatment choice for each subgroup. RESULTS: In all subgroups, changing injection frequency from daily to weekly (independent of the effect of injection frequency on preferences for other attributes) was the most important predictor of treatment choice. Switching from a longer and thicker needle to a shorter and thinner needle and eliminating injection-site reactions were also statistically significant predictors of device choice (P < 0.05). Exenatide once weekly users and those not currently using injections were more likely to choose a treatment with characteristics similar to exenatide once weekly. CONCLUSIONS: The treatment attribute most important to patients choosing among hypothetical injectable treatments for T2DM was injection frequency: patients preferred weekly over daily injections. LIMITATIONS: The primary limitations of this study are that it included only a limited number of attributes that may not reflect the full complexity of patient choices, diagnosis was self-reported, and patients were recruited from an Internet panel and may not be representative of the T2DM patient population.