Patient preferences for treatment in type 2 diabetes: the Italian discrete-choice experiment analysis.

Abstract

Several drug classes are now available to achieve a satisfactory metabolic control in patients with type 2 diabetes (T2DM), but patients' preferences may differ. In a discrete-choice experiment, we tested T2DM patients' preferences for recent antidiabetic drugs, in the event that their treatment might require intensification. The following attributes were considered: (a) route of administration; (b) type of delivery; (c) timing; (d) risk of adverse events; (e) effects on body weight. Twenty-two possible scenarios were built, transferred into 192 paired choices and proposed to 491 cases naïve to injectable treatments and 171 treated by GLP-1 receptor agonists (GLP-1RAs). Analyses were performed by descriptive statistics and random effects logit regression model. Preferences according to dosing frequency, risk of nausea and urinary tract infections (UTls) were similar across groups, age, sex and BMI. Administration route and delivery type accounted for 1/3 of relative importance; the risk of UTIs, nausea and dosing frequency for ≈ 20% each, and weight loss for only 6%. Two significant interactions emerged (p < 0.01): type of delivery × group, and weight change × BMI class. Irrespective of previous treatment, the three preferred choices were injectable, coupled with weekly dosing and a ready-to-use device (first two choices). In a regression model, being naïve or non-naïve changed the ranking of preferences (p < 0.001), and the order was systematically shifted towards injectable medications in non-naïve subjects. Easy-to-deliver, injectable treatment is preferred in T2DM, independently of treatment history, and previous experience with GLP-1RAs strengthens patients' willingness to accept injectable drugs.