Abstract

Background

Adherence to insulin affects real-world health outcomes and may itself be affected by the choice of insulin delivery device (pen or vial/syringe). The choice of insulin delivery device may also have direct effects on effectiveness.

Objective

This study aimed to estimate the effects of insulin adherence and delivery device on real-world health outcomes.

Methods

This study included adults with type 2 diabetes mellitus initiating insulin, with continuous health plan insurance for 6 or more months before initiation (baseline) and 1 or more year after. Measured outcomes included glycosylated hemoglobin (Hb A ) reduction, hospitalization rate, total health care costs, and pharmacy costs over 1 year of follow-up. Adherence (defined as having insulin fills sufficient for the entire quarter), pen or vial/syringe use, and disease-related patient characteristics were assessed in each quarter. To account for the time-varying relationship between adherence, patient characteristics, and outcomes, marginal structural generalized linear models were used to estimate the effect of adherence and device use. Mean outcomes were predicted for different combinations of adherence and device choice.

Results

Among the 13,428 patients (mean age 54 years; 46% women; baseline Hb A 9.3%), adherent pen users had greater reductions in Hb A (−0.35%; P = 0.045), lower hospitalization rates (−0.36; P P = 0.1). Pen use and adherent vial use decreased hospitalization rate and increased pharmacy but not total costs.

Conclusions

Adherence and pen use have beneficial effects on patients’ real-world outcomes, with the most favorable effects attributable to adherent pen use.