Objective: While many patients with type 2 diabetes require insulin to achieve glycemic goals, little is known about patients' persistence with insulin or other injectable antidiabetic therapies. The objective of this study was to evaluate persistence with injectable antidiabetic agents in patients with type 2 diabetes who were naïve to these treatments.
Methods: The study cohort was obtained using administrative and pharmacy claims data from a commercial managed care organization of approximately 1.2 million members with pharmacy benefits. The inclusion criteria were members with type 2 diabetes who had at least one pharmacy claim for insulin glargine, insulin detemir, exenatide, or isophane insulin human (NPH insulin) from January 1, 2006 through June 30, 2006. The first claim for any of these injectable therapies was considered the index prescription. Members were excluded if they filled a prescription (a) for any injectable antidiabetic medication from July 2005 through December 2005; or (b) for short/rapid-acting or mixed insulins during 2006, either prior to the index date or within 30 days following the index date. The primary outcome was persistence with the index drug, defined as number of months between the initiation of therapy (i.e., index date) and either the end of therapy (date of last fill plus days supply) or study period of 12 months. The secondary outcome was the percentage of patients with claims for new antidiabetic agents added after index date. Multivariate regression with life data (survival analysis) was performed with number of months of persistence as the dependent variable.
Results: The cohort consisted of 1769 members with a mean (SD) age of 53.2 (12.5) years and 47.4% were men. Mean (SD) months of persistence for members on insulin glargine, insulin detemir and exenatide were similar at 7.8 (4.1), 7.8 (4.4), and 7.6 (4.4), respectively. Members taking NPH insulin had statistically lower persistence at 5.6 (4.5) months (P < 0.001). Overall persistence was 28.7% for injectable antidiabetics at 1 year among treatment-naïve patients. In a multivariate regression model, patients who were younger (P = 0.025) and who initially received NPH insulin (P < 0.001) were less likely to persist. There was no association between persistence and sex, initial copayment, or number of oral antidiabetic medications at time of index prescription. Members in the exenatide and NPH insulin groups were less likely to receive new antidiabetic agents compared with the insulin glargine group (P < 0.001). Limitations include the use of pharmacy claims to proxy the type of diabetes and patients with gestational diabetes may have been included. Missing or inaccurate claims data, the use of samples and hospitalizations may have occurred and would result in an underestimation of persistence.
Conclusions: Persistence was low for injectable antidiabetics at 1 year among treatment-naïve patients. Patients who received insulin glargine, insulin detemir, or exenatide were more likely to persist than patients receiving NPH insulin. Older patients were more likely to persist, but sex, copayment and number of oral antidiabetic medications at initiation of the injectable antidiabetic were not associated with persistence.