BACKGROUND: Nonmedical formulary switches (NMFS) routinely occur in managed health care plans and involve changing preferred medications for reasons outside of clinical considerations. The cost implications of NMFS are infrequently published and the clinical outcomes rarely assessed.
OBJECTIVE:To assess the real-world clinical and cost implications of an NMFS involving sitagliptin and linagliptin.
METHODS:An NMFS was made to the Geisinger Health Plan (GHP) commercial, health care reform, and Medicaid formularies on February 1, 2018, involving a change in preferred medication from sitagliptin to linagliptin. Claims data from GHP and clinical information from electronic health records of the Geisinger Health System were used to evaluate the cost and clinical impact of this change. Patients aged 18 years or older who were continuously enrolled in a GHP commercial, health care reform, or Medicaid plan throughout the entire study period and had at least 1 fill for sitagliptin during the preswitch phase were included in the study. We investigated the differences in various clinical and economic outcomes from pre-to postswitch among those who switched and remained adherent to the new preferred therapy throughout the 12-month postperiod ("linagliptin switch" group) and patients who did not ("other switch" group). Clinical outcomes included all-cause hospitalization, diabetes-related hospitalization, and glycosylated hemoglobin (HbA1c), while economic measures included changes in per member per month (PMPM) spending. The negative binomial regression model was used to estimate utilization counts. A generalized linear model with a log link and gamma distribution was used to analyze cost data.RESULTS: 1,203 patients met the inclusion criteria. Of these, 501 (41.6%) individuals switched to and remained at least 80% adherent to linagliptin in the postperiod, while 702 (58.4%) did not. No difference between groups was found when comparing the pre-to postswitch change in all-cause hospitalization (incidence rate ratio (IRR) = 1.46, 95% CI = 0.66-3.23, P = 0.3436) or diabetes-related hospitalization (IRR = 1.39, 95% CI = 0.62-3.10, P = 0.4203). Additionally, no difference was found between groups regarding the change in HbA1c 12-month postswitch compared with baseline (difference between groups = −0.10%,