Historical cohort analysis of treatment patterns for patients with type 2 diabetes initiating metformin monotherapy.

objectiveTo describe treatment regimen changes of patients with type 2 diabetes mellitus (T2DM) initiating metformin monotherapy, and assess factors associated with those changes 12 months post-initiation.

methodsRetrospective cohort analysis of medical, pharmacy and laboratory claims of 17,527 Medicare Advantage (MAPD) Humana members aged 18-89, who had ≥1 medical claim with primary diagnosis or ≥2 medical claims with secondary diagnosis of T2DM (ICD-9-CM code 250.x0 or 250.x2) who filled an initial prescription for metformin (GPI code 2725) between 1 January 2008 and 30 September 2011. The main outcome measure was change in metformin monotherapy during the 12 months following initiation. Factors associated with treatment changes during follow-up were examined using Cox proportional hazards regression models.

resultsFifty-nine percent of patients (mean age 69.6 years) remained on metformin monotherapy with no changes. Discontinuation was the most common treatment change (33%), followed by addition (5%), and switching (2%) to other antidiabetics. Of patients who discontinued treatment (median time to discontinuation = 90 days), 61% did not reinitiate any diabetic treatment during the follow-up period. Among patients who added or switched to other antidiabetics, sulfonylureas were the most common addition or replacement agent. Predictors of discontinuation were being female, Black or Hispanic, low-income subsidy eligible, having higher initial out-of-pocket metformin costs, or a diagnosis of depression. Discontinuation was less likely during follow-up if patients had higher pre-index pill burdens or records of a pre-index A1C screening test. A higher risk of discontinuation was observed for patients with low baseline A1C. One study limitation was that exact discontinuation dates could not be determined using claims.

conclusionsThe findings suggest that gender, race, ethnicity, depression, and low income status were contributory factors to metformin discontinuation. More intensive monitoring and treatment adjustments may be warranted for patients newly initiated on metformin. This could ultimately improve morbidity, mortality, and costs associated with poor glycemic control.