A retrospective cohort study of economic outcomes and adherence to monotherapy with metformin, pioglitazone, or a sulfonylurea among patients with type 2 diabetes mellitus in the United States from 2003 to 2005.

OBJECTIVES: The aims of this study were to compare all-cause total health care costs and diabetes mellitus (DM)-specific health care costs between patients who were adherent or nonadherent to monotherapy with metformin, pioglitazone, or a sulfonylurea and to examine whether cost differences varied among patients using these oral antidiabetic drugs.

METHODS: This was a retrospective cohort study using data from the MEDSTAT MarketScan Research Databases. Patients aged 18 to 90 years who were continuously insured between 2003 and 2005 and had > or =2 outpatient claims or > or =1 inpatient claim with a diagnosis of DM (International Classification of Diseases, Ninth Revision, Clinical Modification code 250.xx) in 2003 were eligible for the study. To be part of the final sample, patients had to fill > or =2 prescriptions for metformin, pioglitazone, or a sulfonylurea during 2003, including > or =1 prescription during the last 3 months of the year. Patients were not eligible if they were taking polytherapy or a combination drug. All eligible patients were followed in 2004 and 2005. Adherence was calculated for each year using a medication possession ratio, and was dichotomized at > or =80% as either adherent or nonadherent. Annual all-cause health care costs and diabetes-specific costs were estimated using generalized linear models, adjusting for demographic characteristics, insurance, and comorbid conditions.

RESULTS: A total of 108,592 patients who met the inclusion criteria were identified. Their mean age was 63 years; 49.8% (54,037/108,592) were women. More pioglitazone users resided in the north-central or south regions (81.3% [9364/11,520]) compared with metformin (62.4% [32,550/52,156]) or sulfonylurea (62.6% [28,105/44,916]) users (P < 0.001). Mean comorbidity scores were higher in the sulfonylurea (1.78) and pioglitazone (1.69) group than in the metformin group (1.45) (P < 0.001). Mean adherence ranged from 61.3% to 73.8% during the 2 years of follow-up. After adjustment, all-cause health care costs were $12,412 annually among adherent patients and $13,258 among nonadherent patients (difference, $846 [95% CI, $747 to $945]). Diabetes-related health care costs were $2230 annually among adherent patients and $2284 among nonadherent patients (difference, $55 [95% CI, $33 to $77]). In specific monotherapy groups, adjusted annual all-cause health care costs were $336 higher (95% CI, $216 to $456) for nonadherent metformin users, $1140 higher (95% CI, $793 to $1486) for nonadherent pioglitazone users, and $1509 higher (95% CI, $1339 to $1679) for nonadherent sulfonylurea users compared with adherent users. Compared with metformin users, sulfonylurea and pioglitazone users had greater adherence-related differences in all-cause health care costs (P < 0.05). There was no significant difference in diabetes-specific total costs between nonadherent and adherent patients taking metformin (difference, $6; 95% CI, -$31 to $20). Patients who were nonadherent to sulfonylureas had $271 higher (95% CI, $235 to $307) diabetes-specific costs per year than patients who were adherent to sulfonylureas. Patients who were nonadherent to pioglitazone had $433 lower (95% CI, -$516 to -$350) diabetes-specific costs per year than patients who were adherent to pioglitazone.

CONCLUSIONS: Adherence with metformin, pioglitazone, or a sulfonylurea was associated with overall cost reductions in the patients studied, but these cost reductions varied by monotherapy. Adherence to sulfonylureas or pioglitazone was associated with greater total cost reductions than was adherence to metformin. Health systems that commit resources to improving interventions may be able to achieve a return on investment if adherence to oral antidiabetic agents can be improved.