Brand name or generic? What are the health professionals prescribed for treating diabetes? A longitudinal analysis of the National Health Insurance reimbursement database

Objective: This study aimed to explore whether physicians prescribe more brand-name oral hypoglycemic agents (OHA) for diabetic patients with medical training background (MP) than for general patients (GP). Research design and methods: A longitudinal analysis of 1000000 National Health Insurance cohorts of 1998-2008 was conducted. Univariate and multivariate models were performed to assess the associations of the outcome (the ratio of brand-name/generic odds in the MP group to that in the GP group) and the covariates, including patient medical training background, characteristics of patient, prescriber, and medical settings, and market competition. A generalized estimating equation method was used to control the dependency of longitudinal data. Results: A total of 46850 diabetic patients were prescribed with 2703149 OHA prescriptions during the study period. Compared with GP, MP had 1.37 times greater odds of being prescribed with brand-name instead of generic OHA, among whom pharmacists and physicians had the highest odds ratios of 2.78 (95%CI, 1.05-7.36) and 1.68 (95%CI, 0.99-2.85), respectively. Patients' diabetes severity, prescribers' level of experience, medical settings that were publicly owned, had a higher accreditation level, and were located in a higher urbanized area, lower market competition, and earlier dates of prescription were positively associated with brand-name prescription. Among all medical sub-specialties, cardiologists were more likely to prescribe brand-name OHA. Conclusions: This study is the first to demonstrate how a patients' medical training background, in addition to the characteristics of patients, prescribers, and medical settings, and market competition might influence physicians' prescribing choice of brand-name or generic OHA.