: This dissertation investigates the factors that influence physicians' prescribing behavior and measure their relative effects on drug choice. This dissertation has three objectives. The primary objective is to test the impact of the following factors on drug selection: incentive-based tiered formularies, patient insurance type, pharmaceutical detailing, and pharmaceutical direct-to-consumer advertising. The second objective is to estimate the effects of physician and patient characteristics on physicians' decisions to prescribe a generic over a branded medication. The third objective is to test whether tiered formulary co-payment differences are associated with increased therapeutic substitution of patented branded drugs to generics in the pharmacy. The goal of this research is to advance our understanding of what factors drive physicians' drug choices and to inform policy on prescription drug costs, pharmaceutical marketing, and formularies. With the introduction of Medicare Part D, the government has become a major payer of pharmaceuticals and directly concerned with rising drugs costs. The last decade has seen a dramatic increase in both tiered formularies and pharmaceutical promotion. Incentive-based tiered formularies are financial mechanisms created to influence the selection of lower cost drugs through reducing moral hazard. Pharmaceutical promotion provides information about drugs that can lead to product differentiation and the selection of higher priced drugs or increase competition among drugs, driving prices lower. These potentially competing effects can have important implications for health care costs and patient's quality of care, since higher drug co-payments have been associated with poor health outcomes, especially in the elderly and poor. Generic prescriptions and substitution are important mechanisms for reducing drug costs and research on physician's prescribing behavior of generics and generic therapeutic substitution by pharmacists is limited. This study will contribute to the literature by analyzing what factors lead physicians to prescribe generics and whether co-payments and other patient characteristics are associated with therapeutic substitution of generics since most generic substitutions occur in pharmacies. These results can provide insight into the prevalence of therapeutic substitution and the effectiveness of mechanisms like tiered formularies to promote generic use. This research is one of the first studies to examine physician prescribing behavior using a large physician panel with all primary factors of drug choice at the physician-patient unit of analysis. The study merges unique datasets to construct a nationally representative panel of 3,771 physicians with 941,713 Statin prescriptions and 212,899 patients to study prescription choice in the Statin therapeutic class in 2004. Conditional and mixed logit regression analyses will be used to estimate the effects of formularies and pharmaceutical promotion on drug choice, generic choice, and generic therapeutic substitution. These models will be estimated for all patients and separately for new and switch prescriptions. ? ? ?
