Despite stable growth over the past few years, the pharmaceutical industry has recently been experiencing unique challenges due to shortened patent protection periods and the emergence of low-priced generic drugs. As a result, it is necessary for pharmaceutical companies more than ever to rely on a handful of successful new drugs not only to channel their growth but also to survive. However, new drug development in the pharmaceutical industry is a very risky activity because of the large amount of time and the costs involved and because chances of success are slim. Therefore, there is an increasing need for pharmaceutical managers to assess how successful a new drug in the pipeline will be and what factors drive the market penetration rate of a new drug.
This dissertation develops a pre-launch market evaluation measurement methodology and model framework for patient simulation studies. A patient simulation study is designed to generate market share forecasts and to support marketing decision making for a new pharmaceutical product after the product has been developed but before it has been launched. The heart of a patient simulation study is mimicking the physician prescription choice process by explicitly incorporating patient information, physician information, and competitive marketing activities to enhance the validity of physicians' choice response. Using a prescription choice model in a hierarchical Bayesian framework, a patient simulation study can provide a good basis to understand physician behavior, to formulate better pharmaceutical marketing strategies, and to forecast market shares. Specifically, the patient simulation model allows researchers and managers to develop physician-specific marketing tactics and apply them in the field.
I describe and analyze two empirical applications of patient simulation to demonstrate its managerial value: One is a study conducted by an existing brand anticipating the launch of a new brand, and the other is a study conducted by a new brand. Specifically, the first application is built in the situation where a physician prescribes only one drug to treat a patient; in contrast, the second study analyzes the case where multiple drugs are prescribed simultaneously. The results suggest that patient simulation produces reasonably accurate market share forecasts prior to launch. More importantly, the model provides insights on what drive physicians' prescription brand choice and offers the opportunity to simulate the market under different marketing plans.
Based on the initial validation results, patient simulation can be used to provide researchers and managers with a deeper understanding of the factors that drive physician behavior, and it can be instrumental in developing other decision-support systems that can improve the odds of commercial success of new pharmaceutical products.