Provider behavior and estimation methods in health economics.
详细信息
文摘
The first chapter, Provider Behavior under a Regulatory Drug Ceiling and Its Implications for Taiwan, explores how provider behavior can be influenced by how health care providers are paid, with payment policies affecting both treatment intensity and the timing of visits during each episode of illness. I analyze how regulated drug prices, such as those adopted by Taiwan, affect physician behavior when they have multiple tasks and compare the equilibrium implied by the model to the socially optimal allocation. I examine the theoretical implications of this model and use it to provide insight into current treatment and prescription patterns in Taiwan. In Taiwan, patients make twice as many doctors visits as in other developed countries and drugs are still dispensed primarily in doctor's offices. I also provide policy suggestions that would improve social welfare.;The second chapter, Estimating the Demand for Medical Care Using a Multiple Sample Selection Model, develops a new estimation method which is able to extract useful information from the individual's decision-making behavior that is ignored using aggregated data. Correlations between the different types of services are allowed, which provides a more flexible and realistic framework. I apply this method to forecast the medical cost for 2002 MarketScan Database and find improved performance, especially for the youngest and oldest age groups.;The third chapter, Evaluating Underage Drunk Driving Laws with a Smooth Transition Model, uses a smooth transition model to jointly estimate policy effects and policy lags. When evaluating any program, there often exists an endogenous policy lag due to information acquisition or cost of adjustment. Ignoring this effect may result in a downward bias. The policy change is modeled as a smooth process for capturing the gradual implementation with the smoothness determined endogenously from the data. I apply the method to re-evaluate the impact of underage drunk driving laws from 1984 to 2002 and find a significant transitional period. I show that states adopting the policy later have shorter transition periods.