Dehmer, Steven Patrick2013-05-072013-05-072012-04https://hdl.handle.net/11299/148746University of Minnesota Ph.D. dissertation. April, 2012. Major: Applied Economics. Advisor: Philip Gordon Pardey. 1 computer file (PDF); xi, 179 pages, appendices A-B.Background: Heart disease has been the leading cause of death in the United States since 1921, and although death rates have declined since their peak in the 1960s, it remains a significant burden to the population's health. For most of the second half of the 20th century, changes in lifestyle--such as in diet and exercise--have been the most salient intervention to prevent cardiovascular disease. However, pharmaceutical interventions have gained significant traction in recent decades and may now be supplanting lifestyle change therapies. This study presents a theoretical framework for conceptualizing behavioral feedbacks from new medical technologies, and conducts an empirical case study to investigate whether the introduction of cholesterol-lowering statin drugs can be associated with changes in diet and exercise behavior. Although statins are highly effective in reducing cardiovascular disease risk, lifestyle improvements can garner significant health benefits in addition to lowering cholesterol. Methods: Grossman's model for the demand of health is applied to inform theoretical predictions on health behaviors with the introduction of a new medical technology such as statins. An extension incorporating uncertainty in one's own health status is introduced to motivate the demand for prevention. This extension also allows for the possibility that prevention screening services may send signals of varying quality to an individual regarding one's true health status. If a prescription for a statin sends a stronger informational signal than a high cholesterol diagnosis alone, this could lead to an association of statin use with improved health behaviors. An empirical investigation to test the theoretical predictions is conducted through a causal inference analysis of the introduction of statins on therapeutic lifestyle behaviors (i.e., improved diet and exercise). Longitudinal survey data on a panel of 8,000 individuals from 1995-1998 is linked with statin prescription (claims) data for analysis. Key to causal inference, the timing of this survey coincides with a rapid secular trends in statin adoption. This corresponds with considerable intrapersonal heterogeneity in observed treatment (i.e., statin use) and allows for multiple analytical strategies to account for potential endogeneity bias--namely, panel data and physician/clinic-based instrumental variable methods. To test the practical implications of behavioral changes associated with statin use, multiple simulations are conducted to estimate the economic and health impacts on a U.S.-representative birth cohort, as well as the surveyed cohort from which the empirical results are drawn. An evidence-based Markov microsimulation model is presented for the conduct of this simulation analysis. Results: Preliminary results show an effect that is ostensibly counter-intuitive: a new statin prescription tends to improve health behaviors. Physical activity, in particular, is shown to increase by as much as 30 to 50 percent with statin use. The available data limits inference to short-run impacts (i.e., within four years), but simulations are used to test a range of potential short- and long-term consequences. Short-run impact on health and costs is negligible; however, the predicted long-term impact of improved behaviors on health outcomes can reach 5 to 10 percent reductions in events. For most scenarios, predicted changes in costs are negligible, in part due to their general inverse relationship with longevity. Conclusions: When faced with a newly introduced technology, people are anticipated to respond to any altered incentives. However, if a new technology also serves as a conduit for improved health information, people should be expected to respond to that new information as well. With the introduction of statins, this study shows that a new prescription caused people to choose to exercise more. This result suggests that although both potential modes of action may be in effect, the response to new information may dominate--at least in the short-run. Further empirical and theoretical study will be needed to fully understand the long-term response to statins or similar "new" health technologies.en-USBehaviorsEconomicsHealthR&DStatinsTechnologyThesis or Dissertation