Economic Analysis for Evidence-Based Health Policy: Progress and Pitfalls.

• Author: Baicker, Katherine

It is an enormous honor to deliver this lecture, particularly this year. You have just heard from a number of our profession's luminaries about the monumental effect that Marty Feldstein had on both the world and on their careers. This is personal for me as well, as another of the many students to whom Marty provided such exceptional guidance, support, and investment. Marty has served as a mentor for me throughout my career, from my first week of graduate school through every professional opportunity and decision that I have had, and I am profoundly grateful.

I am privileged to be able to talk with you about one of the many policy arenas that Marty Feldstein helped to shape, from one of his earliest publications, to which this lecture's title alludes. Health system reform is one of the pressing policy issues of the day, and economics has a great deal to contribute to the debate. The economics tool kit is particularly well suited to generating the analytical framework and evidence base needed to inform decisions around the difficult trade-offs inherent to so many aspects of health policy. But, of course, there are a number of challenges to overcome in translating a vast and impressive body of research into policy impact. I will start by trying to define what I mean by "evidence-based health policy" before turning to the current reform debate. I will conclude with thoughts about where, as a profession, we have been more or less successful in achieving that impact and how we might continue to promote the use of the evidence we generate in informing effective policy.

What do I mean by evidence-based health policy? Proponents on both sides of debates often claim to have views grounded in evidence. Having a clear framework for characterizing what we mean by evidence-based health policy is a prerequisite for a rational approach to making policy choices, and may even help focus the debate on the most promising approaches. First, policies need to be well-specified; a slogan is not sufficient. For example, "Medicare for All" isn't really a policy. That slogan masks areas of fundamental disagreement about the role of government and markets, coverage expectations, and cost structure. It may be an effective way to signal a political orientation, but the lack of specificity sidesteps the hard work of assessing the relative effectiveness of the different policies that might fall under that umbrella.

Second, grounding policy decisions in evidence requires a careful distinction between policies and goals. This is important on two fronts: Many different policies may aim to achieve the same goal with very different degrees of effectiveness; and there may be many different goals for a given policy, with different likelihoods of success. For example, consider the financial incentives for physicians to coordinate care. The evidence that this would reduce health care spending--one potential goal--is quite weak, while the evidence that it might improve health outcomes--a different goal--is stronger. Similarly, there may be alternative policies that are more effective in improving health outcomes through coordinating care across sites.

Third, generating the evidence to support policy is an inherently empirical endeavor. Introspection--particularly by economists!--and theory are terrible ways to evaluate policy. For some policies, we have clear conceptual models that suggest the direction of the effect the policy is likely to have, but these models never tell us how big the effect is likely to be. For example, economic theory says that, all else equal, when copayments or deductibles are higher, patients use less care--we're pretty sure that demand slopes down, even for health care--but the theory doesn't tell us by how much. And for other policies, even the direction of the effect is unclear without empirical research, with competing effects potentially going in different directions.

Interpreting the often vast bodies of evidence that speak to a policy question requires judgment and nuance, but it is crucial that interpretation not be flavored by the policy preferences of the researcher or analyst. A given body of evidence can be used to support very different policy positions depending...