5

Zero-Credit Politics: The Government’s Sluggish Effort to Promote Evidence-Based Medicine, 1970s–2008

Imagine a political world in which the value of evidence-based medicine was recognized by policy elites across the political spectrum and had a “taken-for-granted” status, just as it is expected that any nuclear power plant that society chooses to build will reflect accepted engineering safety standards. This world is not so distant from our own. Indeed, despite the increasing politicization of governance, there have been moments in the recent past when the U.S. political system seemed on the verge of developing a bipartisan, technocratic consensus around EBM. A month before the 2008 election, for example, Newt Gingrich (R-GA), John Kerry (D-MA), and Billy Beane (the statistics-loving baseball executive portrayed in the film Moneyball) published an op-ed in the New York Times entitled “How to Take American Health Care from Worst to First.” They pointed out that the United States “spends more than twice as much per capita on health care compared to almost every other country in the world—and with worse health quality than most industrialized nations.” A major reason why U.S. health care is so wasteful and low performing, they argued, is that medical practices are not based on clinical evidence of what works best and what does not. “Remarkably, a doctor today can get more data on the starting third baseman on his fantasy baseball team than on the effectiveness of life-and-death medical procedures.… To deliver better health care, we should learn from the successful teams that have adopted baseball’s new evidence-based methods,” they wrote.¹ To be sure, there are dimensions of EBM that could stimulate political polarization, such as legitimate concerns about government interference with doctors’ clinical judgment and autonomy. As Sheila Jasanoff, a leading scholar of the role of science in public policy, argues, the “boundaries between science and politics … are constructed and maintained through politically inflected ‘boundary work’. Put differently, the balance between reliance on science and reliance on politics is itself a product of social accommodation and power plays.”²

The Gingrich-Kerry-Beane agreement on the need for U.S. health care to become more scientific thus does not demonstrate that the pursuit of better evidence in medicine inevitably “ought” to be a nonpartisan, “good government” issue, but rather points to this potential. The key question we address in this and the following chapter is why a potentially technocratic issue that could be handled pragmatically by experts instead becomes politicized.

In brief, our answer is that the politics surrounding efforts to address the medical evidence problem have been “unhealthy.” This chapter shows that while the pragmatic case for bringing standardized evidence more systematically to bear on decisions in health care is compelling, federal politicians took only modest actions to support steps to identify and eliminate wasteful, unnecessary services between the 1970s and late 2000s. Moreover, politicians (responding to their electoral incentives) were quick to abandon even these incremental reforms when they sparked opposition from providers and the medical products industry. Then, when the Obama administration surprisingly did focus public attention on the need for more federal investment in research on the comparative effectiveness of treatments—an idea that had support among health care experts associated with both parties— the issue became politicized, as the next chapter shows. Republicans accused the administration of “rationing” and interfering with the doctor-patient relationship. In sum, the government’s performance as a problem-solving institution with respect to health care quality and efficiency was doubly disappointing. First, the medical guesswork problem was a low-priority concern, barely registering on the policy agenda; it was an “inside initiative” in which the conversation excluded both politicians and the general public.³ Later, when the medical evidence problem began to generate the broad attention it deserved, proposed solutions became the object of partisan manipulation and political distortion.

Taken together, the two episodes suggest that systemic failures in the performance of our institutions of collective choice are not automatically self-correcting; they can persist for long periods of time. They also point to the political limits of the Progressive reform model of government as a neutral, problem-solving institution built on professional expertise, scientific management, and enlightened public opinion.⁴

Why the Supply of Political Entrepreneurship May Be Too Low

It is useful to begin by stepping back from the details of the medical evidence problem to ask: Why would government fail to address a major, well-documented problem that is causing the nation to perform below potential year after year? The proximate causes of government failure will of course vary across sectors and over time. Just as each unhappy family is unhappy in its own way (to borrow from Tolstoy), so each story of democratic performance failure is distinctive. The starting point for analysis, however, is that pragmatic problem solving and the promotion of the public interest may quite simply offer meager political returns to reelection-minded politicians.⁵

If, following Joseph Schumpeter, we construe societal problem solving as an unintended by-product of partisan competition and the pursuit of power rather than something that the average politician pursues for its own sake,⁶ a tension exists in representative democracy. From the standpoint of social welfare, a new policy should be adopted if the social benefits are greater than the costs, whereas from the standpoint of a politician, a policy should be adopted if the political benefits to the politician are greater than the political costs. Good policies that have large social benefits but small, selective benefits for the politician may not find a political sponsor. Proposing creative, thoughtful solutions to important policy problems may not be the best way for savvy politicians to generate support.

This is so for several reasons. First, the simple act of defining a problem is not straightforward. There may be disagreement over whether some objective societal condition constitutes a “problem” that requires governmental intervention. Is wealth inequality a problem that should be addressed by government action? Is the level of soft drink consumption a social problem? Even if people agree that a problem exists, there will typically be disagreement over what to do about the problem.⁷ Is gun control the solution to mass shootings? Is the answer better-trained police forces or better mental health services? These debates are the very stuff of politics. A legislator who genuinely wishes to do good for society, while also doing well for him- or herself, first needs to frame a problem in a way that brings the public along yet also makes sense from the standpoint of instrumental rationality. Finally, to turn an idea into policy, the politician must then “deliberate, bargain, and compromise in a fishbowl setting in a fashion that can swerve both publics and experts toward emergent solutions.”⁸

None of this is likely to happen in the absence of political entrepreneurship. Political entrepreneurs are creative actors who invest their time, energy, resources, and political reputations to advocate for an idea or proposal—in order to capture a political reward.⁹ They frame issues, create new public demands, build coalitions, and expand the set of issues it is considered legitimate and expected for government to address.¹⁰ To use an economics analogy, political entrepreneurs are “sellers” of problem definitions and proposed solutions who seek to satisfy an unmet social need in exchange for a price, which may take the form of reputational gains or advancement of ideological goals. As legal theorist and federal judge Richard Posner observes, political entrepreneurs can have a large influence on public policy:

The voting public did not know that it wanted social security, conscription, public education, an independent central bank, an interstate highway system, a Presidency opened to divorced or Catholic persons, the North Atlantic Treaty Organization, or the auctioning of rights to the use of the electromagnetic spectrum before those things were proposed by political entrepreneurs, as distinct from run-of-the-mill politicians.¹¹

Taking a step back, the surprising situation of licensed physicians performing procedures and ordering tests not grounded or based on evidence about what works best for patients would seem to be precisely the kind of important problem where a healthy dose of political entrepreneurship is warranted. Yet there is no guarantee that the need for political entrepreneurship will generate its own supply. We develop a theory of “zero-credit politics” to explain why this is so. The core idea is not that the supply of political entrepreneurship will literally be zero—there will nearly always be some actors who are willing to pursue a cause to make a reputation for themselves or further their personal agenda—but rather that the absence of significant political rewards for tackling an important but thorny problem will lead to a much lower level of entrepreneurship in the policy community than is socially optimal.

THE THEORY OF ZERO-CREDIT POLITICS

The role of political entrepreneurs as agents of policy change has received considerable attention.¹² Despite extensive research on the topic, however, few studies examine the role that political entrepreneurship plays in improving governance and system performance. Many political scientists have casually borrowed a market analogy from economics, but few have followed the logic of the analogy or adapted the theory of entrepreneurship to the special features of democratic politics.¹³

Economists have argued that the activities of private sector entrepreneurs make a vital contribution not only to economic change, but to economic growth and improvements in social welfare.¹⁴ Similarly, political entrepreneurship deserves close attention not only because it provides a channel for innovation and the disruption of existing political equilibria, but also because it can shape the problem-solving capacity of a society. In the short run, a political system’s performance depends on the responsiveness of elected officials to extant citizen preferences and needs, but in the long run it depends on the capacity of its leaders and institutions to identify problems, fashion solutions, and adapt to evolving conditions.

Nothing ensures that the supply of problem-focused political entrepreneurship will be adequate from society’s standpoint. Political entrepreneurship entails creating a demand among voters for new policy ideas. This task is far more difficult than the routine marketing of proposals already in circulation. This effort to advocate for and propose new policy ideas and solutions might be viewed as making a risky investment. There are two ways the investment can go sour. First, if the public declines to buy the new policy product, the entrepreneur’s own political reputation may suffer. If a politician pushes a policy idea that conflicts with voters’ beliefs or understanding of reality, it is possible that voters will conclude that the politician is a courageous truth teller and change their minds. But, it is also possible (and perhaps more likely) that voters will instead remain skeptical about the proposal and become skeptical about its advocate as well.¹⁵ There are good reasons for voter skepticism. Sometimes politicians are well meaning but out of step with their constituents. Sometimes they really are snake oil salesmen. When an uninvited visitor comes knocking with promises to help, it is not absurd to think that the first instinct of many citizens is to slam the door. One thing is certain: the opponents of the policy change can be counted on to pounce if they sense an opportunity to denigrate the proposal and thereby undermine the policy innovator’s public standing. And, if the public is initially skeptical, such attacks will have a considerable tailwind.

The second risk arises if the entrepreneur’s efforts to build public support for a new policy approach begin to work. In a commercial setting, a risky investment often enjoys legal protections such as patents and trademarks. In a political setting, however, there is nothing to stop an opportunistic opponent who observes the changes in public opinion produced by a rival’s hard work from proposing a substantively similar proposal of his or her own. If this effort at political mimicry is successful, the initial, true policy innovator will capture, at best, a small share of the credit for the results of his or her efforts. Worse, the second politician, by hanging back until political conditions become more favorable and observing how opinion unfolds, may generate more support for his or her alternative scheme, a copycat plan tweaked to be better tailored to public opinion. In the ruthlessly competitive world of democratic politics, the political entrepreneur and innovative problem solver could even end up worse off for his or her effort.

We call this dynamic “zero-credit politics,” meaning a government intervention or activity that offers no or few “captureable” political returns even though it has large net social benefits.¹⁶ At the extreme, if novel solutions to societal problems are perfectly appropriated, the effect will be to discourage the entrepreneurial investments in the first place.¹⁷

To be sure, the zero-credit dynamic does not always thwart all efforts at innovative policy reform. One can point to examples where political entrepreneurs were willing to invest their time and energy on behalf of the development and advocacy of proposals, including reforms that served a diffuse public interest.¹⁸ An example is airline deregulation. In the 1970s, the federal government controlled the fares airlines could charge and what routes they could fly. Airline service was generally excellent, but ordinary Americans could not find affordable flights to take them where they wanted to go. Although economists argued on the basis of theory and demonstration studies that the freeing of market forces would be highly beneficial for air passengers, the millions of potential winners from airline deregulation were largely unorganized. In contrast, the major airlines that benefited from anticompetitive regulations were mobilized. Nevertheless, in 1978 Congress passed an airline deregulation bill in the face of strong industry opposition. Key to this reform victory was the entrepreneurial activity of legislators such as Senator Ted Kennedy (D-MA), who saw in airline deregulation the potential to enhance his standing with the burgeoning consumer movement.¹⁹ To be sure, expertise alone did not bring about reform; the forces of good policy and good politics had to be brought into alignment through skillful manipulation of the procedural context in which decisions were made.²⁰ There is no escape from politics. Yet airline deregulation, together with cases such as the creation of an efficient market trading system to control the sulfur dioxide emissions that cause acid rain, clearly demonstrates that general-interest reform is possible in American government, even if it is not an everyday occurrence.²¹

But it is a mistake to infer from these canonical examples that the need for political entrepreneurship to improve system performance will automatically lead actors to supply it. The expected return on political entrepreneurship may simply be too low. While James Q. Wilson emphasized that policy entrepreneurs like Ralph Nader could speak on behalf of the unorganized and serve as an effective agent of change,²² the incentives to engage in entrepreneurial activity may be much lower when the status quo is dominated not by corporations or labor unions, whom voters may view as motivated by narrow economic interests, but by members of highly esteemed professions, such as doctors, whom voters may view as public regarding.

Empirical Support for the Zero-Credit Politics Model

We offer support for the zero-credit politics model in two ways. At the end of this chapter, we provide a brief historical narrative of the federal government’s sluggish efforts to promote EBM between the 1970s and 2008. The narrative shows that relatively few members of Congress sponsored or cosponsored legislation to strengthen the government’s role in funding or overseeing comparative effectiveness research to identify what treatments work best. This case study evidence is instructive because lawmakers who did support such proposals were acting as institutional designers. They were not questioning the appropriateness of particular treatments or suggesting that specific medical societies were not behaving responsibly. In most cases, they were merely proposing to create an agency that would conduct research that might someday discredit popular treatments prescribed by doctors. Nevertheless, even tiptoeing into questions about need for greater third-party oversight of doctors’ therapeutic authority was evidently seen by members of Congress as a “bad” investment of entrepreneurial energy—presumably because calling attention to the problems of overtreatment and unnecessary spending implicitly challenges the public’s trust in doctors as guardians of patient and social welfare. As chapter 6 shows, Congress did eventually vote to increase taxpayer funding of health outcomes and created a new, nongovernmental, nonprofit entity (PCORI) to coordinate the work—but that was only because the proposals “hitched a ride” with two virtually must-pass, omnibus Democratic Party vehicles (the 2009 Recovery Act and the Affordable Care Act). The proposals likely would not have gained traction as stand-alone measures. In sum, the policy history of Congress’s role in CER is broadly consistent with the zero-credit politics model.

While a review of past events illuminates what occurred, it tells us less about what could have happened but did not. To appreciate how political forces channel and constrain the incentives for political entrepreneurship, innovation, and pragmatic problem solving, it is also essential to consider what might have happened if politicians had been more willing to support evidence-based medicine, even at the cost of challenging doctors. It is important to view the lack of a robust response by elected representatives to the vast gaps in medical evidence as a genuine puzzle that requires an explanation. Imagine a world in which a politician decides to become a political entrepreneur to address medical evidence gaps. Would this politician be punished or rewarded by voters for his or her efforts?

This counterfactual question cannot be answered definitively, but we can illuminate it through survey experiments that employ vignettes about medical evidence controversies. We find that if politicians challenge doctors and medical societies on the appropriateness of treatments, their electoral standing declines, even if the scientific evidence is on the politicians’ side. To be sure, politicians are quite reluctant to inject themselves into medical evidence controversies in the real world. The scenarios we describe rarely occur. But that is exactly the point. Challenging doctors is (in the terms of the economist) “off the equilibrium path” behavior. That is, certain kinds of efforts to promote evidence-based policy making, which could produce benefits for society, do not occur because of the political reactions or the punishments they would trigger. In short, in their potential roles as both institutional designers of expert agencies to promote health care quality and efficiency and as “position takers” on specific medical controversies, lawmakers have recognized that even attempting to curb unnecessary treatments and impose new monitoring and accountability mechanisms of the medical profession offers few electoral rewards. The take-away lesson is that the lack of effective political oversight of the use of evidence in the U.S. health care system is the product of powerful incentives. Only by taking such incentives into account can durable and effective reforms occur.

SURVEY ON POLITICAL ENTREPRENEURSHIP

Below we highlight results from a July 2015 web-based survey that we conducted on a nationally representative sample of 1,100 U.S. adults. The survey contained a variety of questions designed to examine the plausibility of the theory of zero-credit politics.

Members of Congress have multiple goals, but they cannot achieve any of them unless they win reelection. That is why most political science theories of legislative behavior start from the premise that legislators want to remain in office.²³ But what specific governing tasks will reelection-seeking lawmakers pursue to win votes? Are some activities more highly valued by the public than others? Can lawmakers reasonably expect to build a base of support among voters on the basis of entrepreneurial problem-solving activities, or would they generate greater electoral rewards by placing their attention elsewhere? To address these questions, we asked respondents to evaluate the performance of a politician based on a list of 18 factors individually.

Specifically, respondents were asked, “Below is a list of factors related to a politician’s job performance. Please tell us the extent to which each factor would affect how likely you are to vote for your congressional representative.” The response options ranged from “not at all more likely to vote for” to “a great deal more likely to vote for” on a five-point scale. The 18 items we asked about are listed in table 5.1. They range from factors that we did not expect to affect evaluations of politicians all that much (e.g., “spends a lot of time outdoors”) to those that were meant to capture activities that in theory directly benefit constituents (e.g., “provides high-quality help to constituents who are running into problems with Social Security”) to items that were meant to capture “entrepreneurial activity” related to solving collective problems.

A clear result—unsurprising yet of fundamental importance—is that voters are not clamoring for legislators to devote their time and energy to political entrepreneurship targeted at developing the expertise necessary to address broad societal problems. Only 23 percent of respondents stated that they were much or a great deal more likely to vote for a politician who “is a genuine expert on a policy area that is important to the nation as a whole but is not of special importance to your district,” whereas 45 percent were not at all or a little more likely. In contrast, 52 percent were much or a great deal more likely to vote for a politician who “is an expert on the issues of greatest concern to your district,” and only 19 percent were not at all or a little more likely.²⁴

Note: Question wording (and response options): “Below is a list of factors related to a politician’s job performance. Please tell us the extent to which each factor would affect how likely you are to vote for your congressional representative. (Not at all more likely to vote for; A little more likely to vote for; Somewhat more likely to vote for; Much more likely to vote for; A great deal more likely to vote for).”

Source: July 2015 YouGov/Polimetrix survey.

To be sure, it is possible for elected officials to gain public support while engaging in some entrepreneurial activities. We asked respondents to consider two activities relevant to “marketing” new policy products: (1) developing new policy solutions that have not been considered before and (2) questioning conventional wisdom about public policy issues. Both of these factors received high levels of support, as did general leadership traits such as having the courage to challenge powerful actors and organizations and being effective at advancing bills in the legislative process. Yet while these generic political skills may have a bearing on political entrepreneurship focused on problem solving, they can also feed into symbolic activity or legislating on behalf of narrow geographic or group interests. Where entrepreneurship is most needed—and most likely to be undersupplied—is problem solving on behalf of broad, unorganized groups.²⁵

Our results suggest that the public is less likely to reward this type of political entrepreneurship, especially when it involves becoming a policy expert and not just publicizing ideas. As an illustrative example, we asked if respondents would be more likely to vote for a legislator who has average knowledge about the problems of the district, but who is a genuine expert on U.S. counterterrorism policy in the Middle East. Just 23 percent of respondents said they would be much more or a great deal more likely to vote for such a lawmaker.

An analysis of the cross section between support for national problem solving and generic political entrepreneurship highlights the limited public receptivity for national problem solving. Of the 617 respondents in our survey who said they would be a great deal or much more likely to vote for a lawmaker who has the courage to challenge powerful actors and organizations, only one in three (32 percent) also said they would be a great deal or much more likely to vote for a representative who is a genuine expert on a policy area that is important to the country as a whole but is not of special importance to their district. Similarly, only 40 percent of the respondents who said they were either a great deal or much more likely to vote for a representative who is effective at advancing bills through the legislative process or developing new policy solutions were also a great deal or much more likely to vote for a representative who is a national policy expert.

PUBLIC SUPPORT FOR LAWMAKERS WHO STAND UP FOR EVIDENCE-BASED MEDICINE

The survey results discussed above suggest that the safest course for the reelection-minded lawmaker is to focus on district representation, and if the lawmaker does want to develop a policy reputation to focus on the distinctive concerns of his or her constituency rather than broad national challenges. The would-be entrepreneur contemplating devoting his or her scarce time and energy to developing the expertise required not just to publicize but to actually develop and build support for solutions to broad problems knows that it is, at best, a political investment with uncertain returns.

These conclusions were produced by asking respondents about what they value in a representative in a very abstract context. Do we find similar results when respondents are asked about specific legislative actions taken by a representative? What would happen if a lawmaker who understands the problem of overtreatment and unnecessary care stands up for sound science and the promotion of evidence-based medicine? Will he or she be rewarded or punished for such efforts? To address this question, we conducted two survey experiments.

In the first experiment, respondents read a short vignette about a study published in a medical journal that questioned the effectiveness of a procedure to treat heart disease.²⁶ Respondents (not in the baseline, “study only” condition) were randomly assigned to read about the stances of doctors and/or members of Congress concerning the journal article. After reading the vignette, respondents were first asked to answer a battery of questions concerning use of the procedure, trust in the various entities mentioned in the vignette, and evaluations of a member of Congress (when applicable).

Specifically, all respondents read the following prompt (bolded text was bolded on the screen for respondents):

A recent study published in a leading, peer-reviewed medical journal shows that a common procedure for heart disease is not effective. The study concludes that there are better treatments, and the procedure should be used much less often.

Respondents randomly assigned to the “study only” condition read only this information before answering questions about what they had read. All other respondents were randomly assigned (with equal probability) to one of 10 experimental treatment conditions. We focus on six of those conditions here, each of which is summarized in table 5.2. In short, some respondents read that doctors supported the study’s recommendation that “the procedure should be used much less often” whereas others read that doctors opposed the study’s recommendation. Similarly, some respondents read that politicians (“Representative X”) supported the study’s recommendation whereas others read that politicians (“Representative X”) opposed the study’s recommendation. Importantly, the arguments used were the same, which allows us to measure whether the fact that doctors or politicians are the informational source matters for respondents. In addition, two of the experimental conditions included responses from both doctors and politicians. In one, both doctors and politicians opposed the study’s recommendation. In the other, doctors opposed the study’s recommendation, but politicians supported it.

TABLE 5.2. Summary of Heart Disease Vignette Experimental Conditions
Experimental Condition	Text
STUDY ONLY	A recent study published in a leading, peer-reviewed medical journal shows that a common procedure for heart disease is not effective. The study concludes that there are better treatments, and the procedure should be used much less often.
DOCTORS OPPOSE STUDY	[“Study only” text plus] Doctors from an association of cardiologists (heart doctors) disagree with the study’s findings. They say that they have decades of experience treating patients with heart disease, and the procedure is effective. The cardiologists say that the study is wrong, and they oppose reducing the use of the procedure. They say reducing use of the procedure would cost many lives.
DOCTORS SUPPORT STUDY	[“Study only” text plus] Doctors from an association of cardiologists (heart doctors) agree with the study’s findings. They say this is the first rigorous study of the effectiveness of the procedure for treating patients with heart disease, and the study was well done. The cardiologists say that the study is right, and they support reducing the use of the procedure. They say reducing the use of the procedure would save many lives.
MEMBER OF CONGRESS OPPOSES STUDY	[“Study only” text plus] A number of politicians disagree with the study’s findings. Representative X says that doctors have decades of experience treating patients with heart disease, and the procedure is effective. Representative X says that the study is wrong, and he opposes reducing the use of the procedure. He says reducing use of the procedure would cost many lives.
MEMBER OF CONGRESS SUPPORTS STUDY	[“Study only” text plus] A number of politicians agree with the study’s findings. Representative X says this is the first rigorous study of the effectiveness of the procedure for treating patients with heart disease, and the study was well done. Representative X says that the study is right, and he supports reducing the use of the procedure. He says reducing use of the procedure would save many lives.
DOCTORS AND MEMBER OF CONGRESS OPPOSE STUDY	[“Study only” text plus “Doctors oppose study” text plus “member of Congress opposes study” text]
DOCTORS OPPOSE STUDY; MEMBER OF CONGRESS SUPPORTS STUDY	[“Study only” text plus “Doctors oppose study” text plus “member of Congress supports study” text]

The first outcome of interest we present (in figure 5.1) concerns respondent beliefs about whether the procedure should be used more or less often. Specifically, we asked: “Based on the exchange, do you think the procedure should be used: Less often than it is currently used; At the same rate as it is currently used; More often than it is currently used.” We scored these responses so that “more often” is given a value of 1, “less often” -1, and “the same” 0. Figure 5.1 presents average responses to this question by experimental condition.

Figure 5.1 shows that, on average, respondents in the Study Only condition agreed with the study’s recommendation that the procedure should be used less often—the average response is around -.5. When doctors oppose the study’s recommendation, however, this shifts to around -.25. In other words, when respondents learn that physicians oppose the study, support for using the procedure less often declines (p<.05 for difference between Study Only and Doctors Oppose Study conditions). The member of Congress (Representative X) making the same argument as doctors has a similar effect on respondents (p<.05 for difference between Study Only and Member of Congress Opposes Study conditions); (p>.10 for difference between Doctors Oppose Study and Member of Congress Opposes Study conditions). Neither doctors nor politicians supporting the study’s recommendation, however, moves opinion about how often the procedure should be used much from where it is in the Study Only condition (p>.10 for both comparisons to the Study Only condition, and the comparison of Doctors Support Study and Member of Congress Supports Study conditions). This suggests that although doctors and politicians can both discredit the study (when using the same argument), neither bolster public support for the study’s recommendation.

The most interesting comparison is the one between the two conditions in which both doctors and politicians are mentioned. When both doctors and politicians oppose the study’s recommendation there is no cumulative effect (i.e., the effect of Doctors Oppose Study and Doctors and Member of Congress Oppose Study are statistically indistinguishable from one another,as is Member of Congress Opposes Study and Doctors and Member of Congress Oppose Study). However—and this point is critical—politicians are not able to counteract the opposition of doctors. When politicians support the study’s recommendation to reduce the use of the heart procedure in light of research findings and doctors oppose this recommendation, the average response is no different in this condition compared to when both doctors and politicians oppose the study’s recommendation (p>.10 for comparison of Doctors and Member of Congress Oppose Study and Doctors Oppose and Member of Congress Supports Study).

FIGURE 5.1. Effects of elite positions on public support for use of heart disease procedure after study says procedure should be used less often. Note: Dots represent means for each experimental condition; whiskers are 95% confidence intervals. In the “Study Only” condition respondents were simply told, “A recent study published in a leading, peer-reviewed medical journal shows that a common procedure for heart disease is not effective. The study concludes that there are better treatments, and the procedure should be used much less often.” Other conditions, as displayed in table 5.2, added information for those respondents about whether, for example, doctors opposed (“Dr Oppose Study”) or supported (“Dr Support Study”) the study’s findings. The figure shows that when doctors oppose the study’s findings, public support for using the procedure less often declines. Source: July 2015 YouGov/Polimetrix survey.

In sum, a politician who takes a stand in support of sound evidence has no independent ability to influence public opinion about the use of the treatment if a doctors’ group attacks the recommendation. This would be highly discouraging to any lawmaker contemplating wading into these waters. But will the public reward or punish the member for their courage? What are the political returns to supporting or opposing the doctors?

FIGURE 5.2. Effects of elite positions about use of heart disease procedure after study says procedure should be used less often on public approval of job member of Congress is doing. Note: Dots represent means for each experimental condition; whiskers are 95% confidence intervals. Question wording: “How likely do you think it is that Representative X is doing a good job as a representative?” Scale ranged from 1 (very unlikely) to 7 (very likely). The figure shows that compared to when the position of doctors is not mentioned, when doctors oppose the study’s recommendations the member of Congress’s support of the study’s recommendations results in views that the member of Congress is doing a worse job. Source: July 2015 YouGov/Polimetrix survey.

To find out, we also asked respondents who were assigned to a condition that mentioned a politician to evaluate the member of Congress (Representative X). We asked them, “How likely do you think it is that Representative X is doing a good job as a representative?” Responses were recorded on a seven-point scale ranging from “very unlikely” to “very likely.” Figure 5.2 presents average responses for the four conditions in which Representative X was mentioned. The primary takeaway from figure 5.2 is that when the member of Congress challenges doctors’ opposition to the study’s recommendation, the lawmaker receives the most negative evaluation—about 3.5 on the seven-point scale, which falls somewhere between being somewhat unlikely to think the representative is doing a good job and being undecided. We observe this effect most clearly when we compare the condition in which the legislator supports the study’s recommendation without any mention of doctors (second to left dot in figure 5.2) to the one in which the legislator supports the study’s recommendation but doctors oppose it (far right dot in figure 5.2). Evaluations of the member of Congress are a full point higher on the seven-point scale when doctors are not mentioned in the vignette compared to when their opposition is described (p<.05). Overall, the results of this vignette experiment suggest that politicians face an uphill battle when they face off against doctors in the realm of health care policy, specifically when the use of medical procedures and medical effectiveness is in question.

To see if these findings were robust across contexts, we performed a second survey experiment. Respondents read a short vignette about a recommendation by a task force regarding use of the PSA test. Respondents (not in the baseline, “task force only” condition) were randomly assigned to read about the stances of doctors and/or members of Congress concerning the task force’s recommendation. After reading the vignette, respondents were asked to answer a battery of questions concerning use of the procedure, trust in the various entities mentioned in the vignette, and evaluations of the member of Congress (when applicable).

All respondents read the following prompt (bolded text was bolded on the screen for respondents):

In 2012 a government-appointed, independent panel of national experts in prevention and evidence-based medicine issued a recommendation that men who have no symptoms of prostate cancer should not routinely be given a PSA test, a common test to screen for prostate cancer. The task force concluded that it is likely “that the service has no benefit or that the harms outweigh the benefits.”

Respondents randomly assigned to the “task force only” condition only read this information before answering questions about what they had read. All other respondents were randomly assigned (with equal probability) to one of 11 experimental treatment conditions. We focus on five of those conditions here, each of which is summarized in table 5.3. In this vignette, all respondents read some information about whether politicians (Representative A) supported or opposed the task force’s recommendation “that men who have no symptoms of prostate cancer should not routinely be given a PSA test.” Similar to the design of the previous experiment, when Representative A supported the recommendation of the task force, we also had a condition in which doctors opposed the recommendation. Finally, we had two conditions in which another lawmaker (Representative B) opposed the task force’s recommendation (when Representative A supported it), and doctors either supported (in one condition) or opposed (in a different condition) the recommendation of the task force.

TABLE 5.3. Summary of PSA Vignette Experimental Conditions
Experimental Condition	Text
TASK FORCE ONLY	In 2012 a government-appointed, independent panel of national experts in prevention and evidence-based medicine issued a recommendation that men who have no symptoms of prostate cancer should not routinely be given a PSA test, a common test to screen for prostate cancer. The task force concluded that it is likely “that the service has no benefit or that the harms outweigh the benefits.”
REP. A OPPOSES TASK FORCE	[“Task force only” text plus] A number of politicians disagree with the task force’s recommendations. Representative A says that we should trust doctors to do what is best for their patients, and not base medical practice on the recommendations of so-called experts on a task force. Representative A says that the task force is wrong. He says that recommending against routine screening would cost many lives.
REP. A SUPPORTS TASK FORCE	[“Task force only” text plus] A number of politicians agree with the task force’s recommendations. Representative A says that we should follow the scientific evidence and base medical practice on the task force recommendations. Also, doctors sometimes push for more screening than is medically necessary because they receive money from tests and procedures. Representative A says that the task force is right. He says that recommending against routine screening is the right decision.
REP. A SUPPORTS TASK FORCE; DOCTORS OPPOSE	[“Task force only” text plus “Rep. A. supports task force” text plus] Doctors from an association of urologists challenge Representative A’s support for the task force’s recommendations. They say that we should trust doctors to do what is best for their patients, and not base medical practice on the recommendations of so-called experts on a task force. The urologists say that both the task force and Representative A are wrong. They say that recommending against routine screening would cost many lives.
REP. A SUPPORTS TASK FORCE; REP. B OPPOSES; DOCTORS SUPPORT	[“Task force only” text plus “Rep. A. supports task force” text plus] Representative B challenges Representative A’s support for the task force’s recommendations. Representative B says that we should trust doctors to do what is best for their patients, and not base medical practice on the recommendations of so-called experts on a task force. Representative B says that both the task force and Representative A are wrong. Representative B says that recommending against routine screening would cost many lives. Doctors from an association of urologists weighed in to say that Representative B is wrong. They say that Representative A is right and that the task force recommendation against routine screening is the right decision.
REP. A SUPPORTS TASK FORCE; REP. B OPPOSES; DOCTORS OPPOSE	[“Task force only” text plus “Rep. A. supports task force” text plus] Representative B challenges Representative A’s support for the task force’s recommendations. Representative B says that we should trust doctors to do what is best for their patients, and not base medical practice on the recommendations of so-called experts on a task force. Representative B says that both the task force and Representative A are wrong. Representative B says that recommending against routine screening would cost many lives. Doctors from an association of urologists weighed in to say that the task force is wrong and Representative A is wrong. They say that Representative B is right that recommending against routine screening would cost many lives.

The results of this study are summarized in figures 5.3 and 5.4. As with the heart disease vignette study, the first outcome of interest we present concerns respondent beliefs about use of the procedure (in this case, the PSA test). Specifically, we asked: “Based on the exchange, do you think the screening test should be routinely given to men with no symptoms of prostate cancer?” The response options were, “Yes, screening test should be routinely given to men with no symptoms of prostate cancer”; “No, screening test should not be routinely given to men with no symptoms of prostate cancer”; and “I don’t know.” We scored these responses so that “yes” is given a value of 1, “no” -1, and “don’t know” 0. Figure 5.3 presents average responses to this question by experimental condition.

Figure 5.3 shows that, on average, respondents in the Task Force Only condition were uncertain (at 0) about whether the PSA test should be given to men without symptoms. The position of Representative A does move opinion on this issue—when Representative A opposes the task force recommendation, more people think the PSA test should be given to men without symptoms (around .3 on the scale, p<.05); when Representative A supports the task force recommendation, more people think the PSA test should not be given to men without symptoms (around -.2 on the scale, p=.18). However, the next dot over (fourth from the left) shows that Representative A’s support of the task force recommendation can be completely counteracted by the opposition of physicians, as the average response in that condition is virtually the same as when Representative A opposed the task force recommendation himself (p<.05 for difference between Representative A Supports Task Force and Representative A Supports Task Force and Doctors Oppose Task Force).

The last two entries of figure 5.3 further illustrate the unique power of physicians in this area. In those two conditions, there was a disagreement between politicians. When such disagreements occur, our results suggest that doctors can move opinion. In the face of political disagreement over the task force recommendation, doctors’ support of the task force recommendation moves people in the direction of the task force recommendation; doctors’ opposition to the task force recommendation, on the other hand, moves people away from the task force recommendation (p<.05 for difference between these two conditions).

FIGURE 5.3. Effects of elite positions on public support for use of PSA test after task force recommends PSA test should not be given to men without symptoms. Note: Dots represent means for each experimental condition; whiskers are 95% confidence intervals. In the “Task Force (TF) Only” condition respondents were told, “In 2012 a government-appointed, independent panel of national experts in prevention and evidence-based medicine issued a recommendation that men who have no symptoms of prostate cancer should not routinely be given a PSA test, a common test to screen for prostate cancer. The task force concluded that it is likely ‘that the service has no benefit or that the harms outweigh the benefits.’” Other conditions, as displayed in table 5.3, added information for those respondents about whether, for example, a representative opposed (“Rep. A Oppose TF”) or supported (“Rep. A Support TF”) the task force’s recommendation. The figure shows that compared to the “task force only” condition, when Rep A. supports the task force’s recommendation, respondents are somewhat more likely to think the PSA test should not be given to men without symptoms, but not when doctors oppose the task force’s recommendation. Source: July 2015 YouGov/Polimetrix survey.

We also asked respondents who were assigned to a condition that mentioned a politician to evaluate Representative A. We asked them, “How likely do you think it is that Representative A is doing a good job as a representative?” Responses were recorded on a seven-point scale ranging from “very unlikely” to “very likely.” Figure 5.4 presents average responses for the five conditions in which Representative A was mentioned. The primary takeaway from figure 5.4 is that when there is a political disagreement over EBM (last two dots of figure 5.4), evaluations of the member of Congress are more favorable when the lawmaker is on the side of doctors.²⁷ When doctors are of the same opinion as Representative A (that is, both support the task force recommendation), Representative A’s evaluation is almost 4.5 (somewhere between being undecided and somewhat likely to think the representative is doing a good job). However, when doctors disagree with Representative A and the task force recommendation, Representative A’s evaluation falls to just under 3.7 (nearly a full point, difference is statistically significant, p<.05). In short, being a policy entrepreneur in this area is likely prudent only when doctors are in agreement with your position.

FIGURE 5.4. Effects of elite positions about use of PSA test after task force recommends PSA test should not be given to men without symptoms on public approval of job Rep. A is doing. Note: Dots represent means for each experimental condition; whiskers are 95% confidence intervals. Question wording: “How likely do you think it is that Representative A is doing a good job as a representative?” Scale ranged from 1 (very unlikely) to 7 (very likely). Comparing the two dots on the right hand side of the figure suggests that when politicians (Rep. A and Rep. B) are in disagreement over the task force’s recommendation, the position of doctors determines whether approval of Rep. A is higher (when doctors agree with Rep. A) or lower (when doctors disagree with Rep. A). Source: July 2015 YouGov/Polimetrix survey.

These survey experiments are simple by design, but they capture a fundamental truth about efforts to tighten oversight of the clinical authority ofphy- sicians and medical societies: it is a risky proposition for reelection-minded lawmakers. This, at least in part, helps explain why we have seen relatively little of this activity.

The Government’s Sluggish Response to the Medical Evidence Problem, 1970s-2008

The survey experiments described above necessarily abstract from reality. But what has actually happened in Washington? How much effort have lawmakers devoted over the past several decades to tackling the problems of wasteful spending, overtreatment, and bad science in American medicine?

Below, we present a thumbnail history of the federal government’s efforts to address the medical evidence gap prior to the Obama administration.²⁸ Before recounting this history, however, it is useful to ask why government even has a role in funding research on the comparative effectiveness of treatments. Why can’t the provision of information about the relative benefits of treatment alternatives be left to the private market? Insurance plans have long conducted some evidence-based technology assessments to inform their coverage decisions²⁹—so it is not the case that no evidence will be generated without taxpayer money.³⁰ The reason is that CER is a “public good,” which will be undersupplied without government support. As the Congressional Budget Office (CBO) explains:

The knowledge created by such studies is costly to produce—but once it is produced, can be disseminated at essentially no additional cost, and charging all users for access to that information is not always feasible. As a result, private insurers and other entities conducting research on comparative effectiveness often stand to capture only a portion of the resulting benefits and therefore do not invest as much in such research as they would if they took into account the benefits to all parties. In health plans that do not have exclusive provider networks, some of the benefits probably “spill over” to other health plans using the same doctors, because physicians tend to use a similar approach to care for all of their patients. Even if organizations could keep their findings confidential, so that they captured all of the benefits, some duplication of effort would probably occur. In such a situation, research constitutes a “public good,” and economists have long recognized a role for government to increase the supply of such research toward the socially optimal level.³¹

The existence of a market failure does not mean that government will step in to correct it. The perceived costs of more rigorous scrutiny of treatments, tests, and technologies are concentrated on well-organized groups such as device manufacturers and medical specialty societies. This is a classic “diffuse benefits/concentrated costs” situation that Wilson identified as prone to undermining good policy making.³²

One way to overcome this problem is through “public interest” lobbying. The millions of patients who suffer from cancer, heart disease, and other chronic conditions are dispersed across the nation. Tragically, the patients who are in most desperate need of unbiased information about treatment options are often least able to mobilize politically. “Patient advocacy” groups ostensibly represent the interests of such patients. The number of such organizations has exploded in recent years. Patient advocacy organizations have the capacity to accelerate FDA drug reviews,³³ and they could be an important voice for better use of evidence in medicine. However, research suggests that patient advocacy groups do not act like public interest organizations.³⁴ They often act much more like nonoccupational trade associations. Many patient groups are underwritten by drug companies.³⁵ Whether owing to financial or political pressure or organizational maintenance reasons, the patient advocacy community has not prioritized the need to increase the funding of a generic public good like CER. Patient advocacy tends to be disease specific rather than focused on general patient goals or the overall improvement of the health care system.³⁶

The theory of zero-credit politics helps account for three observable facts in this history. First, it accounts for the relatively low level of interest in the issue among elected officials, with the important exception of a few lawmakers including Tom Allen (D-ME), Jo Ann Emerson (R-MO), and Max Baucus (D-MT). Second, it helps explain why the most dogged and passionate actors who have tried to sell CER have been private citizens (John E. Wennberg), former officials (Gail Wilensky), or political appointees (former Congressional Budget Office and Office of Management and Budget director Peter Orszag), who, compared to elected officials, have not had to worry about pleasing voters and winning elections. Finally, the theory offers insight into why elected officials have often found the payoff from attacking or weakening EBM reforms greater than the payoff from initiating or strengthening them, especially when faced with hostility from doctors and doctors’ organizations.

THE RISE AND FALL OF THE NATIONAL CENTER FOR HEALTH CARE TECHNOLOGY

Between the 1970s and mid-2000s, support among politicians for efforts to promote the integration of hard data into clinical decisions was inconsistent.³⁷ A few politicians did make entrepreneurial investments of time and energy in the promotion of evidence-based medicine, facilitating the enactment of modest reforms, but most policy makers did not see it as a winning issue. Members of Congress frequently served as allies of industry and medical professionals opposed to reform.

As early as the 1970s, both health services experts and parts of the U.S. health policy community were aware that medical technologies were increasingly diffusing into practice before their effectiveness had been evaluated. Some government officials recognized that the overuse or misuse of expensive technology contributed to the problem of rising health care costs and sought to increase the knowledge base about what interventions work best. A limited (and short-lived) effort to improve the medical evidence base was the National Center for Health Care Technology (NCHCT), which Congress established in 1978 as a small bureau (with a $4 million budget and staff of twenty) inside the Department of Health, Education, and Welfare. Its mission was to generate information about the safety and efficacy of medical technologies and, through its dissemination, to advise the Medicare agency on coverage issues. The act creating the center (PL 95-623) won bipartisan passage in Congress. Hopes were high that the center’s work would improve quality and reduce wasteful spending. As the New York Times reported, pressing questions, such as what is the best treatment for breast cancer, would no longer be considered “only by individual practitioners, who may have limited access to the facts, or by special-interest professional groups, who may have prejudiced views.”³⁸ The NCHCT managed a program of research grants and was given responsibility for developing standards and norms concerning the use of particular technologies. In addition, the NCHCT provided information on the safety and effectiveness (but not cost-effectiveness) of technologies to the Health Care Financing Administration (the precursor agency to the Centers for Medicare and Medicaid Services) for its coverage decisions in the Medicare program. Health and Human Services officials emphasized that the opinions of the NCHCT were only “advisory” and that Medicare did not have to accept them.³⁹ Between 1978 and 1981, the NCHCT provided Medicare with 75 such evaluations, of which approximately 40 percent were for noncoverage.⁴⁰

But the NCHCT was short-lived. In 1981, the Reagan administration zeroed out the agency’s budget, and it ceased operations. (Some of its work was shifted to the National Center for Health Services Research.) Office of Management and Budget director David Stockman viewed NCHCT as an example of excessive government regulation.⁴¹ Yet it would be a mistake to regard sincere ideological opposition as the principal cause of the agency’s early demise. Indeed, the conservative Heritage Foundation’s book Mandate for Leadership, which had a major influence on the Reagan administration’s initial budget submissions to Congress, had called for the creation of a research unit to evaluate medical procedures and deny reimbursement under Medicare and Medicaid for procedures that were challenged.⁴² While antiregulatory sentiments were clearly a factor in the NCHCT’s erosion of support, the most direct cause was the growing opposition of doctors and industry groups. In 1981, the AMA opposed the reauthorization of the agency, arguing that it was a threat to the exercise of individual judgment by physicians:

The relevant clinical policy analysis and judgments are better made—and are being responsibly made—within the medical profession. Assessing risks and costs, as well as benefits, has been central to the exercise of good medical judgment for decades. The advantage the individual physician has over any national center or advisory council is that he or she is dealing with individuals in need of medical care, not hypothetical cases.⁴³

The Health Industry Manufacturer’s Association also opposed NCHCT’s continuation, fearing that the center would constrain the industry’s ability to bring new products to the marketplace.⁴⁴

THE RISE AND FALL OF THE AGENCY FOR HEALTH CARE POLICY AND RESEARCH

While the NCHCT ceased operations in late 1981, the lack of solid, scientific evidence for many common treatments remained a glaring problem. By the mid-1980s, health policy makers in both the legislative and executive branches were becoming increasingly aware of the research on geographical variation in medical practice by Dartmouth professor John E. Wennberg and colleagues. Wennberg’s work on Medicare showed that different regions of the United States had marked differences in the use of many health services that could not be explained by age, gender, race, and other population factors, and that high levels of medical services use were not associated with better outcomes such as mortality and functional status.⁴⁵ The implication was that a large portion (perhaps as much as 20–30 percent) of health care in the United States may be wasteful.⁴⁶ Wennberg himself played an entrepreneurial role in educating elected officials about the practice variation phenomenon through congressional testimony.

Wennberg’s research and the need for better information to address unwarranted variation and curb wasteful spending formed an idea that key policy actors judged as plausible, technically feasible, and responsive to the government’s needs. The concept entered what political scientist John W. Kingdon calls the “policy stream.”⁴⁷ When a window of opportunity opened, a proposal to expand the federal government’s role in improving the medical evidence base led to the creation of the Agency for Health Care Policy and Research (AHCPR) in 1989. The mission of AHCPR was to carry out and disseminate research on the effectiveness, efficiency, quality, and outcomes of health services. In addition, it was charged with developing clinical practice guidelines. The hope was that better health outcomes research would identify unnecessary or low-value services and help control the growth of Medicare spending. Although the idea of having the federal government establish standards for medical practice had long been anathema to physicians, several medical associations, including the AMA, endorsed the inclusion of practice guidelines in the legislation.⁴⁸ Because the law establishing AHCPR was folded into an omnibus budget bill, it never had a separate vote in Congress. The legislative record shows, however, that AHCPR was a bipartisan creation. Key supporters included Congressmen Henry Waxman (D-CA), Bill Gradison (R-Ohio), and Fortney H. (Pete) Stark (D-CA); Senators George Mitchell (D-ME) and David Durenberger (R-MN); and William Roper, Medicare program administrator under President George H. W. Bush.⁴⁹ During its first several years of activity, AHCPR sponsored Patient Outcomes Research Teams to study topics such as prostate enlargement and heart attack, and developed 15 clinical practice guidelines. The agency had a low profile, and its work received little attention from politicians. Its budget grew steadily from $115 million in FY 1991 to $159 million in FY 1995.⁵⁰

Although the AHCPR was launched with great enthusiasm, the agency quickly got into political trouble. Both partisan politics and interest group pressures led to the agency’s loss of support. In 1994, Republicans gained control of Congress with an agenda (the Contract with America) to reduce federal spending. Even though the AHCPR was created under a Republican administration, with the backing of key GOP lawmakers, the shift in the political environment left the agency vulnerable to budget cuts. (The Office of Technology Assessment [OTA], an even more prominent congressional support agency, also did not survive this period.⁵¹ The OTA’s downfall reflected a general lack of congressional support for its research mission, but its health technology assessment work was certainly not without controversy and sparked some medical products industry opposition.) Support for AHCPR among Republicans was further weakened by the perception that it had played an advocacy role in the debate over the Clinton health reform initiative in 1993–94, a perception not helped by the agency’s decision to hire several people who had worked on the Clinton plan and Democratic staffers from Capitol Hill.⁵²

In addition, the agency set off a firestorm when it released practice guidelines for treating acute lower back pain. The guidelines said that back surgery benefited only 1 in 100 patients and should be avoided, and that doctors should also refrain from using imaging tests, X-rays, and MRIs at the beginning of a back pain episode.⁵³ The North American Spine Society, the professional group of back surgeons, criticized the study and took their complaints to Congress. Representative Sam Johnson (R-TX) led an attempt to eliminate the AHCPR, arguing that the government should not tell doctors how to practice medicine.⁵⁴

The agency survived the conflict but suffered a sharp reduction in its budget. Moreover, it was stripped of its role as a developer of practice guidelines. Both the continued existence of the agency, which was renamed the Agency for Healthcare Research and Quality (AHRQ), and the narrowing of its mission are significant. That the agency survived at all reflected the fact that it enjoyed residual support among Republican members of the House, such as Bill Thomas (R-CA), John Porter (R-IL), and Speaker Newt Gingrich (R-GA), Senator James Jeffords (R-VT), well-placed committee staffers, and health care expert Gail Wilensky, who had headed the Medicare program under George H. W. Bush. One Republican Hill staffer recalls accompanying his principal to the House Budget Committee and pleading with the legislators not to zero out the budget for the agency: “Are you going nuts?” the staffer recalled saying. “Republicans created this. It is allowing us to figure out what we are spending money on. [Zeroing it out] is just crazy.”⁵⁵ But the scope of the agency’s influence was severely curtailed. No longer would the agency publish guidelines to direct the behavior of clinicians. Instead, its task was reduced to evidence generation, “leaving it to the health care industry to sort what they should be doing.”⁵⁶ As Harvard School of Medicine professor Jerry Avorn observed, “The linkage between research and policy had been formally renounced, and the nation took one giant step away from bringing science to bear on the systematic assessment of therapeutics.”⁵⁷

Efforts to strengthen the role of CER within Medicare during this period also met with political resistance. Under the program’s authorizing legislation in 1965, Medicare pays for “reasonable and necessary” services, which was “understood to reflect the prevailing views of physicians, although there were no formal criteria to define this standard at either the local or the national level.”⁵⁸ In the late 1980s, the Health Care Financing Agency (later the CMS) proposed a requirement that technologies be cost-effective to be covered. The AMA and the Pharmaceutical Research and Manufacturers Association opposed the proposal, and it was withdrawn.⁵⁹

REFORM FORCES IN CONGRESS, 2000-2008

As noted above, to argue that the supply of political entrepreneurship has been below the socially optimal level is not to claim it has been missing entirely. One legislator who did play an important entrepreneurial role was Representative Tom Allen (D-ME). After the failure of the comprehensive Clinton health reform plan in 1994, Allen was searching for more incremental health reforms that could help make health care more affordable and potentially attract bipartisan support. Allen recognized that FDA approval “means less than most people think,” that the pharmaceutical industry often buried adverse research results, and that there was no guarantee that expensive drugs worked better than cheaper alternatives.⁶⁰ Allen became aware of the drug effectiveness review project started by Oregon governor John Kitzhaber, which sought to integrate drug data into the state Medicaid program, as well as comparative effectiveness initiatives underway in Australia and New Zealand. He became convinced that independent research on the relative effectiveness of drugs would help “reduce system cost and improve quality.”⁶¹

In 2003, Allen and Jo Ann Emerson (R-MO) introduced the Prescription Drug Comparative Effectiveness Act (HR 2356). The bill authorized $50 million for the NIH to conduct research and $25 million for the AHRQ to conduct studies on the comparative effectiveness and cost-effectiveness of prescription drugs. There was a conscious effort not to go as far as drug assessment systems employed in Australia and New Zealand, which used comparative effectiveness information to determine whether particular drugs would be included on a formulary and the price at which the cost of the drug would be reimbursed.⁶² The purpose of HR 2356 was simply to generate accurate information on how drugs that treat particular conditions compared to one another and to disseminate this information to patients and providers. Investing time and energy in the development of this issue was not an obvious way to gain political rewards for a new lawmaker. As Allen recollects, “across the country, there was no significant constituency for funding comparative effectiveness research; few even knew what it was.”⁶³

The bill had eight Republican and seven Democratic cosponsors and was endorsed by consumer groups, the AFL-CIO, the AARP, General Motors, and the American Academy of Family Physicians.⁶⁴ Hillary Rodham Clinton introduced similar legislation in the Senate.⁶⁵ The Bush administration’s support for a Medicare prescription drug bill provided an opening to congressional advocates of CER. Allen and Emerson argued that if the government was going to spend a half a trillion dollars on new drug benefits over the next decade, both doctors and patients should have hard data on what drugs are most effective. It was hoped that this medical evidence base would counterbalance the biased information about drugs that consumers received from television advertising and also help reduce doctor’s reliance on pharmaceutical industry marketing in deciding what to prescribe. CER was thus framed as an informational tool that would help doctors push back against the immense influence of drug companies.⁶⁶

During floor debate over the Medicare prescription drug bill, Clinton proposed an amendment authorizing $75 million in research funding. The amendment was defeated 43–52 (Republicans and Independents voted 49–0 against, and Democrats voted 43–3 in favor). Senators Charles Grassley (R-IA) and Mike Enzi (R-WY) spoke against the proposal. Enzi argued that the research would “promote one-size-fits-all medicine,” and “end up as a tool for health care rationing by bureaucrats in Washington.”⁶⁷ Enzi’s statement foreshadowed the major criticisms of comparative effectiveness research that opponents would make in 2008–10. Indeed, it is striking how little the anti-CER arguments changed over the decade.⁶⁸

Yet, our interviews with congressional insiders suggest that the GOP opposition to CER at this stage was less sincere than tactical. Passing the prescription drug bill was a top legislative priority for both President Bush and Republicans in Congress, and party leaders wanted to keep the bill “clean” to make it easier to pass.⁶⁹ CER funding was an unnecessary diversion at this stage of the process, and even supporters, including Senate Majority Leader Bill Frist (R-TN), a physician by training, voted against the Clinton amendment. Once the passage of the prescription drug bill (the Medicare Modernization Act, MMA) was assured, however, the Republican leadership (most likely Frist, according to our sources) tucked a CER provision into the conference agreement.

As signed into law, Section 1013 of the MMA authorized the AHRQ to spend up to $50 million in 2004 and additional amounts in future years to conduct systematic reviews of existing evidence on the comparative effectiveness of drugs and other treatments. PhRMA opposed the CER program but was unable to block its adoption.⁷⁰ The industry was, however, able to get language in the bill prohibiting the AHRQ from using the data to mandate clinical guidelines and stating that the CMS could not use the information to withhold coverage of prescription drugs under Medicare and Medicaid.⁷¹

Allen and Emerson continued to cosponsor legislation to establish a more ambitious and permanent CER program, and expert support for a major CER initiative grew. In 2006, health economist Wilensky published a widely discussed article in Health Affairs calling for a comparative effectiveness center and laying out a range of ideas for how the entity could be organized and financed.⁷² Wilensky’s article demonstrated that CER was not a partisan idea, but rather a technocratic solution to a national problem. Joseph Antos of the conservative Heritage Foundation was on board:

We also ought to work on comparative effectiveness research. Information is a public good, and the government is in the best position to collect information. In fact, Medicare collects information on millions of medical treatments and then doesn’t use it to better understand what works and what does not. That should be fixed.⁷³

The following year, the Congressional Budget Office (CBO), under director Peter Orszag, published a major report on CER.⁷⁴ By all accounts, Orszag viewed the Dartmouth Institute’s research on regional variation in Medicare as compelling evidence that it should be possible to curb wasteful health care spending by generating better information about the costs and benefits of treatment options. He was known to carry the Dartmouth charts to congressional meetings and to preach “to anyone who would listen about the ‘evidence-based’ cure for rising medical costs.”⁷⁵ Under Orszag, the CBO began highlighting the potential for CER to constrain health care costs “without adverse health consequences.”⁷⁶

In May 2007, Allen and Emerson introduced a bill authorizing $3 billion for research in AHRQ over five years (HR 2184). Emerson called the bill a “no brainer” that deserved to receive bipartisan support.⁷⁷ In June 2007, the Subcommittee on Health of the House Ways and Means Committee, chaired by Stark, held a hearing on CER. Allen, Wilensky, and Orszag advocated for increased government investment in CER. The panel’s ranking Republican, Dave Camp of Michigan, likewise expressed the need for a much broader effort on CER. “While we have agencies like the FDA to determine if drugs and devices are safe, we have very little information that compares the actual effectiveness of drug, devices, and medical procedures.” Camp and other Republicans did raise concerns about the potential for the Medicare agency to use the information to limit access to certain treatments because they don’t work for the “average” patient. However, Allen stressed that the aim of the proposal was not to “drive decisions.”⁷⁸

While the Allen-Emerson proposal never made it out of committee, the House folded its language into a larger bill reauthorizing a children’s health insurance program that was nearing expiration.⁷⁹ The bill, nicknamed “CHAMP” (for Children’s Health and Medicare Protection Act) would have created a Center for Comparative Effectiveness under AHRQ. Funding for this center would begin with $300 million from 2008 to 2010 and $375 million per year thereafter. The measure would have created a Comparative Effectiveness Research Trust Fund, fed initially by transfers from the Medicare Trust Funds and over time by a tax on health insurance companies. The CBO projected that the CER provisions contained in CHAMP would reduce total health care spending by half a billion dollars over the 2008–12 period and by about six billion over the 2008–17 period.⁸⁰ The projected savings were modest because the CBO assumed that for larger savings the results of studies would “ultimately have to change the behavior of doctors and patients,” which, in turn, would require actions by Medicare and private insurers to incorporate comparative effectiveness information in some combination of coverage and payment policies. The CBO thought none of this would be easy.⁸¹ Still, advocates were pleased that the CBO had given CER a “score”—it could save money.⁸² But the CHAMP bill did not become law, owing to conflict within Congress, and between Congress and President Bush, over the cost of the children’s health program expansion.

DECLINING GOP SUPPORT FOR CER IN CONGRESS

While the incorporation of the Allen-Emerson proposal into the House-passed CHAMP bill might appear to have signaled growing support for CER on the Hill, the proposal would almost certainly not have passed as a standalone measure. Indeed, the number of cosponsors of the Allen-Emerson bill declined from 15 in 2003 to 7 in 2006 and 5 in 2007. This is not the pattern one expects to see when a clear perception emerges on the Hill that a policy idea’s “time has come.” By 2007, only one House Republican (Emerson) remained as a cosponsor of the legislation.

The immediate cause of the decline in the number of cosponsors was changes in the composition of Congress. Some of the Republicans who had previously sponsored CER bills retired or lost reelection, and more conservative members who were resistant to any new federal spending initiative replaced them. According to our sources, however, pharmaceutical industry opposition also contributed to the erosion of legislative support. Industry lobbyists had been taken by surprise when CER funding was tucked into the MMA and recognized it needed to become more active on the issue. While PhRMA was unwilling to oppose calls for better information on the comparative effectiveness of drugs, industry leaders wanted to broaden the scope of any CER effort so that the full range of treatments options, not just drugs, were evaluated.⁸³ They also “wanted CER to focus solely on clinical issues and not at all on cost effectiveness, and resisted any use of so-called quality-adjusted life year (QALY) indices that measure the value of therapies and devices according to expected life years.”⁸⁴

______

By late 2008, CER was at a legislative crossroads. The Washington community of health policy experts, including economists associated with both political parties, was solidly behind the idea. The CBO report connected the problem of inadequate research on the effects of treatment options to widely shared concerns about the budget and wasteful and excessive health care spending. In the Senate, Finance Committee chair Max Baucus (D-MT) and Budget Committee chairman Kent Conrad (D-ND) jointly introduced legislation (S 3408) to create a nonprofit entity to direct up to $300 million a year in federal funding for CER. The money would come from general revenues, the Medicare Trust Fund, and fees on private insurance firms.⁸⁵

This constituted progress for CER, yet the effort to win bipartisan support for the idea of better research on medical treatments had stalled on the Hill. Only a handful of members, all of them Democrats, showed a serious interest in the issue by late 2008. There was no major lobbying campaign to raise awareness of the problems of waste and uneven quality in health care, to cultivate an enlightened public opinion, or to explain to ordinary Americans how they would benefit from better research on treatment options. Advocates were playing an inside-the-beltway game, but only a handful of members and policy wonks were participating. The ground was clearly not prepared for legislative action that would bring both parties and the general public along when the Obama administration made investment in CER one of its health reform priorities after the 2008 election. Any remaining chance to build a bipartisan coalition for CER then collapsed entirely when the effort got caught up in the highly charged debates over the economic stimulus measure and Obamacare, as the next chapter shows.