While public investment in CER increased significantly during the Obama years, it is unclear whether this policy achievement will lead to durable improvements in the efficiency, quality, and cost-effectiveness of U.S. health care.1 There were ample grounds for skepticism even before Donald Trump’s surprising victory in the 2016 presidential election created new uncertainty over the future of U.S. health policy making. While PCORI spent more than two billion dollars over its first ten years, this is a small amount compared to the research effort needed to fill key evidence gaps. Enacted as part of a law (the ACA) that passed without a single Republican vote, PCORI’s capacity to influence the delivery of medical care, the professional norms of physicians, and the growth of health spending remains uncertain. In December 2019, Congress reauthorized PCORI for another decade with no major changes to its authority or mission (see Preface). As the reauthorization debate unfolded, PCORI was still in the process of building a reputation for producing important, relevant studies. Indeed, even sympathetic observers, such as health economists who are strong advocates for CER, believe that while the agency has sometimes produced good work, some of its grants have funded low-priority research that has generated little insight into the most pressing health care questions.2
What makes the failure of PCORI to generate wide esteem and growing support for the broader evidence-based medicine project all the more unfortunate is that PCORI’s leadership has been at pains to avoid antagonizing stakeholders. They have used their administrative discretion, not to push the envelope, but to stay well inside it. For example, while the ACA permits research that considers costs,3 PCORI’s executive director has publicly stated, “You can take it to the bank that PCORI will never do a cost-effectiveness analysis.”4 Yet this conflict-avoidance strategy plainly did not cure the “unhealthy politics” that surrounds efforts to bolster the scientific foundation and efficiency of American medicine.
From the outset, PCORI’s architects recognized that consolidating political support for an expanded federal role in CER would be a daunting challenge. Memories of Congress’s attacks on the Agency for Health Care Policy and Research back in the early 1990s after it had issued a report questioning the benefits of back surgery were still fresh.5 PCORI’s designers were painfully aware that any new CER entity could be similarly vulnerable. As Gail Wilensky wrote in a 2006 essay laying out the major design options for building a federal comparative effectiveness institute, “The center’s findings might anger various stakeholders affected by the findings, who, in turn, could use the political process to threaten the continued existence of the agency that produces the ‘threatening’ material.”6 It was hoped that PCORI’s design as an “independent, nonprofit, nongovernmental organization,” funded out of an earmarked trust fund, would afford the agency some measure of political insulation. In view of the politicization of health policy and the massive economic stakes in debates over coverage and reimbursement decisions, however, any organization like PCORI will inevitably encounter political risk.
To be sure, it remains possible that technological changes, including the growing use of Electronic Medical Records and computer-assisted clinical decision support tools in hospitals and doctor offices, may help close information gaps and integrate evidence into treatment decisions in the decades ahead. However, early attempts to use these techniques and mechanisms to promote evidence-based decision making have “had poor adoption rates and hence little influence on the clinical environment.”7 In addition, without government support, there will probably still be an undersupply of research on the comparative effectiveness of treatment alternatives.
Ultimately, if the EBM project is to realize its aspirational goal to improve the quality and efficiency of U.S. medical care, it is necessary but insufficient for research agencies like PCORI to endure. In the long run, patterns of medical governance must change. There will need to be stronger mechanisms in place to promote the uptake of evidence in clinical decision making. There is no shortage of plausible reform ideas, including options that would aim to change the behavior of physicians through financial incentives (e.g., value-based payments, changes in malpractice rules to create a safe harbor for evidence-based decision making) or nonfinancial educational strategies (e.g., the “academic detailing” approach to continuing medical education, in which noncommercial, evidence-based information about drugs is provided to doctors) as well as options that would aim to empower patients (e.g., through shared decision making).8 Strategies to reduce the delivery of low-value care will require sensitivity to the public anxieties and concerns about rationing and interference with the doctor-patient relationship identified in our public opinion surveys. As Mark Schlesinger and Rachel Grob observe, building robust public support for measures to reduce low-value care may require “shifting the focus from particular tests and treatments to emphasize, instead, the potential for better communication and more personalized attention if clinicians spend more time talking and less time testing.”9
No matter what approaches are tried, it will be critical for advocates to evaluate the effectiveness of particular strategies through randomized controlled trials or other rigorous methods. One notable randomized controlled trial found that a multifaceted package of educational interventions aimed at making the care of back pain more evidence-based produced a significant, but relatively small (9 percent) reduction in the rate of low-back surgery although it is unclear if these effects endured beyond the 30-month study period.10 The evaluation of the long-term impact of interventions to promote the uptake of evidence and curb overtreatment and unwarranted variation should be a priority of the research community.
Government also has a vital role to play. PCORI (or whatever entity succeeds it) must develop a reputation among key stakeholders for competence, relevance, and impact that causes policy makers to conclude that supporting EBM is in their own political interest. Improving the effectiveness, quality, and affordability of medicine in the United States will thus require more politics, not less.11
Rather than relying on procedural independence and conflict avoidance as the keys to consolidating and deepening support for EBM, we argue here that a more promising approach is to recognize that any worthy reform project must seek to curb the use of low-value treatments and therefore must seek to change how the health care system works. History suggests that the most resilient domestic policy reforms do not leave a light footprint; instead, they reconfigure the political dynamic.12 Accordingly, the task is less to insulate a CER entity from politics, although obviously researchers must be shielded from improper interference. Rather, the aim is to build a robust base of support so that future policy makers will possess an incentive to sustain and deepen the project over time.
How can countervailing pressures to improve the quality and efficiency of the U.S. health care sector be unleashed? In this concluding chapter, we draw on lessons from the literature on U.S. state building to develop strategies to load the dice in favor of the political sustainability and success of the EBM project, including postenactment coalition building, agency reputation building, and lateral network building and strategic partnerships. As a preface to this discussion, we briefly review the challenges of political sustainability that face any new agency or policy.
Why the Political Sustainability of Agencies and Policies Cannot Be Taken for Granted
Given how difficult it is to revise an existing law, it might seem that the durability of an agency or program is assured once Congress establishes it. In his 1976 book Are Government Organizations Immortal?, political scientist Herbert Kaufman argued that “government activities tend to go on indefinitely.”13 More recent empirical research shows, however, that policy constructions—including agencies, statutes, and programs—are regularly subject to modification and even elimination. According to one study, a federal program has a 1 percent chance of death every year in its first 10 years of life, after which the probability of termination slowly begins to decline.14 New policies are trial and error affairs; they do not always pan out. New policies can be killed. An example is the Medicare Catastrophic Coverage Act of 1988, which Congress terminated when seniors soured on the measure.
What is true of policies and programs is also true of bureaucracies. In 1972, Congress established the Office of Technology Assessment (OTA) to give members expert advice on a wide range of scientific issues, including the use of medical technologies. When Republicans were searching for agencies to cut to reduce the budget deficit during the mid-1990s, the OTA found itself on the chopping block. Not only did the OTA lack for powerful defenders, but some of its health technology assessment reports had antagonized organized medicine and the drug and device industries.15
In sum, the process of consolidating support for a new policy or agency— especially ones with general-interest purposes—can be more challenging than winning their enactment in the first place.16 A crucial issue is political sustainability: whether a public policy or bureau possesses the capacity to maintain its integrity and use its core principles to guide its course amid inevitable political pressure for change.17 Many factors clearly shape whether a new construction will “take.” Here we focus on four: the level of support it has at enactment, institutional design, the generation of self-reinforcing policy feedback, and the creative destructiveness of market forces. We first briefly describe these factors and then evaluate how the Obama administration’s CER project stacks up.
Commitment at enactment. Some new policies or agencies enjoy overwhelming support at inception; others pass narrowly in the teeth of intense opposition. There is risk in building “minimum winning coalitions.” Political scientists Forrest Maltzman and Charles Shipan have shown that the greater the roll call opposition when a law is passed, the more likely the law is to be amended by a future Congress.18 A key question is whether partisanship exacerbates the problem of divisive enactment. There are good reasons to believe it might. As David Mayhew argues, a cross party opposition to a policy might fade, but “[a] party that loses on a congressional issue and stays angry may have an incentive to keep the conflict going.”19
Political institutions. Policies (and agencies) become more durable, all else being equal, when they destroy the structural bases of support of their opponents, shift decision-making control to venues in which intended beneficiaries are advantaged, and significantly alter governing capacities, such as by enlarging administrative staffs who possess the technical expertise to implement policies in an effective manner.
Policy feedback. Public policies are not merely the products of politics, but also causes. Policies shape the material resources, civic engagement, and incentives of voters and interest groups and, in so doing, affect governing possibilities going forward.20 Policies may generate increasing returns and path dependence, raising the cost of subsequent policy change. Constituency groups may adapt to existing policies, “get stuck in their adaptations,” and then engage in politics to protect their adaptations.21 For example, the AMA opposed Medicare’s creation in 1965, yet once the program was up and running, physicians began aligning their practice plans to Medicare’s coverage policies and fee schedules. Over time, Medicare’s enormous resource flows even began to shape decisions about what areas of medicine doctors choose to specialize in.22 Policy feedbacks may also influence the behavior of interest groups. When Social Security was created in 1935, for example, the American Association of Retired Persons did not exist. Rather than senior mass-membership groups pushing for the enactment of Social Security, Social Security’s growth transformed seniors into the most active and best organized participatory age group in the country, providing them resources and enhancing their sense of political efficacy.23 Finally, durable policy reforms also recast institutions and upset existing power monopolies; they establish new norms and expectations; they eliminate or reduce the organizational cohesion of coalitions opposed to reform; and most importantly they create new vested interests and stimulate investments whose value is tied to the reforms being maintained. When reforms accomplish these things, officeholders find it impossible or unattractive to reverse course.24
Yet the generation of self-reinforcing policy feedback is not automatic; many public policies have negative or nonexistent feedback effects, increasing their vulnerability to downstream erosion.25 Feedback effects may be weak when policies offer broadly dispersed benefits to the public as a whole.26 The low per capita stakes give ordinary citizens little incentive to mobilize if the reform is challenged by groups who would profit from the measure’s unraveling.
The Tax Reform Act of 1986 illustrates how general-interest reforms can fail to remake politics. A landmark reform, the TRA closed many special interest tax loopholes over the opposition of the oil industry, realtors, and other formidable clientele groups. The act had bipartisan support in Congress and the strong backing of President Ronald Reagan. Nevertheless, no interest group emerged to defend the reform against the (inevitable) post-enactment efforts of politicians and lobbyists to reverse course and reclaim the tax code as a vehicle for particularistic favor provision. The per capita benefits of base broadening were too meager to activate civic participation. As a result of this imbalance of political forces, the reform crumbled. Congress began creating new tax breaks almost as soon as the ink on the 1986 law was dry, and many of the achievements of the law have been lost.27 In sum, even if reformers manage to triumph during the initial enactment battle, they haven’t really won. Concentrated groups—such as firms or professional societies—who feel threatened by a reform are unlikely to throw in the towel just because a vote or two in Congress did not go their way; indeed, they likely will continue to enjoy organizational advantages, allowing them to reassert their influence during the postenactment phase. A reform statute is only words on paper; the real struggle to reconfigure politics has only just begun.28
Creative destructiveness of market forces. It is rarely the case that a reform victory will cause every politician who voted against the reform to lose in the next election; they may well return to fight another day. Although political losers can hang around as long as their constituents are willing to tolerate their losing positions, the market actors who lose market share in a postreform environment sometimes disappear. The surviving firms are the ones who adapt their business models to the new policy context.
Airline deregulation illustrates this dynamic. Beginning in the 1930s, air carriers and their unions benefited from a highly inefficient regulatory system that determined rates and routes and blocked new entry into the market. In 1978, a reform coalition, animated by public concerns about price inflation, was able to overcome the intense opposition of these groups and enact sweeping reform. Over time, the new market system drove some legacy carriers into bankruptcy. New discount carriers appeared, and business actors were compelled to invest heavily in route systems, schedule tools, and other forms of organization compatible with the new market environment. Supporting industries and suppliers (e.g., aircraft equipment suppliers, hotels, rental cars, restaurants, and corporate office parks) began to grow up around the new system. Each had a vested interest in the reform’s maintenance. To be sure, there were still groups that would have preferred a heavier government role in the airline sector. However, the emergence of new carriers postreform greatly lowered the political and economic cohesion of the airline industry. While each airline would like to receive government protections, the airlines increasingly check and balance each other, since more rent for one would reduce the profit of the others. In sum, the politics of the airline sector has been thoroughly reconfigured, making it all but impossible for policy makers to reverse course.29
Is the Obama Administration’s CER Project Sustainable?
Based on this political sustainability recipe, the Obama administration’s CER project appears to lack key ingredients. First, the recovery act and the ACA both passed on narrow, party-line votes. The GOP has continued to argue that CER must not become a vehicle for “rationing” and “cookiecutter medicine.”30 In contrast to the willingness of the GOP to acquiesce to Medicare payments to physicians for counseling patients about their end of life options31—another policy that was caught up in the “death panels” debate—Republicans (and many Democrats) have continued to oppose allowing Medicare to use cost-effectiveness in its coverage decisions. Many lawmakers also supported the repeal of the Independent Payment Advisory Board.32
Second, and importantly, PCORI failed to generate new supportive interests or alter coalitional alignments. Ann C. Keller, Robin Flagg, Justin Keller, and Suhasini Ravi at UC Berkeley School of Public Health performed an insightful study of PCORI’s political strengths and vulnerabilities. Based on elite interviews, content analysis of public comments, congressional hearings, and media and Internet content about PCORI (including trade association press), the authors found that PCORI’s leadership has successfully mobilized patient advocacy groups and researchers in support of the agency’s mission. However, patient advocacy groups and researchers have tended to mobilize within rather than across disease categories, limiting their organizational cohesion and impact in broader debates over the future of CER and EBM. Few new supporters of CER have emerged.33 Moreover, the organizational bases of support of the reform’s skeptics were not destroyed; there remained plenty of opposition to PCORI from medical industry groups and members of Congress. The authors found that legislators who discussed PCORI during congressional hearings were far more likely to view the entity as having a negative regulatory impact than a positive, market-correcting one. In particular, lawmakers expressed concerns that PCORI was harming taxpayers, that the agency was not truly independent from HHS, that CER would impose a regulatory burden on the drug and device industry, and that PCORI would ration care.34
Third, the Obama administration’s investment in CER has not brought about shifts in institutional arrangements. The project’s major structural innovation was PCORI’s establishment as an independent, nongovernmental entity, rather than as an office within HHS. But this design has not shielded PCORI from criticism. At the same time, efforts to require Medicare administrators to use clinical and cost-effectiveness information in coverage decisions were “rebuffed not only by the legislative staffs but by the White House healthcare policy people.”35 In sum, the establishment of PCORI did not bring about durable shifts in wider patterns of medical governance. The creation of PCORI failed to rearrange surrounding institutional authority, as it impinged only lightly on existing federal health programs and bureaucratic activities.36
When new policy frameworks stick, they do more than resist the pressure of actors opposed to the shifts. They induce private firms to alter their business models to comply with the new rules of the game; firms unable to do so are more likely to go broke or merge with firms more able to comply, making the new policy regime self-reinforcing.37
It is unclear whether the new emphasis on CER is stimulating hard-to-reverse financial investments in the medical products industry. To the extent that pharmaceutical firms and device manufacturers believe that their future profitability will hinge on the capacity to demonstrate that their products are superior to alternatives on comparative clinical and/or cost-effectiveness grounds, they should be investing heavily in the development and marketing of such products. As one health industry analyst writes,
Pharmaceutical manufacturers must recognize that the drugs in the laboratory today are those that will be commercialized in 2020. They will very likely be launched into a market that demands evidence of economic and clinical value as the price of entry. Guidelines based on cost effectiveness developed by payers or providers will increasingly determine market access. Consideration of economic and clinical value must be integrated into the entire product development and commercialization process …38
However, this analyst goes on to report that the changes he recommends are “not happening quickly enough.”39 Why not? On the one hand, “private payers in the US are aggressively moving to utilize CER in their decisions regarding pricing and reimbursement.”40 For example, a large health care insurer, WellPoint, has “released its own standardized CER guidelines for use in its evaluations of drug coverage.”41 Another large insurer, United Healthcare, has suggested that CER will foster the broader use of copay structures that discourage patients from seeking higher-cost treatments that offer no real benefit over use of lower-cost drugs.42 On the other hand, the supply of CER studies remains small, and the traditional regulatory procedures for approving new products have not been revised to make a demonstration of superior clinical or cost-effectiveness a priority. For example,
the Food Drug and Cosmetic Act of 1938, as amended in 1962 and subsequently, does not require assessment of comparative effectiveness, and the legislative history in 1962 made it very clear that there was no relative effectiveness requirement. A new drug does not have to be better than, or even as good as, existing treatment.43
For a variety of reasons, including methodological concerns and turfprotection, the FDA’s senior drug review officials have expressed a measure of “skepticism of CER in recent years.”44 While some movement toward greater use of CER can be seen in the decisions of both Medicaid and private insurance company administrators, Medicare’s coverage and payment policies remain largely unchanged.
For now, it appears that drug companies and the medical device industry do not regard a focus on comparative effectiveness as a central reality around which to base major investment decisions. There is little reason to believe that the evolving industrial organization of the medical products industry, or the vested interests of individual firms, will serve as the guarantor of the overall EBM project’s future impact and sustainability.45
Toward More and Better Medical Politics
In “The Political Transformation of American Medicine,” Peter Swenson shows that between roughly 1870 and 1910, American medicine experienced a progressive phase in which professionals and lay activists joined forces to improve the quality, economy, and equality of health care. Their major goals were three: (1) reforming the drug industry and the marketing of its products; (2) creating a cabinet-level national health department; and (3) improving the quality of the nation’s medical schools. The reformers’ victories were only partial, and they made important enemies—including parts of the medical profession. Ultimately, the progressive medical reform movement got too far ahead of the rank and file, especially on issues like compulsory national health insurance. Their achievements generated a backlash that helped give rise to the medical establishment’s turn toward organizational conservatism and self-protection during the 1920s. But the turn of the century was nonetheless an era of progress in medical education, hospitals, and other areas.
A key lesson from this earlier period is that improving the quality, economy, and rationality of medical care requires reformers to embrace politics.46 Progress cannot be made without the leadership and buy-in of the medical profession, but it also cannot be made without building support among policy makers, activists, and educated voters.
Given the context of contemporary American politics, what can be done to promote medical progress and the political sustainability of CER and the overall EBM project? We offer three ideas for consolidating and deepening support—postenactment coalition building, building a reputation for competence and efficacy, and lateral network building and strategic partnerships.
POSTENACTMENT COALITION BUILDING AMONG ELECTED OFFICIALS
The first step is to recognize that the task of building support for a new line of policy making does not end when a policy is enacted; it has only just begun.
One of the ways that policies can unravel after enactment is if bureaucracies charged with implementing them shirk or otherwise fail to serve the goals of the policy’s designers. An influential line of argument developed by McNollgast—the nom de plume of social scientists Mathew McCubbins, Roger Noll, and Barry Weingast47—suggests a solution to this problem: leaders who won the enactment battle can “stack the deck” to establish a context that endures long after the enacting coalition has frayed, ensuring that the coalitions that generated the program’s adoption will also hold sway during the postenactment phase. Deck stacking can be accomplished structurally by insulating agencies from executive control and procedurally by crafting administrative rules (i.e., rules concerning what interest groups can intervene in agency decision making). These strategies can work. For example, research shows that bureaus given agency status are more durable than are bureaus subject to direct presidential control.48
Yet there are limits to deck stacking as a strategy for sustaining policies over time. First, the deck-stacking perspective overstates the ability of “winners” to impose their preferences on “losers,” who often have the opportunity to embed some of their goals into policy designs as well. For example, opponents of efforts to use CER as a tool for cost control won language in the ACA constraining how PCORI could use QALYs in its decision making. As Terry Moe has argued, it is not only a program’s strongest supporters who can seek to embed their structural preferences into agency design, but the organization’s skeptics as well.49
Second, deck stacking is more effective as a strategy for preventing undesired policy outcomes than it is as a strategy for ensuring good government performance. As Daniel P. Carpenter agues, “designing constraints is easier than designing capacities.”50 It is one thing to prevent an agency from taking actions that an enacting coalition would have opposed, quite another to ensure that an agency has the ability to solve problems and benefit from its expertise. History suggests that the greater risk is not that a CER agency will make decisions that its original designers would not have supported, but that it will lack the capacity to catalyze reform of medical governance.
Third, the deck-stacking hypothesis is static. It fails to recognize that contexts and coalitions mutate over time, and that design-stage choices can be nullified by decisions made during the implementation phase. As William N. Eskridge and John Ferejohn argue, the deck-stacking model treats the preferences of actors as exogenous, but such preferences are also “a product of deliberation and feedback, not anterior to it.”51 From the standpoint of political sustainability, the issue is not only whether the goals of the original enacting coalition will persist, but also whether a policy causes new supportive coalitions to emerge during the postenactment phase.
But perhaps the most important limitation of the deck-stacking perspective is that it assumes the enacting coalition provided an adequate base of support. The seminal articles by McNollgast on structural politics were written during the 1980s and early 1990s. This was an era in which Democrats had a seemingly permanent majority in Congress. Many laws had bipartisan support, at least on final passage. What members of Congress feared is that their legislative creations would be undermined after enactment by presidents and bureaucrats, who often responded to different constituencies and incentives.52 So the key challenge was to lock in the political bargain and prevent its downstream erosion.53
By contrast, as we have seen, lawmaking today takes place in a highly competitive, polarized environment in which bills often pass by narrow, temporary partisan majorities. The two parties compete at relative parity, meaning that a shift in party control could be only two years away.54 These changes can have significant effects on the politics of policy development. According to a careful statistical study of more than 2,000 federal domestic programs established between 1971 and 2003, changes in the partisan composition of Congresses have a strong influence on program durability. Program life spans are shortened when the Congress that inherits a program is different in partisan terms from the Congress that created it.55 The political sustainability challenge today is less to preserve a policy accomplishment from being undone than it is to generate a bipartisan base of support among officeholders in the first place. Only by broadening the enacting coalition can a policy be sustained in an era of polarization and electoral uncertainty. Long-range thinking remains crucial, but some of the entrenchment strategies employed in the past may no longer be quite so effective.
What can be done? Although there are no silver bullets, we suggest several ideas. First, high-level appointments to public interest purpose agencies like PCORI should be made with an eye not only to technical competence, integrity, and experience—which obviously remain important—but to the imperatives of postenactment coalition building. Strategic leaders should aim to appoint well-qualified officials who can broaden, rather than merely preserve, the agency’s base of support and signal that the policies the officials will be responsible for carrying out have broad public purposes, not merely partisan goals. Richard Kronick, the distinguished former head of the highly valuable yet politically vulnerable health services research agency AHRQ, demonstrates the limits of a conventional appointment strategy. Kronick was an extraordinary well qualified expert with a stellar reputation. Yet Kronick had worked in HHS on implementation of the ACA and before that was a health policy maker in Massachusetts. As one author notes,
That background did not bode well for bonding with the GOP majority in the House, and today Republicans control the Senate, too. Republican opposition to ‘Obamacare’ has often been frenzied, fanatical and unfair, but the job of an agency head, particularly at an embattled one, is to get along with those who control your budget.56
In 2015, House Republicans threatened to zero out AHRQ’s budget.57 In the end, the agency survived with a moderate size (about 8 percent) budget cut—but not without having to wage an exhausting campaign to avoid termination.
To be sure, presidents have long appointed a handful of cabinet officials from the other party. This personnel strategy is often used to burnish a president’s reputation as “president of the whole country” rather than designed to widen support for important yet contested policy goals. Typically, presidents appoint officials from the other party to agencies that are already above the partisan fray. For example, George W. Bush appointed Democratic congressman Norman Mineta to be secretary of transportation, and Barack Obama appointed Republicans Ray LaHood and Robert Gates to be his secretary of transportation and secretary of defense, respectively. But politicians support transportation projects primarily for constituent or geographical representation reasons, and defense policy (while not uncontroversial) does not directly implicate partisan debates over the role of government in the economy.
In today’s environment, in which general-interest reform projects like EBM can become politicized, it would be worth considering more strategic use of cross aisle appointments to build support among opponents. To be sure, there are limits to what even the most brilliant appointment can accomplish. The forces that generate partisan conflict over policies are rooted in both electoral incentives and fundamental value conflicts. Nonetheless, leaders make a difference.
With the benefit of hindsight, it might have been a more prudent strategy to appoint someone who would have signaled a broadening of the political base, someone like Gail Wilensky, to be the inaugural PCORI director. (We hasten to add that we do not know if Wilensky was considered or if she would have accepted if asked; furthermore, it should be noted the power to appoint the executive director belonged to the board of governors of PCORI rather than to the president or Congress.) In offering this idea, we do not intend to denigrate the person who was appointed to this role, Joe Selby, who had an ideal background if the sole requirements for the job were technical expertise, integrity, and experience. (Selby is a family physician, clinical epidemiologist, and health services researcher who was with Kaiser Permanente for three decades.) But his appointment did little to change the political image of PCORI or forge linkages to Republicans or conservative think tanks. Our claim is not that the appointment of a prominent Washington health policy veteran like Wilensky would have completely insulated PCORI from attack, merely that it would have signaled that the entity’s mission draws support from health experts on both sides of the partisan aisle. Strategic appointments like this might help distance the EBM project from continuing debates over the ACA and give Republican officeholders the space to reconsider their positions on CER over time.
BUILDING A REPUTATION FOR COMPETENCE AND EFFICACY
No public organization can be effective if it lacks the capacity to chart its own course and pursue its core mission without constant harping or second-guessing by politicians. Americans, however, have traditionally viewed bureaucratic power with skepticism, fearful that unelected administrators will shirk, trample on the liberties of citizens, or pursue illegitimate goals unsanctioned by their democratically elected principals. Hence, federal agencies are often subjected to tightly drawn rules and procedural constraints designed to keep agencies on a short leash.58
Some federal agencies, however, have managed to gain significant power and the capacity to leverage their authority to exercise vast influence not only over the activities of government, but on private sector actors as well. Key to gaining this organizational power is reputation. When a bureaucracy cultivates a reputation for effectively providing unique services and promoting the public welfare, it can gain the admiration of the general public and the respect of elite audiences, such as scientific and professional organizations. It then becomes costly for politicians to ignore or resist the agencies’ ideas, further increasing the agencies’ autonomy. The most powerful agencies exert their influence less by imposing their will on recalcitrant actors than by defining the “basic terms of debate, essential concepts of thought, learning and activity.”59
In Reputation and Power, Daniel Carpenter illustrates this argument through a detailed historical study of the Food and Drug Administration.60 While the FDA has had a smaller budget than many other federal agencies, it nonetheless became one of the most powerful regulatory agencies in the world during the late twentieth century.61 The FDA exercises gatekeeping authority over the U.S. pharmaceutical marketplace. It has the power to limit advertising and product claims, to determine what cures are available or unavailable to patients, and to define the very scientific concepts used by experts in the medical field. To be sure, the FDA’s power has limits. Under pressure from doctors’ groups, the FDA has largely steered clear of the regulation of medical practice. It has not cracked down on “off-label” usage of drugs and has generally eschewed comparative efficacy judgments, which threaten the authority of the medical profession over standards of care.
Despite these limitations, the FDA’s perceived performance during the 1950s and 1960s offers vital lessons for EBM advocates who would like a CER entity that builds reputation among both the public and elites for promoting sound science, protecting patients against the risks of overtreatment, and policing the medical profession when it fails to police itself. During the postwar era, the FDA developed increasingly rigorous scientific standards for evaluating the effectiveness of drugs—standards that went well beyond its original statutory mandate to test drugs only for safety. What permitted the FDA to build its capacity was its growing reputation as a competent and vigilant agency, symbolized by its decision to reject the drug thalidomide (which was intended to prevent morning sickness in pregnant women) in 1962. When it was discovered that thalidomide (which was available in Europe) caused severe birth defects, the agency’s image as an essential guardian of public welfare was further enhanced. The episode received extensive media coverage and made a celebrity out of Frances Kelsey (the FDA regulator who stood her ground despite pressure from both industry and some actors within the agency to approve the drug), giving advocates an opportunity to ratify the FDA’s expanded regulatory powers though the Kefauver-Harris Amendment. As political scientist Steven Teles observes, while critics have chipped away at the FDA’s image in recent years, the agency’s “powerful reputation gave it a great deal of insulation when an antiregulatory chill took hold in the 1980s.”62
PCORI has so far failed to develop a reputation for doing important work. In January 2014, the Center for American Progress, a liberal think tank closely tied to the Obama administration, issued a harsh assessment of PCORI’s early record.63 The report pointed out that most of the studies funded by PCORI were broad and focused on methodology, education initiatives, and communication tools, rather than on questions directly relevant to clinical practice. During the time frame of the study, PCORI failed to issue a single CER study of medical devices, launched only a few CER studies of drugs, and produced few studies that focused on the priority areas identified by the Institute of Medicine.64 PCORI’s work did accelerate over the ensuing years. In 2016, the Center for American Progress issued a revised evaluation. It found that PCORI had made some progress, allocating 58 percent of its grant funding to CER.65 The agency commissioned new studies to address a range of important topics, including the “use of radiation therapy for breast cancer, effective treatments for bipolar disorder, and lifestyle interventions versus drug therapy for diabetic patients.”66
Yet PCORI continued to fund some projects that experts viewed to be as not of central clinical significance (such as a study of hoarding behavior)67 and remained virtually unknown to the general public. There has been almost no mainstream media coverage of PCORI’s work. Even more damaging, key stakeholders did not perceive CER as advancing rapidly. In a 2015 survey of 122 medical sector players—insurers/health plans, government officials, employers, researchers/thought leaders, business coalitions and associations, 81 percent of respondents said that over the past year, CER had “no effect” on health care decision making or led to a “slight” improvement,” while 19 percent said it had led to a “moderate” or “substantial” improvement.68 Wilensky’s assessment is particularly damning: “PCORI seems to have become almost invisible. Maybe they think that’s the best way to stay under the political radar screen,” she said, adding that the institute has yet to “offer much value.”69
PCORI’s leadership faced many challenges. They had to get a new organization up and running, establish procedures for awarding grants, set priorities, and disseminate hundreds of millions of dollars in research support. The “patient-centered” aspect of PCORI’s mission—added at the behest of the medical products industry—required PCORI to spend considerable time developing methods to include patients in the research process and educating researchers about the need to incorporate patient concerns into their studies.70 Yet the most critical task facing PCORI’s leaders was not technical, but political: demonstrating its value to members of Congress before its funding expired in 2019. While PCORI developed a strategic plan focused on “funding and conducting highly relevant research that is likely to change practice and improve patient outcomes,”71 the agency struggled to find an effective strategy for burnishing its reputation and winning reauthorization. In 2012, Dr. Harold Sox (who was later hired as a senior adviser to PCORI) implored the agency to implement “a strategy to make the largest possible impact before its day of reckoning.”72 The agency emphasized that it is “funding studies to answer questions about common, serious conditions like cardiovascular disease, cancer, and mental illness, which affect millions of Americans.”73 That is as it should be, but we offer three more specific suggestions.
First, it should be recognized that the biggest single threat to the political sustainability of CER is its association with “rationing.” To overcome this association, PCORI (or future CER entities) should actively seek to distinguish the EBM project—which at its core is about better science, better information, and higher quality care—from service denial. One way to do this is to support and publicize research on the comparative effectiveness of worthy treatments that are being underused. Headlines associating the agency with research demonstrating the need for more medicine will help build the agency’s reputation as a scientific body, rather than a cost-cutting board.
The United States can here learn from the U.K. experience. When the National Institute for Health and Care Excellence, which makes recommendations for covering medical interventions and treatments on the basis of cost-effectiveness analysis, was established in 1999, one of the main rationales was to end the “postcode” lottery in which the availability of services and treatments under the National Health Services depended on where a British citizen happened to live. (In the U.K. single-payer system, such geographical variation is considered to be a problem on equity grounds, not just efficiency grounds.) NICE’s conclusions (including the denial of some cancer drugs available in the United States) generated controversy in the U.K., but it helped that NICE was created during a period when the U.K. was seeking to increase access to medical services and that following NICE’s guidance usually meant spending more, rather than less.74
Another thought is that strategic leaders should link CER not only to geographic variation, but also to the investigation of the heterogeneity of treatment effects across groups, such as children, women, and the elderly. This approach could quell the refrain that CER creates insights only about population averages and does not help guide MDs in treating specific patients. If PCORI publicized findings that helped target specific groups, it could blunt this criticism.
Finally, PCORI needs to put “points on the board” to make the benefits of evidence-based medicine far more salient to ordinary Americans. That means studying treatments that affect millions of Americans and producing research that is relevant and helpful to physicians, that is understandable to patients and policy makers, and that warrants extensive (and favorable) media coverage.
LATERAL NETWORK BUILDING, FEDERATED DESIGN, AND STRATEGIC PARTNERSHIPS
Because public administration is inescapably political, all public agencies require advocates. One of the most common mistakes made by institutional designers is to believe that the only political support that matters resides “inside the beltway.” In America’s fragmented governance system, however, the support that counts most is often locally rooted.75
Consider the Veterans Administration. Despite recent scandals, a deserved reputation for bureaucratic sluggishness and incompetence in some quarters, and calls for organizational reform, the VA has survived over many decades and even expanded its clientele base in recent years. Key to the VA’s durability has been its federated design. The VA has a Washington-based bureaucracy, but it also entered into strategic partnerships with powerful private organizations, including academic medical centers at prestigious universities. The VA thus has a major presence in a large number of congressional districts around the country. These lateral network ties and public-private partnerships give the VA a geographic and distributive foundation. In addition, they allow the VA to “borrow” the reputation of other revered institutions.76 A broadly analogous story of networked capacity building can be told about the success of the Department of Agriculture in bringing policy expertise to bear on its decisions during the New Deal through its close ties to state land-grant colleges and experiment stations.77
PCORI does not perform medical studies in-house; it distributes grants to researchers across the country. But the entity’s visibility in local communities—and before local elite audiences—is much weaker than it could be. Imagine how much more robust the agency might be if instead of one Washington-based organization, Congress had created a dozen (smaller) comparative effectiveness institutes around the country, based at academic medical centers or major research hospitals. Perhaps a Center for Comparative Effectiveness and Cardiology could have been established at the University of Alabama, a Center for Comparative Effectiveness for Breast Cancer at Duke, and so forth. Each time each locally rooted center issued a study, a press conference would be held, attended by the state’s congressional delegation.
Most importantly, the entity should develop close ties to leading doctors and local medical societies in communities around the nation, who could work with national organizations such as “Choosing Wisely” and consumer groups to help patients understand evidence-based medicine, the meaning of research findings for patient decision making, the significance of regional variation, and the problems of over- and underutilization. While the American medical profession clearly must lead at the national level as well, our physician survey results (chapter 4) suggest that rank-and-file doctors have strong views about the role of their medical societies. Until most doctors become well-informed about the waste and inefficiency of U.S. health care and politically engaged on these issues, little is likely to change. Indeed, physicians often appear to take the opposing stance and rally against efforts to promote the efficient use of health care dollars. For example, a recent proposal to test new payment models for Medicare Part B drugs, which would have reduced incentives for doctors to prescribe expensive drugs that generate higher reimbursement instead of cheaper drugs that are just as effective, generated intense opposition from medical societies.78 Yet, local doctors enjoy the individual trust of their patients and could be an important counterweight to national policy makers who oppose EBM reform initiatives.79
Finally, PCORI could develop strategic partnerships with payers and providers participating in accountable care organizations (ACOs) around the country to implement PCORI’s findings into their clinical management decisions. By collaborating with ACOs, PCORI can
ensure that a subset of clinical decision makers will translate their recommendations to clinical practice. PCORI can then follow this subset of recruited ACOs, track the clinical outcomes before and after they begin following their recommendations, and compare outcomes with nonparticipating ACOs and providers.80
To be sure, a decentralized strategy is not without risks. There are more opportunities for agency heads to lose control over managing controversies, as in the recent case of VA hospitals.81 In the American federated system, however, there are huge advantages to locally rooted policy.
______
The progressive desire to root out waste, inefficiency, and bad science in the U.S. health care system is well-founded, but the project needs deeper grounding in America’s local communities and political culture, greater bipartisan support, and much stronger and more visible public leadership from the medical profession. We have shown that powerful political forces will need to be harnessed to improve the quality, rationality, and efficiency of American medicine. The work will be difficult, but the first step is to recognize that this is primarily a political challenge, not a technocratic one. Forging new linkages between expertise, power, and democratic accountability is vital to the durable reform of medical governance—and crucial to American government’s performance as an effective problem-solving institution in the twenty-first century.