Medicare's Requirement for Research Participation as a Condition of Coverage: Title and subTitle BreakIs It Ethical?

Policy makers agree that the US health care system needs new and better ways to ascertain the risks, benefits, and costs of new medical technologies as they are introduced into practice.^{1 - 3} One central policy mechanism to develop such evidence is to link insurance coverage with a requirement that patients participate in relevant research, including registries and clinical trials. Medicare has introduced its own version of this approach, termed coverage with evidence development (CED), and has used it in several high-profile coverage decisions.⁴ But the policy has been criticized as coercive and unfair.⁵ The future of CED at Medicare and of similar approaches among private insurers may well depend on whether such approaches are viewed as ethical.

In this article, we argue that CED should not be viewed as coercive because Medicare patients are not entitled to new technologies that would not receive coverage in the absence of CED. We discuss the ethical role of cost concerns in selecting new technologies to receive coverage through CED; the importance of embedding CED in a coverage decision-making process that sets explicit standards for evidentiary review; and whether CED can play an important, ethically legitimate role in balancing needs for better evidence of effectiveness with the needs and desires of patients for rapid access to new innovations.

On November 23, 2005, by coincidence the day before millions of Americans would sit down to Thanksgiving dinner, the Centers for Medicare and Medicaid Services (CMS) released its draft National Coverage Decision memorandum for bariatric surgery to treat morbid obesity.⁶ In its draft memo, the CMS stated that the scientific evidence appeared sufficient to support coverage for several procedures currently being used to treat morbid obesity among Medicare beneficiaries younger than 65 years. However, due to evidence of higher short-term mortality rates among older patients, the CMS indicated that it would deny coverage for bariatric surgery for all patients aged 65 years or older, the vast majority of Medicare beneficiaries. But the draft memo suggested that another option was being considered. The CMS made clear that it wished to receive public comment on whether coverage should be extended to patients older than 65 years—but only for patients participating in a registry or clinical trial.

Consideration of the CED coverage option in the bariatric surgery decision stimulated a long-simmering debate within the CMS and across other components of the US Department of Health and Human Services. Was CED research? If so, was it ethical for the CMS to require participation in research to receive insurance coverage? Was this approach good for patients and the public—or was it unfair and coercive as some noted ethicists were already claiming?⁵

This concern about the ethics of CED poses a fundamental challenge to the premise of many fledgling efforts, both in the United States and internationally, that seek to balance the need for better information on the risks and benefits of medical innovations with the increasing demand for rapid access to these services.^{7 - 8} Coverage with evidence development has been prominent among these efforts even though it has been explicitly used in only 5 national coverage decisions, including coverage for positron-emission tomographic scans in the management of Alzheimer disease, the off-label use of certain chemotherapeutic agents for colorectal cancer, and most prominently the 2003 decision expanding coverage of implantable cardiodefibrillators.⁹ The CMS designed CED as a coverage mechanism that could be used when promising evidence suggested that patients might benefit from a new technology, but additional evidence was needed to determine with confidence that the technology met Medicare's statutory standard for coverage—that no payment may be made “for any expenses incurred for items and services. . . . [that are] not reasonable and necessary for the diagnosis or treatment of illness or injury. . . . ”¹⁰

Before CED, therefore, evidence that did not quite reach the CMS interpretation of “reasonable and necessary” routinely led to a denial of coverage for a new technology. Patients who wanted such technologies would either have to pay out of pocket or enroll in a clinical trial if one was available. By contrast, CED was designed so that patients would gain earlier access to promising but as yet unproven technologies, industry would receive payment for innovations that would not have been covered otherwise, and all health care decision makers would benefit from the generation of better evidence on the true risks, benefits, and costs.¹

Although CED is clearly well-intended, it raises several important ethical questions. First, do any or all forms of CED constitute research, and if so, what patient protections and regulatory oversight should be required? Some CED decisions have required participation in a clinical trial, but most so far have linked coverage to entry into a clinical registry. Should participation in all registries be considered research? In a registry, all eligible patients receive the treatment, and in most registries patients face minimal additional research burdens while often benefiting from the information gained. According to one well-known definition, many such registries would be classified as quality improvement and not research.¹¹ However, there is wide variation in attitudes on the boundary between quality improvement and research activities.¹²

Regulatory implications give this boundary great practical importance. If a data-gathering process is considered research, federal regulations require that a variety of procedures must be followed to protect the participants involved.¹³ For example, active informed consent from every patient in a registry has been required by some institutional review boards but not by other organizations, and there is active debate and disagreement among experts over whether registries and other forms of health care services research require full, partial, or no informed consent.^{14 - 15}

The questions regarding which elements of CED are research and which are not have plagued the policy discussions at the CMS, but it is acknowledged that some CED decisions, such as the coverage of off-label uses of chemotherapeutic agents for colorectal cancer, have linked coverage to studies that all would acknowledge are research. It is useful, therefore, to discuss the 2 major issues that have divided opinion on the ethics of CED—coercion and fairness.

Although many articles have discussed coercion to participate in research, no published articles have addressed coercion in relation to insurance coverage. Coercion occurs when a threat of some harm compels a person to act in a manner that he or she would not otherwise choose.^{16 - 17} An example is that of a kidnapper demanding ransom. The kidnapped victim's family may be coerced into giving up money to avoid the threatened harm to their loved one.

It might be argued that CED is coercive because some patients might feel that to get the treatment they “need,” they have no other option but to participate in research, although they would not choose to do so if insurance coverage did not require that participation. But feeling compelled to participate in research does not constitute coercion. Some patients with terminal cancer, because of their dire prognosis, may feel compelled to enter phase I research studies, but making such a choice, even though it is done under difficult circumstances and with limited options, does not constitute coercion.¹⁸ Coercion requires a person, organization, or policy to threaten a specific “harm.”

It could be argued, however, that CED threatens patients with a specific harm—the withholding of unrestricted insurance coverage. If obtaining coverage for a new technology without any requirement for research participation is the “best” option for patients, any policy short of unfettered access might represent a harm and be coercive. However, this argument is unsound. The fact that patients might feel entitled to the “best” option does not mean that they are entitled. Withholding coverage for a technology counts as a harm only if patients are morally or legally entitled to unrestricted coverage for that technology.

Medicare explicitly conveys no entitlement to insurance coverage for all new technologies, only to those technologies judged by the CMS to be “reasonable and necessary.” Whether the Medicare legislation implies a moral entitlement to coverage outside of these constraints is a separate argument. Following the legislative language, coverage has always been routinely denied for technologies that fail to meet Medicare's interpretation of its evidentiary standards. Without CED, coverage denial would thus be the common fate of these technologies. Because CED does not propose to deny coverage to a technology to which patients are entitled, it does not threaten them with any harm. Since there is no threat of harm, concerns that CED is coercive are mistaken or misplaced. To be sure, patients seeking access to the new technology covered by insurance only under CED have an inducement to participate in research; however, this is no different than seeking access to an experimental treatment only available in clinical trials. Patients contemplating their treatment options under CED may face tough choices and may feel compelled to participate in research, but they do not do so under a cloud of coercion.

If CED is not coercive, may it yet be unfair? The registries or trials in which patients would have to participate under CED may not be readily available in all geographic regions, thereby creating or worsening unequal access to new technologies. For example, Medicare recipients in rural areas may have less access to a new technology than urban residents just because no approved research program exists near where they live. Inequality in access to research programs is a regrettable practical reality but does not constitute an injustice. Unequal access to CED-linked research programs would present a significant problem of fairness only if patients were entitled to the technology. But without CED there would be no coverage at all, so inconsistent access to the technology after CED, while not ideal, is not unethical. Unequal access is not remedied by denying opportunities for all. The CMS and sponsors of similar research-coverage programs should seek links to registries and clinical trials that offer as wide availability as possible to prospective participants, but perfectly equal access is not mandatory for an ethical policy and will rarely be achievable.

To have scientific and ethical legitimacy, a coverage policy must be grounded in a clear hierarchy of standards of evidence.¹⁹ Coverage with evidence development is based on identifying technologies for which the supporting evidence is not strong enough to warrant an entitlement to unlimited coverage. In fact, the ethical justification of CED, as the CMS has designed and described it, requires the evidence on a new technology to be appraised as falling somewhere in the middle of an evidentiary hierarchy.²⁰ There must be enough evidence to suggest a potential benefit for Medicare beneficiaries, but technologies being considered for CED must not meet the usual standard for the strength of evidence at which they would automatically be considered “reasonable and necessary.”

Allegations of coercion and unfairness for programs such as CED reveal an underlying bias that new technologies are beneficial and that early access to them will enhance patient outcomes. A full analysis of the ethics of CED must not ignore the possibility that CED could result in the premature and inflated use of technologies that have underappreciated risks for many patients.²¹ Therefore, the reliable and ethical application of CED requires that clear criteria exist by which technologies can be identified as fitting into an evidentiary middle ground, one that might be called “promising but unproven.” This middle ground must not be so broad as to include almost any new technology, nor can it remain vaguely determined. Although the CMS has since its inception used the term “reasonable and necessary” without being able to define this threshold more explicitly, its adoption of the CED approach creates additional pressure to bring greater clarity to the evidentiary standards it uses for coverage decisions.

The CMS does not currently consider costs in its coverage decision-making process, and CED is expressly not a proxy for cost-effectiveness or other techniques that might help control Medicare expenditures.¹⁰ However, cost should be a practical and ethical element in prioritizing among the many technologies that could be eligible for CED. Medicare and private insurers alike share a moral obligation to seek the prudent use of health care resources. If a new technology with “promising” evidence would be extremely expensive if used widely without further definition of its true risks and benefits, further ethical weight would be added to the rationale for requiring evidence development as a condition of coverage. Cost considerations should not prejudice the interpretation of the existing evidence on risks and benefits for new technologies; however, since CED cannot for practical reasons be used as the coverage approach for all “promising” technologies, cost considerations should play a valid and honest role in selecting which technologies should be prioritized for CED.

This ethical analysis has identified strong reasons to reject global assertions that programs linking insurance coverage to needed evidence development are coercive or unfair. Coverage with evidence development should be embraced as a bold initiative by the CMS to use its considerable power as a public insurer to promote efforts to improve the evidence available on critical clinical questions for many health care decision makers.

In assessing the ethics of CED, one more aspect can be named in its favor. Societal interest in greater knowledge to guide health care decisions is considerable, particularly when the risks and benefits of very expensive medical innovations are poorly known. Patients who share in the benefits of society, and who ask society to pay for these benefits, arguably should share a willingness to contribute to the body of evidence that will improve the quality and value of the health care of tomorrow. Coverage with evidence development is a fledgling effort to recognize these ethical values and find a reasonable way to embed them in the coverage process. If insurance coverage and properly designed efforts to develop improved evidence cannot be linked together usefully because of mistaken concerns about the ethics of CED, much will have been misunderstood, and much will be lost.