Extending Medicare Reimbursement in Clinical Trials (2000)

Clinical trials have become an essential component of modern medical care. The breakneck speed of medical advances and the increased effort to base clinical decisions on reliable evidence place clinical trials in an ever more prominent position between medical innovation and medical practice. Expanding the evidence base for health care interventions is clearly in the interest of both taxpayers who support Medicare and beneficiaries who receive services.

The impression is widespread that some patient care in clinical trials is not reimbursable under Medicare. But except in the case of certain investigational medical devices and a few instances of "coverage with conditions," the Health Care Financing Administration (HCFA) has never explicitly laid out exactly what should and should not be reimbursed. This omission has led to varying interpretations of HCFA's intent by its fiscal intermediaries and carriers who process claims, as well as by providers submitting claims.

A large proportion of patient care provided in clinical trials is routine—care that would be eligible for reimbursement if delivered outside of a trial. Although the evidence is limited, it appears that claims for much of this care are submitted to HCFA (and other insurers, for non-Medicare patients) without acknowledgment that the patient is in a clinical trial, and they are paid in the normal course of business by HCFA's contractors. But not all such costs for clinical trial patients are paid.

Increasingly over the past five years, uncertainty about reimbursement for routine patient care has been suspected as contributing to problems enrolling people in clinical trials. Clinical trial investigators cannot guarantee that Medicare will pay for the care required, and they must disclose this uncertainty to potential participants during the informed consent process. Since Medicare does not routinely

"preauthorize" care (as do many commercial insurers) the uncertainty cannot be dispelled in advance. Thus, patients considering whether to enter trials must assume that they may have to pay bills that Medicare rejects simply because they have enrolled in the trial.

This report recommends an explicit policy for reimbursement of routine patient care costs in clinical trials. It further recommends that HCFA provide additional support for selected clinical trials, and that the government support the establishment of a national clinical trials registry. These policies (1) should assure that beneficiaries would not be denied coverage merely because they have volunteered to participate in a clinical trial; and (2) would not impose excessive administrative burdens on HCFA, its fiscal intermediaries and carriers, or investigators, providers, or participants in clinical trials. Explicit rules would have the added benefit of increasing the uniformity of reimbursement decisions made by Medicare fiscal intermediaries and carriers in different parts of the country. Greater uniformity would, in turn, decrease the uncertainty about reimbursement when providers and patients embark on a clinical trial.

In the Balanced Budget Act of 1997 (P.L. 105–33, Section 4108), Congress directed HCFA to enter into a contract with the National Academy of Sciences for a "Study on Preventive and Enhanced Benefits" under Medicare. Five specific items were to be studied:

1. nutrition therapy services, including parenteral and enteral nutrition and the provision of such services by a registered dietitian;

2. skin cancer screening;

3. medically necessary dental care;

4. routine patient care costs for beneficiaries enrolled in approved clinical trial programs;

5. elimination of time limitation for coverage of immunosuppressive drugs for transplant patients.

Three committees were established to carry out the tasks, including this one to focus exclusively on the clinical trial question.

The clinical trial committee is aware that the question of reimbursement for care in clinical trials is not a new issue. Clinical trial investigators, patients, and potential volunteers have increasingly seen as a problem the lack of coverage for routine patient care that would be covered if the patient were not in the trial. Cancer activists and organizations, including cancer centers, were the most active agents in bringing this issue into public view. Several draft bills have mandated that Medicare cover routine care costs in clinical trials. Some have

named only ''cancer trials," while others have pointed to "cancer and other life-threatening diseases." Agreements to pay for treatment in cancer trials have been drawn up between the National Cancer Institute (NCI) and the departments of Defense (DoD) and Veterans Affairs (VA); and United Health care, a major managed care organization, does pay for treatment (at varying levels) in selected cancer trials.

In this report, we define the term "clinical trial" as a formal study carried out according to a prospectively defined protocol. Clinical trials are intended to discover or verify the safety and effectiveness in human beings of interventions to promote well-being, or to prevent, diagnose, or treat illness. Other definitions are more expansive, including even the first use of a new intervention without a formal plan or any type of comparison. Our definition is limited to the activities that could be eligible for having at least some patient care costs reimbursed under Medicare. This definition includes:

• interventions to prevent, diagnose, and treat disease;

• drugs and devices; surgical, manipulative, and other procedures; diagnostic laboratory tests, scans, and examinations; dietary, behavioral, and psychological techniques;

• interventions associated with any illnesses or conditions (not limited to specific ones such as cancer, AIDS, and heart disease);

• new interventions, as well as "standard" interventions that have been used in a limited way (or extensively, but about which not enough reliable information is available).

This definition does not include a new intervention applied by a single practitioner to a single patient in what might be the earliest phase of innovation. It applies only after a protocol describing the intervention, the types of patients, the endpoints, and other details has been developed to find out whether an intervention is safe and effective for a given condition. For all types of interventions, the definition encompasses the comparative trials that are needed to produce definitive evidence, and for drugs and devices, in particular, the definition also includes early trials that may be focused mainly on safety and have only one intervention group ("single-arm trials," i.e., they do not compare outcomes in one group versus another; they simply observe what happens when the intervention is given).

The central importance of research to medical practice is relatively new. In the past, when few effective medical interventions were available and most cost relatively little, the lack of precise information about their effects made less difference to the public's health or wealth (although people were adversely af

fected by interventions that were not only ineffective but harmful). What is often cited as the first deliberately randomized clinical trial took place in the late 1940s to determine the efficacy of treating tuberculosis with a newly invented antibiotic, streptomycin. Over the years, occasional tragic complications associated with new drugs or devices led Congress to authorize regulatory agencies to mandate clinical trials to determine safety and efficacy before drugs and devices could be marketed in the United States. Although at least a minimal level of evidence from clinical trials is required for the legal marketing of drugs, biologics, and medical devices, information on whether a new drug or device works better than an old one is not required by law. And there is no such legal requirement to demonstrate safety or efficacy, to say nothing of superiority over existing procedures, for a new procedure that does not involve new commercial products. As a consequence, although many decisions are being made on the basis of sound evidence from clinical trials, the use of many medical interventions, old and new, does not rest on solid evidence. The new emphasis on evidence reflects the realization that intelligent decisions require substantial information that properly conducted clinical trials can provide.

The legislation establishing the Medicare program states:

Since the inception of the Medicare program in the mid-1960s, the phrase "reasonable and necessary" has guided Medicare reimbursement. Although little explicit policy has been issued on the topic, this clause has been the basis for excluding reimbursement for at least some routine patient care in clinical trials. This Medicare interpretation has historical roots in the private insurance sector, whose policies in the 1960s and still, in 1999, exclude coverage of services in clinical trials (GAO, 1999). Most private insurance plans have excluded coverage of services in clinical trials on the basis that the treatment is "experimental" or ''investigational," although the language does not explicitly mention clinical trials (GAO, 1999).* However, Medical Directors report that they often approve

payment for care in clinical trials on a case-by-case basis. In addition, private insurers have been involved in supporting specific trials (e.g., the Blue Cross/Blue Shield Association was instrumental in initiating a trial of high-dose chemotherapy with bone marrow transplant rescue for women with advanced breast cancer).

Despite the lack of an explicit Medicare policy excluding reimbursement for routine care in clinical trials, HCFA has signaled its intent in several ways in recent years. In 1993, HCFA asked the Office of the Inspector General (OIG) of the Department of Health and Human Services (DHHS) to investigate whether hospitals were billing Medicare "improperly for millions of dollars worth of surgical procedures involving unapproved medical devices," specifically investigational pacemakers, defibrillators, and other cardiac devices in clinical trials. In a 1996 hearing of the Subcommittee on Investigations of the Senate Committee on Governmental Affairs, an official of the OIG reported their finding that most of the 130 hospitals they investigated had, in fact, improperly billed Medicare for implanting investigational devices. The Inspector General urged HCFA to recover these "overpayments" from the hospitals (Hartwig, 1996).

What might not be clear from the OIG account is that it was not only payment for the investigational devices themselves, but for the implantation procedures, as illustrated by comments of others, including at least one HCFA official.*

By the time of the hearing, HCFA had already changed its policy to allow reimbursement for patients in certain trials involving investigational devices (FDA "Category B" devices, which are refinements of, or very similar to, approved devices) but not for trials of other types of interventions. The clearest indication that routine patient care is not reimbursable in other types of trials is found in a 1997 report by the General Accounting Office (GAO, 1997) on reimbursement by HCFA for Medicare beneficiaries in cancer clinical trials. GAO found that reimbursement was, indeed, occurring without HCFA's knowledge. In responding to GAO's draft report, HCFA reported that their actuaries had "nearly doubled their estimates of the extent to which Medicare mistakenly reimburses claims for routine patient care costs. Under HCFA's current policy, any reimbursement for care associated with a cancer clinical trial would be made in error" (GAO, 1997).

HCFA has not issued any new language to change clinical trial reimbursement policy since the 1995 change for trials involving Category B medical devices, and no HCFA statements contradictory to what is presented here were found in the course of this study.

The 1995 agreement between HCFA and FDA constitutes the only formal statement of policy about reimbursement of routine patient care costs in clinical trials, authorizing reimbursement for those costs in most trials of investigational medical devices.

There is relatively little information about how the costs of patient care in clinical trials are actually paid, and the extent to which insurers are paying these costs, either knowingly or unknowingly. What information is available suggests that a sizable proportion is paid for by insurers, including HCFA. This conclusion derives from:

• direct evidence from one study that HCFA has paid unknowingly for most routine care of Medicare beneficiaries in certain cancer trials (GAO, 1997),

• evidence that, in the past (before the 1995 change in policy) HCFA unknowingly paid millions of dollars in reimbursement for Medicare beneficiaries in medical device trials (Hartwig, 1996),

• interviews with clinical trial investigators conducted for this study, in which they uniformly acknowledged submitting claims for reimbursement to HCFA and other insurers for routine patient care in trials and getting them paid,

• interviews with private-sector clinical trial sponsors conducted for this study who stated that, while they do cover the costs of "protocol-induced" services, in general they do not provide money to pay for routine patient care; they expect providers to bill insurers for those costs,

• deduction, given the lack of another obvious source of payment for most routine care in trials, and

• lack of evidence from any source that HCFA and other insurers are not reimbursing for this care.

Providers violated no clear rules in billing for routine patient care costs in clinical trials because no such rules were ever codified. But the gap between the impressions—and statements of responsible HCFA officials—regarding reimbursement rules on the one hand, and reimbursement practices on the other hand, should be ended.

This principle applies to payments for physicians and other providers, roufine laboratory and other diagnostic tests, and any other services that comprise routine care for a given patient. All coverage and medical necessity rules and all other restrictions that apply to patients not in clinical trials would apply to care in clinical trials.

The committee recommends a broad definition of clinical trials—including all phases and legitimate designs and all sources of sponsorship (government, industry, or other)—all of which should be equally eligible for reimbursement. This definition does not mean, however, that any treatment simply called a "clinical trial" would qualify for reimbursement. To qualify, a clinical trial must have a written protocol that describes a scientifically sound study and have been approved by all relevant IRBs before participants are enrolled. HCFA should articulate criteria for an acceptable trial and IRB review, which investigators would apply to determine whether their studies are eligible for reimbursement. (HCFA could state the criteria in terms of "current NIH standards," e.g., rather than stating specific study characteristics.) The committee recognizes that controversies surround both the quality of current clinical trials and IRBs, but holds that these issues are being addressed in various ways by DHHS and other sectors of government, and should not be addressed routinely in HCFA's reimbursement decisions.

Medicare should reimburse routine patient care costs, but not all costs in clinical trials. Medicare should not reimburse the costs of experimental inter

ventions (except category B devices for which reimbursement is allowed under agreement with FDA, and certain procedures as described in recommendation 2), of data collection and record keeping that would not be required but for the trial, or of other services to clinical trial participants necessary solely to satisfy data collection needs of the clinical trial ("protocol-induced costs"). These costs should remain the responsibility of research sponsors, private and public.

Medicare should continue its current practice of reimbursing costs of treating conditions that result as unintended consequences (complications) of clinical trials.

Under the current interpretation of Medicare reimbursement rules, the committee believes that surgeons and others performing surgical or other procedures in trials might not be eligible to be reimbursed for those services. Therefore, the committee recommends that procedures that have become widely accepted as a part of standard medical practice, but which, as part of a clinical trial, are being rigorously evaluated, or are being modified or applied for new indications to determine the incremental risks and benefits, should be eligible for reimbursement at the rate for the standard procedure. Conversely, types of procedures for which initial questions of safety and efficacy have not been resolved would not be eligible for reimbursement.

Unlike the basic recommendation regarding routine patient care costs, which applies to all clinical trials, this recommendation would limit reimbursement to randomized trials (the equivalent of "phase 3" trials for drugs and devices). The committee believes this limitation is appropriate in order to avoid providing reimbursement for uncontrolled experimentation by practitioners. The introduction of new drugs and devices is governed by FDA under a formal system that involves phased trials. In contrast, the introduction of new procedures is not governed by any regulatory authority. In their early phases, procedures are modified or tried for different indications in clinical practice, but rarely in formal trials. However, once a new or modified procedure has been defined and developed to the point that it is distinct enough from the predicate procedure, it may be tested against the standard treatment (the predicate procedure or other accepted treatment) in a formal randomized trial. Medicare should provide reimbursement to the surgeon or other practitioner for treating patients in such trials.

Further clarification may be needed to make clear the committee's intent with regard to reimbursement for procedures on patients in clinical trials. The committee is expressing no judgments about when trials of procedures should or should not be carried out, or who should be involved in them if they are. This

recommendation is not intended to influence the criteria or processes HCFA uses to decide on coverage of new procedures under usual medical care. It applies only when a trial of a procedure is being done—for all the reasons that trials are done—and claims for reimbursement for the procedure are submitted by practitioners.

HCFA's initial task in implementing this recommendation will be to develop definitions for classifying procedures analogous to "category A" and "category B" devices. These definitions describing what is and is not allowed will be applied in the field when claims are submitted. HCFA should not be required to rule routinely on the eligibility of procedures before bills may be submitted. In the same way that providers are responsible for following reimbursement rules for all services under Medicare, they will be responsible for applying the rules appropriately in the case of procedures in clinical trials. Fiscal intermediaries and carriers audit these interpretations by providers in clinical trials, as they now audit bills from providers who are not in clinical trials. Advice or an interpretation could, of course, be requested of HCFA at any time. In addition, HCFA would retain the right to initiate its own review, without being asked, if it believes there is an issue to be explored, to carry out a random check, or for another reason.

The committee recognizes that creating definitions that neatly separate "category A" and "category B" procedures will not be simple, and disagreements are inescapable about where the line between "A" and "B" should be drawn in specific cases.

Wherever the separation lies, some procedures will fall into a "gray zone." HCFA can narrow the gray zone by applying the definitions to a wide range of real and hypothetical procedures, and stating whether the procedures would or would not be eligible for reimbursement. To deal with cases in which uncertainty remains, HCFA should set up a process to role quickly on reimbursement eligibility. With accumulated experience, the number of gray zone cases should decline, as has been the case with FDA classification of devices into categories A and B.

The committee has not attempted to specify an institutional mechanism under which HCFA might carry out the tasks required by this recommendation. However, the committee notes that the new Medicare Coverage Advisory Committee* might provide the needed expertise for the task of defining categories A and B and ruling quickly on "gray zone" cases that arise.

Practitioners and institutions would be expected to submit reimbursement claims for services to patients in clinical trials under rules outlined in Recommendations 1 and 2. With a clear statement of reimbursement policy, such claims should pose difficulties no different from those arising in the administration of coverage and reimbursement rules for claims for care outside of trials.

Investigators and providers would not be routinely required to submit documentation about the trial to HCFA, but HCFA could, at any time, request such documentation to confirm that the clinical trial meets current standards for scientific merit and has the relevant IRB approval.

This recommendation is not one that can be enforced as part of a reimbursement policy by HCFA; however, the committee believes it is an important principle that could be adopted by clinical trial sponsors and investigators. It also could be incorporated in any legislation passed to implement the committee's recommendations.

Nearly one Medicare beneficiary in five belongs to a capitated plan—an HMO or some other form of managed care. That number is likely to increase over time. It is vital, therefore, that the committee's recommendations carry over to patients served outside traditional Medicare. Managed care plans must provide all benefits offered under traditional Medicare. (Most offer additional benefits, including coverage of outpatient drugs.) Accordingly, the committee recommends that managed care plans be required to offer Medicare beneficiaries access to clinical trials involving services available within their networks. If, for example, a plan routinely covers a particular drug, it should cover it in a trial, as well. If the plan limits the choice of drugs to those listed on a formulary, the plan should not be required to cover a nonformulary drug in a trial.

Under point-of-service plans, patients should have the right to go outside the managed care network to participate in a trial, under the terms stipulated in the plan for point-of-service care, but no such right should be inferred under plans that limit enrollees to the plan's providers. This recommendation is not comprehensive, but is suggestive of a policy for managed care. Full implementation will require additional thought when HCFA adopts a clinical trial reimbursement policy, but the committee urges that the new policy not create obstacles to clinical trial enrollment for beneficiaries in managed care.

Researchers should be able to apply to HCFA for reimbursement above routine rates in cases meriting special treatment. Such trials could include some interventions that do not qualify under the basic recommendations, such as "category A" procedures, primary and secondary screening, diagnostics, and interventions not usually covered by Medicare (e.g., behavioral interventions). The rationale for extending coverage is straightforward. HCFA has a large stake in determining whether more effective or less costly alternatives to current interventions may exist, preventing ineffective procedures from becoming common practice, and facilitating the identification of innovations that would benefit the Medicare population.

For example, a behavioral intervention, which normally would not be covered, might replace a more expensive drug or surgical intervention to the benefit of both patients' health and Medicare finances. HCFA should have sole authority to decide whether to extend coverage and should make such determinations expeditiously. The committee assumes that only a few trials would be appropriate for such exceptional treatment each year.

In the case of interventions of particular importance to the Medicare population, HCFA should collaborate with the National Institutes of Health (NIH) or others to see that appropriate trials are fielded. HCFA should cover routine patient care costs for these trials along the lines of the committee's basic recommendation. It could also fired other costs as well, under the exceptions procedure described in the preceding paragraph. But the objective is to encourage trials, not necessarily to pay for them. Such an active role would not be new for HCFA. This recommen

dation follows the model of the ongoing study of lung volume reduction surgery, which grew out of collaboration among HCFA, NIH, and the Agency for Health Care Policy and Research (AHCPR), at HCFA's initiative.

Reimbursement claims should bear an identification number assigned by the registry. A registry will facilitate ex post audits of reimbursement claims, HCFA's main tool for monitoring clinical trial coverage and detecting potential abuse. But identification of a claim as part of a clinical trial should not be relevant to the reimbursement decision.

The committee recognizes that implementation of this recommendation will necessarily take some time. Therefore, the committee's recommendations regarding reimbursement of routine patient care costs do not hinge on the existence of a clinical trials registry. Until a registry is in operation, reimbursement claims for interventions associated with a clinical trial should be denoted on the form, in a manner HCFA specifies. However, a registry would contribute to uniform administration and permit HCFA and others to carry out analyses of clinical trials and the costs of implementing the recommendations put forward here. It should, therefore, be put in place as quickly as possible.

Ideally, such a registry should include all publicly and privately sponsored trials before they begin accruing patients, thereby providing a link to all claims for Medicare patients in clinical trials. If the goal of creating such a registry is accepted, the practical question of how best to achieve it must be addressed. It would be possible to build upon the registry currently under development at NIH by broadening the definition of trials to be included and consulting widely on how to present data. Or a separate registry could be created.

Whether the registry should operate within NIH or elsewhere merits consideration. The design of the NIH registry has been underway for some time. Its designers claim that it will be functioning, at least in part, by early 2000. Redirecting any ongoing effort will be difficult for reasons that are well understood. If it were concluded that converting this limited registry into an inclusive national registry would be needlessly cumbersome, the creation of a separate comprehensive registry, serving objectives beyond those of NIH, or even of HCFA, should be explored. The committee urges the Secretary of DHHS to examine this issue promptly, set a timetable for completion of a registry, and seek adequate funding for it.

Implementation of the committee's recommendations would likely cause some increase in administrative costs to HCFA. In making its recommendations, the committee strove to minimize potential administrative burdens. It is the committee's assertion that any added administrative costs required by institution of this reimbursement policy will be small.

Effects of the committee's recommendations on benefit costs are more important and far more uncertain. For several reasons, the cost impact of these recommendations is likely to be quite small. First, the recommended reimbursement policy is designed to limit payments for an individual to roughly the cost of "standard care" for which he or she would be eligible if not enrolled in the trial. This limitation does not imply that each individual would have chosen standard care that cost the same as the care in the clinical trial, so in individual cases, the cost of actual care in the trial might be higher or lower than forgone care outside the trial. Although the incremental cost of routine care in clinical trials is not known with certainty, it is almost surely small in comparison to the costs otherwise incurred by Medicare. Some clinical trial groups claim that the costs of treating patients in some trials may be less than treating them outside of trials. Second, clinical trials hold the long-term prospect of identifying ineffective interventions, which would fall out of favor in the clinical community, or could be excluded from coverage, in some cases saving Medicare dollars.

Finally, only a tiny fraction of Medicare patients participate in clinical trials. No accurate count of clinical trial participants at any point in time exists, but the Lewin Group has estimated that about 265,000 people in the United States participate in clinical trials each year, including about 161,000 Medicare beneficiaries—less than 0.5 percent of the 38 million Medicare enrollees in 1997 (Dobson and Sturm, 1999). Clearly, the proportion of Medicare beneficiaries in clinical trials is quite small. The available evidence suggests that Medicare already pays 50 to 90 percent of routine patient care in such trials. These estimates take into account both costs for which no reimbursement is sought, and claims that are submitted and rejected.

The largest effect on Medicare costs could come from the speedier determination of the efficacy of innovative or experimental procedures, drugs, and devices. Some will be cost increasing. Others will be cost reducing. Whether the net effect is to raise or lower total Medicare spending, the speedy determination of what works and what does not work will benefit the Medicare population and the nation as a whole.

Clinical trials are integral to modern medical care and to the progress of medical science. Although HCFA has issued little explicit policy about pay

ment for routine care for patients in clinical trials, the Medicare statute has been widely interpreted to exclude reimbursement for such care. However, evidence is ample to suggest that providers submit claims for routine care for Medicare beneficiaries in trials without noting the existence of the clinical trial, and HCFA's financial contractors usually pay them. The thrust of the committee's recommendations is that nothing should be done that would materially curtail Medicare's reimbursement for routine patient care costs for patients in clinical trials. On the contrary, HCFA should encourage such trials and even extend reimbursement in a limited number of specifically approved exceptional cases. To achieve these goals, the committee believes that HCFA should assure patients in clinical trials the same reimbursement of routine patient care that is available to patients who are not in trials. Extending reimbursement to certain procedures that represent modifications of current practice and distinguishing those from procedures for which risks and benefits are largely unknown will require some additional effort by HCFA, but it is an essential component of the committee's recommendations. The fundamental recommendation—reimbursement independent of trial participation—should be implemented relatively easily.