| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Ann Thorac Surg 2001;72:330-333
© 2001 The Society of Thoracic Surgeons
a Division of Cardiothoracic Surgery, Washington University School of Medicine, St. Louis, Missouri, USA
Address reprint requests to Dr Cooper, Division of Cardiothoracic Surgery, Washington University School of Medicine, 3108 Queeny Tower, One Barnes Hospital Plaza, St. Louis, MO 63110
e-mail: cooperj{at}msnostes.wustl.edu
Two articles in this volume relating to lung volume reduction surgery (LVRS) address a critical issue for surgeons and patients: namely, how can innovative surgical procedures be ethically introduced in a timely and responsible fashion, while protecting patients and payors from abuse? The review by Stirling and coauthors [1] concludes that in selected cases LVRS is safe and effective with sufficient data available to verify benefit at least out to 2 years. The authors conclude, however, that "higher level evidence is required in the form of randomized control trials comparing LVRS with medical management to establish the worth of this technique in clinical practice." The second article, by Berger and coworkers [2], cites extensive evidence as to the efficacy of LVRS and concludes that there is no justification for continuing to deny the procedure to Medicare patients, whose only current option for coverage is enrollment in the National Emphysema Treatment Trial (NETT), which consigns half of the eligible LVRS Medicare patients to a controlled, nonsurgical arm for an indefinite period, the very type of trial recommended by Stirling and associates. They conclude that "the nonpayment policy [of HCFA] is no longer justifiable."
The two groups present the same data yet with differing conclusions. One recommendation represents the epidemiologist’s quest for measurement of the true "worth" of an effective procedure. The other recommendation seeks to eliminate barriers to the provision of promising new procedures to patients.
At the core of the issue is the role of what is known as evidence-based medicine and its application to new surgical procedures. There are at least three components: When, and under what circumstances should a payor provide access for its beneficiaries to a new procedure? How should scientific validation of new interventions be obtained? How important is it to determine the true "value" of a new procedure, and how should this affect medical decision making?
As the article by Berger and associates documents, there is overwhelming evidence that, for selected patients, LVRS produces benefit unobtainable by any nonsurgical treatment. But measurement of the true "worth" as advocated by Stirling and associates can only be determined by a randomized trial in which half the patients are denied the procedure. This is akin to the so-called "parachute experiment" in which 10 people are pushed out of an airplane, 5 with parachutes and 5 without. The fate of both groups must be determined to establish the true worth of the parachute. The point of this hyperbolic example of course is that randomized trials are inappropriate under many circumstances.
Coverage for a new procedure is usually provided by a payor once there is preliminary evidence of efficacy, especially if no alternative method of therapy is available. Coverage is not denied or delayed until scientific proof of long-term efficacy has been established. New procedures or technologies that fail to fulfill their anticipated promise are soon dropped from our surgical armamentarium.
It is the confusion between two separate issues, namely a coverage decision by an insurer on the one hand, and scientific assessment and validation by the medical community on the other hand, that has produced the unfortunate political, ethical, and professional angst associated with the NETT. The trial highlights a major deficiency regarding the process by which Medicare coverage is determined for its beneficiaries. Medicare is one of the few government agencies permitted to make decisions, such as coverage decisions, without first promulgating the methodology and criteria by which such decisions will be made. Having no such process, and possibly influenced by unsubstantiated projections that LVRS might be applied to millions of Americans with astronomic cost to the Medicare budget, HCFA withdrew coverage for the procedure citing insufficient evidence by which to judge the procedure. HCFA then offered to pay the clinical costs for a trial of LVRS under the auspices of the National Institutes of Health (NIH), but only if the trial design was restricted to a randomized one in which half of the patients were denied access to LVRS for the duration of the trial in order to serve as the control group. Previously, the Agency for Healthcare Policy Research (AHCPR) had carried out a nationwide evaluation of LVRS and recommended "a prospective trial of LVRS under uniform protocol requirements with comprehensive long-term postoperative follow-up data" [3]. Specifically, the AHCPR report did not call for a randomized clinical trial, but HCFA chose to ignore the recommendation, and in a recent letter to a US Senator in defense of the NETT, a high-ranking HCFA official misrepresented the AHCPR report as recommending "the need for a randomized clinical trial of LVRS." Presumably, a long-term randomized trial would serve Medicare’s goal of once again delaying coverage for a new procedure often for up to 10 years after private insurers have instituted coverage.
An unfettered NIH trial may indeed be one of the best methods for scientifically evaluating and validating a new surgical procedure over the long term. It is not, however, the appropriate method for making a timely coverage decision by a healthcare provider. The NETT has been hamstrung by its dual and sometimes conflicting goals, namely to provide sufficient evidence of efficacy for insurance coverage by Medicare on the one hand, and to conduct an independent long-term scientific evaluation on the other hand. The scientific mission of the NETT has been hampered by the restrictions imposed by HCFA as a requirement of their financial support.
From a clinical standpoint, it is important to confirm the efficacy of LVRS, determine the magnitude and duration of the benefit, refine patient selection including risk stratification, and define the morbidity and mortality associated with the procedure. These are, in fact, among the specified goals of the NETT. None of these goals requires that all patients be randomized. Broadly speaking, patients with end-stage emphysema can be separated into three categories. The first is those for whom LVRS has been demonstrated to provide significant benefit. In general, this includes patients with a heterogeneous pattern of emphysema, usually with the most severe destruction located in the upper lobes. The value of LVRS for this group has been confirmed by over 200 scientific publications and five randomized clinical trials, as cited by Berger and coauthors. A second group, representing the majority of severe emphysema patients, is not suitable for the procedure because of the uniform, severe destruction in all parts of the lungs. Virtually all would agree that such patients should not be subject to the risks of the procedure even in context of a randomized trial. A third, intermediate group of patients fall somewhere between ideal candidates and inappropriate candidates, and for this group, a randomized trial might be ethically and scientifically appropriate. The principal investigators of the NETT (including the author) proposed to exclude the first group from randomization and enroll them in the type of prospective trial recommended by the AHCPR with careful preoperative evaluation, uniform selection criteria, standardized operative procedure, and complete and accurate collection and evaluation of follow-up data. By this method, each patient serves as their own control and the magnitude of improvement due to the operation can be identified. Randomization would then have been employed for the third group, those for whom the risk/benefit ratio of LVRS was truly uncertain. This proposal by NETT investigators was denied. So too was a compromise proposal that NETT patients be prospectively stratified into "ideal" and "uncertain" groups even if both were to be randomized. This would allow early review of patients in the "ideal" group with elimination of randomization if the anticipated benefits were indeed observed early on.
The two remaining goals of the NETT trial (which are in fact defined as the primary goals) are more problematic, specifically, determination of the effect of LVRS on long-term mortality and measurement of the true "value" of the procedure, namely how much better the patient is at a given time point after LVRS compared with how an identical patient would be at the same time point without the operation. Only a long-term randomized trial could scientifically provide the answers to these specific questions. This was imposed on all Medicare candidates, undermining the credibility of the trial, creating an ethical dilemma, and inviting the analogy to the "parachute experiment."
The primary goal of LVRS is to improve the quality of life for a period of time for a group of patients whose lives are miserable and who are willing to accept the risks of the procedure in the hopes of alleviating their distress. For these patients, LVRS is not employed as an alternative to any other therapy but as "end of the line" treatment when all other measures have failed to provide an acceptable quality of life. It would not be unexpected if the functional improvement caused by LVRS also translated into a survival benefit. We recently reported the long-term mortality in 87 Medicare patients selected for LVRS in 1995. Sixty-six patients received the operation while 21 were denied the operation when Medicare abruptly halted funding after the patients had been selected for the procedure and were in the preoperative rehabilitation phase. Both groups of patients were virtually identical in their demographic and baseline data [4]. At 5 years, survival of the LVRS group was 72% versus 41% for those denied the procedure (p < 0.02 by log rank) [5].
The second justification for randomizing half of all NETT patients to a nonsurgical control group also rings hollow. Even precise quantitation of the "value" of the procedure, such as in quality-adjusted life years, would not resolve the debate over Medicare coverage. After all is said and done, who will determine what magnitude of value is required for provision of coverage? The confirmation of scientific validity is the responsibility of the medical community, but the question as to how much value is required to provide coverage is not a scientific issue but a societal and economic one. From the standpoint of the surgeon and the patient, the decision must be individualized. The estimated benefit, the estimated risk, and the preferences of the patient must be considered. Operation by formula or prescription is alien to our sense of professional responsibility.
Long-term randomization of half of the ideal LVRS candidates to nonsurgical treatment has actually weakened the ability of the NETT trial to achieve its most important goals. Half of the potential data as to the effects of the operation have been lost because half of the eligible patients are denied the operation. This has also had the effect of undermining the attractiveness and the credibility of the trial. As noted by Berger and associates, the NETT was designed to randomize 4,700 Medicare patients to either best medical treatment or best medical treatment plus LVRS. To date, after more than 4 years, the NETT has enrolled approximately 1,000 patients, and this in spite of a recent two million-dollar advertising campaign designed to improve accrual. Furthermore, the original selection criteria have been substantially relaxed in the hope of improving accrual. We have evaluated 34 Medicare Patients initially enrolled in the NETT who sought to receive LVRS at our center, usually after they were randomized to the nonsurgical arm. Using our own, longstanding validated criteria, we found only 18% of these patients eligible for bilateral LVRS (the procedure specified by the NETT trial). This raises serious questions both as to the validity of the trial design and the generalizability of its eventual conclusions.
The greatest fault with the NETT, however, is the ethical dilemma it has created for the surgical community. Only after the selection process for the NETT centers had been completed were they informed that all surgeons who participate in the NETT would be prohibited from performing LVRS on any patient not enrolled in the trial, including non-Medicare patients. Notably, this prohibition included all surgeons associated with the trial but exempted most participating pulmonary physicians, except those acting as the principal investigator. This interference with the right of surgeons to offer their services to nontrial patients is unprecedented and an intolerable intrusion on the patient-physician relationship. The investigators overwhelmingly passed a resolution seeking to overturn this prohibition, but the NIH overruled the investigators, suspended the trial, and appointed a committee of ethical and legal consultants to consider the matter. While acknowledging that the limitation on patient choice was a "serious situation," the majority of the committee nonetheless felt that "it was appropriate to assign a higher priority, in this case, to the successful completion of the NETT study than to freedom of individual patients to choose a certain physician for whatever reason" [6]. One ethicist, however, dissented, noting that "even though patients often have limited access to the physician of their choice, this fact does not provide justification for the federal government to deprive patients of access to certain physicians solely because these physicians are participating in a research project." The ethics consultants also "indicated that if a state of clinical equipoise was not recognized by the experts within a particular institution then that institution should not participate in the trial." On this basis our institution withdrew from the trial. One NETT center has indeed restricted the application of LVRS at their institution exclusively to NETT patients. Almost all of the centers, however, were placed in a difficult position, as they wished to continue offering LVRS to appropriate candidates who were not part of the NETT. These centers were forced to create two separate programs, each with distinct surgeons: one for the NETT trial and one for all others. The investigators from these centers did not make the trial rules. It was Medicare who called the tune, requiring half of all LVRS candidates in the NETT to be indefinitely randomized to the nonsurgical arm and refusing to allow crossover to surgery on compassionate or any other grounds. The investigators were given no opportunity to consider either nonrandomization for the ideal group or subsequent crossover to surgery for any patients on a compassionate basis. Presumably, any acknowledgement that LVRS is beneficial would undermine HCFA’s denial of coverage outside of the restrictive and coercive NETT trial.
One cannot argue against evidence-based medicine, but the nature of the evidence and the manner in which it is obtained is highly relevant to surgical innovation. Randomized clinical trials are appropriate for some but not all circumstances. Furthermore, participation in clinical trials, be they randomized or otherwise, can be a potential source of conflict between the interest of the investigator and the interest of the subjects. This was recently highlighted in an editorial in The New England Journal of Medicine by Donna Shalala [7]. In another NEJM editorial comment, Dr Larry Bonchek sounded the warning to surgeons about the potential perils of government-sponsored clinical trials. His comments are highly relevant today: "The impetus for proliferation of early, randomized surgical studies has come from the financial support offered by the government. Once such studies begin, they generate their own essential bureaucracy and peripheral constituency, and changing their course is like turning a battleship around. Abandonment of a flawed or obsolete randomized study involves an admission of error and a loss of financial support—a form of academic suicide" [8].
In an article from the MacLean Center for Clinical Medical Ethics, entitled "Ethics, Randomization, and Technology Assessment," Lantos observed that "our statisticians encourage clinical investigators to design trials that make their lives a little too easy by making our lives more difficult. Rather than constrain human choices to meet the needs of statisticians, I think we should encourage the statisticians to devise methods of evaluating the sorts of choices that free and intelligent citizens make all the time" [9]. This is the challenge: to find a way to reconcile the laudable goals of both of the articles in this volume, to temper the scientific process with compassion for the patient.
The NETT has served neither our patients nor the surgical community well. It has engendered ill will, set one surgeon against another, and has poorly served our profession, our patients, and our pride and credibility as surgical scientists. The fault lies primarily with HCFA, which has once again found a means to deny coverage for a procedure long after it has surpassed the standards for provision of coverage set by other insurers. The NETT was the mechanism chosen by Medicare this time around to delay coverage for a new procedure as long as possible. The role of the NIH in scientifically evaluating and validating new procedures should be encouraged. The NETT, for all its faults, would be perfectly acceptable if it were the usual type of voluntary clinical trial in which willing and informed patients are recruited, randomized, and evaluated without coercion and with no penalty for withdrawal. I am unaware of any precedent for coercing a specific group of patients into a randomized clinical trial as their only means of obtaining access to a procedure that at the same time is readily available to all other candidates on a nonrandomized basis. Medicare has used its reimbursement policy to de facto regulate the medical decision-making process between patients and physicians. This does not establish a favorable precedent for the introduction of other new surgical procedures or for the rights of patients who might benefit from them. The leadership of our national surgical organizations should take this matter seriously, explore what went wrong with the NETT, and take the lead in developing a process to avoid this dilemma in the future. A mechanism should be developed whereby a promising procedure can initially be offered under controlled circumstances at specialized centers of excellence. This will speed up the development, foster appropriate scientific inquiry, protect the public purse, and prevent vulnerable patients from being victimized by unproven or inappropriate interventions. Randomized trials should be encouraged when appropriate with clearly defined inclusion and exclusion criteria. Bureaucratic interference in the surgical decision-making process by means of arbitrary reimbursement decisions should be condemned. HCFA should provide assistance in obtaining relevant data, but the design of clinical trials and the evaluation of data should be left to the scientific community. The next time around, someone other than Medicare should be calling the tune.
References
This article has been cited by other articles:
|
|
 |
|
  T. Treasure The evidence on which to base practice: different tools for different times Eur. J. Cardiothorac. Surg., December 1, 2006; 30(6): 819 - 824. [Full Text] [PDF]
|
|
|
|
 |
|
 S. D. Ramsey and S. D. Sullivan Evidence, Economics, And Emphysema: Medicare's Long Journey With Lung Volume Reduction Surgery Health Aff., January 1, 2005; 24(1): 55 - 66. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
