Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
9 Autologous Bone Marrow Transplantation: A Microcosm of the U.S. Health Care System WILLIAM T. McGIVNEY~ The controversy over the appropriate utilization of high-dose chemotherapy- autologous bone marrow transplantation (HDC-ABMT) in the treatment of can- cer epitomizes the debate in the United States over increasing expenditures for the application of health care technology (drugs, devices, procedures, and tech- niques). The debate includes all imaginable constituencies patients, physicians, hospitals, payers, employers, lawyers, economists, and the mass media. The issue is fascinating, because it continually presents new twists and turns. The major question raised is whether or not the utilization and payment for such expensive, potentially high-volume technologies should proceed only after rigor- ous outcomes data concerning its use for a particular indication are available. In the last decade, the concept of basing clinical and coverage policies on cold, hard data has become an axiom in medicine that is widely quoted yet often ignored. The manner in which the HDC-ABMT issue is resolved (or not resolved) will presage the manner in which similar pressing issues are addressed in the l990s and the century beyond. This chapter reviews the controversy surrounding HDC-ABMT, discusses the consensus on the use of outcomes data in health care decisionmaking, de- scribes one payer's process for the difficult coverage decisions associated with HDC-ABMT, and proposes a system for the evaluation and diffusion of signifi- cant new technologies. 1 The views expressed in this chapter are those of the author and are not necessarily those of Aetna Health Plans. 109
110 WILLIAM T. McGIVNEY THE CONTROVERSY SURROUNDING HDC-ABMT The high expense of HDC-ABMT is a major reason for the attention that has been given to decisions regarding its clinical application. Charges for the process of bone marrow harvest, high-dose chemotherapy, and bone marrow reinfusion range from $150,000 to $200,000, with most being in the vicinity of the latter value. The use of peripheral blood as the source of stem cell support in combina- tion with the use of colony-stimulating factors has the potential to reduce these charges significantly. Nevertheless, the cost of an individual procedure is magni- fied by the approximately 1 million new cases of cancer diagnosed each year, including 135,000 to 150,000 new cases of breast cancer. The potential for high- volume use of HDC-ABMT is being realized by an expanding list of indications and by the application of HDC-ABMT earlier and earlier in therapeutic regi mens. HDC-ABMT continues to be used for the established indications of acute leukemia, Hodgkin's disease, non-Hodgkin's lymphoma, and stage III and IV neuroblastoma. It also is being applied for the treatment of stage II, III, and IV (metastatic) breast cancer and multiple myelomas. Finally, a variety of new applications of HDC-ABMT has been observed, including brain tumors in adults and children, ovarian cancer, testicular cancer, Ewing's sarcoma, and metastatic melanoma. The shift toward using HDC-ABMT in earlier stages of the disease process is exemplified by the use of the regimen in stage II and III breast cancer and earlier stages of multiple myelomas. Additional factors influencing utiliza- tion include the use of regimens that involve tandem transplants, repeating HDC- ABMT at some predetermined interval after the first treatment and applying HDC-ABMT to treat the recurrence of a cancer anywhere from one to three years after the initial HDC-ABMT treatment. A final issue involves the appro- priateness of the harvest and storage of bone marrow for patients who at that time are not candidates for HDC-ABMT but who in the future may be, if a cancer progresses or if the outcomes associated with HDC-ABMT improve. Thus, the high expense, the size of the potential patient population, rapidly expanding applications, and the scarcity of outcomes data clearly identify HDC- ABMT as a controversial clinical issue with substantial policy implications. These factors and that of critical patient need raised the issue to the national public policy level. The courtroom often has become the forum for this debate in the context of challenges to payer denials of HDC-ABMT. In a clear majority of these cases, the plaintiff has prevailed. In most of these cases, court decisions have turned not so much on the merits (e.g., data) available to support the use of HDC-ABMT as on the coverage decisionmaking process and the accurate trans- lation of the terms (e.g., investigational) and criteria used in this process into specific contractual language (see chapter 10, this volume). Outside the courtroom, there have been hyperbole and posturing suggesting that payer cooperation with the National Cancer Institute (NCI) in trials on breast
AUTOLOGOUS BONE MARROW TRANSPLANTATION 111 cancer represents an attempt on the part of payers to avoid payment for a larger number (e.g., 40,000) of cases for breast cancer. When viewed from a societal perspective, one must ask if the expenditure of $7 billion (40,000 x $175,000 per procedure) for the use of HDC-ABMT in breast cancer is justified by the avail- able outcomes data. Furthermore, such posturing ignores the fact that the trials would not be sponsored by the NCI if outcomes data had convincingly demon- strated a therapeutic advantage for HDC-ABMT over conventional therapy for the treatment of breast cancer. Indeed, if such an advantage had been shown, it would be unethical for such trials to continue. Rather, the NCI trials represent attempts by responsible parties to obtain objective data that can be incorporated into their processes for making these difficult decisions. OUTCOMES-BASED DECISIONMAKING The decade of the 1980s will be remembered as a period when the health policy community came to intellectual grips with the fact that the widespread utilization of some health care technologies was based upon no more than the "expert opinion" of a handful of proponents. The sentinel study of Wennberg and colleagues (1988) vividly illustrated the need for the practicing medical com- munity to more firmly ground clinical decisionmaking in outcomes data. A1- though much homage has been paid to this axiom, there has been sporadic appli- cation of the concept. The heart of the debate over the expanded use of HDC-ABMT proceeds from lingering and justifiable concerns over whether HDC-ABMT improves the final health outcome (e.g., survival) in comparison with that from standard che- motherapeutic regimens. For example, the results for improvement in survival from metastatic breast cancer overlap in large measure for HDC-ABMT and repetitive conventional-dose chemotherapy. In April, 1992, David Eddy published a review and analysis of all published studies comparing the benefits and harms associated with HDC-ABMT and with conventional doses of chemotherapy in the treatment of metastatic breast cancer. His analysis concluded that: 1. HDC-ABMT had a "higher treatment mortality and morbidity rate than conventional dose chemotherapy"; 2. HDC-ABMT had "higher complete response and overall response rates than conventional dose chemotherapy"; 3. analysis of available data does not indicate that "median disease-free sur- vival, median overall survival, or actual survival is superior with HDC-ABMT versus conventional dose chemotherapy"; and 4. available evidence did not permit conclusions about the effectiveness of the treatment compared with its alternatives (Eddy, 1992~. Publication of that analysis occurred at a time of rapid diffusion of HDC-ABMT
2 WILLIAM T. McGIVNEY in general oncologic practice, that is, rapid diffusion based upon a modicum of long-term effectiveness data. PAYER PERSPECTIVE Requests for coverage for highly expensive, investigational treatments, such as HDC-ABMT in terminally ill cancer patients, are among the most difficult decisions to which payers must respond. The difficulty of the decision to use HDC-ABMT on an investigational basis is compounded by the often desperate situation of the patient and by the aforementioned general lack of data regarding the safety and effectiveness (e.g., survival) of HDC-ABMT compared with those of conventional procedures. Additionally, challenge of payer denials by patients has generally resulted in adverse decisions by the courts and adverse portrayals of payers by the mass media. Thus, payer denials have often lacked clinical, legal, and societal defensibility. JEtna Health Plans has implemented a process to directly address questions regarding the appropriateness of and coverage for the use of investigational treat- ments in terminally ill patients with cancer. This process has worked extremely well and has given JEtna Health Plans a thorough, scientific, equitable, and ob- jective process for reviewing these trying cases. Indeed, AItna's process may serve as a prototype for the involvement of payers and independent expert clini- cians in a cooperative decisionmaking process that best serves the needs of des- perately ill patients. HDC in combination with allogeneic or autologous bone marrow transplan- tation or with peripheral stem cell transplantation most often is the treatment in question. JEtna no longer automatically denies the use of these investigational technologies in terminally ill patients. Rather, AItna has recognized that both the clinical and the coverage decisionmaking processes really constitute risk-benefit analyses. The sicker the patient is, the less the degree of certitude about the effectiveness of a technology and the greater the risk of harm the patient, physi- cian, and payer may be willing to accept. To address this continuum of care, A3tna has determined that if an investigational treatment for a terminally ill pa- tient is investigational yet "promising," then that treatment is eligible for cover- age. A promising treatment is defined as a treatment that "is effective for that disease or shows promise of being effective for that disease as demonstrated by scientific data." citation?? The process for determining whether or not an investigational treatment (e.g., HDC-ABMT) is promising is described below. First, it should be point- ed out that AItna's process affords every opportunity for a decision favorable to the patient if the treatment is appropriate. Basically, Aetna has extended its previous decisionmaking process by adding a clinical review step that relies on two independent and outside sources of expert medical information and advice. These are NCI and the Medical Care Ombudsman Program.
AUTOLOGOUS BONE MARROW TRANSPLANTATION 113 This model process represents a just and equitable solution to a difficult situation. However, it is a short-term solution because it does not address the major problem identified at the outset of this discussion; that is, health care technologies with significant implications for patient well-being and for the sub- stantial expenditure of heath care dollars often diffuse widely into medical prac- tice before there are sufficient outcomes data available to demonstrate the safety and effectiveness of the technology. On the basis of our experience at A?tna, the following proposal describes a system for guiding the introduction and use of health care technologies.2 PROPOSAL A national advisory body should be established to oversee the conduct of evaluative outcomes research on important new technologies. The use and study of these new technologies would be restricted to a network of designated medical centers. Reimbursement for the use of new technologies (e.g., procedures, proce- dures involving Food and Drug Administration EFDA]-approved devices, and drugs used beyond FDA-approved indications) would be provided but would be restricted to use under the study protocol and within the identified academic centers. In return, outcomes data would be collected and analyzed under the auspices of the independent advisory body. When the outcomes data collected from the study were judged to be sufficient, a comprehensive evaluation would be undertaken, culminating in definitive conclusions about the safety and effec- tiveness of the technology. Once the national advisory body concluded that a technology was safe and effective for the specified indication, the technology would be allowed to diffuse into practice. At this point, individual payers would decide whether they would pay for the specified use of the technology. RATIONALE The rationale for the proposal is that outcomes data should be available to demonstrate the safety and effectiveness of a particular technology (e.g., HDC- ABMT) for its intended use before widespread use of the technology occurs. Assurance of reimbursement would permit earlier access for use as "treatment" for patients in need. The "hassle factor," which is a factor for payers, providers, and sometimes for patients, would be mitigated, if not eliminated. The outcomes data so desperately needed to make the informed risk-benefit determinations inte- gral to the formulation of sound coverage policy and clinical decisionmaking would be generated, collected, and analyzed in a cooperative scientific environ- ment. Inadequate investigator participation in clinical trials would be rectified by 2 Interestingly, a similar proposal was made in the early 1980s (Bunker et al., 1982). However, the time now seems riper for implementing such a system than perhaps it was then.
4 WILLIAM T. McGIVNEY the opportunity to use the most advanced technology and the financial incentive of early reimbursement. Inadequacies in patient accrual to trials would be ad- dressed by the concomitant expansion of the size (number of studies) and by restriction of general diffusion. The permissible designs used in the studies also would be expanded. Studies and patient care would be conducted in institutions where rigid quality assurance such as that carried out by NCI helps to ensure consistency, competence, and experience. Payers and society, in general, would achieve answers to important medical issues in a much more efficient and expe- ditious manner. Finally, as an added bonus, the sagging infrastructure of aca- demic medicine would be buttressed by the restriction of payment for new tech- nology to the setting where its initial diffusion and study belong, that is, in institutions dedicated to the deliberate evaluation of and advancement of patient care. DISCUSSION HDC-ABMT crystallized the issues and concerns that have surrounded the debate over ways to deliver high-quality health care while controlling the rate of rise in health care expenditures. With over 1 million new cases of cancer diag- nosed every year and with an approximate cost of $175,000 per procedure, HDC- ABMT is a rapidly diffusing technology that has significant implications for the expenditure of health care dollars. One would then expect that with such a large attendant cost, the cancer care community would proceed cautiously with the use of this technology for investigational indications on a limited basis until scientif- ic data convincingly demonstrated the safety and effectiveness of HDC-ABMT for these indications. Actually, just the opposite seems to be occurring; that is, the use of HDC-ABMT seems to be growing almost exponentially, far outpacing the ability of clinical research to generate data to support such use. How much more rational, responsible, efficient, and valuable to patients, physicians, and payers it would have been if HDC-ABMT were introduced into medical practice through an approach such as the one described above. For example, a hypothesis that extends the use of HDC to an investigational indica- tion would be generated. A network of academic medical centers would be established to evaluate the safety and effectiveness of the procedure. Using the NCI model, after initial dosing and tumor activity studies had been successful, studies would be set up to define the clinical effectiveness of the treatment in comparison with that of the state-of-the-art treatment. The "gold standard" of a randomized controlled clinical trial would be utilized, but alternative designs (e.g., controlled case series) would be available to those patients who did not wish to be randomized. HDC-ABMT, however, would be available only to patients participating in these studies in the defined network of institutions. During these studies the patient care costs would be covered by the payers, whereas the research costs would be borne by the nonprofit research institution.
AUTOLOGOUS BONE MARROW TRANSPLANTATION 115 Patients would be enrolled in the study only if they fulfilled the criteria estab- lished for the study. Payers would abide by the clinical judgments of the partici- pating investigators. Thus, there would be no individual case management re- garding the appropriateness of patients as candidates for the procedure. Outcomes data would be collected and analyzed by an objective panel, with representation given to practicing physicians and also to payer medical directors. When sufficient data were available, the independent panel would reach a defini- tive conclusion and make a recommendation as to whether the technology should be permitted to diffuse into general practice. Individual payers would then for- mulate their own coverage policy regarding the use of the technology. The proposed systematic approach seeks to establish a logical and orderly mechanism for the diffusion and utilization of important new technologies. This mechanism would be based on solid outcomes data and expert consensus. Payers would cover the cost of important new technologies early on within a designated network of academic medical centers in exchange for data collection and analysis and definitive conclusions about the safety and effectiveness of the technology. Implementation of this proposal would contribute to the enhancement of societal good by (1) supporting the use and study of technologies that will lead to either definitive support for widespread use or agreement that the technology should be discarded, (2) enhancing the availability of important new technologies to needy and appropriate patients early on, and (3) shoring up the sagging infrastructure of the academic medical center. Critics will be quick to pounce on the plan and claim that great harm will be done to innovation in technology and access to needed care by patients. In my opinion, however, the proposal at hand will, in the long run, facilitate the innova- tion, introduction, and diffusion of new technologies by establishing a broader base of financial and scientific support early on and by financing a rapid general diffusion once the value of a technology has been established. Indeed, this pro- posal recognizes and affirms the import of a technological imperative in medi- cine, but seeks to avoid a technological Armageddon. In the near future, HDC-ABMT may achieve the distinction of becoming the technology that annually expends the greatest amount of health care dollars. What will happen to a new technology like HDC-ABMT in the year 2000? Will it be guided expertly through an orderly scientific process with expanded patient access and, then, into full clinical practice? Will it proceed as HDC-ABMT is doing now, with limited scientific study paralleled by rapid and somewhat chaot- ic diffusion that results in irretrievable patient outcomes data? Or, will HDC- ABMT be prioritized as, for example, technology no. 798, with national health care funding available only for technologies 1 through 612? A workable solution to maintaining the delicate balance between sustaining the technological imperative and avoiding technological gluttony must be achieved. All sides will have to give in on something. The present proposal and others underscore the fact that major parties are willing to subordinate individual
116 WILLIAM T. McGIVNEY interests to improve a health care system that has served this county and the world very well. REFERENCES Bunker, J. P., Fowles, J., and Schaffarzick, R. 1982. Evaluation of medical-technology strategies: Effects of coverage and reimbursement (first of two parts). New England Journal of Medicine 206:620~24. Eddy, D. M. 1992. High dose chemotherapy with autologous bone marrow transplanta- tion for the treatment of metastatic breast cancer. Journal of Clinical Oncology 10:657-670. Wennberg, J. E., Mulley, A. G., Jr., Hanley, T. D., Fowler, F. J., et al. 1988. An assess- ment of prostatectomy for benign urinary tract obstruction. Journal of the American Medical Association 259:3027-3030.
