fects of programs or technologies are not valued in monetary terms but rather are measured in some other unit (such as years of life gained). A CEA, therefore, does not result in a net monetary value for a project. Instead, it produces a measure of the cost involved in attaining some desirable health-related effect. Conceptually, FINDINGS General Most of the specific findings presented below relate to the two major, general findings of the OTA assessment: • Performing an analysis of costs and benefits can be very helpful to decisionmakers because the process of analysis gives structure to the problem, allows an open consideration of all relevant effects of a decision, and forces the explicit treatment of key assumptions. • CEA/CBA exhibits too many methodological and other limitations, however, to justify relying solely on the results of formal CEA/CBA studies in making a decision. Thus, although CEA/CBA is useful for assisting in many decisions, it should not be the sole or prime determinant of a decision. CEA/CBA is viewed by different parties as ranging in usefulness from obfuscating the pertinent issues in a decision process at one extreme to illuminating and synthesizing these issues so well that the technique can be used to make decisions at the other extreme. There is, however, a middle position which maintains that the technique could be helpful in structuring information and that this information should be only one of several components of a decision process. The OTA case study on the artificial heart, for example, lays out many of the factors to be considered in decisions regarding continued funding of R&D of this technology. But those decisions will also be dependent on a number of other political and social factors that are difficult to analyze systematically. Both extreme positions mentioned above pertain to the use of CEA/CBA as a formal, struc CEA permits direct comparison of only those programs or technologies that share similar objectives. This OTA assessment uses the term CEA/CBA to refer to both of these techniques because the findings below apply generally to both. tured analysis which is oriented toward a bottom-line answer, such as a cost-benefit ratio. Such a bottom line, however, often hides many important value judgments, thus providing a seemingly unambiguous answer that may rest on ambiguous data or assumptions. Advocates of the middle position propose that CEA/CBA be used within the context of accepted principles of analysis in order to illuminate the costs and the benefits of a decision but not necessarily to aggregate and weigh them. The findings of this and any other assessments of the past usefulness or current potential of CEA/CBA in health care decisionmaking should be kept in perspective: Because there has been little experience with the use of CEA/CBA, these findings are based on very little definite information. Therefore, any of the findings or projections of the usefulness or implications of CEA/CBA in this report could be proven wrong as more experience accumulates. In fact, one of the priorities for future examinations of the role of CEA/CBA in health care should perhaps be small, controlled experiments, or demonstrations, of its potential use (see Option 10 below). Methodology There is no set combination of specific analytical elements that form a standard CEA or CBA methodology. A standard or rigid methodology, however, is to be neither expected nor desired. OTA found a wide variation in the forms of CEA/CBA analyses actually conducted to date. Most of the analyses reviewed seem to be academic exercises, infrequently connected to the policy process. In addition, the literature on CEA/CBA indicates that a great many of the analyses conducted tend to be on the "net cost" end of the CEA/CBA continuum, that is, analyses in which health outcomes were not taken into explicit account. In part, this may be a reflection of the paucity of data on the efficacy, safety, and appropriate use of medical technology. OTA's assessment does find, however, general agreement on a set of 10 principles of analysis that could be used to guide the conduct, evaluation, or use of CEA/CBA studies. OTA believes that use of generally accepted principles is important, largely because of the basic methodological limitations of CEA/CBA. For example, although the results of a CEA/CBA are often dependent on the discount rate chosen, there is no general agreement on what discount rate should be used under what circumstances. Furthermore, as primarily an efficiency-oriented technique, CEA/CBA is weak in the areas of equity and other ethical considerations. Finally, CEA/CBA must often address issues of great uncertainty, such as disease progression, patient compliance rates, differing responses to technology by differing population groups, and so on. These and other limitations must be kept in mind and must to the maximum extent feasible be compensated for by techniques such as sensitivity analysis (see #8 below). The 10 basic principles of analysis are: 1. Define Problem.-The problem should 2. State Objectives.-The objectives of the 4. Analyze Benefits/Effects.-All foresee- when possible, should be measured. Also, when possible, and if agreement can be reached, it may be helpful to value all benefits in common terms in order to make comparisons easier. 5. Analyze Costs.-All expected costs should be identified, and when possible, should be measured and valued in dollars. 6. Differentiate Perspective of Analysis.When private or program benefits and costs differ from social benefits and costs (and if a private or program perspective is appropriate for the analysis), the differences should be identified. 7. Perform Discounting.-All future costs and benefits should be discounted to their present value. 8. Analyze Uncertainties.-Sensitivity analysis should be conducted. Key variables should be analyzed to determine the importance of their uncertainty to the results of the analysis. A range of possible values for each variable should be examined for effects on results. 9. Address Ethical Issues.-Ethical issues should be identified, discussed, and placed in appropriate perspective relative to the rest of the analysis and the objectives of the technology. 10. Discuss Results.-The results of the analysis should be discussed in terms of validity, sensitivity to changes in assumptions, and implications for policy or decisionmaking. In addition to conforming to these 10 principles, all quantitative analyses should specify data sources and be written as clearly and nontechnically as possible. They should also be subjected to review-including public scrutiny when appropriate-especially regarding assumptions upon which their outcomes may rest. OTA found that many of the methodological limitations of CEA/CBA are often hidden by the practice of deriving a cost-benefit or costeffectiveness ratio-that is, a numerical bottomline. Therefore, the possibility of not aggregating the often complex sets of calculations should be investigated. Instead of aggregating, analyses might be done by explicitly listing or ARRAYING all the elements which are included in or would be affected by a decision. Where costs and effects could be quantified, that would be done; where they could be combined, that would be done. But no effort would be made to arrive at a single combined value when one or more important nonquantifiable variables that would have to be left out or relegated to a footnote. Such an array method of analysis would mean that decisionmakers would have a greater number of elements to consider, but it would also mean that intangible or nonquantifiable factors would be more explicit and, thus, more likely to be taken into consideration. The findings of this assessment, especially methodological ones, focus primarily on the practical implications of CEA/CBA for health policy. The report is not written for the academic research community. The general principles above apply to analysis for policy use. In no way, however, should the findings be taken to mean that complex, sophisticated analysis is always unnecessary or superfluous. Advancement of the state-of-the-art of CEA/CBA requires ongoing and sophisticated research. Current research on the development and validation of sets of indexes of the health status of a population, for example, appears very valuable. Use of such complex aspects of CEA/CBA, however, may require greater technical capabilities than most health programs currently possess. Addressing a more limited approach to analysis, then, seems appropriate for the goal of this assessment. But it does not diminish the need for more complex approaches in research or other specialized circumstances. Applying CEA/CBA in Health Care As stated above, CEA/CBA can be more valuable to health care decisionmaking when it serves as a problem structuring process than when it becomes the primary factor, with numerical results, of a decision. Furthermore, CEA/CBA potentially can be more valuable for decisionmaking under a constrained budget, when tradeoffs have to be made directly, than when constraints are nonexistent or very indi rect. In neither case, however, would CEA/ CBA necessarily function as an effective costconstraining mechanism or tool. Under the budget system, the budget itself would be the constraining mechanism. Under the nonconstrained system, since no direct tradeoffs are required, no direct limit on expenditures is set. CEA/CBA might, however, change the mix of expenditures. Especially under a budget system, technologies might be substituted for one another as a result of analysis. The context within which a decision will take place must be specified before any judgment of the usefulness or applicability of CEA/CBA can be made. For example, does the decision relate to a technology at an early stage in its lifecycle such as bone marrow transplants? Or does it concern an established technology, such as cervical cancer screening? Is the technology in question a diagnostic technology, such as the CT scanner, or a therapeutic one, such as renal dialysis? The possibility of affecting the course of a technology's diffusion and use might be greater in early stages of its development, but the uncertainties about its health effects and its costs will generally be greater. Thus, it may be possible to do a more valid or certain CEA/ CBA later in the technology's lifecycle, but the information gained may be less valuable for public policy. The tradeoff required will vary depending on the specifics of the technology and the policy decision to be made. Similarly, diagnostic technologies are often more difficult to study than other types because of the uncertainties involved in linking their use to health outcomes. Thus, studies of diagnostic technologies often tend toward the "net cost" end of the CEA/CBA spectrum, where the measures of outcome or benefit may be numbers of tests performed or levels of diagnostic accuracy. Applying CEA/CBA in Specific Health OTA examined the current and potential use of CEA/CBA or related techniques in six health care programs (see chs. 5 through 10). Although informal and often implicit analysis of costs and benefits frequently takes place in health policy formulation, OTA found very little formal use of CEA/CBA in the programs studied. In several of the areas, cost itself has played little or no role in policy decisions. Reimbursement programs such as medicare and Blue Cross/Blue Shield, when deciding what technologies will be covered, concentrate on criteria that generally do not explicitly include costs (e.g., stage of development of the technology and acceptance by the medical community). Under medicare, initial responsibility for identifying questions about whether a technology should be covered lies with a system of local contractors who administer the program. When not resolved at the local level, a question is referred to the Health Care Financing Administration (HCFA), which may seek a recommendation from the Public Health Service (PHS). PHS has traditionally used four criteria in its recommendations: efficacy, safety, stage of development, and acceptance by the medical community. Other health insurance programs, such as Blue Cross/Blue Shield, operate similarly. The possibility of expanding coverage criteria to explicitly include costs or cost effectiveness is being examined by HCFA and PHS. The first question to be answered is whether there is a legal basis for any such inclusion. Currently, the language of the Social Security Act requires the medicare program to pay for services that are "reasonable and necessary." There is no definitive interpretation of whether that language means that the relative "cost effectiveness" of a particular technology might make it unreasonable or unnecessary. Most current reimbursement programs are examples of programs without direct budget constraints. Each reimbursement coverage decision does not involve a tradeoff. Approval of one technology does not mean that another will not be covered. In a very real sense, the existing reimbursement system is an open-ended system of financing medical care. The PSRO program was enacted to assure that health services provided under medicare, medicaid, and certain other programs are medically necessary, meet professionally recognized standards of care, and are provided at the most economical level consistent with quality care. CEA/CBA approaches have theoretical applicability in three areas: 1) the development of standards of care against which actual practices are judged, 2) the internal management of individual PSROs, and 3) evaluations of the national PSRO effort. Cost-effectiveness criteria have not been directly incorporated into standards of care except in a few instances (e.g., respiratory therapy), but it appears possible for PSROS to do so, although the studies would most likely have to be conducted elsewhere. PSROS generally do not have the analytical capabilities for CEA/CBA. Net cost techniques have been used to evaluate whether the savings achieved outweigh the costs of administering the review activities of the overall PSRO program. These analyses, which often reach contradictory conclusions, do not, however, examine costs in relation to changes in health outcomes that may result from PSRO reviews. Nor do they address the fact that even the most optimistic reports of savings represent an infinitesimal portion of total medicare and medicaid expenditures. Because individual PSROs operate under a program management budget, incentives exist for them to use cost-effectiveness-like approaches in choosing areas in which to concentrate their review activities. In contrast to PSROS, in the area of health planning, the National Health Planning and Resources Development Act, with its amendments, explicitly states that resources are to be allocated in a more efficient manner and that health planners should weigh both costs and benefits in their decision processes. The Health Resources Administration (HRA) is emphasizing a more analytical approach to health planning, especially in regard to capital budgeting. OTA found that State health planning and development agencies (SHPDAs) and health systems agencies (HSAs) are for the most part still primarily health "needs" oriented. An OTA survey of planning agencies indicated that few agencies, even those in the vanguard of using economic analysis for allocation of resources, are going beyond the practice of considering only capital costs. There is a slight trend, however, for the most analytically advanced staffs to consider the marginal (or incremental) costs associated with changes in use of a technology. The analysis that took place around the CT scanner is a good example of that. OTA discovered no HSAs that explicitly balance costs with health benefits in, for example, certificate-ofneed recommendations. Thus, although there appear to be no legal barriers to its use, CEA/ CBA has not been much applied. In health planning, as in reimbursement, there is no direct budget constraint, i.e., the area served by an HSA is not operating with a fixed or predetermined amount of resources to be spent on health care. Few pressures, therefore, act on planning agencies to force consideration of how to get the most health effect for the fewest dollars. Market approval for drugs and medical devices, by the Food and Drug Administration (FDA), is an area where Congress has specified the decision criteria, which explicitly do not include costs. FDA is required to regulate the market introduction of drugs and devices on the basis of effectiveness (efficacy) and safety. FDA has not formally used cost-effectiveness or any other economic criteria to evaluate drug or device applications. The primary purpose of FDA is to protect the public from unsafe and inefficacious products. Although the Agency's processes do have indirect influence on the way resources are allocated, there are several factors arguing against the incorporation of explicit economic criteria in its decision processes. Perhaps the most important of these is that by incorporating such criteria, FDA might be extending its regulatory functions beyond those envisioned by Congress. Further, the administrative burden and demands of doing so would be enor mous. The Federal health R&D effort encompasses a broad spectrum of activities and involves several Federal agencies. At one end of the spectrum is biomedical and behavioral research which is supported by the National Institutes of Health, and at the other end is health services research such as that supported by the National Center for Health Services Research, the National Center for Health Statistics, and HFCA. Lying somewhere in the middle of the spectrum and in corporating both end points is research supported by the National Center for Health Care Technology. The Federal agencies involved in health R&D rarely use explicit CEA/CBA considerations to set research priorities, to allocate research resources, or to evaluate the results of research. The uncertain end products of much research makes it difficult to conduct CEA/ CBA. Technologies being assessed or to be developed are often at an early stage of the lifecycle. A CEA/CBA-like approach, with no requirement for aggregation of variables, may be more applicable for R&D than is rigid, traditional CEA/CBA. The more basic the level of research, the less applicable the techniques become, owing to the increased uncertainties. Ironically, it may frequently be more desirable to assess a technology, including its cost implications, earlier in its development rather than later. HMOs are both insurers and providers of medical care. They are an example of a program operating under a constrained budget and have a direct economic incentive to provide "cost-effective" care. However, a preliminary OTA examination of HMO's decisionmaking criteria indicates that, at least in part because HMOs operate in a predominantly fee-for-service environment and must compete for enrollees, these organizations do not commonly weigh health benefit against cost in deciding what medical services to offer. The actual analytical approach used by HMOs seems to be related to "net cost" techniques. Although HMOs do not always consider health benefits, their use of a "net cost" approach suggests that they view the provision of technology in terms of efficiency. They seek ways to provide benefits comparable to fee-forservice medicine at the lowest cost feasible. The current potential for use of CEA/CBA by HMOs, weighing health benefits against costs, does not appear to be as large as the existence of direct budget constraints would predict. That situation may change, however, as more experience is gained with CEA/CBA and as HMOs increasingly encounter competition with each other in addition to fee-for-service health care providers. 66-220 0 - 80 - 2 |