The Nearly Good, the Bad, and the Ugly in Cost-Effectiveness Analysis of Health Care

Article excerpt

As the national debate about health care access, cost, and quality continues, the citizen-participant in democratic self-governance will continue to be bombarded by contradictory scientific discoveries about medical practices. This exacerbates the problems faced by government, health care providers, and individuals as they try to decide what medical procedures are worth the money spent on them. This article examines one approach that promises some solutions, Cost-Effectiveness Analysis (CEA). Other forms of evaluation of medical outcomes are also being discussed, but CEA appears to be the most popular because it combines indications of results (effectiveness of a treatment, for example) with the cost of achieving the result. The more ardent advocates of this kind of analysis, or variations on it, think that its use may improve efficiency of resource allocation for health care not only of treatments for specific ailments, but across the totality of health programs [Torrance 1986, 5].

Theoretically, the policy maker could use utility measures [to give patient-based weights to medical outcomes] to make efficient resource allocation decisions and to at least recognize the equity implications of the efficient solutions. Choosing among programs and interventions that might produce many different kinds of health outcomes, a single dimension such as quality-adjusted life years (QALY), fungible across programs and outcomes, might be used to assure that resources are allocated to maximize QALYs and thereby maximize collective well-being [Mulley 1989, S570; emphasis added].

The appeal of a systematic way of evaluating medical costs and benefits is enhanced because many citizens may become directly involved in allocating resources for health care, as were Oregonians a few years ago. That state set its Medicaid spending priorities partly on the basis of public opinion rankings of nearly 700 broadly defined procedures [Connell 1993]. The rankings were obtained through surveys and public hearings [Stason 1991, 2238; Daniels 1991, 2234] but not, apparently, with much attention to CEA, perhaps because of the disjointed and fragmentary nature of the body of health care CEA. Nevertheless, future health policy deliberations cannot avoid dealing with CEA, and it is therefore important that the meaning of CEA be clearly grasped, that citizens be aware of the conceptual weakness of many CEA analyses, and that they be encouraged to be critical of their content. Because much that is being written about CEA is by people whose understanding of the technique is inadequate, the content of many analyses often fails to advance warranted knowledge upon which to base democratic decisions; indeed, it sometimes reduces warranted knowledge.(1)

This article seeks to establish a context for assessing the value of the many CEA articles appearing (especially) in medical journals. That context does not require that all readers of these articles master every fine point of neoclassical microeconomic theory, but it does impose a significant burden of competence on those who publish the articles. Given the serious uncertainties of outcomes (or "effectiveness") of medical procedures [see Hildred and Beauvais 1995] and the continuing weaknesses of current cost accounting practice in the field, it should be clear that CEA is unavoidably weak and should not be the only or even the dominant basis for deciding resource allocation and other public health policy issues. The (unhealthy) power of physicians, and even more importantly, of corporate financial officers, in decisions about provision and prices of health care services might well be augmented by uncritical acceptance of the legitimacy of CEA findings.(2) (Recent consideration in the Congress of the United States about requiring Cost-Benefit Analysis of federal regulations for nearly every environmental and workplace safety rule should suggest additional grounds for fear of the misuse of this kind of economic analysis. …