Frank A. Sloan
In a world with unlimited resources, it would be unnecessary to have methods to determine the best way to allocate those resources among alternative uses. But resources are limited: in 1994, U.S. health spending will exceed $1 trillion for the first time and spending is projected to grow 50 percent faster than gross domestic product for the remainder of the decade (Burner, McKusick, and Waldo 1992). Similar trends are evident in other countries (Schieber, Poullier, and Greenwald 1993). Throughout the world, there are pressures on public budgets. Policymakers and the public have begun to recognize that every dollar spent on health care is no longer available for spending on education, crime control, or infrastructure improvement. Further, much of what we now spend is wasted on care that does not improve our health or yields small improvements in health at exorbitant cost (Siu et al. 1986; Winslow et al. 1988; Enthoven and Kronick 1989; CBO 1992a).
In the public sector, even in the face of mounting cost pressures, the balancing of benefits and costs has typically not been a criterion in deciding whether medical services should be made available or should be covered under public insurance. At one extreme, policymakers have focused exclusively on costs, rejecting higher cost alternatives without any consideration of their potential benefits (e.g., some state Medicaid programs in the United States). Others have focused exclusively on benefits while ignoring the costs of achieving these benefits. Even today the U.S. Food and Drug Administration and Medicare do not use cost-effectiveness as a criterion for making what are, in essence, decisions about resource allocation. In the private sector, hospitals, health maintenance organizations (HMOs), insurers, and other decision makers may be more willing to use such criteria, but formal analysis of costs and benefits does not appear to be widespread. In the face of tighter budget constraints, public and private decision makers have made forays in the direction of adopting formal analysis of benefits versus costs. Examples in the public sector include the Oregon priority-setting method (Hadorn 1991; OTA 1992b; Fox and Leichter 1993; Chaps. 3 and 10, this volume) and Australia's guidelines for subsidized Pharmaceuticals (Drummond 1992; Johannesson 1992a; Chap. 10, this volume). The news media report private health decision makers' greater use of studies that evaluate both costs and benefits of Pharmaceuticals and other medical technologies (Winslow 1992; Tanouye 1993). Although there is greater demand for information about returns to investment in health care by numerous parties, including employers and public and private insurers, questions have been raised about the quality of many of the studies conducted to date (Lee and Sanchez 1991; Udvarhelyi et al. 1992).
Cost-benefit analysis has been used in public finance and therefore has had a social perspective in that calculations have weighed social costs and benefits. Such studies were used when individual decisions alone did not lead to an optimal allocation of resources (market failure). In other words, these studies were used by public decision makers to supplant the market. In most situations, however, individual consumers could be relied upon to weigh costs and benefits prior to making a purchase, with the consequence that the sum of individual wants produced the correct allocation of resources.
Formal methods were also applied to health care, but with several important differences. First, there was an overwhelming reluctance to attach money values to benefits. Second, costs largely were limited to those that appear in public or private budgets and that are more narrowly defined than social costs. Third, because of the pervasiveness of consumer ignorance and, more recently, uncertainty among health care providers about the effectiveness of many preventive, diagnostic and therapeutic interventions, there has been demand, even among private decision makers, for information about outcomes that may be obtained from such interventions. Thus, many of the clients for such analyses are in the private sector. Added to such private demand is the increasing public demand as additional medical services are collectivized either through an explicit tax-and-subsidize mechanism or through mandates on private parties.
Major objectives of this book
As in any field for which there is a substantial growth in demand, two concerns about quality may well be raised. One is that parties will skew the results of formal analyses without users of such information being able to detect it. The second is that even well-intentioned authors of such analyses may go astray because they are unaware of alternative methodological approaches. Fearing this, public policymakers may tend to specify details of methodology in a regulatory format or to perform such analyses themselves. As a reading of this book will indicate, at each step along the way there are so many plausible alternatives, each with its own strengths and weaknesses, that standardization or regulation of methods is likely be counterproductive.
A major objective of this book is to summarize the current state of the art in cost-effectiveness/cost-benefit analyses as they are applied to medical problems and to discuss aspects of these methods that have received comparatively little attention in these applications, such as measuring quality of life, costs, discounting, downstream treatment effects, and the sensitivity of findings to underlying assumptions. Several different audiences will find this book useful. First, the book serves as a reference for those who are or aspire to be users of cost-effectiveness/cost-benefit analysis. It is not a `how-to` manual in these methods; such exist elsewhere (e.g., Weinstein and Stason 1976; Warner and Luce 1982; DHHS 1992b). Our goal is to systematically review the methods now in use, assess their underlying conceptual foundations, and summarize their relative strengths and weaknesses. For example, the reader who wants to know how to structure time trade-off questions or use Markov chains for analyses of long-term effects of health care interventions will have to look elsewhere but can use this book to find out where to look. Second, the book provides some guidance for methodologists who have an interest in extending the analytical frontiers on which cost-effectiveness/cost-benefit analyses are based. Experts are likely to be well versed in some aspects of such analyses but fairly uninformed about aspects outside their own discipline. We have not attempted to push the current methodological frontiers but to clearly identify current boundaries that may suggest questions for further research. Third, the book can assist policymakers in showing the strengths and limitations of current methods. Chapter 10 in particular focuses on the problems and pitfalls of using these methods in policymaking.
Though not limited to Pharmaceuticals, many of the examples included in this book come from pharmaceutical studies. The superior evidence on efficacy makes it easier to conduct cost-effectiveness or cost-benefit analysis. At the same time, the relative ease of conducting such studies increases the expectation that such information will be provided as part of the coverage decision-making process. Aside from the relatively greater availability of data on efficacy of phannaceuttcals, the evaluation of Pharmaceuticals is not inherently different from the evaluation of medical devices and other medical procedures...