Introduction

The past two decades have been as hard on science as they have been on many other once venerated institutions, and for at least one of the same reasons: The idea that it consists exclusively of delineation of truths about the state of the world has been successfully challenged.

In some instances, scientific "facts" have been revealed to be deliberately falsified. Apart from tarnishing science's image, falsification has other consequences. If a cancer researcher fakes results, others may spend months or years and large amounts of public money attempting to replicate them. If a psychologist fabricates IQ studies of twins, the data may be used in other similar studies and reviews, and the cumulated erroneous data may be influential in formulating a worthless (at best), or even a repressive and openly racist (at worst), social policy.

It is one of the axioms of the scientific enterprise that the norms and internal controls of science guarantee that such abuses will ultimately be exposed. Even if that is true for flagrant violations, however, it is not clear that it is true or even can be true for the much more subtle distortions inherent in the daily practice of science. These generally involve nothing so egregious as deliberate falsification of data, but rather much less obvious effects of the researchers' biases and value commitments, many of which are unconscious.

In general, the argument that science's valuefreedom is illusory has fallen less harshly on the "hard" or natural sciences than on the "soft" or social sciences. Economics, however, is almost in a class by itself. Since its methods are quantitative almost by definition, economics appears to be relatively "hard" compared to many of the other social sciences. Partly because of this, economics has been one of the least criticized social sciences.

NOTE: This appendix includes nearly all of a report prepared by The Hastings Center under contract to OTA. As with all reports published by OTA, its inclusion does not imply endorsement by the members of the Technology Assessment Board or by any advisory panels. Although the report focuses on CBA, the discussions apply in many instances to CEA as well. The report should not be considered an exhaustive treatment of ethical issues involved in the conduct and use of CEA/CBA.

In recent years, that situation has begun to change. Largely as a result of the "dismal science's" dismal record of forecasting, economics has been subjected to increasing criticism. That record has been increasingly frustrating for citizens and policymakers, both of whom must somehow cope with continuing uncertainties about inflation, unemployment, and other economic indicators. Deciding between big bang and steady state theories of the universe may ultimately have profound effects on our view of mankind, but most people are more concerned and more immediately affected by theories of economics. The impact of economics on government policy and daily life underscores the importance of accurate predictions and conclusions.

Among the currently fashionable methods of economics are CEA and CBA. A major text in the field of CBA was written by the economist E. J. Mishan, who believes economists doing CBA are really asking a question similar to that asked by a company accountant (410). Instead of asking whether the company will be better off engaging in one activity as opposed to another, the economist asks whether society would be better off by undertaking the project under study. In Mishan's view, CBA applies standard economics to help advance the social welfare. The rationale is based on the Pareto improvement,1 which occurs if some economic arrangement makes people better off without making anyone worse off.

According to Herbert Klarman, a pioneer in using CBA in the health field, its use by experts, and the idea that it should influence public decisions, is only a generation old (317). There are few references to it in the literature before 1958. More recently, CBA has become a favorite tool among policymakers and has been applied in many areas of public life. Along with systems analysis, CBA has been widely used by the military and for major public works projects.2

Noting that investment in human capital was a popular theme in postwar economics, Mishan suggests that it was an easy step from investment in human productivity to investment in health (410).

'Named for the Italian economist and sociologist Vilfredo Pareto (1848-1923). "An equilibrim is said to be 'Pareto-optimal' if (and only if) there is no possible movement from it that could make everyone better off" (531).

2One famous and very controversial British study, for example, used it to help find a site for a third London airport.

The rationale is simple: Disease and premature death are costly to an economy, so investment in medical care can bring significant economic rewards by increasing human capital.

An increasing number of cost-benefit and costeffectiveness studies have been done in the health field in the last decade, although how influential such studies have been in the forging of policy is not clear. Rashi Fein has pointed out that the attractiveness of economic arguments derives from a belief that economics is value free and objective (194). That makes it seem particularly appropriate for fields dealing with social policy, such as health care, which are fraught with value conflicts and otherwise appear to lack rigorous guidelines for decisionmaking.

Sociologist Duncan MacRae suggests that the historical development of economics has narrowed the discipline's outlook, making it highly specialized and scientific (394). Whereas economists like Adam Smith all made contributions to disciplines other than their own, today's economists are either interested exclusively in economics, or they are interested in extending economic theory to other fields like political science or sociology.

Thus, MacRae argues, welfare economics is a much more closed system than its predecessors, more impervious to values that come out of other aspects of its practitioners' lives. According to R. M. Parish, this kind of narrow specialization is exactly what CBA needs least (475). Good CBA, he says, "seems to call for a combination of subtle theoretical analysis, imagination, and a lot of hard, pedestrian slugging-the capacities to do which are seldom combined in one individual."

The need for economists with diverse capacities may be especially great in the area of health. As noted by R. N. Grosse, there has been little analysis of the relationship between expenditures for and outcomes of social programs (265). At the very simplest level, we really do not know how medical services affect our health, in part because "health" is an abstract and fuzzy concept. Health programs often really cannot be evaluated. We lack agreement on their objectives and on how to achieve them. We find it hard to estimate measures even when we agree on them. We cannot sort out cause and effect relationships because there are many inputs and outcomes. Finally, we have very little information about programs whose effects occur over a long time.

To pick a program goal and then attempt to evaluate how well the program is achieving that goal requires normative assessments, evaluations, and assumptions. CBA, however, is simply descriptive. Norms, cutoff points, decisions about where to draw the line between good and bad outcomes, all are in

the mind of the policymaker and need to be imposed on the data from outside. Norms are also in the analyst's mind, although analysts' commitment to the disciplinary ethos of quantification and objectivity may make them genuinely and sincerely unaware that norms are an inevitable part of their description.

In a book-length critique of CBA, Peter Self asserts that CBA depends on value judgments more than most types of information (556). Further, some of the major figures in economics, including many of that field's most thoughtful practitioners, share substantial agreement that economics in general and CBA in particular are anything but norm and value free.

This paper lays out major values underlying the quantitative techniques of cost-benefit and cost-effectiveness analysis in the health field. It shows how those values can result in ethical problems, and makes some suggestions for avoiding the difficulties the values engender. Though based on extensive reading of both the theoretical and technical literature in this field, the discussion is not exhaustive. Our brief for this exercise has been well stated by Fein, who observed that it is incorrect to suppose that measurement is neutral. "Cost-benefit analysis," he says, "is too important to be left to analysts or economists" (194).

Working Definitions of Cost-Benefit and
Cost-Effectiveness Analysis

Definitions of CEA and CBA in the literature vary. OTA has defined them as follows:

Cost-Effectiveness Analysis (CEA): An analytical technique that compares the costs of a project or of alternative projects to the resultant benefits, with costs and benefits/effects not expressed by the same measure. Costs are usually expressed in dollars, but benefits/effects are ordinarily expressed in terms such as "lives saved," "disability avoided," or "quality-adjusted life years (QALYs) saved," or any other relevant objectives. Also, when benefits/effects are difficult to express in a common metric, they may be presented as an "array."

Cost-Benefit Analysis (CBA): An analytical technique that compares the costs of a project or technological applications to the resultant benefits, with both costs and benefits expressed by the same measure. This measure is nearly always monetary. E. J. Mishan says (410):

The general question that a cost-benefit analysis sets out to answer is whether a number of investment projects, A, B, C, etc., should be undertaken and, if investible funds are limited, which one, two, or more, among these specific projects that would otherwise