The economic appraisal of health care interventions has a long history. Twenty years ago, Culyer and colleagues [1] compiled a 70-page bibliography on the subject, and manuals of methods and principles have been available for more than a decade [2]. The evaluative techniques themselves are older still [3].

What is quite new, as noted by the Task Force on Principles for Economic Analysis of Health Care Technology in this issue [4], is the "widespread use of economic analysis ... (in) support(ing) the pricing and marketing of new interventions." A pseudo-discipline, "pharmacoeconomics," has been conjured into existence by the magic of money, with its own practitioners, conferences, and journals. There are a lot of drugs, and there is a lot of money, so the "field" is booming [5].

But drug marketing is even older than economic evaluation. So why now? Why have the marketers so recently become so interested in these techniques?

Perhaps the interest has developed because the environment in which drugs are purchased is changing, at least in the United States [6]. Control over decisions on drug purchasing is allegedly shifting to the health care plan managers who must actually pay for the drugs. Marketing systems aimed by drug companies at physicians are thus bypassed. And indeed, drug costs have increased less rapidly in recent years [7]. This transition may explain the otherwise puzzling statement by T.R. Franson of Eli Lilly [5]: "Our greatest collective failing as a medical establishment, not just the pharmaceutical industry, is that we have failed to communicate to patients and policymakers the value of our services and of our drugs."

If purchasing decisions are shifting, then the "communications" strategy must also shift, changing its style and content to address the new audience. But the purpose of this strategy remains the same—to move the product.

Hence, the dilemma of "pharmacoeconomics." Since the new sponsors of research have such an obvious commercial interest in the outcomes, why should we believe them? And how can researchers take advantage of the new publication outlets, career opportunities, and money, without being suspected of being in a sponsor's pocket? Economics offers ample opportunities for choosing assumptions and techniques so as to reach a preselected result.

Thus, the primary concern of the Leonard Davis Institute Task Force report is with lack of appropriate independence for researchers, that is, undue influence by marketers over the design, conduct, or reporting of evaluations. The Task Force offers guidelines for appropriate conduct for both parties. At one point, these are referred to as "voluntary," but later the Task Force "recommend(s) that scientific organizations adopt these guidelines as minimum requirements ... (and that) sponsors, researchers, and editors explicitly indicate their adherence" [4].

But who is to enforce the rules, and how? Marketers face powerful incentives to push to or beyond the limits of any constraints. "Minimal requirements" soon become sufficient compliance. And undue influence need not leave a track. Researchers know what sponsors want to hear and what will affect the probability of subsequent support. "The milkman's horse does not need to be told where to stop."

Most of the Task Force recommendations essentially say that those who undertake economic analyses of technology should do good and abstain from evil. They might provide support for a researcher faced with improper pressure, or guidance to one in doubt. But, in the end, they offer procedural solutions to what is fundamentally a structural problem. The sponsors have a large economic interest in research outcomes, and the researchers have a professional (and economic) interest in doing research. Bias is inherent in such a structure: Telling each party (how) to be good may be of limited usefulness.

Hence, the secondary concern of the Task Force: the standardization of methodology. If only there were a consensus on rules for doing economic analyses, one could readily determine whether researchers had played by them. Did they use the correct procedures for evaluating changes in mortality and morbidity, the correct discount rate, the correct rules for valuing nonmarket goods and services? Those who tried to bias their findings could easily be identified and dismissed.

But there is no such consensus on rules. The Task Force gets on the wrong track early on, however, describing "lack of consensus about methods" as a source of bias and implying that, presumably through further research, unbiased "correct" methods could be identified and agreed on. Certainly, some choices of method are simply mistaken, and some current methodologic controversies may be resolvable. But the long history of classic questions, such as those about the valuation of life, or the discount rate, warn us that the real sources of disagreement lie deeper.

Different methods may actually be answering different questions. For example, "What is the value of a human life?" is not one question but many. Which is relevant to the problem at hand? Analysts will disagree, because they have different interpretations of the context in which the question is being asked—interpretations both of how the world works and of the "right" bases for valuation. Underlying many of the methodologic alternatives referred to by the Task Force are different theories of how economies function, or conflicting ideological prejudices as to how they should function. These are not converging.

A decision by "sponsors, researchers, and editors" to adopt and try to enforce particular methodologic approaches (or parameters—a 5% discount rate independent of point of view or expected inflation rate!?) would be as ineffective as it is unlikely. The authors of the Task Force report wisely do not take sides in the various methodologic controversies they mention, recommending only that researchers make clear and justify their choices, and that editors provide the necessary publication space.

How, then, do these recommendations restrain the marketers? One can certainly sympathize with the Task Force in saying, at so many points, "be good" and "show your work." But that leaves plenty of scope for influencing the results by choice of assumptions and variables. It is also recommended, at the end of the report, that editors select knowledgeable reviewers—hard to argue with that. But then why do we need guidelines?

Indeed, why do we need them? Apart from the concerns of drug marketers and pharmacoeconomic researchers, the public has a real interest in valid information on new drugs and interventions. In particular, those who must pay for drugs, whether managed care systems in the United States or public reimbursement agencies elsewhere, ought to be enthusiastic about guidelines, if the guidelines make it easier to identify cost-effective drugs.

Yet, curiously, the Task Force was supported only by a coalition of pharmaceutical companies. Government and other agencies declined to participate. This observation does not, in itself, cast doubt on the bona fides of the authors. But presumably the sponsors anticipate that this activity will eventually promote their sales. Otherwise, they were being irresponsible with their shareholders' money. Did drug purchasers hold back support because they foresaw no corresponding benefit?

The traditional justification for regulating drug marketing is a belief that, left unregulated, marketers will mislead clinicians and patients will suffer physical and economic harm from inappropriate prescriptions. Clinicians do not have the resources to evaluate marketing claims. Furthermore, they do not pay for the drugs and treatments they order, and indeed they may be rewarded for ordering more.

But in the new world of managed care, do purchasers need to be protected? They can always hire their own evaluators. If an economic evaluation looks dubious, they don't have to pay for the product. This new balance of power may explain why the Task Force was supported by pharmaceutical firms. The new marketing techniques need a scientific appearance to be more effective. Recalcitrant buyers or uncooperative regulators may be outflanked by appeals to patients, politicians, or the courts, using "scientific" findings of economic benefit. Consensual standards adopted by an emerging "discipline" of pharmacoeconomics could be quite useful for this purpose, particularly if they leave plenty of room for selective methodology.

The Task Force on Principles for Economic Analysis of Health Care Technology is only one part of a wider program, supported by the pharmaceutical industry, to promote guidelines for evaluative research [8]. That the coordination of conduct will be promoted by those who expect to benefit is not a particularly contentious idea, at least among economists. But it is not yet clear whether the wider scientific (or economics) community should cooperate.

Effectiveness studies—does this drug do more good than harm for these patients?—are very important to clinicians (and patients) who are typically not researchers. For them, a set of appraisal guidelines may be valuable. And regulators will need to set their own standards for valid evidence of effectiveness. But the findings of an economic evaluation of the "nth + 1" drug, which in any case are likely to be specific to a particular service environment, are of primary interest to purchasers and sellers.

If the results of such studies are offered to the wider research community, they can and must be assessed in the same rigorous way as all other science. Certainly, transparency regarding methods and assumptions is a good thing and obscurity regarding interested bias is not. We did not have formal guidelines before "pharmacoeconomics" emerged, and nothing in the Task Force report suggests that guidelines will help us now.