Education And Debate Personal paper

Has health economics lost its way?

BMJ 1998; 317 doi: (Published 18 July 1998) Cite this as: BMJ 1998;317:197
  1. D P Kernick (stthomas{at}, general practitioner
  1. St Thomas' Health Centre, Exeter EX4 1HJ
  • Accepted 5 May 1998

Until recently, medicine was fairly straightforward. The doctor had limited therapeutic options, patients did as they were told, and mortality was an unequivocal outcome. Now medicine has become health care, an amalgam of psychology, physiology, anthropology, epidemiology, education, management theory, and politics. Economics has been the latest candidate for the melting pot, on the back of a simple message: when resources are limited, relating the cost of an intervention to its benefits can facilitate the difficult choices that have to be made between competing options. In Britain, with the development of a primary care led service, 1this approach will be of particular relevance to general practice.

Although the difficulties of applying traditional economics to health care and the uncertain relation between health care and health is well recognised, the introduction of internal market reform in Britain in the 1980s gave health economists a chance to establish their credentials and assert their influence on the way health care is delivered. But the early optimism that health economics could provide an explicit framework which could facilitate an ethical approach to the inevitable rationing of health care has proved illusory. And although the importance of economic information is clearly recognised,23 economic studies have little impact on healthcare decisions, which continue to be made with little or no evidence of cost effectiveness.4Why should this be?

Summary points

  • Relating the outputs of a health intervention to the resources that are used is important in choosing between competing interventions

  • Health economics has not made a substantial impact on decisions in health because of problems with acquiring evidence, recognising the relevance to the decision maker, and implementing the message

  • Approaches should be developed that are acceptable to and assessable by end users and which reflect local circumstances and context

The process of utilising research evidence has been described as a series of five stages5 acquisition of evidence; dissemination of information; recognition of validity of information; relevance to the decision maker; and understanding, adoption, and implementation of the message. All these areas have been tested from the point of view of primary care and found to be wanting.

Acquisition of economic evidence

Although there is agreement on a number of basic principles erd5553.gif”> governing the design of economic studies, many issues remain unresolved.6 Ideally, evidence should be obtained from trials that can deliver unbiased and unambiguous answers in generalisable settings, but these criteria are rarely met. Like clinical trials, economic evaluations are often contentious, and disagreement can arise over method, presentation, and interpretation.7 Often there is conflict between the clinical and economic requirements for statistical power–and usually clinical demands will prevail.

Guidelines on the conduct of economic studies help to maintain consistency and comparability in an area where uncertainty can arise from variability in sample data, generalisability, and the analytical methods used,89 and some countries have developed regulatory standards for the conduct of the economic studies which all new drugs are required to undergo.10 11 But guidelines continue to have a large theoretical component across which there remains a lack of consensus among health economists. Maynard has said that they disguise “where analysts are coming from in a fog of pseudo consensus.”12

The failure to agree on how economic evidence is acquired and integrated does not inspire confidence in end users.

Dissemination of information

Pharmacoeconomic evaluations form the majority of published economic evaluations. They are often commercially funded and like clinical trials can suffer from publication bias.1315

The Department of Health has assembled a list of independent cost effective studies relevant to the NHS that contains 200 economic evaluations, of which 147 were considered suitable to be published in a register of cost effectiveness studies.16 An analysis of these data found a wide range of quality and concluded that methods and data were not reported in a way that would facilitate dissemination and decision making by end users.17

Recognition of validity to decision maker

If a study is to achieve internal validity, both its costs and outcomes must accurately reflect what they set out to measure. The introduction of the internal market in the early 1990s revealed the paucity of cost data within the NHS, and the wide range of estimates that are found by various studies probably reflect inconsistent methodology rather than true differences in efficiency. For example, a recent review of 20 studies that derived the unit cost of a consultation with a general practitioner found a range of between $3 and $11, depending on the method of costing used.18 Indirect costs such as productivity losses often form a major component of a study, but again, there is no consensus on the best approach to this area.19

Traditionally, measurement of health outcomes has concentrated on mortality and morbidity, but with the development of a broader concept of health, other domains have been included. These multidimensional outcomes are often difficult to quantify, and it may be difficult to attribute them to specific interventions. Even with this broader approach, other sources of benefit and disbenefit that affect health status can be overlooked,20 and no satisfactory approach has been devised to integrate the disparate outcomes of health interventions, particularly in the complex environment of primary care.

Relevance to decision make

The viewpoint of an analysis defines which costs and benefits of an intervention are relevant. Researchers often generate information within a political framework, and this may influence the alternatives explored and the method of presentation. The perspective of the individual patient, the general practice, the hospital, the purchasing authority, the NHS, or society in general can all be considered, and different answers may be obtained for each approach. Drummond has argued that the relevance of individual costs and benefit will be a function of each decision making setting.21 He concluded that even perfect standardisation would not necessarily permit simple comparisons or generalisability across different settings, where particular analytic viewpoints may differ. The perspective of many studies may not be relevant to the general practitioner.

Understanding, adoption, and implementation of message

The health economist's aim is for a universal method where a seamless theory can be applied to all healthcare decisions, integrating all outcome measures into a single unit of measurement which can be weighed to take equity considerations into account.22 This population ethic of efficiency sits uneasily with the individual ethic of effectiveness, and it may be difficult to resolve the conflict of perspective between the individual patient and society that often occur in an economic analysis. Often, probabilistic findings do not coincide with the need to make choices for a particular patient.

Doctors will be influenced by advocacy for their patients, with an emphasis on decisions taken jointly on the basis of full information; economic practitioners will be driven by a value system based on cost and efficiency. Considerations on equity will sit uncomfortably between these two perspectives. Although the message may be clear, implementation may prove difficult.


Economic evaluation has been introduced into health to provide a framework on which rational decisions can be made, as without it choices may be made on the basis of “politics, emotion, and unsubstantiated advocacy.” But over the past decade, this approach has made little impact on the delivery of health. Maynard has argued that resources used to “re-invent the wheel with guideline reiteration and quasi-consensus statements” should be targeted where they could produce the greatest return.12 He calls for a retreat to basics and a closer relationship to academic departments–a return to a broader perspective where the formulation and execution of health policy can be influenced, rather than the narrower confines of economic evaluation where success has been limited.23 This is a move in the wrong direction.

Overstandardisation following a recipe blindly may be counterproductive. There is a difference between “cookbook” and “toolkit” approaches to economic analysis.21 General practitioners have made implicit economic choices for many years, and fundholding has shown that they can accommodate a more explicit approach. Health economists should develop simple tool kits that complement and support a pragmatic system of health delivery; help local providers evolve satisfactory rather than optimum solutions; work more closely with those they seek to influence; and develop simpler rather than more complex evaluation systems that are accessible and acceptable to end users.

Health economists have lost their way. They have failed to grasp that decisions will continue to be distanced from government and focused on end users where judgments are often taken in a broader context. But all is not lost. Rather than retreating into their academic laagers and analysing why they have made so little impact, health economists should scrutinise their baggage: in a primary care led health service, it may be best to travel light.


Funding: St Thomas' Medical Group is a research practice and receives funding from the NHS Research and Development Executive.

Conflict of interest: None.


View Abstract

Sign in

Log in through your institution

Free trial

Register for a free trial to to receive unlimited access to all content on for 14 days.
Sign up for a free trial