Commentary: Possible road to efficiency in the health serviceBMJ 1994; 309 doi: https://doi.org/10.1136/bmj.309.6957.784 (Published 24 September 1994) Cite this as: BMJ 1994;309:784
- C Donaldson
- Health Economics Research Unit, Department of Public Health, University of Aberdeen, Aberdeen AB9 2ZD.
The NHS at last seems ready to use economics as a framework for setting priorities. Cohen points to two main reasons why this is the case. Firstly, the advent of the purchaser-provider split has, among other things, led to a clearer definition of roles; explicit priority setting is now less easily avoided by purchasers. Secondly, some sensible person in the Welsh Office has offered guidance on how to set priorities. This process requires the use of economics techniques. Will England and Scotland follow suit?
The data free environment of the NHS should not prevent the use of economics as a framework within which less tangible costs and benefits are included alongside those which are tangible. This allows explicit observation of the trade offs made as a result of decisions to expand or contract a service. Cohen's work is an important example of such use of the economic framework.
The economics approach is not, however, free of problems.
Marginal analysis takes time. A team approach is required, with a change in focus from firefighting to more considered analysis carried out in enough time to be relevant to setting contracts.
Marginal analysis is multidisciplinary. This is a strength as all perspectives are considered. Responsibility for such exercises, however, must be allocated to one or two people. Otherwise, each discipline will revert to focusing on day to day tasks within its own function, and the exercise will flounder.
Disaggregation of data
Collection of information in the NHS is not geared up for marginal analysis. It is often difficult to disaggregate data. The fact that marginal analysis exposes this is useful, and, as Cohen has shown, good estimates can still be obtained without delay. It is also important to note that although data on needs, current distribution of expenditure, and local knowledge are often unsophisticated, such data are used only to compile investment and disinvestment proposals. Subsequent marginal analysis would entail extracting more detailed data on these proposals, at least on costs.
Furthermore, this more detailed analysis may show that some proposals for disinvestment will be dropped and that not all proposed investments will be implemented. To be on the “wish list,” therefore, is no guarantee of expansion or reduction. This seems to have been what happened in Mid Glamorgan, as shown by the starred options in the box.
Asymmetry of information
There is an asymmetry of information between purchasers and providers. This can have two effects. The first is to question whether and how providers should participate. The Mid Glamorgan group seems to have worked well, but there is no explicit reference to group dynamics. The extent and quality of such participation will vary geographically and according to the problem addressed.
The second effect is an overreliance on published work as a source of evidence, particularly on outcomes. Often, no such data relating to local issues are available. We are a long way from change based on outcomes in the NHS. The important thing about marginal analysis in this context is a framework. At best, this simply means a description of the possible outcomes of each option assessed, but it does not diminish the importance of these outcomes or the need to make decisions.
Allocative versus technical efficiency
Marginal analysis can be used to identify ways of improving technical efficiency. This means that the same group of patients will receive care but in a different way. An example of this is day surgery. Resources may be saved which can then be released for another type of care.
Improving technical efficiency is useful, but it can be carried only so far until some people's outcomes are worsened to improve those of others. The exercise in Mid Glamorgan came up against the usual reluctance of group members to consider such disinvestment in beneficial activities.1 Cohen implies that this problem was overcome, but it is not clear whether the four unstarred disinvestment proposals in the box do in fact entail reductions in benefit or improvements in technical efficiency. More description of these proposals would have been useful. A further point is that resource allocation across programmes - for example, child health versus palliative care - is not addressed. However, the within programme exercise is the starting point, not the end point.
Broad issues identified
To say that we need more counselling is useful, but measurement of costs and benefits requires some notion of how much counselling is to be introduced or expanded. This was done in Mid Glamorgan, but, for purposes of confidentiality, only the broad areas are listed in the box. This is unfortunate, as the move to specific proposals represents progress over many purchasers' vague plans and strategies.
Beyond efficiency: involving the public and analysing equity
The confidentiality referred to above is worrying for another reason. It seems to imply that consumers have no role in the process. Are they not to be consulted on the proposals? If not, why not? This, of course, is a problem not of the use of economics but of the NHS in general. There is no inbuilt incentive to involve the public.
Equity as well as efficiency is important. Cohen points out that other factors have to be taken into account. For equity purposes, it is important to know who incurs the costs and who receives the benefits of any decision to invest or disinvest. Marginal analysis is still required for this.
This paper is a significant contribution to the use of economics in setting priorities in health care. The exercise was seen as useful by all parties. Marginal analysis addresses the relevant issues within the correct framework. Perhaps those involved in purchasing should undertake a marginal analysis of their own current activities. Potential areas for disinvestment could be firefighting and needs assessment, with a proposed expansion in marginal analysis.