In a traditional cost-effectiveness analysis there is a single objective which is usually measured by a common physical unit (i.e, number of breast cancers detected, number of chronic schizophrenics who return to full employment), and we want to achieve either a fixed level of objective at minimum cost, or as much objective as we can for a fixed cost. There is, therefore, some sort of outcome which is explicitly stated for each service or method of treatment, and the analysis will compare the cost of achieving the goal by various means. It is not easy to find examples of such analyses in psychiatry, since it is unusual to think of the effectiveness of treatment in such crude terms. Patients who have received a particular treatment cannot be usefully divided into’ still ill’ and’ cured’ ; and indeed we may wish to examine the effects of a particular treatment on many different aspects of psychosocial adjustment. It is also sometimes necessary to evaluate services which do not reduce disability at all: for example, services for mentally handicapped children, or for severely demented old people. Fortunately, it is possible to modify cost-effectiveness analysis in such a way that useful conclusions about psychiatric services can still be drawn.

The basic strategy of such modified studies is to compare two alternative forms of treatment service as comprehensively as possible in terms of the quality of care provided and the effects which they have on their patients. No attempt should be made to express the value of these effects in monetary terms. An overall comparison is made of all the ways in which these two services compare with one another, and this is examined to see whether one service is clearly superior to the other, or whether no clear superiority emerges. The economic analysis is carried out separately, and addresses itself to what each service costs. A service which is more expensive than another should be able to justify the increased cost in terms of the higher quality of the services provided. A cost-effectiveness analysis provides us with a rough framework for effecting a comparison between expenditure on health care and the quality of the resulting service.