Messages from Mid Glamorgan: a multi-programme experiment with marginal analysis.

This paper discusses the issues and problems arising from the first attempt to apply marginal analysis across a full range of health care activities within a UK Health District. Following a brief description of the two stage process undertaken in Mid Glamorgan, the paper focuses on the importance of a conducive environment to a successful application of the approach and to three key issues of group composition, group dynamics and openness/explicitness. The paper also addresses two parts of the Mid Glamorgan process which might be of questionable importance; the need for programme budgets prior to the exercise and the relevance of stage II which attempted to interfere with the expressed priorities identified by the expert groups in stage I. The main message is that marginal analysis works in practice as well as being attractive in theory and that much can be achieved even when only crude data on marginal costs and benefits are available. It is more important that priority setting be addressed within an appropriate framework than with accurate data. Marginal analysis provides such a framework.