The robustness of the London Health Planning Consortium model.

The reports from the London Health Planning Consortium (LHPC) Towards a Balance' and Acute Hospital Services in London2 advocate a reduction in the number of acute beds in London hospitals by 1988. Compared with 1977 "reductions of some 2300 beds (23 0o) are indicated in the teaching health districts." A reduction in the non-teaching districts is also proposed, giving a net reduction in the total number of acute beds in the Thames regions of about 11 00. These bed reductions are based on assumptions about future populations, patient flows, hospital admission rates, and average durations of inpatient stay. Much of the data used by the LHPC in its calculations of future bed provision in London are open to questions that cast doubt on their reliability. When there is uncertainty about the quality of data that are available for planning it is best to employ a robust approach-that is, one in which plans are not unduly sensitive to the unreliability of the data.3 The data used in the methodology include projected 1988 populations, 1977 Hospital Activity Analysis (HAA) data as an estimate for 1988 population flows, projected shorter durations of stay, and lower hospital admission rates (based on 1962-75 Hospital Inpatient Enquiry (HIPE) data) and retrospective mortality data (1974-6) weighted by utilisation (1975 HIPE) as an indicator of the future level of morbidity. The projections of bed requirements for 1988 are made at district level and clearly stated in terms of numbers of beds required for each specialty. The DHSS has provided us with projected bed requirements based on alternative population projections, using the LHPC model. The South-east Thames Regional Health Authority has also explored the effects of varying some of the other assumptions one at a time. Though the other assumptions make appreciable differences for some districts I have concentrated on discussing the effects of population projections for which plausible alternatives have been suggested. It should not be overlooked, however, that data on hospital use, particularly HAA, HIPE, and SH3 forms (statistical returns for the hospital service), have been found to be of questionable reliability.5 One of the central LHPC assumptions is that the population of Inner London will decline by 12-50, between 1977 and 1988, while that of London as a whole will decline by 8 0°/ over that period. These assumptions are based on the Office of Population Censuses and Surveys (OPCS) mid-1977-based projections, which in turn are based on the 1971 census now nine years out of date. We will not have accurate population data until 1982, when the results of the 1981