Forecasting demand for regional healthcare

Abstract

Trends in developed nations point to increased demand for acute inpatient care across most age groups, though especially at older ages. Demand growth has also been differentiated by specialty, with evidence of a major rise in demand for medical rather than surgical hospital care. This growth has occurred despite new emphasis on siting appropriate care in primary and community settings. Building on existing work relating to spatial perspectives on changing health demand, the present analysis develops a Bayesian approach to modeling the generation of health demand in a region and its allocation to health providers. At the first stage projections of acute demand are made by specialty, patient age and area of residence, with the allocation to providers then determined by a gravity model. A case study considers trends in health demand during the 1990s as a basis for forecasting during 2000-2010. The study region comprises North East London and South Essex. In this application, projections of specialty referral rates (usage rates) by age are based on national data, and are applied to regional population projections to give a forecast health demand. The allocation of this demand to hospitals then takes account of projected changes in the configuration of hospital beds in the region.