POPULATION PREDICTORS OF COMMUNITY-HEALTH AND SOCIAL-SERVICE USE IN NORTHERN-IRELAND

Study objective-To investigate the characteristics of elderly populati ons associated with variations in their use of community health and pe rsonal social services and to test the hypotheses that the variations are related to: (a) the age structure of an elderly population; (b) th e population's socioeconomic composition, including the level of depri vation; and (c) household or living arrangements. Design-A common file of 1991 population census and 1994 NHS community trust operational va riables was constructed for 67 postcode sectors, with the independent variables describing the age-sex groups to be studied. Clear criteria for the exclusion of ''empty'' sectors were developed. Relationships u sing bivariate and multivariate correlation and stepwise multiple regr ession were explored. Setting-Eastern Health and Social Services Board area, Northern Ireland (Belfast and hinterland). Participants-Populat ion of statutory pensionable age; in aggregate, younger and older age bands. Main results-The age structure or mean age of the elderly popul ation had only a weak association with the community health and social service client rate, but there were strong associations with socioeco nomic variables, particularly the percentage of those living alone who were without a car and the percentage of pensioner households that in cluded an adult of below pensionable age. Parsimonious multiple regres sion models accounted for between 46% and 80% of the variation in the NHS community trust client rate. Greater explanations were achieved fo r the young elderly population than for those aged 75+ years and, when the population was divided between young and old age bands, for men t han for women. Conclusions-Community health and social services for el derly people in eastern Northern Ireland were focused on those with a low income and those who were not co-resident with adults of working a ge. When local elderly populations are compared, per capita morbidity and dependency are often higher where the mean age is low, and vice ve rsa, because of the inverse relationship between socioeconomic status and survival in old age. Capitation scales for resource allocation wit h positive age weighting will be of little use if no account is taken of the relative prevalence of need in the youngest or base age group.