Socio-economic factors that predict psychiatric admissions at a local level

Abstract

Background: The aim of the study was to confirm the predictive relationship between socio-economic factors and psychiatric admissions at a fine grain geographical level. The strength of association was compared with those of other studies that have looked at separate diagnostic groups.

Method: Psychiatric admissions were from electoral wards of the County of South Glamorgan, which encompasses the capital city of Wales, Cardiff. Standardized psychiatric admission ratios (SAR) for different diagnostic groups were calculated for a 5-year period. The ecological association with deprivation indices and with single variables at the level of electoral ward was examined. Of a total of 15266 psychiatric admissions, 11296 were analysed.

Results: Psychiatric morbidity, reflected in SAR was inversely related to socio-economic deprivation for both sexes. This applied to all diagnostic groups except organic disorders. The relationship was most marked for schizophrenia, delusional disorders and substance abuse, closely followed by personality disorders, and less for affective and neurotic disorders. Little difference existed between three composite indices of deprivation (Carstairs, Jarman, Townsend), but the marginally best predictor was that designed by Jarman. However, low rates of car ownership and high unemployment were as good at predicting SAR as any of the compound indices.

Conclusion: Socio-economic factors account for almost 50 % of the variance in psychiatric admission rates between electoral wards. The degree of association between psychiatric morbidity and deprivation varies between diagnostic groups, arguing against a common factor linking deprivation and psychiatric admissions generally. Frequently updated unemployment figures provide nearly as useful and more immediate information than 10-yearly Census data used to calculate the deprivation indices. These figures may be used for needs assessment and targeting resources at a local level.