DIFFERENCES IN-HOSPITAL CASEMIX, AND THE RELATIONSHIP BETWEEN CASEMIXAND HOSPITAL COSTS

Background The aim of the study was to examine the relationship betwee n hospital costs and casemix, and after adjustment for casemix differe nces, between cost and institutional size, number of specialties, occu pancy and teaching status. Methods A retrospective analysis of all adm issions to nine acute-care NHS hospitals in the Oxford region during t he 1991-1992 financial year was undertaken. All episodes were assigned to a diagnosis-related group (DRG) and a cost weight assigned accordi ngly. Costs per finished consultant episode, before and after adjustme nt for casemix differences, were analysed at the hospital and specialt y level. Results Casemix differences were significant, and accounted f or approximately 77 per cent of the difference in costs between provid ers. Costs per casemix-adjusted episode were not significantly associa ted with differences in hospital size, scope, occupancy levels or teac hing status, but sample size was insufficient to investigate these rel ationships adequately. Specialty costs were poorly correlated with spe cialty casemix. This was probably due to poor apportionment of special ty costs in hospital accounting returns. Conclusions Casemix differenc es need to be taken into account when comparing providers for the purp oses of contracting, as unadjusted unit costs may be misleading. Altho ugh the methods used may currently be applied to most NHS hospitals, w idespread use would be greatly facilitated by the development of indig enous cost weights and better routine hospital data coding and collect ion.