Data from the Oxford Record Linkage Study between 1976 and 1986 were used to analyse statistical profiles of hospital care, taking account of multiple admissions per patient. Admission rates were higher in the elderly than the young, they were higher for men than women, and they increased over time. Episode-based admission rates increased by 15.5 per cent and person-based rates by 11.6 per cent. Most of the increase therefore represented a real rise in the number of people treated rather than an increase in multiple admissions per person. In the latest year of the study there was a mean of 125 admissions per annum per 100 male patients admitted and of 118 admissions per annum per 100 female patients admitted. The mean length of stay per admission and the total time spent in hospital per person per annum declined substantially from 1976 to 1986. Admission rates for prostatectomy, colectomy and varicose vein surgery increased significantly and those for appendicectomy, cholecystectomy and peptic ulcer decreased significantly.