Following the abandonment of geographical restrictions to admission, case records of all patients admitted to a geriatric unit during a six month period were examined. Over one quarter of admissions came from outside the catchment area. These patients were indistinguishable from those residing within the catchment area with regard to age, problem precipitating admission, length of stay, readmission rate, mortality and destination on discharge. Cross border referrals were more likely to be admitted directly from the general practitioner. There was no evidence to suggest that an unrestricted admissions policy selectively attracts 'problem patients' for whom designated services are inadequate. There were marked differences in the mix of problems arriving under the care of individual consultants. Junior staff produce a greater selective effect on the type of patient a consultant sees than that produced by area of residence of mode of admission.
