A retrospective review of all adult medical in-patient hospital record
s in one health district (adult population 378 000) was performed over
three consecutive years. Yearly age-standardised rates for emergency
admissions were calculated and compared between sections of the popula
tion with differing access to hospital beds. Confounding and other exp
lanatory variables were examined with a logistic regression model. Eme
rgency medical admission rates were consistently higher in the populat
ion whose general practitioner had access to community hospital beds,
as compared with those whose general practitioner had no access, (46.1
per thousand population vs 39.3 per thousand in the year 1994-95, dif
ference significance, P < 0.05). Multivariable analysis suggests that
in addition to supply factors, age, sex, morbidity and socio-economic
circumstance influence admission rates.