SYNOPSISEnvironmental as well as individual socio-demographic and illness characteristics are related to the risk of admission. This paper addresses the problem of the interrelationships of these factors to admission rates. Using the Groningen Psychiatric Case Register, admission rates (during 1986 and 1987) from 34 administrative areas were calculated. Logit models were fitted in order to test the relationship between the relative risk of being admitted and sex, age, marital status, diagnosis, urbanization and distance from facilities. The effect of urbanization remains under the control of the other independent variables. The concept of ‘need for care’, related to ‘true’ and ‘treated’ incidence, is discussed.