Our findings support a hypothesis that readmission risk is multifactorially determined and that interactions have to be considered.

Stratifying on gender, diagnostic group and hospital revealed a general pattern of associations except for age.

The problem addressed in this paper is how continuity of care is related to characteristics of psychiatric services, previous events in a patient's pattern of care and patient characteristics. The present paper is a part of a Nordic Comparative Study on Sectorized Psychiatry in seven catchment areas in four Nordic countries. One-year-treated-incidence cohorts were used. Each patient was followed for 1 year after the first contact with the psychiatric service. Continuity of care was measured by the time from discharge from hospital to the first subsequent day-patient or outpatient contact. Notable findings were large differences in the continuity of care in the seven services, high proportions of discharges without any aftercare contacts and long time lags between discharges and aftercare contacts in most of the catchment areas. A Cox regression analysis revealed that aftercare following hospitalisation seems to be more probable if the outpatient services are located geographically close to the patients, if the hospitalisation lasted between 2 and 4 weeks, if there was a community care contact shortly before the hospital admission and if the patient is not retired and not divorced. Staff resources were not related to continuity of care.

There was a great variation in rates of compulsory care. No consistent rural-urban pattern in rates of commitment was found. It is discussed whether a formal referral procedure to the psychiatric service is associated with higher rates of compulsory care.

In the present paper a sample of patients using psychiatric in-patient care only is characterized and analysed with regard to characteristics of the psychiatric services. This paper forms part of the Nordic Comparative Study on Sectorized Psychiatry, designed to investigate contact rates and use of psychiatric care by new patients in 7 catchment areas in 4 Nordic countries during a 1-year follow-up. One-year treated incidence cohorts were used. The logistic regression analysis revealed that the variable 'psychiatric service' was one of the statistically significant determinants of using only in-patient care during the follow-up. The diagnostic groups with the highest probability of using only in-patient care were dependencies and functional psychoses. The following factors were associated with a high risk of using only in-patient care: older age, being referred by another psychiatrist, having received previous psychiatric in-patient care, male sex, being retired, and not living with one's parents or a partner. Patients who used only in-patient care had fewer admissions and days in in-patient care than others during the 1-year follow-up period. Lack of 24-hour emergency services in out-patient care correlated positively with the use of only in-patient care.