A retrospective study of factors predicting readmissions and follow-up treatment was undertaken of all first-ever episodes of inpatient care (age under 65), excluding psychotic and organic mental disorders, during 1987 and 1988 in University Psychiatric Clinic in Turku City Hospital, Finland. The cohort consisted of 64 subjects, 24 men and 40 women. The study was carried out in the end of 1993, thus allowing about 5 years of follow-up. The diagnosis of personality disorder did not predict readmission. The only factor predicting readmission nearly significantly was not having a relationship. The incidence of the revolving door syndrome, defined as 4 or more admissions within 5 years, was 12.5%. Women had a greater risk of readmission, but not that of the revolving door. Patients who had psychotherapy as follow-up treatment showed a frequency of 8% for 4 or more admissions, whereas patients who had no follow-up treatment had a frequency of 21% for 4 or more readmissions. The only factor significantly predicting follow-up treatment arrangement was previous treatment contact.