Admission to an Italian Forensic Hospital (OPG) results in formal exit from psychiatric care provided by NHS community based psychiatric services. The length of stay in such facilities is often perceived as a factor negatively affecting the opportunity for reintegration in the community. Method – Factors predicting length of stay in OPG were investigated by means of a survival analysis carried out on a cohort of 118 inmates of three OPGs (Castiglione delle Stiviere, Reggio Emilia and Montelupo Fiorentino), who represent the whole forensic population from 3 different geographical areas at 30.06.97; all discharges occurred in the following 18 months were examined. Results – In survival analyses conducted on individual predictors, five variables predicted a longer stay: type of offence (homicide: 706.6 weeks vs. 307.1 for minor offences and 194.7 for grievous bodily harm, log-rank =31.8, p >0.001), type of admission (RR=0.98, CI 95% 0.97-0.99, p>0.001), the diagnosis of schizophrenia (621.9 weeks vs. 398.9 weeks or less for other diagnoses; log rank=9.08, df=3, p=0.028), BPRS thought disturbance score (RR=0.89, CI 98% 0.81-0.98, p>0.01), hospital of stay (314.6 weeks in Montelupo Fiorentino vs. 706.6 for Reggio Emilia and 621.9 for Castiglione delle Stiviere; log-rank = 9.64, df=2, p>0.001). In a Cox linear regression model three significant factors were selected: type of offence, type of admission, diagnosis of schizophrenia. Conclusions – Judicial factors are relevant in determining the length of stay in OPG. The diagnosis of schizophrenia seems to play an independent role in predicting a longer stay.