RESUMEN
BACKGROUND: A new forensic mental health law was enacted in Japan in 2003, enabling development of specialist services. Before their establishment, it is important to determine the nature, frequency and correlates of the problems they are designed to ameliorate. AIMS: To establish rates of violent recidivism among mentally disordered offenders before the new legislation, and examine associated risk factors. METHOD: Data were extracted from one complete annual (1980) national cohort of people judged by the Court or prosecutor to be without responsibility for a criminal offence, or of sufficiently diminished responsibility for it to be diverted for psychiatric treatment. The outcome measure was violent recidivism after community discharge and before 1991. RESULTS: Fifty-two (10%) of 489 in the cohort were arrested or convicted of further violent offences. Violent recidivism was most strongly associated with a substance-related disorder, but histories of violence, homelessness and short index admissions were independently related. CONCLUSIONS: Violent recidivism was so unusual that, on this outcome, it could take many years to show any effect of the new service. Desistance from substance use, compliance with treatment and maintenance of stable housing may be better indicators of success, and their achievement a good preventive strategy.