Data were collected from the hospital regarding those whose psychiatric admission was their first, as well as from criminal justice records. The follow-up period ranged from 6 months to nearly 14 years and averaged 6.6 years. The locations of 209 participants were known at the end of the study on October 31, 1994. Twenty-four percent of the participants had at least 1 subsequent conviction; 12 percent were convicted of a serious offense. The 56 recidivists were responsible for 267 reconvictions; violent and sex offenses were overrepresented in these reconvictions. Findings suggested that these figures were an underestimate of the true rate of recidivism, because many participants spent a large proportion of the follow-up period in an institution and thus had reduced opportunities for further offending. The rates were broadly comparable to those emerging from other follow-up studies of forensic psychiatric individuals, given factor and allowing for methodological differences. In addition, these recidivism rates were much lower than recidivism rates for ordinary prisoners. The analysis concluded that recidivism rates are of little use in measuring outcomes in forensic psychiatric populations. Instead, health-based outcome measures are needed that include the occurrence or violence or other offensive behavior and that do not rely on the need for a criminal conviction as an indicator of recidivism. Tables and 25 references (Author abstract modified)