Factors which influence the reconviction after release from medical prison were investigated, including psychiatric factors, demographics, criminal career, type of crime, behavioural problems, etc. Subjects were 145 people who were released from medical prison. After release these subjects were followed during a period of 2 to 7 years. By using Relative Risk Trees for Censored Survival Data Cox Regression, the following nine risk factors were revealed to be significant predictors of reconviction, age of first offence (less than 25.5 years), the lack of need for treatment in special seclusion room, the individual's experience within a criminal institution, anxiolytics medication, maintaining insight, verbal threat perpetuated by individual while in prison, previous illegal drug offenses (volatile substance), age of current offence (exceeds 29.5 years of age), and offences committed in the period of a suspended sentence. This type of research has scarcely been done in Japan until now, the results of this research will be useful to select the candidate for the inmate reeducation program.
