type:text

[要約] 2018年に法務省は，Risk-Need-Responsivity モデルを基盤に，保護観察中の人の保険統計的再犯リスク，動的再犯誘発要因及び動的保護要因のアセスメントツールを含むCase Formulation in Probation/Parole (CFP) を開発した。本研究は，CFPと再犯の静的リスク要因を査定する既存のアセスメントツールの再犯予測力を検証した。2018年10月から2019年３月までに保護観察となった655人が既存アセスメントツール実施群とCFP実施群に無作為に分けられた。法定保護観察期間が2020年５月末を超える485人（既存アセスメントツール実施群205人，CFP実施群280人；法的性質が異なる刑事施設仮釈放者と保護観察付一部猶予者を除く）を分析対象とした。同月末までの再犯率は12.0%だった。再犯を従属変数としてCox回帰分析を行った。既存アセスメントツール実施群は静的リスク得点，CFP実施群は保険統計的再犯リスク，動的再犯誘発要因及び動的保護要因の得点を独立変数にした。CFPの動的再犯誘発要因得点は再犯の有意な正の要因であり，動的保護要因は負の有意傾向を示した。CFPがより適切なアセスメントツールと考えられた。

[SUMMARY] The Ministry of Justice in Japan developed the Case Formulation in Probation/Parole (CFP) based on the Risk-Need-Responsivity model in 2018, which includes tools to assess actuarial risk for recidivism, dynamic criminogenic needs, and dynamic protective factors among individuals under supervision. This study examined CFP’s predictive ability of recidivism, as well as that of the current assessment tools designed to assess only static risk factors for reoffending. Six hundred fifty-five individuals placed under supervision from October 2018 to March 2019 in Japan were randomly allocated to either Group A (assessed by the current tools) or Group B (assessed by CFP). From the total number, data from 485 participants with supervision periods ending after May 2020 were analyzed in this study; we excluded data from individuals put under supervision after release from prisons because their legal status and supervision systems for them are different. The participants’ recidivism rate at the end of May 2020 was 12.0%. Cox regression analysis was performed using recidivism as a dependent variable. The static risk scores were used as independent variables for Group A participants, and the actuarial risk, dynamic criminogenic needs, and dynamic protective factor scores were used for Group B participants. The results showed that the dynamic criminogenic needs scores were significantly and positively associated with recidivism. Additionally, there was a marginally significant negative association between dynamic protective factor scores and recidivism. In conclusion, CFP showed a better predictive ability of recidivism for individuals under supervision and was considered more applicable than the current assessment tools.