Dynamic Risk Scales Degrade Over Time: Evidence for Reassessments.

Area of Science:

• Criminology
• Forensic Psychology
• Risk Assessment

Background:

• Dynamic risk assessment tools are crucial for evaluating recidivism risk in justice-involved populations.
• Frequent reassessments are recommended, but empirical data on their impact on predictive accuracy over time is limited.
• The ACUTE-2007 and STABLE-2007 are widely used tools for assessing sexual recidivism risk.

Purpose of the Study:

• To investigate how the predictive accuracy of dynamic risk assessments, specifically the ACUTE-2007 and STABLE-2007, changes over time.
• To determine the effect of reassessment frequency on the accuracy of sexual recidivism risk predictions.
• To provide evidence-based recommendations for the optimal timing of risk assessment updates.

Main Methods:

• Utilized two independent samples of men on community supervision (NStudy 1 = 795; NStudy 2 = 4,221).
• Analyzed the predictive accuracy of the ACUTE-2007 and STABLE-2007 risk tools across different time intervals.
• Examined outcomes including sexual, violent, and any recidivism (including technical violations).

Main Results:

• Reassessments consistently improved predictive accuracy for all recidivism outcomes.
• The most recent assessment, temporally closest to the recidivism event, demonstrated the greatest predictive power.
• The findings indicate a decay in predictive accuracy over time without reassessment.

Conclusions:

• Regular reassessments are vital for maintaining the accuracy of dynamic risk tools.
• Recommended reassessment intervals: ACUTE-2007 every 30 days; STABLE-2007 every 6 months or post-significant life changes.
• These findings support timely updates to risk assessments to enhance public safety and inform correctional interventions.