Abstract
Few studies have assessed acute dynamic risk repeatedly among paroled offenders to investigate the relationship between changes in acute dynamic risk and recidivism in crime. The present study investigates whether one-month changes in ten stress-related acute dynamic risk factors, collected through automated telephony while the participants were still in prison and over 30 consecutive days following parole, predict one-year criminal recidivism, including its predictive accuracy. The study also investigates whether a brief feedback intervention in conjunction with the daily assessments reduces recidivism compared to an assessment-only control group. Changes in five risk factors were found to be associated with increased risk of criminal recidivism after controlling for the results in prison, the initial value after parole, and the intervention. The predictive accuracy is marginally accurate: Summary score (AUC) = .666; Level of stress (AHSS) = .644; Psychiatric symptoms (SCL-8D) = .641; Anxiety symptoms = .673; Severity of most stressful daily event = .690. No differences in one-year recidivism rates were established between the intervention group and the control group. The study shows that daily assessments can usefully be made of dynamic risk factors in paroled offenders.
Acknowledgements
For participation we would like to acknowledge all participants in the present study. For the programming of the automated telephone system, we would like to acknowledge Milo Fryling, TeleSage, Chapel Hill, NC, USA. For statistical advice we would also like to acknowledge Håkan Löfkvist, PhD/medical statistician, Unit for Medial Statistics and Epidemiology, Skåne University Hospital, Sweden, and Ingvar Rosendahl, PhD, Karolinska Institutet, Department of Clinical Neuroscience, Centre for Psychiatric Research, Stockholm, Sweden.
Disclosure Statement
No potential conflict of interest was reported by the authors.
Notes
1. Parole refers to the mandatory release and community supervision of a prisoner once two thirds of his prison sentence has been served.
2. This study follows consensus-based guidelines for what should be routinely reported in risk assessment research (Singh, Yang, Mulvey, & the RAGEE Group, Citation2015).
3. The number of parameters that a growth model can estimate is one less than the number of occasions (n − 1). In a linear model with one covariate there will be three estimates – one slope, one intercept, and the covariate – so at least four points in time are needed.