Risk assessment tools
Viljoen, Cochrane and Jonnson (2018) set out to measure rates of reoffending, by comparing the use and application of risk assessment and management tools. Due to study limitations they were not able to use conventional measures such as reoffending rates, as the sample was largely composed of individuals who have not committed any crimes i.e. were patients on a hospital ward. Instead, the review used the percentage of violent incidents for pre- and post-exposure to assessment and management tools (in hospital environments). Only two crime-related studies mentioned reoffending rates.
The review did not find enough high quality evidence with a reoffending outcome to under-take a meta-analysis, and therefore doesn’t establish whether the use of assessment tools is associated with reductions in reoffending.
What is it?
Risk assessment and management tools are utilised widely in the criminal justice system, primarily as a means of managing both risk and violent offending. Viljoen et al’s (2018) systematic review included both male and female individuals, from conventional prison populations as well as psychiatric settings, with tools being used to devise patient care plans and risk assessments. Many tools are now in their fourth generation, with over 400 assessment tools currently in practice and adopted by a range of professionals such as judges, jurors, to probation workers and police officers, they are regarded as the gold standard of offender management. The systematic review by Viljoen and colleagues is an to attempt to evaluate risk assessment tools’ assumed effectiveness, both in its predictive validity of risk and apparent reductions in aggression.
Should it work?
The current model (the risk-need-response model; RNR) is the dominant thinking governing the use of such assessment tools in offender treatment. Models that adhere to these three RNR principles have been shown to be more effective in reducing the rates of offending and correctly assigning levels of risk (Bonta & Andrews, 2017). Of the 34 assessment tools used, 28 were previously validated as reliable risk and assessment tools, in line with the three aspects – evaluating and mitigating risk, assessing their level of criminogenic need, and the correct response i.e. responding with an appropriate intervention. These tools are used to devise programmes that are both evidence-based, and also uniquely catered to that individual.
Does it work?
The systematic review found 1588 studies from initial searches, and after screening included 73 studies. 12 were suitable for reoffending comparisons and of these three were randomised controlled trials; the remaining nine were based on nonrandomised experimental groups.
As the included studies contained a wide range of designs, the researchers judged that a meta-analysis of their results was not feasible. The review also required a high level of descriptive detail i.e. their methodology (e.g. search terms, databases). The review then created a suitability table for each study, with all the research questions and specific outcomes, etc. but were too diverse to combine. The authors therefore decided to undertake a narrative synthesis, which highlighted some of the complications and difficulties practitioners faced in using of-fender assessment tools. It was clear there was tremendous variation in both their type, application and the professionals themselves who were utilising them. In each study the assessment tools’ effectiveness was evaluated, in reducing violence/offending, correct implementation and judging appropriate level of offender risk. These tool evaluations were scored on a 1-10 scale, with overall ratings of low, moderate and high being based on this scale. The results of this analysis were not consistent across the studies, ranging from low to high.
How strong is the evidence?
The author expressed concerns throughout the review concerning methodological limitations, that ‘there was insufficient evidence from nonrandomized studies to conclude that risk as-sessment tools reduce violence and offending’ (Viljoen et al., 2018, p.20). Broadly speaking, this review attempted to evaluate the professional perceptions of the tool’s utility, how much they guided decisions and risk assessments, if they matched up with the RNR principles, and what, if any, tangible effect there was on reductions of violence. In all domains the results were ‘mixed’ with little consistency. There were few offending related studies that utilised comparison groups; of those that did, the review found many of the experimental groups had higher baseline levels of aggression than the control groups. For example, in the three RCTs that measured reoffending, adherence to the use of the tool was rated by the authors as ‘unclear’. The settings of the studies themselves, prison and psychiatric populations wards are also not directly comparable and may have considerable differences in how these assessment tools were used.
Is it worth it?
The authors did not include a cost and benefit analysis in their systematic review. However, the authors mentioned the high cost of implementing offender management tools, and that these are likely to be too heterogeneous to calculate across studies.
Can it be implemented?
There is clear evidence for the continued use and implementations of such devices to manage and reduce reoffending. However, many notable points were raised by the systematic review, concerning offender assessment difficulties and their correct implementation going forward. Viljoen and colleague address that the scope of these tools appears to be expanding such as, in helping judges decide between different sentence types or parole board rulings. There are significant differences in how risk assessment tools are seen and used by different professionals, which will need to be clarified if these tools are to be correctly used and have the maxi-mum effect upon rates of recidivism.
What's missing from the evidence?
Wide variations in the findings of the studies reviewed suggest that a number of unknown factors (e.g. other interventions running alongside assessment tools) may be having an effect on the study outcomes. It was clear that the evaluators did not anticipate the lack of rigorous studies necessary to make an earnest comparison of risk assessment and offender reduction tools. What is missing is a significant number of RCT trials as the basis for making causal inferences about the effect assessment tools on reoffending. Should a standardised programme of trials be established, then the tools themselves can be compared, to observe their effectiveness, with their unique differences in mind (e.g. dynamic versus static criminogenic elements), to clarify which factors are more salient and also what context-dependent components matter in offender management.