Building a bridge between assessment and practice: Risk assessment tool and RNR principles adherence in probation case plans

Heather L. Dyck, Glenrose Rehabilitation Hospital, Edmonton, Alberta, Canada and University of New Brunswick—Fredericton
Mary Ann Campbell, University of New Brunswick—Saint John
Julie L. Wershler, University of New Brunswick—Fredericton

The risk–need–responsivity model (RNR; Bonta & Andrews, 2017) has become a leading approach for effective offender case management, but field tests of this model are still required. The present study first assessed the predictive validity of the RNR-informed Level of Service/Case Management Inventory (LS/CMI; Andrews, Bonta, & Wormith, 2004) with a sample of Atlantic Canadian male and female community-supervised provincial offenders (N = 136). Next, the case management plans prepared from these LS/CMI results were analyzed for adherence to the principles of risk, need, and responsivity. As expected, the LS/CMI was a strong predictor of general recidivism for both males (area under the curve = .75, 95% confidence interval [.66, .85]), and especially females (area under the curve = .94, 95% confidence interval [.84, 1.00]), over an average 3.42-year follow-up period. The LS/CMI was predictive of time to recidivism, with lower risk cases taking longer to reoffend than higher risk cases. Despite the robust predictive validity of the LS/CMI, case management plans developed by probation officers generally reflected poor adherence to the RNR principles. These findings highlight the need for better training on how to transfer risk appraisal information from valid risk tools to case plans to better meet the best-practice principles of risk, need, and responsivity for criminal behavior risk reduction.

risk-need-responsivity model, case management, Level of Service/Case Management Inventory, best practices

“The risk–need–responsivity (RNR) model provides guidance for effective offender risk assessment and case management. This model reflects an integrated theory of criminal behavior drawn from personality, cognitive, and social learning approaches and contains three foundational principles: the risk, need, and responsivity principles. The risk principle states that supervision and intervention intensity should match the individual’s recidivism risk (more intensive services for high-risk offenders). The need principle calls for the identification, and targeted intervention, of criminogenic needs (e.g., procriminal thinking, substance use, family problems) driving the criminal behavior. The responsivity principle outlines guidelines for how to provide intervention services; namely by means of cognitive-behavioral skill building techniques rooted in social learning theory that are individualized to match the offender’s characteristics (e.g., age, ethnicity, learning style, motivation).” (p. 258)

“Generally, correctional programs and case plans that more strongly adhere to the RNR model show decreased levels of recidivism in males and females, youth and minority offenders, and in community and custodial settings with recidivism reductions of 10–50%. Stronger adherence to the RNR model is associated with decreases in substance abuse relapses and a variety of criminal behaviors, including nonviolent, violent, gang-related, and sexual offenses. […] Research has demonstrated that the RNR model is cost effective when compared to traditional sanctions for criminal behavior, with a $2 cost for each 1% decrease in recidivism compared with $40 for each 1% decrease when using sanctions. These results indicate that correctional services that implement a RNR model framework can significantly reduce recidivism without a significant increase in cost.” (p. 259)

“A key aspect of the RNR model and evidence-based case management is the use of risk assessment measures. Using a risk assessment instrument to inform decision-making increases the agreement in case management plans across probation officers. Contemporary risk tools are advancing into a framework that integrates the risk appraisal process with case management planning and progress assessments until discharge. The Level of Service/Case Management Inventory (LS/CMI) is one instrument that aligns with this integrative case management process. The LS/CMI is designed to assist professionals in justice, forensic, correctional and crime prevention agencies with the management, supervision, and case planning of offenders over the age of 16 years. The LS/CMI was developed as an extension of the Level of Service Inventory– Revised (LSI-R). […] The LS/CMI has demonstrated moderate to high predictive validity for general recidivism with male and female offenders in both incarcerated and community-based settings, and is moderately predictive of violence.” (p. 259)

“Despite the importance of evidence-based assessment for informing how best to reduce recidivism risk, risk assessment is not without its criticisms. Some scholars have noted the inherent challenges with holding someone accountable for something they “might” do in the future, for the failure of group data to apply to individuals and the risk of harm of such approaches to marginalized groups, for the failure of group data to apply to individuals and the risk of harm of such approaches to marginalized groups, and greater need for culturally informed risk assessment. Thus, the value of risk assessment must be contextualized within these concerns.” (p. 259)

“Real-world use of risk instruments and their integration into case planning is often less than ideal. […] Challenges with RNR adherence may depend on the quality of the risk assessment tool used to inform the case plan. […] Given that the RNR model was foundational to the practices of the community supervision agency assessed in the current study, we took the opportunity to examine the degree of RNR adherence in their case management plans. These plans were informed, as per policy, by completion of the LS/CMI. Thus, consistent with past research, we expected the LS/CMI to have strong predictive validity for general recidivism for males and females. Furthermore, in accordance with the RNR model, we predicted that high-risk cases would receive higher intensity intervention and support services (e.g., more sessions, longer duration) than lower risk offenders. For all offenders, regardless of risk level, it was predicted that offenders with LS/CMI identified criminogenic needs would be referred to services that appropriately targeted these needs. Finally, it was predicted that the intervention services to which offenders were referred would follow the responsivity principle— use of effective intervention strategies for offenders (e.g., cognitive–behavioral strategies) and tailored delivery to the client’s strengths and weaknesses (e.g., motivation level, cognitive ability).” (p. 259–260)

“Data were drawn from case files generated between 2007 and 2012 by a provincial community correctional service in Atlantic Canada. Cases were randomly selected for inclusion, resulting in a sample of 136 offenders (101 male, 35 female; M age = 30.0 years, SD = 9.86; age range 18.05 – 70.07 years). Most (80.1%) were sentenced to probation, 9.6% were serving a conditional sentence, 2.2% were on house arrest, and 8.1% were on a combination of community supervision types. Thirty-one percent had a preindex history of criminal behavior. The sample was primarily Caucasian (93.4%), with the remaining 6.6% comprised of African Canadian, Asian, and Indigenous persons. A majority (78.7%) completed high school and/or had more advanced training/education.[…] The total sample available for pre–post comparisons in LS/CMI scores was 78 [cases].” (p. 260)

Materials included Level of Service/Case Management Inventory (LS/CMI), Adherence to the risk–need–responsivity (RNR) model (higher intensity treatment for higher risk individuals, primary treatment goals were related to identified criminogenic needs, case plan was tailored to individual’s strengths and limitations and when evidence-based interventions were used), and Index offense and recidivism data.

“Consistent with past research, the LS/CMI was a valid predictor of general recidivism for both male and female offenders in the current study. As expected, LS/CMI-identified high-risk cases reoffended at a significantly faster and higher rate than low-risk cases. Thus, the LS/CMI is useful for appraising recidivism risk in community supervised adult offenders. […] LS/CMI total scores did not significantly change from intake to the time of reassessment for the subgroup of cases for which such data were available. Reassessment cases did not differ from clients without reassessment on age, gender or LS/CMI intake risk scores. Nevertheless, being reassessed for risk was associated with a lower risk of general recidivism (52.6%) relative to clients only assessed at intake (67.2%). Thus, although risk scores did not change, the rate of reoffending may have been impacted by the reassessment. Although causation cannot be inferred, reassessment offers the opportunity to adjust case plans to mitigate the real-world expression of the identified risk (i.e., better risk management).” (p. 264)

“Despite use of a valid risk assessment instrument, the present study’s data did reflect challenges with the transfer of assessment information into RNR-informed case management planning. Notably, only 44.1% of reviewed plans met the risk principle and only 59.6% met the need principle. Other studies also identify problems with the translation of risk assessment information into RNR-informed case plans. Collectively, these results highlight the need to provide better training to facilitate the integration of these two processes. When stronger adherence to the RNR model was achieved, the rate of general recidivism was lower in the current study, as has been found by others. Adhering to both the risk and need principles in the current sample was associated with a 26.5% rate of reoffending, whereas this rate was 60.6% when neither of these two core principles were met. However, it was the lower risk case files that received the strongest RNR adherence ratings, which may bias interpretation of these differential recidivism outcomes.” (p. 265)

“For need principle adherence, there was little indication of overintervening with low-risk offenders. The exceptions were in the criminogenic need domains of employment/education and family/marital issues where interventions tended to be provided despite being assessed as low need areas. Probation officers may have viewed these areas as strengths to build on, but this is not clear from the current data. The importance of incorporating assessment of individual strengths or protective factors into risk assessment and case planning is increasingly being recognized and advocated for in the RNR model. Although some probation officers may have considered protective factors, it was not clearly documented. In addition, the way in which individual strengths were incorporated into case plans was likely inconsistent across probation officers without a structured protocol to inform which factors to consider and how to integrate them into the case plan.” (p. 265)

“In line with RNR recommended practice, high risk offenders in the current study were referred to a significantly greater number of services than low-risk offenders.[…] It is the quality of these services that matter more to risk/need adherence than the quantity of services per se. […] We found that high-risk cases were less likely to have their criminogenic needs targeted for intervention relative to low-risk cases. It may be that probation officers had difficulty teasing apart criminogenic and noncriminogenic interventions when working with higher risk cases, and the criminogenic focused interventions may not have been prioritized appropriately. Understanding of the process of intervention and how to target criminogenic needs in supervision meetings is a new responsibility for probation officers.” (p. 265)

“The responsivity principle is the most understudied component of the three main principles of the RNR approach, but is generally viewed as a valuable component. Unfortunately, the responsivity principle could not be meaningfully examined in the present study due to the condition of the case files. Limited information relating to the responsivity principle is consistent with previous community research.” (p. 265)

“The present study provides additional evidence for the use of the LS/CMI in community-supervision environments, especially in Atlantic Canadian populations. Furthermore, it highlights the difficulty with real-world application of the RNR model. The present study also provides evidence of challenges probation officers may experience when asked to alter their roles from a focus on supervision to a therapeutic/intervention focus. Field training with ongoing mentorship and fidelity checks is essential for ensuring appropriate implementation of the RNR model.” (p. 266)

“Transitioning from a role in which supervision of offenders was the primary task into one that balances supervision and intervention, as advocated by the RNR model, is one that takes time for employees to achieve.” (p. 265)

“Changes in probation officers’ approach to intervention are expected to be gradual as they adapt to a new system of practice. It is through the integration of the assessment and rehabilitation processes that the strongest outcome results are obtained. […] Quality field training with the application of the RNR model is essential and should include educational workshops about the model, why adherence to it matters for risk reduction and enhanced client outcomes, and intervention skill learning (e.g., cognitive restructuring, motivational interviewing). Over time, individual mentorship, booster sessions and skill upgrading will be key as the field of evolves. The Strategic Training Initiative in Community Corrections model is an example of this type of training, with effective results for reducing criminal risk in the clients supervised by probation officers trained in this manner relative to clients supervised by untrained probation officers. Further evaluation of such training models is needed.” (p. 265–266)

“A major strength of the present study was its use of offender case files that reflect the real-time practices and implementation of the RNR framework. However, using data gathered from records maintained by professionals without a focus on research has its limitations. First, although Section 1 of the LS/CMI had strong predictive validity for general recidivism events and time to first recidivism, the majority of files did not have the other LS/CMI sections completed despite being intended to inform case management and supervision practices. […] Furthermore, many files did not have a follow-up LS/CMI completed or any information on the program(s) the offender was referred to (e.g., intervention orientation, main intervention goals). […] Examining RNR adherence with a larger sample size may reveal greater nuances in practices and limitations, including the value of reassessment. Another limitation of the current study was the limited information available for the offender rehabilitation programs. […] Finally, it should be acknowledged that the sample used in the current study was modest relative to other risk prediction studies using the LS/CMI which tend to exceed 500+ cases for analyses, often into the thousands. Although our smaller sample may influence the generalizability and reliability of the current findings, our risk prediction and RNR adherence results are fairly consistent with the broader research literature on the LS/CMI and RNR adherence use in the field.” (p. 266)

“Future research should focus on the integrity of implementing the RNR principles in real world case management as these principles, and more widely the RNR model, become foundational for more organizations. The most fruitful research may come in the form of full prospective program evaluations during times of organizational transition and policy changes.” (p. 266)

As always, please join the discussion below if you have thoughts or comments to add!