The Level of Service/Case Management Inventory (LS/CMI) is one of the Level of Service Inventory (LSI) instruments. It is an example of a fourth generation offender risk/need assessment scale and is distinguished from previous generation instruments by the fact that it was not only designed to predict offender recidivism but also includes a built-in case management component that translates the need portion of the assessment to an active intervention plan for offender supervision, treatment, and programming. Like the LSI-Revised (LSI-R), the LS/CMI was pilot-tested on a large scale for almost a decade prior to its general release. This article describes the origins of the LS/ CMI, places it historically in the sequence of LSI instruments, itemizes the differences between it and its predecessors, and summarizes the research that it has generated.
History and Evolution
While the LSI-R was enjoying widespread success and was being implemented in Canada, the United States, and numerous European and Asian countries in the mid-1990s, it became evident to the authors, researchers, administrators, and end users that it could be expanded. Specifically, there was a growing sense that the scope of the assessment should be broadened and its linkage to offender supervision and intervention should be strengthened.
Consequently, Donald Andrews, along with James Bonta and Stephen Wormith, began a systematic review of the research and practical experience that had accumulated by that time. There was already extensive, system-wide experience with the LSI-VI (later known as the LSI-R) in Ontario, where correctional authorities were considering extending its use into its provincial prisons. They analyzed LSI assessment and recidivism data provided by the Ontario government. They also reviewed the early and rather crude, by today’s standards, meta-analyses on offender’s risk factors and recidivism. This review coincided with the emergence of the Central Eight domains of offender’s risk/need factors and helped to reorganize and reformat the next version of the LSI.
The authors also consulted with LSI assessors, primarily probation officers and psychologists, about their clinical experiences with the offenders they had assessed. They asked what more should the LSI provide to users in the field. They met with various levels of correctional officials, from senior administrators responsible for corporate policy to staff supervisors in the field, professional associations, and the provincial parole board. Routinely, they asked, what more would the various stakeholder groups like to see in the instrument to improve it, to ease its use, and to increase its value to the organization.
In 1995, the LSI: Ontario Revision (LSI-OR) was introduced into Ontario’s provincial correctional services, in both its probation offices and its correctional institutions. Then, in 2004, the LS/ CMI was released for more general use by correctional agencies around the world. It included organizational and stylistic modifications from the original LSI-OR, but none of the actual items in the core risk/need assessment portion were changed. In other words, as was the case with the LSI-VI and the LSI-R, the LSI-OR served as a thoroughly monitored and researched precursor to the LS/CMI.
Structure and Organization
Research and consultation led to nine critical enhancements to the LSI-R that are found in the LS/CMI. The first is a reduction of items from 53 to 43 that were grouped into eight, instead of the original 10, domains or subscales in the general risk/need section of the LS/CMI. These eight domains represent the Central Eight risk factors for offender recidivism that have emerged from various meta-analyses. They include criminal history, employment/education, leisure/recreation, family/marital, companions, criminal attitudes, substance abuse, and antisocial pattern. The accommodation and financial subscales from LSI-R were eliminated as were some redundant items. The emotional/personal section in the LSI-R was changed to a new domain called antisocial pattern. The antisocial pattern domain is somewhat different than the other subsections. It is also reminiscent of Robert Hare’s criminal versatility in the Psychopathy Checklist–Revised but is not limited to criminal behavior per se, but a general lifestyle, including antisocial thoughts, actions, and personality. As such, many of the concepts that are captured in its items come from the other Central Eight domains in the general risk/need section of the LS/CMI.
The redesign of the general risk/need factor section enabled a second new feature of the LS/CMI. Greater attention is given to the eight subsections of the LS/CMI than was ever afforded to the 10 categories of the LSI-R. For example, the subsections are scored and graphed in the same way that one plots a profile of subscale scores on many of the standard measures of intelligence and personality. The resulting graph allows the practitioner to visually inspect the results of the general risk/need assessment along the Central Eight risk factors and to make the links with case management, particularly programming and supervision, more easily.
A third new aspect of the LS/CMI is that the number of risk factors was increased to five (very low, low, medium, high, and very high) from the three or four risk levels in various representations of the LSI-R. It should be noted that the number of risk levels in any scale or instrument is decided arbitrarily by the developer or the agency using the instrument. The number of risk levels of a risk scale usually depends on the developers’ confidence in the instrument’s ability to discriminate between groups on the basis of small differences in scores. An accurate scheme with few levels of risk classification essentially loses some of its precision by collapsing offenders into larger categories. Since numerous LSI studies have produced generally a linear relationship between the LSI score and probability of recidivism, a five-level risk scheme was used so that the decision maker or case manager would be working with a more accurate description of the offender in terms of his or her risk to reoffend. The five levels of risk were also applied to the eight risk/need domains, thus enabling the graphic risk/need profile described earlier.
A fourth enhancement found in the LS/CMI was the introduction of offender strengths. One of the criticisms of traditional offender risk assessment instruments, including the LSI-R, was that they focus exclusively on risk, or the negative characteristics demonstrated by the offender, and ignore any positive attributes that he or she might present. The concept of strengths, characteristics that might serve as protective factors from a lifestyle of criminality, was introduced. Any of the risk/need domains may be declared as a strength if the assessor believes that a particular domain represents an area in which correctional workers could use it when taking on some of the more problematic need areas in the offender’s profile or could focus on it to build offender motivation, confidence, and rapport with the offender. The presence of strengths may also be used to adjust the offender’s overall risk rating if it is believed that a particular strength may mitigate a corresponding risk/need, such as a strong, prosocial parental role model neutralizing criminal acquaintances. These aspects of the LS/CMI are often overlooked by critics who characterize it strictly as a risk/need assessment.
A fifth enhancement of the LS/CMI is that it adds a supplementary list of specific risk/need items that might be considered in an offender’s overall rating and profile. This addition was a result of input from clinical professionals and other practitioners who were concerned about specific features of an offender that might not be captured in the Central Eight risk/need domains and their 43 items. The specific risk/need factors section consisted of 21 items grouped into two subsections, personal problems with criminogenic potential and history of perpetration. Many of these items, particularly in the second subsection, were designed to capture items that might relate to risk of violent recidivism including sexual recidivism. These items are not scored or added quantitatively. Rather, they serve as a checklist of possible problems that should be considered, both in assessing or adjusting risk and in case planning. Nonetheless, studies have demonstrated that the number of items endorsed in this section is positively correlated with recidivism, including violent recidivism.
A sixth enhancement relates to the manner by which assessors may override the original risk level of an offender (i.e., that which is derived directly from the general risk/need assessment score). At the end of every LS/CMI assessment, the assessor is asked specifically whether he or she elects to exercise a clinical or administrative override. This feature was added to the LS/CMI in large part to accommodate the concerns of practitioners. It also routinizes the override, a concept that is related to one of Andrews’s basic principles of effective correctional treatment, namely professional discretion. Specifically, assessors are asked whether there are sufficient grounds to modify the score-based risk level. Factors to be considered in making such a determination are the presence of strengths (to lower risk) and supplementary risk factors (to increase risk). An administrative override was also allowed to accommodate correctional agencies that require certain offenders to be classified in a specific manner, as a matter of policy. The value of the override, however, at least empirically, has not been demonstrated.
A seventh enhancement of the LS/CMI introduced other clinical issues to the traditional risk/need assessment protocol. These include social, health, and mental health needs of the offender. This section extends the LS/CMI beyond its traditional risk/need function to a more comprehensive offender assessment. Its items are also considered for professional and ethical reasons. It is perhaps a little ironic that, after a decade of trying to convince professionals to consider criminogenic needs (the Central Eight) as opposed to traditional clinical issues relating to client well-being, the LS/ CMI brings the assessor back to these traditional social- and health-related matters (e.g., homelessness, suicidal, HIV, self-esteem, and victimization). However, to call the items in this section non-criminogenic appears to be erroneous as studies that have examined this section’s score have found a modest positive correlation with recidivism.
Related to the previous item, the eighth enhancement to the LS/CMI introduced the concept of specific responsivity into the assessment process. The responsivity principle is one of three key aspects in the Risk-Need-Responsivity (RNR) model of effective intervention with offenders.
Specific responsivity pertains to particular characteristics of the offender that should be considered when planning and delivering case management practices (supervision and intervention) to an individual offender. Generally, they include demographic characteristics (e.g., age, gender, and race or ethnicity), personality and mental health attributes (e.g., anxiety, antisocial characteristics, mental disorder, and motivation), and cognitive attributes (e.g., intellect and various kinds of disability). The LS/CMI asks assessors to document these characteristics and encourages case managers and other users of the assessment to attend to them in the delivery of services to the offender. Thus, by considering specific responsivity characteristics of the offender, along with offender strengths and other, non-criminogenic needs, the LS/CMI extends the assessment beyond the traditional offender risk/need assessment protocol.
To accommodate the diversity of contexts to which the LS/CMI might be applied, the ninth enhancement to the LS/CMI constituted a new section devoted to institution (jail and prison) behavior. Separate subsections capture information pertaining to both previous incarcerations and the current term in custody. Items focus on security classification, disciplinary incidents, and special accommodation needs, such as segregation. Because of the inclusion of these institutional items, the LS/CMI is now being used as part of an inmate classification and case management system in some jurisdictions.
Addition of Case Management
The final enhancement to the LS/CMI entails a concrete effort to link the offender risk/need assessment to offender case management, hence the CM component of LS/CMI. The combination of a risk/need assessment and a case management procedure in the same document was triggered by consultation with experts in the field and by some early empirical investigations that revealed a disconnect or, at best, a weak connection between risk/need assessment and offender case management that fell short of expectations. In retrospect, it was apparent that instructing correctional practitioners to “just do it!” was unrealistic. An important study that demonstrated a need for risk assessment to be more directly related to offender planning and supervision was conducted by Bonta with probationers in Manitoba, Canada. Bonta and colleagues revealed that the probation officers’ supervision conformed much more strongly to the formal conditions of the probation order than the offender’s risk/need assessment. Similar findings were unearthed in Texas, where directions from an instrument, the Client Management Classification system, were ignored more often than they were followed to the detriment of offender’s success in the community.
Consequently, the LS/CMI added three more sections devoted to case management of the offender: (1) the case management plan, wherein interventions were developed in the context of the offender’s criminogenic needs, non-criminogenic needs, and responsivity characteristics; (2) the progress record, wherein positive and negative changes in the status of the offender are documented in the context of their LS/CMI assessment profile; and (3) a discharge summary, which reflects on the offender’s status upon release vis-à-vis his or her assessment upon entry. Because numerous correctional organizations use their own internal case management record, a truncated version of the LS/CMI, the LS/RNR, which simply omits the three case management sections of the LS/CMI, was published in 2008.
In sum, the LS/CMI is a more comprehensive assessment than the LSI-R and other third generation risk/need assessment tools and is more directly linked to the subsequent and ongoing case management (supervision and service delivery) of the offender. As such, the LS/CMI was deemed a fourth generation offender risk assessment tool.
Research
As is the case with other versions of the LSI, a substantial body of research has been produced documenting the predictive validity of the LS/CMI with various types of offenders, in various legal contexts, and in various jurisdictions. Although some differences in the magnitude of the relationship between LS/CMI score and recidivism differ along these dimensions, Canadian LS/CMI studies generate a stronger LS/CMI relationship with recidivism than studies conducted in other jurisdictions (outside North America, followed by the United States). In addition, some, but not all, studies have suggested that there are differences by gender and race. Nonetheless, the overall robustness of the LS/CMI is well established. This is illustrated in a meta-analysis that summarized 30 years of LSI research. Examining 151 offender samples, Mark Olver and colleagues found that the LS/CMI and LSI-OR tend to perform as well or better than its LSI predecessors.
The Future of the LS/CMI
The LS/CMI has emerged among a growing number of offender risk assessment tools as one of the most widely used instruments, worldwide, slowly surpassing its predecessor, the LSI-R. In designing the LS/CMI as an assessment tool that includes both traditional risk/need components and other non-criminogenic needs and responsivity features and then combines it with offender case management, it was anticipated that it would force correctional workers and their agencies to apply the RNR principles on offenders, one by one. The link between risk/need assessment and RNR-based intervention may be further strengthened in automated versions of the LS/CMI, as implemented, for example, in Oregon. Moreover, LS/ CMI validation research continues across multiple types of offenders, including driving while impaired and child pornography offenders, and in a wide range of jurisdictions that extend beyond North America and Europe to countries such as Chile, Singapore, Hong Kong, China, and Pakistan. Yet, empirical demonstrations of its capacity to change the post-assessment supervision and counseling practices of correctional workers remain few.
References:
- Andrews, D. A., & Bonta, J. (1995). The Level of Service Inventory–Revised: User’s manual. Toronto, Canada: Multi-Health Systems.
- Andrews, D. A., Bonta, J., & Wormith, J. S. (1995). Level of Service Inventory–Ontario Revision (LSI-OR): Interview and scoring guide. Toronto, Canada: Ontario Ministry of the Solicitor General and Correctional Services.
- Andrews, D. A., Bonta, J., & Wormith, J. S. (2004). The Level of Service/Case Management Inventory (LS/ CMI). Toronto, Canada: Multi-Health Systems.
- Andrews, D. A., Bonta, J., & Wormith, J. S. (2006). The recent past and near future of risk and/or need assessment. Crime and Delinquency, 53, 7–27. doi:10.1177/0011128705281756
- Andrews, D. A., Bonta, J., & Wormith, S. J. (2008). The Level of Service/Risk-Need-Responsivity (LS/RNR). Toronto, Canada: Multi-Health Systems.
- Bonta, J., & Andrews, D. A. (2017). The psychology of criminal conduct (6th ed.). New York, NY: Routledge.
- Bonta, J., Rugge, T., Scott, T.-L., Bourgon, G., & Yessine, A. K. (2008). Exploring the black box of community supervision. Journal of Offender Rehabilitation, 47, 248–270. Retrieved from https://doi.org/10.1080/10509670802134085
- Gendreau, P. T., Little, T., & Goggin, C. (1996). A meta-analysis of predictors of adult offender recidivism. Criminology, 34, 401–433. doi:10.1111/j.1745-9125.1996.tb01220.x
- Harris, P. M., Gingerich, R., & Whittaker, T. A. (2004). The “effectiveness” of differential supervision. Crime and Delinquency, 50, 235–271. Retrieved from https://doi.org/10.1177/0011128703258939
- Olver, M. E., Stockdale, K. C., & Wormith, J. S. (2014). Thirty years of research on the Level of Service scales: A meta-analytic examination of predictive accuracy and sources of variability. Psychological Assessment, 26, 156–176. doi:10.1037/a0035080
- Oregon Department of Corrections. (2010). LS/CMI: Level of Service Case Management Inventory for community corrections. Salem: Oregon Department of Corrections, Community Corrections Branch. Retrieved from https://www.oregon.gov/doc/CC/docs/ pdf/lscmi_manual.pdf
- Wormith, J. S. (1997, January). Research to practice: Applying risk/needs assessment to offender classification. Forum on Corrections Research, 9(1), 26–31. Retrieved from http://www.csc-scc.gc.ca/ research/forum/e091/e091f-eng.shtml