Transitional Case Management for Reducing Recidivism of Individuals With Mental Disorders and Multiple Misdemeanors

The study compared arrest records for TCM participants between the 12 months preenrollment and the 12 months postenrollment as well as across groups: participants diverted from jail who completed the program, diverted participants who did not complete the program, and voluntary participants. Participants were compared on case management sessions completed and linkage to long-term services. Eligible participants were individuals with mental disorders who had been diverted by the court or voluntarily enrolled after completing the Day Custody Program (DCP).

Participants were identified through a structured screening process before arraignment in the criminal court. Participants could be diverted from arraignment into TCM. People who voluntarily entered TCM did so through the DCP (13). Also operated by CASES, DCP is a three-day alternative sentence for individuals with multiple misdemeanors. Once individuals complete DCP, they can voluntarily enroll in TCM. Defendants diverted to TCM receive a three- to five-session case management program to satisfy the court and then have the option to continue in the program and can participate for up to six months. Voluntary participants entered TCM for up to six months after completing the DCP.

TCM was staffed by a psychologist responsible for court-based screening and project coordination, a social work supervisor, a substance abuse case manager, and a part-time forensic peer specialist. TCM provided weekly community-based case management services. Services started immediately after release from court or the day after the participant completed the DCP. The social work supervisor and case manager had caseloads ranging from 12 to 16 participants, with additional support from the peer specialist.

Official records on arrests within the State of New York were obtained for the sample. Participants were compared on the measure of arrests for the 12 months preenrollment and postenrollment. Paired-samples and independent-samples t tests were computed in SPSS for comparison of means. A negative binomial regression model for arrests in the 12 months postenrollment was completed in SPSS, version 20. The negative binomial model was selected over Poisson regression because of overdispersion of the dependent variable. Program records were reviewed for data on use of case management services.

This study was a program evaluation that relied on data obtained from deidentified administrative databases and was exempt from review by an institutional review board.

The study enrolled 178 participants into TCM from July 1, 2007, through November 30, 2010. TCM served 125 participants diverted from court, with the remainder voluntarily entering the program (N=53), primarily from DCP. Among diverted participants, 103 (82%) completed the required sessions. The mean±SD age of study participants was 40.0±7.0 years. The gender distribution was 22% (N=39) women and 78% (N=139) men. Fifty-six percent (N=99) of the sample was African American, 25% (N=45) Hispanic or Latino, 12% (N=21) Caucasian, and 7% (N=13) other. The primary diagnoses were bipolar disorder for 38% (N=68) of the sample, depressive disorder for 20% (N=35), and schizophrenia for 19% (N=34). Approximately half (N=95) of participants were homeless at enrollment, and 89% (N=158) had a co-occurring substance use disorder. TCM participants had 27.2±1.9 arrests over the lifetime and 3.6±2.6 arrests in the 12 months preenrollment.

We first examined arrests for all participants between preenrollment and postenrollment using t tests and negative binomial regression, which controlled for demographic characteristics, psychiatric diagnosis and substance use, homelessness, TCM group, case management sessions completed, and lifetime arrests. We then examined changes in arrests within and across diverted and voluntary groups in order to understand whether method of participation affected arrests. Finally, we report on the use of case management sessions and linkage to long-term services.

In the 12 months postenrollment in TCM, 72% (N=129) of participants were arrested at least once. For all 178 participants, the mean number of arrests in the 12 months postenrollment (2.5±3.2) decreased by 32% compared with arrests in the 12 months preenrollment (3.6±2.6) (t=5.1, df=177, p<.001). Factors relating to arrests in the 12 months postenrollment were examined in a negative binomial regression model. Incidence rate ratios are reported in Table 1 for seven person-level variables and two program-level variables.

With a paired-samples t test, arrests were compared in the 12 months preenrollment and postenrollment among the 125 diverted participants. The analysis indicated a 31% decline from the 12 months preenrollment (3.6±2.4) to the 12 months postenrollment (2.5±3.0) (t=4.29, df=124, p<.001). Similarly, participants who voluntarily enrolled in TCM (N=53) experienced a decline in arrests, from 3.6±2.1 arrests in the 12 months preenrollment to 2.5±3.5 arrests in the 12 months postenrollment (t=2.69, df=52, p<.01). In the 12 months postenrollment, the diverted participants who completed TCM (N=103) were arrested 2.2±2.9 times, compared with 3.6±3.2 arrests for diverted participants who did not complete TCM. This difference was not significant in an independent-samples t test.

With regard to face-to-face case management sessions, the 125 diverted participants completed a mean of 17.5±22.0 sessions, and the 53 voluntary participants completed a mean of 12.5±18.0 sessions, although the difference was not statistically significant. The mean difference in the number of case management sessions received by the diverted participants and the voluntary participants in TCM may be associated with the courts’ requiring diverted participants to perform a specified number of case management sessions, ranging from three to ten, in addition to any voluntary sessions received after completion of the court mandate. Using an independent-samples t test and controlling for court-mandated sessions, we found that the diverted group received a mean of 13.1±22.3 TCM sessions, compared with 12.5±18.0 sessions for the voluntary group, although the difference was not statistically significant. Therefore, the number of TCM sessions did not vary as a function of voluntary versus court-mandated status. Of all 178 participants, 25% (N=45) were linked to long-term services by TCM, and 39% (N=70) were connected to services at the time of entry. The other 36% (N=63) were not connected to long-term services by TCM.

Individuals with mental disorders and multiple misdemeanors who were enrolled in TCM, through jail diversion or voluntary enrollment, experienced 32% fewer arrests between the pre- and postperiods of the study. Age and number of lifetime arrests affected the change in arrests for TCM participants between the two periods. Postenrollment arrests varied for diverted participants who completed the program, diverted participants who did not complete the program, and voluntary participants. TCM was designed for individuals with a history of repeated arrests, and improvement is marked by fewer arrests. Obviously, case management services are not the sole component of reducing arrests; access to services, such as housing, substance abuse treatment, and income support, is also essential.

A limitation of this study was the absence of a comparison group within the design. The use of a single-study group of TCM participants limits the conclusions that can be reached regarding the effectiveness of TCM. Moreover, TCM lacks a structured cognitive-behavioral approach focused on recidivism reduction, such as in the “Thinking for a Change” intervention (14). Evidence indicates that programs must specifically address factors relating to criminal behavior and connect people to needed services, not simply to mental health treatment, in order to achieve reductions in criminal recidivism for people with or without mental disorders (15).

This study suggests that individuals with mental disorders and multiple misdemeanors, when engaged in a targeted intervention, may have fewer arrests and better access to behavioral health services and supports. More research is required to determine the effectiveness of the intervention when it combines behavioral health services with services that target risk of criminal behavior and need.

This research was partly supported by grants from the Jacob and Valeria Langeloth Foundation, the New York City Mayor’s Office of the Criminal Justice Coordinator, the New York City Department of Correction, the van Ameringen Foundation, the Charles and Mildred Schnurmacher Foundation, and the U.S. Bureau of Justice Assistance.

The authors report no competing interests.