Our review of the Department of Corrections and Rehabilitation's (Corrections) use of Correctional Offender Management Profiling for Alternative Sanctions (COMPAS) highlighted the following:
The Department of Corrections and Rehabilitation (Corrections) intends to use the Correctional Offender Management Profiling for Alternative Sanctions (COMPAS) software to help identify factors that cause inmates to commit crimes, so they can participate in such rehabilitative programs as substance abuse treatment or vocational education to reduce their likelihood of reoffending, thereby reducing overcrowding in the State's prisons. California's high recidivism rates and difficulties with prison overcrowding are well documented. In its October 2010 outcome evaluation report, Corrections reported that 67.5 percent of all felons released during fiscal year 2005-06 returned to prison within three years. Further, in May 2011 the U.S. Supreme Court issued a ruling upholding the authority of a lower court to require that California reduce its inmate population to 137.5 percent of the design capacity of its correctional institutions. As of June 30, 2011, Corrections had more than 144,000 inmates in its various institutions, which were designed to accommodate only 80,000.
However, the prospects that COMPAS will play a meaningful role in helping Corrections ultimately reduce prison overcrowding and lower its recidivism rates are, at best, uncertain. Corrections uses gender-specific versions of two different COMPAS assessments. The COMPAS core assessment identifies the needs of inmates entering the prison system, while the COMPAS reentry assessment evaluates inmates who are about to reenter society on parole. Our review found Corrections' use of COMPAS during its parole planning process is not consistently enforced, while its use in reception centers—where inmates are initially evaluated and assigned to a prison—does not appear to affect decisions on prison assignments and, by extension, the rehabilitative programs inmates might access at those facilities.
Corrections' process at its 12 reception centers for assigning inmates to prisons is complex and considers factors such as an inmate's history of violence, medical needs, gang affiliations, and the available bed space at suitable facilities that can accommodate the inmate's security requirements. Our observations at one reception center and discussions with Corrections' staff at seven others revealed that prison assignments are often not based on COMPAS. Instead, the inmate's security level and the weekly placement restrictions imposed by Corrections' Population Management Unit—the unit responsible for coordinating inmate movement within the prison system—are the primary determinants of prison assignment.
Our discussions with staff at the reception centers also suggested that some do not see value in using COMPAS. For example, one classification staff representative—who is responsible for the final decision on where an inmate will be sent—told us that she rotates to different prisons each week and that there is a lack of buy-in on COMPAS across the institutions, noting that the tool does not seem beneficial. The classification staff representative also indicated Corrections has not trained staff in her role on how to use COMPAS. Corrections' project manager for COMPAS confirmed that these individuals have not been trained, explaining that reception center staff are still in the process of developing procedures for using COMPAS. According to the project manager, this lack of training would preclude classification staff representatives from consistently considering COMPAS results. An assistant warden at another reception center questioned whether COMPAS identified inmate needs that his staff had not already gleaned from inmate files. Furthermore, we noticed that Corrections has established underground regulations concerning COMPAS's use because they were not adopted in accordance with the California Administrative Procedures Act, nor has it discussed COMPAS in its operating manual.
The value of using COMPAS to assess an inmate's needs is also limited because few programs exist within prisons to treat those needs. The COMPAS core assessment identifies up to five different needs; however, Corrections has rehabilitative programs that address only two—its academic/vocational education and substance abuse treatment programs. It has no programs to address criminal thinking, anger and violence, or family criminality. Corrections' secretary acknowledged that limited rehabilitative programs could cause some to question the value of using COMPAS but maintained that doing so was still worthwhile, since it avoids guessing which inmates need what services and potentially wasting limited resources on the wrong population. However, the secretary's comments contrast with the fact that more inmates with COMPAS core assessments have substance abuse problems when compared to the number of those with academic/vocational needs, yet the inmate capacity for academic/vocational education programs is five times that of substance abuse treatment programs. Considering that Corrections has instructed its staff to consider COMPAS assessment results only when placing inmates in its substance abuse treatment program, it is questionable whether COMPAS is a valuable tool for determining what rehabilitative programs Corrections needs.
Corrections has also cited budget cuts from fiscal year 2009-10 as a main reason for limited rehabilitative programs. In the 2009 Budget Act, the Legislature required Corrections to cut an unspecified amount from its rehabilitative programs and submit a report on its plan to absorb these cuts while maintaining effective rehabilitative programs. Its progress report to the Legislature in April 2010 indicated that it shortened the duration of its substance abuse treatment program, which had been from 12 to 36 months, to three months and reduced the number of prisons at which the program is offered. Corrections also reported it developed new education models that prioritize inmate placement based on their risk of reoffending and the time remaining until their release. These budget cuts caused spending on rehabilitative programs to be reduced by $69.1 million during fiscal year 2010-11.
Our review also raised questions as to whether Corrections was actually putting inmates in the correct rehabilitative programs based on COMPAS. All 11 of its institutions with a substance abuse treatment program treat only a limited number of inmates with moderate to high needs as determined by COMPAS. In February 2011 Corrections had nearly 2,600 inmates with moderate to high needs for substance abuse treatment as determined by COMPAS, but only 800 with moderate to high needs were assigned to substance abuse treatment during that same month. An additional 740 inmates who had no COMPAS assessment and 310 who were assessed as having a low need for substance abuse treatment were enrolled in the program that month. The fact that Corrections' staff can place inmates in the substance abuse treatment program without a COMPAS assessment raises more questions as to whether COMPAS is a valuable screening tool to identify inmates' needs. In addition, Corrections has not developed a plan to measure COMPAS's impact on reducing recidivism, and thus its value is uncertain.
Using the COMPAS reentry assessment as part of the parole planning process has potential benefits for inmates but may not yield lower recidivism rates. Within 240 days of an inmate's release to parole, Corrections' staff administer a reentry assessment that results in a case plan to assist the inmate with transitioning to life outside of prison. The case plan provides the inmate with tasks and goals to pursue after being released, to address identified problems. In addition, the case plan provides the inmate with contact information for programs and resources in the community where he or she will be paroled. Reentry assessments can also potentially provide value to parole agents, eliminating the need to create case plans.
However, the value of using COMPAS reentry assessments on inmates being paroled is questionable, since the inmates are not required to adhere to the goals and tasks in the case plans and parole agents' use of COMPAS has not been consistently enforced. One parole agent indicated he ignores the COMPAS assessments because he has a better knowledge of the area and can recommend programs and services that are free, or at least more affordable. According to the associate director of the Division of Adult Parole Operations (Parole Operations), parole agents need more in-depth training in how to apply COMPAS assessments to a parole plan that will ultimately reduce recidivism. Parole Operations is currently engaged in a pilot project that clarifies how parole agents should use COMPAS. Some enhancements of the pilot project include requiring—as a condition of parole—that parolees follow the goals and tasks identified in the COMPAS case plan as modified by the parole agent. Corrections has also developed a form that requires parole agents to specify goals and how much time parolees should spend over a certain period of time on meeting those goals. However, the ultimate success of COMPAS relies on staff's willingness to use the assessment tool.
Finally, Corrections has not tracked the actual costs of deploying COMPAS. Our discussions with Corrections' budget and accounting staff revealed that they are unaware of any accounting codes in their systems to track actual costs related to COMPAS deployment. Corrections' deployment of COMPAS has occurred in waves, with its parole units and reception centers beginning to use COMPAS in 2006 and 2007, respectively. Corrections' 33 prison institutions began using COMPAS in March 2011. Our discussions with the COMPAS project's controller and technical project manager indicated that they have not reported the actual costs associated with prison deployment to the California Technology Agency (Technology Agency) because Corrections' accounting system is not set up to separately track and report such costs, including staffing costs for Corrections' information technology personnel associated with deployment of COMPAS core assessments to the 33 prison institutions. Furthermore, Corrections could not support a total of $14.6 million in actual COMPAS costs that it reported to the Legislature in the governor's proposed budgets for fiscal years 2008-09 and 2009-10. As a result, Corrections cannot demonstrate accountability for its spending on COMPAS or explain how it monitored the project for any cost overruns.
To ensure that the State does not spend additional resources on COMPAS while its usefulness is uncertain, Corrections should suspend its use of the COMPAS core and reentry assessments until it has done the following:
Once Corrections resumes its use of the COMPAS core and reentry assessments, it should take the following steps to better ensure that COMPAS is a valuable inmate assessment and planning tool:
To ensure transparency and accountability for costs associated with information technology projects such as COMPAS, Corrections should take the following actions:
Corrections disagrees with our recommendation to temporarily suspend its use of COMPAS. Although Corrections acknowledges that it should do more to communicate the value of COMPAS to its employees and provide ongoing training to key staff, it believes suspending COMPAS will result in the loss of the progress it has made, including five years of training, and that it would take years for it to regain its momentum. Corrections did not provide responses to our specific recommendations, but indicated that it would provide additional information in its 60-day, six-month, and one-year responses.