Role of Mandated Community Treatment for Justice-Involved Individuals With Serious Mental Illness

Mandated community treatment has received increased attention in recent years, as states struggle to provide mental health treatment to individuals who have difficulty accessing these services or who resist seeking treatment (1–5). As of 2019, 47 states had some type of involuntary outpatient commitment law, although many states rarely use it. Most studies that have examined the efficacy of involuntary commitment measured outcome in numbers of hospitalizations and days in the hospital (6–8). However, for justice-involved individuals with serious mental illness, the most salient outcomes are reductions in arrests, incarcerations, and violence, especially if symptoms of mental illness were related to the prior offenses. In a review of the outcomes of New York’s outpatient civil commitment law, Appelbaum (9) documented decreases in violent behaviors (47%) and substantial decreases in arrests (83%) and incarcerations (87%). Other studies evaluating mandated community treatment found similar results, with arrest rates decreasing significantly among patients who participated in enforced treatment (10–12).

Oversight of treatment and behavior in the community is critically important with individuals found not guilty by reason of insanity (NGRI). Although the criteria for establishing insanity vary among states, all NGRI statutes require that the defendant have a major mental disorder and that the disorder be directly related to the commission of the offense. Once the defendant is found to be NGRI, requirements for treatment vary among states, although many states require inpatient treatment, with a potential for transition to the community. Typically when defendants found to be NGRI are released to the community, the court imposes additional sanctions and supervision, termed conditional release. Although conditional release is a form of mandated community treatment, the sanction is imposed after the commission of an offense, often a violent felony offense. Mandated community treatment is a civil, not a criminal, commitment procedure. The goals of outpatient civil commitment procedures are to provide treatment and support to individuals with major mental disorders in the community to prevent hospitalization and reduce dangerousness. The prevention of offending is a secondary, albeit important, goal. Conditional release, on the other hand, has the expressed goal of helping individuals to reside safely in the community without further offending; rehospitalization is often viewed as a method for maintaining community safety (13).

Ample research has demonstrated that defendants who are found to be NGRI and conditionally released are not a substantial risk to the community if they are provided adequate community treatment (14, 15). Numerous studies have shown success in conditional release programs, with low rearrest rates and maintenance of community treatment without rehospitalization (13, 16–20). In contrast, the literature is clear that without court supervision, many patients are less likely to use mental health programs after discharge, even though they may have access to the same types of programs available to individuals under court supervision (5, 7, 9). Among justice-involved individuals, this lack of treatment may elevate their risk of future offending and potentially their risk of violence (13, 21).

As with the initial commitment, procedures for conditional release of individuals found to be NGRI vary among states. For example, in both Oregon and Connecticut, psychiatric security review boards (PSRBs) monitor these individuals immediately upon commitment and after their release into the community. The PSRB decides the type of outpatient treatment required by the individual and when he or she can be released from inpatient treatment. The PSRB also provides oversight of outpatient treatment, which is tailored to the needs of the patient (15, 20). In California, a state-funded agency provides oversight of community treatment. Supervision is generally highest immediately after release and is gradually decreased to the extent tolerated by the patient. One requirement of the conditional release program is that the participant reside in approved housing. Services are individualized for each person but always include mental health treatment and, if necessary, substance use disorder treatment. Other terms and conditions of release are dictated by the conditional release program. Failure to abide by these restrictions can lead to rehospitalization and, potentially, court-ordered revocation of conditional release.

Research has suggested that patients can live safely in the community while under conditional release. However, few studies have described whether these outcomes are maintained after mandated treatment ends. Our study examined this question by focusing on a single outcome, defined as rearrest, in patients who were found to be NGRI and who were released to the community.

Methods

This research received approval from the local human subjects committee, the state of California Committee for the Protection of Human Subjects, and the University of California, Davis, Institutional Review Board (IRB).

Participants

This study was conducted in a large state psychiatric hospital in northern California (DSH-Napa) treating primarily patients committed through the criminal justice system. A total of 295 patients participated in the study between July 2002 and July 2013, although only 261 had complete data. There were no statistically significant differences in demographic factors between the patients who did and did not have complete data for all risk assessments and follow-up, including gender, race, marital status, commitment offense, level of education, primary diagnosis, or diagnosis of a personality disorder. The sample comprised both individuals found to be NGRI (N=208) and those found guilty of their offense and paroled to the state inpatient mental health system (N=53). Only patients committed under the NGRI statute were included in this study. Of the 208 patients committed as NGRI, 94 were released to the community during the study period.

Procedures

The study was designed as a prospective evaluation of the patients who were provided treatment at DSH-Napa in the study period. It was descriptive in nature and followed patients through their hospitalization and release into the community. Data collected included a diagnostic assessment, assessments of psychiatric symptoms and violence risk, self-report measures of anger and impulsivity, and behavioral indicators of aggression. A thorough record review was conducted to document demographic information. All patients provided consent to participate in follow-up after release. All discharge recommendations were made independent of the research, and research staff were prohibited by the IRBs from providing information to treatment providers that might affect decision making about release. Data for the current study included demographic information gathered during the patient’s hospitalization and arrests, if any, after release.

In California, the NGRI statute requires that participants remain in the conditional release program for at least 1 year. After this year, either the patient or the conditional release program can petition the court for “restoration of sanity” (a term in the penal code used to indicate that the person is no longer considered dangerous because of a mental disorder). If approved by the court, the individual is released with no further court-mandated conditions. There also is another mechanism for unconditional release in California. All individuals found to be NGRI are given a sentence commensurate with the statutory requirements for the commitment offense. If the patient’s commitment has been continued beyond this determinate sentence length, the hospital can decide that the patient does not meet criteria for continued commitment, as long as the offense does not carry a life sentence. If the committing judge agrees, the patient is released to the community unconditionally. Furthermore, after they have been hospitalized for at least 6 months, patients can file a writ of habeas corpus at any point in their commitment; if they succeed in convincing the court that they are no longer a danger because of their mental disorder and are past their determinate sentence length, they can be released unconditionally.

During the study period, the decision to release a patient to the conditional release program was made by representatives of the hospital in concert with representatives from the program responsible for patient’s care in the community. No clear guidelines were established by either agency about how to make these decisions, although stability of mental illness and risk to the community were always considered. As with unconditional release, the final decision for release was made by the committing court.

Three groups of patients were examined: those discharged to the conditional release program and remaining in conditional release at the end of the study (N=49), those initially released to the conditional release program but restored to sanity with no further court supervision (N=12), and those released from the hospital by the court with no court-imposed conditions for remaining in the community (N=32). One individual was excluded from this study because his or her status was unclear.

For participants released to the conditional release program, quarterly reports were obtained documenting progress in the community or need for rehospitalization. For participants released unconditionally, either on initial discharge from the hospital or via restoration of sanity, there was no formal method for tracking outcomes. On occasion, these individuals were rehospitalized at DSH for a legal commitment after another offense, and the rehospitalization was documented in a database maintained by the state hospital system. To augment this information, Internet search engines were used in an effort to determine whether participants released unconditionally had been rearrested and if so, the date and nature of the offense, if available. The name, date of birth, and middle name of the individual were used to verify that he or she was a participant in the study. No further information could be ascertained for this group. As such, the arrest rates for these groups may be an underestimate of actual reoffending. The average length of follow-up (date of discharge from the hospital to end of study) was 4.83 years, although length of follow-up differed between the three groups.

Results

As shown in Table 1, there were very few demographic differences between the three groups. Compared with nonwhites, whites were more likely to be released to the conditional release program and were more likely to be restored to sanity. For clinical characteristics, patients released unconditionally were more likely than patients in the other two groups to have a personality disorder diagnosis. There was a trend for primary diagnoses to differ between groups, with patients remaining under conditional release slightly more likely to have a schizophrenia spectrum disorder diagnosis.

Table 2 provides the length of treatment for the three groups. As seen in this table, the follow-up period for patients restored to sanity was significantly longer compared with those released unconditionally or those remaining under the supervision of the conditional release program. There was no difference between the latter two groups in length of follow-up. There were also differences between the groups in length of stay prior to discharge for the current hospitalization, with patients restored to sanity having a shorter stay than those remaining in the conditional release program. Length of stay in the hospital did not differ between patients restored to sanity and those released unconditionally.

Length of time between the commitment offense and discharge from the current hospitalization was significantly shorter among patients restored to sanity compared with the other groups. There was no difference in the time between commitment offense and discharge for patients released unconditionally and those remaining in the conditional release program. There was no difference between groups in length of time in the community between release and subsequent arrest or between commitment offense and arrest after discharge. Patients restored to sanity spent an average of almost 4 years in the conditional release program.

As seen in Figure 1, patients released from the hospital by the courts with no requirement for community treatment reoffended at a high rate, with almost 44% (N=14) arrested for an offense in the study period. Patients under the supervision of the conditional release program had low rates of reoffense (N=4, 8.2%). In contrast, once individuals were restored to sanity and released unconditionally, arrest rates increased, with 25% (N=3) arrested in the period studied. Of the three patients restored to sanity who were known to have reoffended, only one was recommended for restoration by the conditional release program; the other two were restored to sanity by the court against the recommendation of the conditional release program. In contrast, of the nine patients who were not known to have reoffended, seven were recommended for restoration by the conditional release program, and records were unclear for another, leaving only one patient who was restored to sanity by the court against the conditional release program’s recommendation. These differences approached statistical significance but fell short because of the small number of individuals who were restored to sanity.

When each group was examined individually, there was a statistically significant difference in arrest rates between patients released with and without mandated treatment (χ²=14.18, df=1, p=0.001). The difference in arrest rates between patients restored to sanity and those remaining in conditional release approached significance, whereas the difference in arrest rates between patients released from the hospital with no supervision and those restored to sanity was not statistically significant. Patients who were not under court supervision had higher odds of rearrest compared with patients who were under supervision (odds ratio [OR]=8.75, 95% confidence interval=2.54–30.19).

Table 3 presents the results of two stepwise logistic regression analyses. As seen in this table, only type of release and length of follow-up contributed to rearrest for the total sample, and only for the comparison between patients with and without mandated treatment. To examine this difference more closely, a separate stepwise logistic regression analysis was conducted that excluded patients restored to sanity. The results of this analysis show that mandated community treatment was highly relevant to outcome. When the analysis controlled for length of follow-up, the OR for risk of arrest among individuals without mandated treatment increased from 8.75 to 10.71.

Discussion

The most striking finding of our study was the observed differences in arrest rates between the three types of release. Almost half of the patients released to the community without court-mandated supervision were arrested for another offense, compared with a very small percentage of patients released under mandated supervision. The unadjusted OR suggests that among individuals found to be NGRI, those released to the community without mandated supervision were almost nine times more likely to reoffend than those released with mandated supervision.

Even after accounting for length of follow-up, the analyses showed that receipt of mandated treatment had the strongest effect on whether the patient was arrested. More striking, after an average of almost 4 years of stability in the community, individuals who were restored to sanity experienced an increase in arrest rates once mandated supervision was discontinued. After mandated treatment ended, arrest rates for this group were not statistically different from those of individuals initially released from the hospital without supervision. Because of the small number of patients who were restored to sanity, the difference between their arrest rate and that of patients who remained in the conditional release program was not statistically significant, although it approached significance.

Abundant research shows that conditional release is effective in assisting defendants found to be NGRI to maintain positive outcomes in the community on a variety of indices. Conditional release programs are effective in managing symptoms, reducing hospitalizations, and reducing numbers of arrests and incarcerations (14–16, 20). However, most of this research reports on patients maintained in conditional release. Only a small number of studies have evaluated arrest rates after release from court-imposed conditions. Norko and colleagues (20) demonstrated that arrest rates were 2.3% for persons under the supervision of the PSRB. After discharge from the PSRB, arrest rates increased to 16%. Although we had no information on what, if any, mental health treatment was received by patients released with no court-mandated treatment, our data suggest that court oversight of psychiatric treatment was effective in reducing rearrests. This result is similar to outcomes for involuntary outpatient commitment (9–11). Although involuntary outpatient treatment is an entirely different mechanism for requiring psychiatric treatment, many studies evaluating its effectiveness use arrest as an outcome measure, presumably because of the increased likelihood of arrest associated with living in the community with an untreated mental illness. These studies showed that participation in mandated community treatment was effective in reducing arrests, at least while mandated treatment was in effect. Although a recent review of community treatment orders reported equivocal outcomes (6), the measures studied were rate of readmission, time in the hospital, and community service use.

Our data suggest that patients who are found not to be criminally responsible because of symptoms of mental illness may need continued support, even after symptoms have remitted. This finding is not surprising, given the increasing evidence that justice-involved individuals with serious mental disorders frequently do not access mental health treatment if it is not mandated (4, 7). One study found that almost 50% of defendants who were found not competent to stand trial had received no mental health services in the 6 months prior to arrest, and, when services were accessed, they frequently were crisis related (22).

Recent research has demonstrated that patients’ attitudes toward mental health treatment are critical in whether they access treatment. Specifically, individuals with a serious mental disorder must believe that they need treatment, that seeking treatment is beneficial to them in some way, and that they have the ability to access treatment (1, 2). Unfortunately, the same research demonstrated that individuals who have been mandated to receive mental health treatment in the past are less likely to seek treatment, perhaps because of the fear of court-imposed restrictions (3). Given our finding that arrest rates are much higher for individuals without court supervision, even after periods of stability, it seems critical to develop methods for engaging patients in long-term treatment to help them remain in the community and out of the criminal justice system. Because engagement in treatment is at least one goal of mandated community treatment, it is included as one component in the model developed to outline critical treatment points for justice-involved individuals with a severe mental disorder (the sequential intercept model, intercept 5, community corrections), presumably to emphasize the need for ongoing supervision for justice-involved individuals with severe mental disorders. Our data suggest that this may be particularly salient for individuals who have been released unconditionally after a finding of NGRI.

In other reports of offense rates during conditional release, observed rates were between 1% and 10% (13, 20), which are consistent with our data. This low rate of arrest is likely related to continued receipt of mental health treatment. In research examining the effectiveness of outpatient treatment (pharmacological and routine outpatient intervention) in reducing posthospitalization arrests among adults with serious mental illness, the authors found that risk of arrest was reduced when these individuals were in possession of their prescribed pharmacological medication (23). More specifically, individuals who had medication for 90 days after discharge from hospitalization had fewer arrests compared with those who had medication for only 30 or 60 days. The authors also observed a reduction in the likelihood of arrest among individuals who utilized outpatient services in addition to adhering to their medication regimen. It appears that the most salient factor in helping justice-involved individuals with mental illness to remain stable in the community and out of the criminal justice system is providing ongoing mental health treatment, an opinion that has been suggested by others (24).

The observed shorter length of hospitalization and time between commitment offense and conditional release for individuals who were ultimately restored to sanity was intriguing. Although differences between groups in primary diagnosis only approached significance, it appears that patients ultimately restored to sanity were less likely to have a diagnosis of a schizophrenia spectrum disorder and were more likely to have a mood disorder diagnosis. There is ample literature that schizophrenia spectrum disorders are more difficult to treat and may require more extended treatment to achieve remission of symptoms, whereas mood disorders typically respond fairly rapidly to pharmacological intervention (25).

Our data also suggest that the conditional release program was able to determine who was ready to be released without court supervision with a fair amount of accuracy. Of the 11 patients for whom this information was available, three were released by the court against the recommendation of the conditional release program and two of those three reoffended. Last, the racial differences between the three groups were interesting. It is unclear why the group that was restored to sanity was almost uniformly white, but it may reflect known racial disparities in the criminal justice system.

There were several limitations to this study. The primary limitation was the small sample size, especially for individuals restored to sanity. Additionally, when patients were unconditionally released from the state hospital or from the conditional release program, there was no systematic way to monitor their outcome in the community, nor was there a mechanism for determining whether they were receiving mental health treatment. Internet searches were the primary method for determining whether these individuals had reoffended, but this method may underestimate offense rates. Furthermore, we were not able to determine the offense for many individuals who reoffended and thus could not distinguish between violent or nonviolent offenses. Moreover, this was not a randomized controlled trial (RCT), and, as such, patient characteristics in the three types of release were not entirely comparable. Unfortunately, it is not possible to conduct an RCT for individuals found to be NGRI, given that courts dictate oversight of these individuals. Finally, although we have drawn comparisons between our research on patients in the conditional release program and research on outpatient civil commitment, these types of commitment are distinctly different; it is unclear whether our results can be extrapolated to the latter group. There is a paucity of research examining outcomes after mandated community treatment ends and many, if not all, studies are conducted with relatively brief follow-up periods (26).

Conclusions

Our study suggests that in a sample of criminally committed patients, court oversight may be necessary on an ongoing basis to help justice-involved individuals with a serious mental disorder avoid the criminal justice system and remain engaged in community treatment, even after they have shown an extended period of stability. Research suggests that when implemented appropriately, outpatient civil commitment laws can help individuals with mental illness live safely in the community during the time of their commitment without further arrests; it is unclear whether these gains are maintained after the commitment ends. More research is needed to determine whether individuals who have been civilly committed would follow the same path as our criminally committed sample, requiring some type of continued support in the community after discharge.