Abstract
Community-based alternatives to conviction and imprisonment for adult offenders with severe mental illness are receiving increasing attention from researchers and policy makers. After discussing the justifications that have been offered in support of community-based alternatives, this article reviews the current empirical evidence relevant to such alternatives. The authors use the sequential intercept model as a guide and summarize the existing research at several points along the criminal justice continuum. They conclude by highlighting the gaps in existing research and discussing the need for further research in several key areas.
Our society has legitimate interests in prosecuting and incarcerating criminal offenders. At the same time, however, there is growing attention to community-based alternatives to conviction and imprisonment of certain subgroups of individuals who are involved in offending. One of these subgroups is adults with severe mental illness. 1 This article provides a review of the current empirical evidence relevant to community-based alternatives to the prosecution of individuals with severe mental illness. We begin by summarizing the justifications that have been provided for such alternatives. Next, using the sequential intercept model (Munetz & Griffin, 2006) as a guide, we summarize the existing research on diversion (prior to conviction) and postincarceration return to the community. Finally, we discuss the gaps in existing research and identify the implications for further research needs in light of this summary.
Justification for Community-Based Alternatives
Various reasons have been offered in support of the growing interest in community-based alternatives to traditional arrest, prosecution, and incarceration of individuals with severe mental illness. The cost of providing needed treatment to those with severe mental illness in jails and prisons can be enormous, relative to the costs of outpatient care in the community. For example, California is currently spending approximately $2 billion a year for prison health care, or $11,600 per inmate (Kiai & Stobo, 2010), with much of this cost attributable to the provision of treatment for inmates with mental illness. Estimates like this do not include the costs to local jails for inmate health care. To the extent that community-based alternatives for treating and supporting individuals with mental illness in the community can be provided without increasing the risk to public safety through higher criminal recidivism rates, then state and local jurisdictions can save substantially. To demonstrate that community-based alternatives have both saved money and enhanced public safety, it is important to obtain financial and criminal justice outcome data.
The second kind of justification for community-based alternative sanctions concerns the appropriateness of treatment that can be provided in communities, contrasted with what can be provided in jails and prisons. Secure settings with heterogeneous populations must prioritize security and provide rehabilitation opportunities to individuals with a wide range of needs while containing costs. These priorities are generally less consistent with the needs of offenders with severe mental illness (e.g., medication, co-occurring disorder treatment, psychosocial skills, and recovery) than are interventions provided in the community to a narrower target group. Data describing the range of interventions available and delivered in the community are needed to demonstrate this “appropriateness of treatment” justification, as are data on the outcomes of diverted versus nondiverted individuals (or other comparison groups).
Finally, community-based alternative sanctions are sometimes justified on humanitarian grounds. Certainly some offenders with severe mental illness have risk-relevant needs that are similar to those of a general offender population (Skeem & Eno Louden, 2006). Others, however, become involved with the criminal justice system through atypical behavior strongly influenced by an actively symptomatic condition or through the attempt to obtain food or shelter when their mental health condition interferes with their capacity to obtain such necessities in a prosocial, legal fashion. It is for these individuals in particular, who would be classified as low risk/high need in risk–need–responsivity terms (Andrews & Bonta, 2006), that alternatives to standard criminal justice sanctions appear easiest to justify in humanitarian terms.
Community-Based Alternatives and the Sequential Intercept Model
It is useful to consider the sequential intercept model (Munetz & Griffin, 2006) as a guide to the description of community-based alternatives to standard criminal justice processing for individuals with severe mental illness. This model describes the following five points at which the process of arrest, conviction, incarceration could be interrupted (or “intercepted”), yielding a different procedure for the affected individual. These points are as follows: (a) law enforcement and emergency services; (b) postarrest: initial detention or initial hearing and pretrial services; (c) post–initial hearings: jail or prison, courts, forensic evaluations, and commitments; (d) reentry from jails, prisons, and forensic hospitalization; and (e) community corrections and community support. The first point refers to the initial contact between the individual and police officers or other first responders. Much of the work in this area has involved approaches to specialized police responding, such as crisis intervention team (Compton, Bahora, Watson, & Oliva, 2008) and comparable efforts to provide police officers with a greater range of approaches to interacting with individuals with behavioral health disorders. Goals of diversion at this stage include providing treatment alternatives to arrest for minor offenses and interaction strategies that decrease the risk of harm to officers or the individuals themselves stemming from these encounters. The second intercept, following arrest at the time of first appearance, focuses on interventions that would interrupt the standard prosecution process after the person had been arrested but before he or she proceeds to trial or enters a plea. Although diversion to a specialty court such as a drug court, mental health court, or (more recently) veterans’ court or prostitutes’ court could occur at this intercept, this kind of specialty court diversion is more typically done at the third intercept. This is also the stage at which an individual can receive some specialized form of probation.
Following an individual’s conviction and incarceration (or acquittal by reason of insanity and commitment to secure forensic hospitalization), that individual will typically return to the community at some time. The fourth intercept thus does not focus on diversion but rather applying community-based interventions to facilitate successful reentry planning and initial steps. Finally, the sequential intercept model describes the stage at which community corrections (including specialized conditions or intensiveness for parole) might be applied to individuals with severe mental illness.
Using these five intercepts as a framework, we now describe the available empirical evidence on community-based specialized interventions for this justice-involved population. Empirical studies were selected for review using several criteria. These included the nature of the design (experimental or quasi-experimental strongly preferred), the operationalization of variables using sound measures, the use of a control or comparison group, and the collection of relevant and meaningful outcome data. Descriptions of programs or qualitative designs were not included. The strength and consistency of this evidence will guide our subsequent discussion of implications for research, practice, and policy.
Law Enforcement and Emergency Services (Intercept 1)
The relevant research in this area is summarized in Table 1. There is a modest number of original studies (n = 11, with Compton et al., 2011, reviewing a total of 20 studies) considering the status of the crisis intervention team (CIT) approach, which is the focus of most of this research. 2 Of these original studies, 7 include a control or comparison group and 4 have samples of substantial size (N > 200). Of the 11 original studies, 7 used some kind of statistical significance testing to measure differences.
Research Summary: Law Enforcement and Emergency Services (Intercept 1)
Note. CIT = crisis intervention team.
The research addresses three broad domains: characteristics and knowledge of officers who have been trained; characteristics of individuals who have been diverted; and “effectiveness” of specialized police response as measured by outcomes such as the number of individuals diverted, their time incarcerated, and the services delivered to diverted individuals, as well as case disposition and subsequent arrests of those diverted. The great majority of these studies focused on some kind of effectiveness measurement, with one study each devoted to officer perceptions, officer knowledge, and characteristics of diverted individuals.
Police trained in CIT, particularly in Memphis, reported feeling most prepared to handle encounters with mentally ill individuals in crisis (Borum, Williams, Deans, Steadman, & Morrissey, 1998), although relevant knowledge of officers declined in the months following training (Compton & Chien, 2008). But most of the studies’ findings are relevant to specialized police response effectiveness. CIT-trained officers were more likely to link individuals to psychiatric services (Compton et al., 2008) and less likely to use physical force (Compton et al., 2011). Such physical force was more likely to be used with individuals who appeared at higher risk for violence themselves but was still used conservatively by CIT-trained officers (Skeem & Bibeau, 2008). Diverted participants spent less time in jail, did not differ from nondiverted individuals in subsequent arrests over a 12-month outcome period, and cost less in criminal justice funding but more in treatment funding (Steadman & Naples, 2005). The absence of differences in subsequent arrests was also observed by others (Teller, Munetz, Gil, & Ritter, 2006; Watson et al., 2010).
Research in this area was generally supportive of the effectiveness of CIT or other specialized police responding in diverting mentally ill individuals in crisis from arrest and linking them with services. More limited data suggested that CIT also results in less application of force during the initial encounter with police. There is not yet evidence to suggest, however, that such specialized responding by itself is associated with lower rates of arrest when individuals are followed over periods of 12 months and compared to other individuals who have not been diverted. Finally, some data indicate that the costs of follow-up with individuals diverted in this way are shifted from the criminal justice system to the behavioral health system, resulting in lower criminal justice costs but enhanced treatment costs.
Postarrest: Initial Detention or Initial Hearing (Intercept 2)
There is a somewhat smaller number of studies addressing diversion at the second intercept. Table 2 summarizes 10 studies, including 9 original studies and 1 review. Of these studies, 6 of the 9 employ a control or comparison group and use significance testing for differences, and 4 have a sample of substantial size (N > 200).
Research Summary for Postarrest: Initial Detention or Initial Hearing (Intercept 2)
Note. ER = emergency room; CSI = California Symptom Inventory; NGRI = Not Guilty by Reason of Insanity; MCS = Mental Component Summary subscale of the SF-12; SF-12 = 12-Item Short Form Health Survey; PCS = Physical Component Summary subscale of the SF-12; SMI = Serious Mental Illness; MH = Mental Health; SA = Substance Abuse.
Like Intercept 1, this intercept focuses largely on efforts to divert individuals with mental illness from standard prosecution—but in a postarrest context. Studies in this area compare the characteristics of individuals who are diverted versus those returning to the community through reentry. Investigators address the relation between diversion and outcomes such as service use, substance use, mental health, crime, and quality of life (Broner, Lattimore, Cowell, & Schlenger, 2004; Broner, Mayrl, & Landsberg, 2005). Other outcomes include jail days (Hoff, Baranosky, Buchanan, Zonana, & Rosenheck, 1999), service utilization and mental health functioning (Lamberti et al., 2001), mental health symptomatology (Shafer, Arthur, & Franczak, 2004), and housing (National GAINS Center, 2002).
All but one of the studies reviewed found differences between diverted and nondiverted individuals. This single study (Bertman-Pate et al., 2004) cited no differences in revocation outcomes. Another study found differences that included time in the community and treatment participation (diverted participants were higher on both; Steadman & Naples, 2005). A similar “time in the community” finding was also noted by others (Broner et al., 2005; Hoff et al., 1999; Lamberti et al., 2001), with diverted individuals also described as having fewer hospital days in the community (Lamberti et al., 2001). Participation in a diversion program was associated with fewer arrests and less homelessness over a 1-year follow-up period (National GAINS Center, 2002), with another study citing more emergency room contacts but fewer subsequent arrests (Shafer et al., 2004).
There does appear to be a pattern suggesting that diverted individuals are likely to spend more time in the community, participate in various kinds of treatment activity more frequently and intensively, and possibly be rearrested less frequently than nondiverted individuals. This should be considered cautiously, however, for several reasons. First, the number of studies is relatively small. Even fewer of these studies used a substantial sample size and a control or comparison group. This means that the inevitable variability associated with differing designs and measures is obscuring whatever patterns might be detected through a larger number of studies. Second, the variability stems not only from study design and measures; investigators are also focusing on a number of differing approaches to “diversion” at this stage. These studies are reviewing programs that differ in terms of approach, consistency, intensity, and goals. Hence the challenges in researching this area are twofold: Researchers must use reliable approaches that accurately measure what programs are attempting and must also account for program differences in considering how well certain outcomes have been achieved.
Post–Initial Hearings: Jail or Prison, Courts, Forensic Evaluations, and Commitments (Intercept 3)
Research on Intercept 3 is summarized in Table 3. This is clearly the area in which the most research has been conducted, as it encompasses specialty courts such as drug courts, mental health courts, and community courts. (Additional specialized problem-solving courts focusing on other groups, such as veterans and prostitutes, have been implemented, but there has been less empirical research to date because of their relative recency of implementation. Some such studies are described in this section under community courts.) Research in all three of these areas—drug courts, mental health courts, and community courts—is summarized in Table 3.
Research Summary: Post–Initial Hearings: Jail or Prison, Courts, Forensic Evaluations, and Commitments (Intercept 3)
Note. OLS = ordinary least squares; QOL-SF = Quality of Life Scale–Short Form; MHC = mental health court.
The research designs for the studies on problem-solving courts are consistently stronger than most of those described for the previous two intercepts. Most of the studies on mental health courts, all of the studies on drug courts, and about half of the studies on community courts employ a control or comparison group, for example. Sample sizes are consistently large (N > 200) among these studies. Significance testing is almost always employed. A range of questions is addressed, including client characteristics and outcomes, process variables (participants’ perceptions of respect and fairness), and support by the community for such courts.
Major findings are described in consecutive paragraphs for drug courts, community courts, and mental health courts, respectively. About half of drug court participants complete these programs, and 60% remained in treatment for at least 1 year (Belenko, 1998), with the drug courts described in this review of 30 studies resulting in savings from jail and other justice-related costs, as well as reduced offending rates during participation in the drug court process and subsequently for graduates and, in many cases, participants who did not graduate (Belenko, 1998). Relapse rates for drug use were lower for graduates (and often participants) as well (Belenko, 1999). Drug courts were associated with lower costs and offense recidivism and increased treatment in some studies (Belenko, Patapis, & French, 2005; Bhati, Roman, & Chalfin, 2008; Gottfredson & Exum, 2002; Gottfredson, Najaka, & Kearley, 2003), with lower costs and criminal recidivism replicated but mixed evidence regarding lowered rates of drug use relapse (Government Accountability Office, 2005). The modest reduction in criminal recidivism was increased when drug courts targeted high risk and young offenders, according to a meta-analysis (Lowenkamp, Holsinger, & Latessa, 2005). The overall size of the offense reduction associated with drug courts was estimated by another meta-analysis at 26%; using a single model (pre- or postplea) may be more effective than using a mixed model (Wilson, Mitchell, & Mackenzie, 2006).
Community courts were considered by participants to be fairer than traditional courts, with perceptions of the judge and quality of the courtroom communication the most important indicators (Frazer, 2006). They imposed more alternative sanctions and less incarceration (Hakuta, Soroushian, & Kralstein, 2008). Compliance with sanctions that were imposed was higher, with substantial decreases in targeted outcomes such as prostitution and illegal vending arrests and slightly higher costs per case ($700 over 7 studies and 4 courts; Kralstein, 2005). A high proportion of community court participants thought that this kind of court was a good idea, working in a productive way, assigning useful community service, treating participants equally, and adding to the favorability with which their neighborhoods were rated (Justice Education Center, 2002).
Mental health courts were implemented with a wide variability of approaches (Erickson, Campbell, & Lamberti, 2006). A number of studies addressed the outcome of criminal recidivism after participation in mental health court. Findings on such outcomes included fewer arrests and jail days (Case, Steadman, Dupuis, & Morris, 2009), lower recidivism rates (McNiel & Binder, 2007; Moore & Hiday, 2006; Steadman, Redlich, Callahan, Robbins, & Vesselinov, 2011), reduced recidivism rates for both graduates and participants who did not graduate (Hiday & Ray, 2010), and possibly lower costs over time relative to traditional courts (Ridgely et al., 2007). Many such courts are operated on the “preadjudication” model, with record of arrest expunged on successful completion. However, as mental health courts have expanded to include felony cases, there has been an increase in the number of courts using postadjudication models, even including jail incarceration (Redlich, Steadman, Petrila, Monahan, & Griffin, 2005).
Research on Intercept 3 of the sequential intercept model is the most mature and consistent of the intercepts described thus far. Specialty courts of different kinds have been studied, with the general focus on the particular characteristics of participants, the perceptions of favorability on the part of such participants, and outcomes such as cost, the nature of appropriate services delivered, and change in justice-relevant outcomes such as rearrest and subsequent incarceration. Although there is mixed evidence regarding whether drug courts consistently reduce the drug use relapse rate, the evidence on the delivery of appropriate services, the perception of favorability on the part of participants, and the reduction of the incidence of subsequent arrest and incarceration seem largely favorable for drug courts, community courts, and mental health courts.
Reentry from Jails, State Prisons, and Forensic Hospitalization (Intercept 4)
Research in this area (see Table 4) can be considered in three domains: programs derived from assertive community treatment (ACT), programs derived from intensive case management (ICM), and correctional reentry programs. About half of the original studies in the first two domains (ACT and ICM, respectively) included a control or comparison group or incorporated samples with more than 200 participants. All of the ICM studies and 2 of the 6 ACT studies used significance testing. By contrast, only one study in the area of correctional reentry was located, and it did not use either a control or comparison group or significance testing.
Research Summary: Reentry From Jails, State Prisons, and Forensic Hospitalization (Intercept 4)
Note. FACT = Forensic Assertive Community Treatment; ACT = assertive community treatment; TAU = treatment as usual; ER = emergency room; QOL-SF = Lehman Quality of Life Scale–Short Form; CSI = Colorado Symptom Index; MCS SF-12 = Mental Component Summary subscale of the SF-12; PCS = Physical Component Summary subscale of the SF-12
Note that the report is not clear about how many participants were included in each outcome comparison (i.e., the reported percentages are slightly off when dividing the reported participant counts by the total reported sample size).
Follow up data were not available for all participants, so the sample size varied on outcomes.
When there was a control or comparison group, it consisted of those who did not participate in an ACT-based program but rather were involved in standard reentry procedures (termed “treatment as usual,” or TAU). These studies focused on comparing outcomes such as criminal justice involvement (e.g., any booking, felony booking, any conviction, felony conviction) and quality of life indicators (e.g., alcohol problem, global functioning, homelessness, employment) between groups. Other studies considered pre–post program differences for individuals participating in a specialized reentry program, using similar outcomes. Research questions were similar in the ICM and correctional reentry domains. Using outcomes in the broad areas of criminal justice, mental health treatment and functioning, and time in the community, these studies generally compared the performance of a specialized program to that of standard (TAU) intervention, or the pre–post differences in outcomes among those participating in a specialized program.
In the ACT domain, there were several major findings. Enhanced treatment within the ACT model was associated with better criminal justice outcomes (any booking, any conviction, mean jail time), better improvement in substance abuse problems, and greater improvement in global functioning and economic self-sufficiency, relative to those receiving TAU. Furthermore, greater fidelity to the ACT model was also associated with better outcomes, relative to those seen in lower-fidelity programs or comparison groups (California Board of Corrections, 2005). Consistent with this, ACT-model program participants had a lower number of hospital days, as well as arrests and incarceration days, relative to the year prior to their participation in the program (McCoy, Roberts, Hanrahan, Clay, & Luchins, 2004; Thresholds State, County Collaborative Jail Linkage Project Chicago, 2001; Weisman, Lamberti, & Price, 2004). As with mental health courts, there has been significant variability in the implementation of ACT-derived programs (Lamberti, Deem, Weisman, & LaDuke, 2011; Lamberti, Weisman, & Faden, 2004; Morrissey, Meyer, & Cuddleback, 2007), and efforts are currently under way to standardize these reentry models.
Programs built on the ICM model showed similar results. Relative to the 6 months prior to program involvement, participants in one program measured at 6-month follow-up showed significant decreases in legal problems and improvements in life situation and symptoms; however, there were no significant changes in income or residential situation (Godley, Finch, Dougan, McDonnell, & McDermeit, 2000). Another program also reflected more time in the community for participants (relative to TAU individuals), but participants were significantly more likely to report emergency room visits and hospitalizations (Steadman & Naples, 2005). The great majority of case management services were provided to mentally ill offenders while these individuals were in jail. Although a high proportion (72%) of these individuals were rearrested during follow-up, those receiving even a modest amount of case management (1–59 minutes per month) had lower odds of rearrest, particularly for violent offenses (Ventura, Cassel, Jacoby, & Huang, 1998). More favorable criminal justice outcomes (fewer jail days) were seen in the group receiving assertive case management services relative to those released straight to the community, with these favorable outcomes seen at 6, 12, and 18 months (Wilson, Tien, & Eaves, 1995).
Finally, there was a single study using correctional reentry more generally (Hartwell, 2010). In a Forensic Transition Team Program, a 3-month follow-up after participation reflected that 47% of participants were engaged in community service, 21% were hospitalized, and 18% were reinvolved with the criminal justice system (14% of participants could not be located at follow-up).
Findings in this area reflect several trends. As with research at some of the other intercepts, the number of studies is somewhat limited. However, the available evidence does support better criminal justice outcomes (whether rearrest or reincarceration) for participants in programs based on either the ACT or ICM model. Evidence on mental health outcomes is also generally favorable although more mixed; some studies show consistently favorable mental health and community adjustment outcomes, whereas others suggest that certain health outcomes (e.g., hospital days) may actually be greater for those in specialized programs. The evidence in this area is generally promising, however, subject to the caveat that we clearly need more studies that are well designed (including comparison or control groups and large samples) to support these conclusions with more confidence.
Community Corrections and Community Support (Intercept 5)
The evidence on the final SIM intercept is summarized in Table 5. As with the previous intercept, the research can be categorized in three domains: specialty mental health programs, either parole or probation; specialty probation; and specialty parole. In the first domain, one of the two original articles and some among those reviewed in the review article featured a control or comparison group, which was also seen in three of the five articles under specialty probation and in the single article under specialty parole. Slightly more than half of these studies used large samples, and half employed significance testing. As in much of the research seen for previous intercepts, the questions addressed in these studies focused on the effectiveness of specialty interventions in promoting improved criminal justice outcomes (e.g., rearrest, reincarceration), delivering appropriate mental health services, and promoting better mental health outcomes.
Research Summary: Community Corrections and Community Support (Intercept 5)
Note. PMI = probationer’s with mental illness; PO = probation officers; PCP = probation’s with co-occurring mental health and substance abuse problems; APD = antisocial personality disorder.
There were some distinctive questions addressed among the studies in the broad area of specialty mental health programs. For instance one study focused on whether reincarcerations of mentally ill people on probation or parole over a 15-month period are more likely to occur for reasons of noncompliance with treatment or for new criminal activities. Other reasons investigated for possible reincarceration include whether clients with case managers who seek stipulations for treatment and housing as a condition of release are more likely to return to jail, whether those with identified primary problems of substance abuse or criminal behavior rather than mental illness are more likely to be reincarcerated, whether psychiatric decompensation is related to reincarceration, and whether case managers who involve parole or probation officers are more likely to have their clients returned to jail.
Officers from specialty agencies must make decisions about allocating their time, so some research focused on questions such as the optimal number of contacts per month for a mentally disordered probationer, how specialty agencies respond to probationers’ misbehavior, and what strategies are used in responding to common violations. For instance, what are the distinctive components of specialty case management, and how practical and effective are they? What are the perceived effects of these methods on probationers’ experiences of coercion, treatment adherence, and outcomes? Finally, the one article located on specialty parole services focused on whether mentally disordered individuals are at greater risk for serious offending that demands a criminal justice response—and whether early onset of criminality and antisocial personality disorder are related to the risk of offending in this population.
There were several major findings from these studies. Broadly, a detailed review suggests that specialty agencies hold considerable promise for improving clinical and criminal outcomes for probationers and parolees with mental illness (Skeem, Emke-Francis, & Eno Louden, 2006). Psychiatric medication and broader treatment motivation played an important role in the risk of reincarceration for a new charge or technical violation, as participants perceiving medication to be less helpful or having low treatment motivation were much more likely to be reincarcerated for a new charge or technical violation. When this phenomenon is combined with the more intensive monitoring that comes with an ACT model, it may explain the finding in one study that an ACT team’s clients were more likely to be returned to jail than individually managed clients (Solomon & Draine, 1995a).
Both traditional and specialty parole or probation officers used graduated sanctions, but traditional officers generally respond to noncompliance with more punitive strategies than did specialty officers (Eno Louden, Skeem, Camp, and Christensen, 2008). This may help to explain the finding in another study: The treatment group was more likely to be charged with a probation violation, whereas control group participants were more likely to be arrested for a new criminal offense (Perez, 2009). It also relates to the distinctive functioning of specialty agencies, which were more likely to focus on monitoring medication and treatment attendance, more likely to use problem-solving strategies, and less likely to use threats of incarceration (Skeem et al., 2006). This approach may be particularly important with probationers with co-occurring mental illness and substance abuse, as these individuals are described as having poor relationships with professionals and being more likely to feel coerced into treatment and less likely to attend treatment sessions (Skeem, Eno Louden, Manchak, Vidal, & Haddad, 2009).
Offenders with mental disorders were given fewer opportunities for early release on full parole and when released were significantly more likely to receive suspension warrants or have their supervision revoked without commission of a new offense. It is interesting that the presence of antisocial personality disorder generally was not related to postrelease outcomes (Porporino & Motiuk, 1995).
The role of the relationship between case manager or specialty parole or probation officer and the individual under supervision is one of the major important findings to emerge from research on this intercept. Individuals with mental illness do appear to respond favorably to specialized interventions emphasizing treatment and rehabilitation, but there are several cautions that must accompany this conclusion. First, it becomes increasingly difficult to emphasize rehabilitation and avoid using coercion (e.g., return to jail) as the size of the caseload increases. Second, the attitude toward treatment—perception that psychotropic medication is helpful and willingness to participate in treatment—is strongly related to whether the individual will remain in the community without incurring criminal justice sanctions. Third, there is a subpopulation of individuals that experiences co-occurring disorders of mental illness and substance abuse, generally has poorer relationships with treatment providers, engages in riskier behavior, and is more likely to feel coerced into treatment.
Discussion
Community-based alternatives to standard criminal justice prosecution for individuals with severe mental illness represent an approach that has grown considerably during the past decade. But there are substantial differences between various approaches, and the sequential intercept model (Munetz & Griffin, 2006) is a useful conceptual tool for integrating and comparing the different stages at which specialized intervention might be attempted.
This review does clearly demonstrate the relative newness of many serious approaches to specialized intervention. Focusing on empirical research rather than conceptual, professional, or advocacy-based literature, the review shows that most of the studies across all intercepts have been conducted within the past decade. As a consequence, there is often a limited number of well-designed studies on which to base judgments about the effectiveness of different strategies and how various approaches might be deconstructed to identify the most effective elements.
Nonetheless, there is a promising core of work that supports the potential effectiveness of specialized interventions across multiple dimensions. First, public safety does not appear adversely affected—and often may be enhanced—when considering the criminal justice outcomes such as rearrest and reincarceration for those who go through specialized interventions. Second, although there are limited data on the cost of specialization, it appears that diversion and postincarceration specialized reentry services for mentally ill individuals may be revenue neutral to somewhat more costly than traditional approaches (depending on variables such as intensiveness of services, level of staffing, and size of caseload) in the community but less expensive than residential placements such as jails, prisons, and forensic hospitals. Third, there can be little doubt that the services provided to mentally ill individuals as part of such specialized initiatives are appropriate to their needs and better meet their liberty interests when delivered in the community.
Despite this promise, however, there is considerably more well-designed empirical research needed across the five intercepts, particularly in developing and standardizing community-based intervention models. One of our conclusions at each of the intercepts was that the number of available studies was too limited to allow us to draw firm scientific or policy conclusions. If this were to change through further research over the next decade, then the field could begin to conduct meta-analyses that allow the gauging of effect sizes more broadly. Prototypical effective approaches can be made clearer, both through such meta-analysis and through multiple studies considering within-group effects (pre–post differences) as well as between-group effects (differences between a specialized treatment group and a control group). Important questions such as the role of fidelity to a particular approach can be addressed: Do we have a model that is most appropriate for guiding specialized interventions at a certain intercept, or are there broader guidelines for effectiveness that do not require fidelity to a particular model? There should be more effort to conduct studies that include multiple jurisdictions, to manage the difficulty of site-specific idiosyncratic findings. Likewise, studies that include moderate to large samples both minimize the impact of chance and provide a more stable estimate of effect size.
The sequential intercept model provides a useful theoretical framework for organizing the empirical literature on community-based alternatives to standard criminal prosecution. That is how it has been used in this article. It should be noted, however, that in practice the five different intercepts are often not used as distinctly as the present discussion might imply. Community-based programs may accept referrals from multiple sources and at different points of intercept. Hence, community-based interventions may operate as “multipoint” rather than single-point strategies. For example, some ACT-based programs accept referrals from pretrial services (second intercept), mental health courts (third intercept), and prisons (fourth intercept).
There are certainly some limitations that should be noted as associated with this kind of review. Given the breadth of the review, covering five intercepts that are linked conceptually, there are certainly some differences among the kinds of interventions that are described across intercepts. Specialized police intervention, problem-solving courts, and specialized parole services have as a major common element the goal of providing appropriate treatment services to those in need in a justice context without increasing the cost to society or threat to public safety. But there are differences as well, and studies conducted in some of these areas may apply marginally (or not at all) in other areas. Available research was typically limited, and even studies selected for their empirical basis and design were often limited by sample size or the absence of an appropriate basis for comparison. These limitations should underscore the importance of continued empirical research in this area, as the empirical base is strengthened for scientific and policy applications in the coming years.
Footnotes
This article is part of a special issue titled “Diversion from Standard Prosecution”, edited by Kirk Heilbrun and David DeMatteo of Drexel University.
Notes
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
