< Previous20 PERSPECTIVES VOLUME 44, NUMBER 2 Importantly, the authors found that the population which is examined matters a lot when examining the link between psychosis and violence. The link is strongest when examined in community samples— samples designed to represent the entire population of community residents. The link is weakest when examined in corrections samples—groups of people involved in the justice system, including clients of probation and parole agencies. The odds of violence on the part of offenders with psychosis are only 27% higher than the odds of violence by those without psychosis, an effect size that is “smaller than small” (see Chen, Cohen & Chen, 2010). Why is the link between psychosis and violence trivial in corrections samples? It is because there are strong, competing risk factors for violence to contend with in these samples. For example, features of antisocial and psychopathic personality disorder are commonly found in corrections populations, and these predict violence much more strongly than mental illness, even within samples of offenders with mental illness (Bonta, Blais, & Wilson, 2014). MYTH: CLIENTS WITH MENTAL ILLNESS HAVE HIGH RE-ARREST RATES Recidivism has different definitions that bear directly on clients with mental illness, and re-arrests are not always based on criminality. Recidivism in the sense of committing a new offense differs meaningfully from recidivism based solely on technical violations, as the latter reflects an arguably less serious failure to follow the rules of probation or parole, such as failure to pay fees, complete community service, maintain employment, or meet with the supervising officer. Studies generally indicate similar rates of new offenses for offenders with and without mental illness. For example, an examination of a subset of Washington State’s probation population found that probation clients without a mental illness The odds of violence on the part of offenders with psychosis are only 27% higher than the odds of violence by those without psychosis, an effect size that is “smaller than small.”21 AMERICAN PROBATION AND PAROLE ASSOCIATION were almost as likely to re-offend as probation clients with a mental illness, with recidivism rates of 38% and 41%, respectively (Gagliardi, Lovell, Peterson, & Jemelka, 2004). These findings are in line with most of the other research on rates of new offense for disordered and non-disordered offenders on community supervision (Feder, 1991; Lovell, Gagliardi, & Peterson, 2002; McShane, Williams, Pelz, & Quarles, 2005; Porporino & Motiuk, 1995; but see Eno Louden & Skeem, 2011, for an exception). At the same time, studies also generally show that offenders with mental illness have higher rates of technical violations than those without mental illness (Eno Louden & Skeem, 2011; Porporino & Motiuk, 1995). There are multiple reasons why this is the case. First, people with mental illness on community supervision tend to have more conditions placed on them, in particular participation in mandated mental health treatment (Skeem, Emke-Francis, & Eno Louden, 2006). The more conditions individuals have to meet to successfully complete community supervision, the more chances they have to fail. Complying with probation or parole conditions can be difficult when one of those conditions involves navigating an overburdened mental health care system that is not designed to meet the needs of people involved in the justice system (see Skeem et al., 2006), especially for people whose functioning is compromised by symptoms of mental illness (Skeem, Encandela, & Eno Louden, 2003). In addition, community corrections officers are quicker to revoke community supervision for clients with mental illness compared to other clients. We examined this in a series of experiments, which is the only type of research design from which conclusions regarding cause and effect can be made (Shadish, Cook, & Campbell, 2002). For example, Eno Louden and colleagues presented 89 probation officers with a description of a hypothetical client with mental illness, and a second “control” client without mental illness. When presented with a scenario in which each of the clients violated the terms of probation by not showing up for work and having a positive urinalysis, officers were more likely to report that they would seek revocation for the client with mental illness than for the client without mental illness (Eno Louden, Manchak, Ricks, & Kennealy, 2018). Further, probation officers reported that the client with mental illness was higher risk than the non-disordered client, even though the two clients had risk ratings based on a structured risk assessment tool and therefore in reality had the same likelihood of reoffending. Officers’ desire to revoke the client with mental illness was directly related to their inflated ratings of risk for the client with mental illness (Eno Louden et al., 2018; see also Eno Louden, 2009). 22 PERSPECTIVES VOLUME 44, NUMBER 2 Many officers worry they will be held responsible for a client with mental illness committing a serious offense in the community, so they may err on the side of revocation at the first sign of misbehavior as a form of risk management ( Lynch, 1998; Skeem et al., 2003). Other officers may file for revocation in an effort to get the offender into treatment—in jail (Lynch, 1998). As such, some technical violations recorded for offenders with mental illness may be driven by risk aversion or a misguided desire to help on the part of the supervising officer. Given that recidivism among offenders with mental illness is driven largely by technical violations rather than new offenses, it is important to attend to how community corrections officers make decisions regarding revocation for this group. MYTH: SYMPTOMS OF MENTAL ILLNESS DIRECTLY CAUSE RECIDIVISM FOR THOSE WITH MENTAL ILLNESS Perhaps the most surprising finding of recent research relates to the effect of mental illness on offending. Historically, policymakers and practitioners held the belief that symptoms of mental illness are the primary driver of offenses for justice-involved people with mental illness (Council of State Governments, 2002). This idea is intuitively appealing, and the logic behind it formed the basis for many types of correctional interventions aimed at reducing recidivism for this group, such as specialty probation caseloads and mental health courts (Skeem, Manchak, & Peterson, 2011). However, as with the other myths already described, research leads to a more complex view. The study of the effect of mental health symptoms on offending is a relatively new area of research, but a few high- quality studies have revealed consistent results. An exemplar study was conducted by Peterson and colleagues (2014), who conducted interviews and file reviews for 143 offenders with mental illness drawn from probation offices and a mental health court. The interviews elicited a timeline of criminal behavior and the circumstances surrounding each offense. Participants’ probation records were reviewed to corroborate the offense data and to confirm diagnoses. The interviews and records were painstakingly coded to determine the extent to which symptoms of psychosis, depression, or bipolar disorder preceded and directly caused each criminal act. Of the 429 offenses that were coded, only about 10% were directly related to symptoms (e.g., a person with paranoia is violent towards someone believed to be plotting against them) and another 28% were moderately related to symptoms (e.g., a person gets into a fight because of being agitated due to hearing voices earlier in the day, but hallucinations were not present at the time of the fight). The remaining offenses 23 AMERICAN PROBATION AND PAROLE ASSOCIATION were unrelated to symptoms (e.g., a person with mental illness steals food due to lack of money) (Peterson et al., 2014). These findings align with the other studies on the link between symptoms and offending (Junginger, Claypoole, Laygo, & Crisanti, 2006; Peterson, Skeem, Hart, Vidal, & Keith, 2010). Notably, Peterson and colleagues (Peterson, Skeem, Kennealy, Bray, & Zvonkovic, 2014) found variability within individuals, as those who committed one offense due to symptoms did not consistently commit offenses due to symptoms. This study also found differences in the extent to which symptoms caused offenses based on the offender’s diagnosis: Individuals with bipolar disorder (which is rare, even in correctional samples) were much more likely to commit crimes due to symptoms than were people with major depression (a much more common disorder). Because symptoms do not explain most offenses for most offenders with mental illness, symptoms should not be the primary focus in correctional treatment. BEYOND THE MYTHS TOWARDS EFFECTIVE SUPERVISION The above discussion tried to shed light on pervasive myths pertaining to people with mental illness who are involved in the criminal justice system. Those in this group are highly stigmatized, and the public has strongly held negative attitudes towards them. Clearly, such stigmatizing attitudes are in part rooted in inaccurate beliefs about people with mental illness, including the mental illness-violence relationship discussed earlier (Link et al., 1999). Since they too are members of the public, those in the probation and parole profession likely have some degree of stigmatizing attitude towards their clients with mental illness, whether they realize it or not. Even among people who consciously report non-stigmatizing attitudes, exposure to negative stereotypes regarding people with mental illness can affect decision- making towards those in this group (Stier & Hinshaw, 2007). As long as stereotypes consciously or unconsciously affect perceptions and are manifested in behavior, such stereotypes and myths have the potential to bias decision-making and negatively affect officers’ ability to engage in evidence-based correctional practices, thereby undermining practitioners’ efforts to rehabilitate their clients with mental illness. Taking the above into account, we will provide recommendations on best practices for the supervision of clients with mental illness. Contemporary research suggests that best practices for clients with mental illness should be informed by the Risk-Need-Responsivity model of correctional supervision (Bonta & Andrews, 2017). This model has been extensively studied among diverse types of offenders (e.g., women, youthful offenders) and 24 PERSPECTIVES VOLUME 44, NUMBER 2 Agency policy and associated officer training should clearly articulate the need to implement the risk assessment with fidelity and reinforce the idea that a mental health diagnosis alone does not automatically warrant placement in a high-risk category. when implemented with fidelity, meaningful reductions in recidivism can be achieved (Bonta & Andrews, 2017). There are some special considerations for applying this model to clients with mental illness (McCormick, Peterson-Badali, & Skilling, 2015; Morgan et al., 2012; Skeem et al., 2011), which we describe below, presented in chronological order based on which point in supervision they are relevant. The first step towards effective supervision involves using a validated risk and needs assessment to inform the intensity of intervention and how to target those interventions for each client (the Risk principle) (Bonta & Andrews, 2017). Because offenders with mental illness share the same criminogenic needs as those without mental illness, no special risk assessment tool is needed with this group (Bonta et al., 2014). The instrument must be implemented with fidelity, meaning that each officer is trained to administer and score the measure in a uniform manner, as deviation from the tool’s protocol undermines the tool’s ability to accurately predict risk (Andrews, Bonta, & Wormith, 2006). Implementing a risk assessment tool with fidelity is particularly important with clients with mental illness because officers’ intuitive judgments tend to overestimate their risk of recidivism (Eno Louden et al., 2018; Ricks, Eno Louden, & Kennealy, 2016). Officers often override the results of a risk assessment tool when the risk classification offered by the tool does not align with the officer’s judgment (Miller & Maloney, 2013), but overrides should be done only in rare cases in which the officer has information relevant to the risk of the client that is not adequately captured by the risk assessment tool (Bonta & Andrews, 25 AMERICAN PROBATION AND PAROLE ASSOCIATION 2017). Agency policy and associated officer training should clearly articulate the need to implement the risk assessment with fidelity and reinforce the idea that a mental health diagnosis alone does not automatically warrant placement in a high-risk category. If officers are made aware of the research showing decreased utility of risk assessment tools when overrides are inappropriately used, they may show increased buy-in for implementing the tool with fidelity, as officers who believe that their own judgment is superior to a risk assessment tool are more likely to deviate from the tool than are officers who have more faith in risk assessment tools (Schaefer & Williamson, 2018). Another helpful component of the intake process is assessment of clients’ mental health needs, so that interventions can be responsive to the mental health needs of the individual offender (a key component of the Responsivity principle) (Bonta & Andrews, 2017). Relying on offender self-report of mental health diagnosis or treatment need may provide agencies with an underestimate of the scope of mental health needs among their clients, because not all clients who have a mental illness will have received treatment in the past. A brief screening tool can be implemented at intake to identify clients who have symptoms of mental illness so those clients can be referred for further assessment by a mental health professional to determine whether a mental illness is present and, if so, the severity of mental health need (Chandler, Peters, Field, & Juliano-Bult, 2004; Eno Louden, Skeem, & Blevins, 2013). Case planning should be informed by the assessments of criminogenic and mental health needs, and Skeem et al. (2011) present a useful model for doing so. For clients with high levels of mental health treatment needs (e.g., pronounced symptoms and functioning impairments), intensive community-based mental health treatment, such as integrated dual- diagnosis treatment or forensic assertive community treatment may be warranted. Those with less pronounced mental health needs can be effectively served by high-quality outpatient psychiatric treatment. Psychiatric rehabilitation, which includes a multipronged approach to helping people with mental illness increase functioning (such as medication, family psychoeducation, and supported employment), has shown promise among justice-involved people (see Morgan et al., 2012). Because many offenders with mental illness have co- occurring substance use disorders, and substance misuse is a strong risk factor for recidivism, substance abuse treatment is a key component of rehabilitation for this group (Bonta & Andrews, 2017; Hartwell, 2004). 26 PERSPECTIVES VOLUME 44, NUMBER 2 The intensity of correctional interventions should align with the principles in the Risk-Need-Responsivity model (Bonta & Andrews, 2017), where intensive interventions for criminogenic needs are reserved for high-risk clients. Low-risk clients should receive less intensive supervision and intervention. Matching the intensity of interventions with the level of client need is crucial, because providing too much intervention to low-risk clients can actually increase their likelihood of offense (Bonta & Andrews, 2017). As with judgments of risk, officers’ judgments of need should take into account individual differences among clients. Not all clients with mental illness will have a high degree of need for mental health services, and in fact they may have stronger needs in other areas. Officer training should address the myth regarding the link between symptoms and offending as well as research demonstrating that mental health treatment alone does not reduce recidivism for offenders with mental illness (Skeem et al., 2011). Such training should also include a review of the research suggesting that most offenders with mental illness commit crimes for the same reasons as non-disordered offenders: criminogenic needs such as antisocial personality traits, pro-criminal thinking patterns, and substance abuse (Bonta & Andrews, 2017; Bonta et al., 2014; Skeem, Winter, Kennealy, Eno Louden, & Tatar, 2014). In accordance with the Need principle, officers should refer clients with these needs to interventions that target them, particularly interventions based on principles of Cognitive Behavioral Therapy (Bonta & Andrews, 2017). When routine supervision is underway, officers should strive to develop high quality “firm but fair” relationships with their clients characterized by trust, caring, and an authoritative (not authoritarian) approach (Skeem, Eno Louden, Polaschek, & Camp, 2007). This is a key component of Core Correctional Practice, a part of the Risk-Need-Responsivity model (Bonta & Andrews, 2017; Dowden & Andrews, 2004). Firm yet fair relationships foster rehabilitation among clients with and without mental illness (Kennealy, Skeem, Manchak, & Eno Louden, 2012; Skeem et al., 2007), whereas relationships characterized by surveillance, punishment, and mistrust hinder rehabilitation (see also Paparozzi & Gendreau, 2005). Strategies for building firm but fair relationships are a crucial part of training in evidence-based correctional practices (see Bonta et al., 2011). Because biases towards offenders with mental illness may hamper officers’ ability to build high- quality relationships with them, these biases should be addressed in training that targets relationship building. The research on interventions to alleviate mental health stigma provides some ideas on best practices for doing this, such as breaking down the myths surrounding 27 AMERICAN PROBATION AND PAROLE ASSOCIATION mental illness and providing examples of individuals who defy common stereotypes about people with mental illness (Corrigan et al., 2001; Sadow & Ryder, 2008). A key point where officers can make a difference during supervision is when the client struggles to comply with the terms of probation. Problem-solving discussions to identify and eliminate barriers to compliance are a key feature of many effective correctional supervision programs (Bonta et al., 2011; Robinson et al., 2012). As noted earlier, officers often find it frustrating to work with clients with mental illness, and this can be a particular challenge when the client has trouble meeting the conditions of supervision. It should be recognized, however, that clients with mental illness may be particularly sensitive to punishment-based strategies. A recent study found that probationers with mental illness whose supervising officers used threats of sanctions had higher violation rates than similar probationers whose officers avoided these negative strategies (Manchak, Skeem, Kennealy, & Eno Louden, 2014). Importantly, many agencies lack formal policies regarding how officers should respond to violations for clients with mental illness (Eno Louden, Skeem, Camp, & Christensen, 2008). Without formal policies in place, officers’ intuitive judgments regarding risk and negative attitudes towards offenders with mental illness affect their decision-making when it comes to filing (or not filing) a revocation request for a noncompliant client (Eno Louden, 2009; Ricks et al., 2016). Policies regarding when to file for revocation should be informed by risk assessments and can ensure that officers within an agency are responding to clients in a uniform manner and eliminate biased responses that may artificially inflate revocation rates for clients with mental illness. Although research on this practice among clients with mental illness is lacking, sanction matrices specifying the type of response to client misbehavior, taking into account client risk level and past noncompliance, have been used in settings with other challenging clients, such as drug courts (Guastaferro & Daigle, 2012). An additional tool available for agencies to explore is specialty mental health caseloads, which have been implemented by many agencies across the United States (Skeem et al., 2006). In these agencies, officers with expertise in mental health issues supervise relatively small caseloads, integrate resources available to them within their agency with resources from other agencies, and use problem-solving to address noncompliance rather than threats of incarceration. Rigorous studies of these caseloads suggest they are successful in reducing recidivism for clients with 28 PERSPECTIVES VOLUME 44, NUMBER 2 As noted earlier, symptoms don’t drive most offenses for most people with mental illness, so simply treating symptoms will make only minimal reductions in recidivism. serious mental illness (Manchak et al., 2014; Wolff et al., 2014). However, it should be noted that researchers are still working to disentangle the active ingredients of specialty caseloads, and the evidence thus far suggests that enforcement of mandated mental health treatment is not the primary mechanism by which these programs are effective (Manchak et al., 2014; Skeem et al., 2011). As noted earlier, symptoms don’t drive most offenses for most people with mental illness, so simply treating symptoms will make only minimal reductions in recidivism. Importantly, close monitoring of clients without addressing criminogenic needs may only serve to increase the likelihood of discovering bad behavior and do nothing to decrease recidivism (Petersilia & Turner, 1990). In fact, the research thus far suggests that the factors most related to recidivism reduction in specialty caseloads are those that overlap with the Risk-Need-Responsivity model (e.g., firm but fair relationships) (Manchak et al., 2014). Clients with mental illness can be taxing on agencies’ resources and officers’ patience. As stated by an officer who participated in a focus group regarding supervision of offenders with mental illness, “No, [we haven’t found anything that works] . . . we’re stalling. We’re baby-sitting until we get them off of our caseload whether we’re stalling them out, throwing them in and out of jail to get them through their minimum [sentence] or we’re ignoring them or we’re handing them off to different officers.” (Skeem et al., 2003, p. 442). Since that focus group was conducted more than 15 years ago, the research base has accumulated more knowledge regarding what agencies and officers can do 29 AMERICAN PROBATION AND PAROLE ASSOCIATION to successfully rehabilitate this group of clients. Supervision informed by research rather than myths regarding mental illness will likely lead to improved outcomes for clients and less frustration for practitioners. REFERENCES Andrews, D. A., Bonta, J., & Wormith, J. S. (2006). The recent past and near future of risk and/ or need assessment. Crime & Delinquency, 52, 7-27. Bonta, J., & Andrews, D. A. (2017). The psychology of criminal conduct (6th ed.). New York, NY: Routledge. Bonta, J., Blais, J., & Wilson, H. A. (2014). A theoretically informed meta-analysis of the risk for general and violent recidivism for mentally disordered offenders. Aggression and Violent Behavior, 19, 278–287. Bonta, J., Bourgon, G., Rugge, T., Scott, T., Yessine, A., Gutierrez, L., & Li, J. (2011). An experimental demonstration of training probation officers in evidence-based community supervision. Criminal Justice and Behavior, 38, 1127-1148. Chandler, R. K., Peters, R. H., Field, G., & Juliano-Bult, D. (2004). Challenges in implementing evidence-based treatment practices for co-occurring disorders in the criminal justice system. Behavioral Sciences and the Law, 22, 431-448. Chen, H., Cohen, P., & Chen, S. (2010). How big is a big odds ratio? Interpreting the magnitudes of odds ratios in epidemiological studies. Communications in Statistics - Simulation and Computation, 39, 860–864. Corrigan, P. W., River, L. P., Lundin, R. K., Penn, D. L., Uphoff-Wasowski, K., Campion, J., ... & Kubiak, M. A. (2001). Three strategies for changing attributions about severe mental illness. Schizophrenia Bulletin, 27, 187-195. Council of State Government (2002). Criminal justice/Mental health consensus project. Retrieved from Douglas, K. S., Guy, L. S., & Hart, S. D. (2009). Psychosis as a risk factor for violence to others: a meta-analysis. Psychological Bulletin, 135, 679-706. Dowden, C., & Andrews, D. A. (2004). The importance of staff practice in delivering effective correctional treatment: A meta-analytic review of core correctional practice. International Journal of Offender Therapy and Comparative Criminology, 48, 203-214. Eno Louden, J. (2009). Effect of mental disorder and substance abuse stigma on probation officers’ case management decisions. University of California, Irvine. Eno Louden, J., Manchak, S. M., Ricks, E. P., & Kennealy, P. J. (2018). The Role of Stigma Toward Mental Illness in Probation Officers’ Perceptions of Risk and Case Management Decisions. Criminal Justice and Behavior, 45, 573-588. Eno Louden, J., & Skeem, J. (2011). Parolees with mental disorder: Toward evidence-based practice. Bulletin of the Center for Evidence-Based Corrections, 7, 1-9. Eno Louden, J., & Skeem, J. L. (2013). How do probation officers assess and manage recidivism and violence risk for probationers with mental disorder? An experimental investigation. Law and Human Behavior, 37, 22-34. Eno Louden, J., Skeem, J. L., & Blevins, A. (2013). Comparing the predictive utility of two screening tools for mental disorder among probationers. Psychological Assessment, 25, 450- 461. Eno Louden, J., Skeem, J. L., Camp, J., & Christensen, E. (2008). Supervising probationers with mental disorder: How do agencies respond to violations? Criminal Justice and Behavior, 35, 832- 847. Feder, L. (1991). A comparison of the community adjustment of mentally ill offenders with those from the general prison population. Law and Human Behavior, 15, 477-493.Next >