Keywords

Shay’s (1995) observations of the impact of combat trauma aligns with a common archetype of a military veteran in criminal justice system—a “wounded warrior” programmed for violence and exposed to trauma and other moral injury during his or her service, who now carries the scars of battle into a civilian life marked by recklessness and illegal activity. Certainly, this scenario is applicable to some veterans who become involved in the criminal justice system and is one that is commonly used in media reports of veterans who act out violently. However, this archetype fails to acknowledge the resiliency demonstrated by most veterans who successfully readjust to civilian life. Further, as illustrated by the second quote, involvement in the criminal justice system can serve as the impetus for rehabilitation for many veterans who struggle with addiction and other mental health issues. To facilitate such efforts, relevant stakeholders (e.g., policymakers, criminal justice personnel, treatment providers) should be aware of the unique criminogenic risks and needs that characterize veterans in the criminal justice system.

The most current estimates indicate that approximately 181,500 veterans are housed in jails and prisons, which represents 8% of the total incarcerated population in the United States (US; Bronson, Carson, Noonan, & Berzofsky, 2015). Importantly, nearly 70% of the US correctional population is supervised in the community on parole or probation (Kaeble & Glaze, 2016); thus, estimates on the number of incarcerated veterans are a fraction of the total number of veterans who are involved in some stage of the criminal justice system (i.e., arrest and initial detention, courts, community supervision, jails, prisons). We use the term justice-involved veterans to denote this larger population.

The past decade has seen a surge of empirical research on criminal justice involvement among veterans. The goal of this chapter is to provide an overview of this literature and highlight what is known, and what gaps in knowledge remain, regarding criminal justice involvement among veterans of the US military. First, we describe the most current data on the rate of criminal justice involvement in veterans and the types of offenses that are most common in this population. When available, we focus on the evidence of characteristics that distinguish veterans from their civilian counterparts in the criminal justice system. Second, we review studies that have examined associations between mental health problems (including substance use and trauma) and criminal justice involvement among veterans using a range of approaches. Here, the prevalence of mental health problems among veterans involved in the criminal justice system is highlighted, as are differences in the rate and type of these problems between justice-involved veterans and non-veterans. Third, we review research relevant to criminal recidivism among justice-involved veterans and use a leading model of offender rehabilitation—the Risk-Need-Responsivity (RNR) model—to frame the discussion. Fourth, we summarize research examining the impact of criminal justice involvement on veterans’ housing and employment statuses. Finally, we end by highlighting directions for future research to address gaps in knowledge regarding criminal justice involvement among veterans.

Prevalence and Type of Criminal Behavior Among Veterans

Incarcerated Veterans

Some of the most comprehensive data on the prevalence and type of criminal behavior among justice-involved veterans are for those incarcerated in jails and prisons. Such data are based on the National Inmate Survey , which is administered by the Bureau of Justice Statistics (BJS) , most recently in 2011–2012 (Bronson et al., 2015). BJS data are reported separately for inmates of jails and prisons, given that these populations differ in conviction status, offense distribution, and average length of stay. The estimated 181,500 veterans (8% of all incarcerated adults) in 2011–2012 is a decrease from the estimates reported by the BJS in 2004 (206,500; 9% of all incarcerated adults) and 1998 (225,700; 12% of all incarcerated adults).

In terms of their military history, most incarcerated veterans reported receiving an honorable military discharge, with only 5% receiving a dishonorable or bad conduct discharge. A little less than half of all incarcerated veterans served less than three years in a military branch, most commonly in the Army. The majority of incarcerated veterans in both jail (66%) and prison (67%) were discharged from the military after the Vietnam era but prior to conflicts in Iraq and Afghanistan (i.e., between 1974 and 2000). Veterans who were discharged during the most recent military era (Operation Enduring Freedom/Operation Iraqi Freedom/Operation New Dawn; OEF/OIF/OND) accounted for only 25% and 13% of inmates in jails and prisons, respectively. This accords with other research indicating that, across multiple age groups and within different race/ethnic groups, OEF/OIF/OND veterans are less than half as likely as veterans from other service eras to be incarcerated (Tsai, Rosenheck, Kasprow, & McGuire, 2013). However, compared to other incarcerated veterans, OEF/OIF/OND veterans are more likely to report combat exposure and have a diagnosis of posttraumatic stress disorder (PTSD; Tsai et al., 2013). While differences in trauma exposure between incarcerated veterans of different service eras are notable, only 31% and 25% of veterans in jails and prisons, respectively, report combat exposure during military service (Bronson et al., 2015). These rates are slightly lower than the rate of combat exposure in a nationally representative survey of US military veterans (e.g., 38%; Campbell, Wisco, Marx, & Pietrzak, 2017). Further, it suggests that difficulties coping with combat trauma may not be the most common pathway to criminal justice involvement for most incarcerated veterans.

Comparisons Between Veterans and Non-Veterans

Demographic Characteristics

The BJS data also afford comparisons between veteran and non-veteran inmates in prevalence of justice involvement and other characteristics. Among their differences, the rate of incarceration for veterans (855 per 100,000 veterans in the US, or 0.86%) is lower than the rate of incarceration for non-veterans (968 per 100,000 US residents, or 0.97%). One factor that may explain this difference is that incarcerated veterans are older than non-veterans, and risk of crime decreases with age (Sweeten, Piquero, & Steinberg, 2013). On average, veterans incarcerated in jail were 11 years older than non-veterans (43 years vs. 32 years), and veterans incarcerated in prison were 12 years older than non-veterans (49 years vs. 37 years). In terms of other demographics, incarcerated veterans comprise a smaller proportion of Black (non-Hispanic) and Hispanic inmates (44% in jails, 38% in prisons) relative to incarcerated non-veterans (59% in jails; 63% in prisons). A significantly smaller proportion of incarcerated veterans have never been married (32% in jails; 24% in prisons) compared to non-veteran inmates (61% in jails; 57% in prisons). Incarcerated veterans are also more educated than their non-veteran counterparts. For example, at the time of the survey, fewer incarcerated veterans in jail (22%) and prisons (28%) had not yet received their high school diploma or GED, relative to non-veteran inmates of jails (56%) and prisons (61%).

Offense Profiles

Incarcerated veterans can also be distinguished from their non-veteran counterparts with respect to offense profiles (Bronson et al., 2015). For example, veterans in prisons are less likely than non-veterans to have been convicted of property, drug, or driving under the influence/driving while intoxicated offenses. These differences are also observed for property and drug offenses among jail inmates. In addition, prison sentence lengths are longer for incarcerated veterans than non-veterans. In jails, the proportion of those not sentenced is nearly equivalent for veterans (53%) and non-veterans (52%), and among those serving sentences, the length of the sentence is not significantly different between the two groups. One of the most robust differences between incarcerated veterans and non-veterans is their likelihood of committing a violent offense. In both jails and prisons, incarcerated veterans are more likely to have committed a violent offense than non-veteran inmates; after adjusting for differences in age and race/ethnicity, 64% of veterans incarcerated in prisons had been sentenced for a violent offense, compared to 52% of non-veterans. A similar offense profile has been observed for justice-involved veterans from the OEF/OIF/OND era. Among a national sample of OEF/OIF/OND veterans who had a contact with re-entry services from the Veterans Health Administration (VHA), the most common offense was a violent offense (35%), followed by property (25%) and drug (24%) offenses (Tsai et al., 2013). Notably, among justice-involved veterans contacted by VHA in jails or courts, public order offenses (e.g., weapons offense, public intoxication) were most common (29%), followed by violent (25%) and drug offenses (22%) (Department of Veterans Affairs, 2012). These latter figures suggest that the higher rate of violent offending among incarcerated veterans (vs. non-veterans) in the BJS survey could reflect the fact that veterans who are non-violent offenders may have been diverted from prison.

Within the category of violent offenses, two types of offenses have been highlighted as driving this difference between veterans and non-veterans: sexual offenses and intimate partner violence. In the BJS data, differences in the rate of violent offending are due to a higher proportion of veteran inmates convicted of a violent sexual offense (35% vs. 23% for non-veterans). Regarding intimate partner violence, through deployments and reintegration, military service often puts stress on the family unit. When combined with other sequalae associated with military service (e.g., head trauma, PTSD, substance use), such stress can increase risk for physical conflicts between veterans and their partners (Gierisch et al., 2013). In accordance with this, a review of studies on intimate partner violence among veterans found the prevalence of this offense type to range from 14% to 58%, with rates highest among veterans with mental health problems (Marshall, Panuzio, & Taft, 2005). These rates are up to three times higher than those found among non-veterans (16%; Straus & Gelles, 1990).

Although comprehensive, the BJS only provides data on the prevalence and type of criminal behavior among incarcerated veterans. Population-based data can also be used to compare the rate of criminal justice involvement among civilians and veterans in the general population. For example, data from the National Study on Drug Use and Health from 2002 to 2014 suggests that military service members and veterans have higher lifetime rates of arrests than their civilian counterparts (Snowden, Oh, Salas-Wright, Vaughn, & King, 2017). However, the effects were driven largely by individuals who were only in the military briefly, rather than those with more extensive military careers. This aligns with the BJS data (Bronson et al., 2015), which indicates that nearly half of all incarcerated veterans served in the military less than three years. By comparison, an estimated 28% of active duty soldiers from 2004 to 2009 had less than three years of military service (Ursano et al., 2015). Thus, to the extent that veterans are at increased risk of criminal justice involvement relative to civilians, this may be driven by those who did not “fit” in the military culture and discharged early via self-selection or were removed administratively.

Mental Health Problems Among Justice-Involved Veterans

Associations between mental health problems and criminal involvement among veterans have been examined using several different approaches and sample types. The 2011–2012 data from the BJS provides estimates of mental health problems for incarcerated veterans. Approximately one-half of all veterans in prison (48%) or jail (55%) reported that they had been told by a mental health professional that they had a mental health disorder. These prevalence rates were significantly higher than those for non-veterans in prison (36%) or jail (43%). Among veterans in prison who were told they had a mental health disorder, the most common disorder was depression (27% vs. 24% for non-veterans), followed by PTSD (23% vs. 11%), bipolar disorder (17% vs. 16%), personality disorder (16% vs. 13%), anxiety disorder (12% vs. 11%), and schizophrenia or another psychotic disorder (10% vs. 9%). Among prison inmates who had a mental health disorder, veterans were significantly more likely to have PTSD or a personality disorder; all other disorder prevalence differences between veteran and non-veteran prisoners were not statistically significant. Similarly, among veterans in jail who were told they had a mental health disorder, the most common was also depression (34% vs. 30% for non-veterans), followed by PTSD (31% vs. 15%), bipolar disorder (27% vs. 23%), personality disorder (17% vs. 15%), anxiety disorder (19% vs. 17%), and schizophrenia or another psychotic disorder (13% vs. 14%). As was observed for prison inmates, among jail inmates who had a mental health disorder, veterans were significantly more likely to have PTSD or a personality disorder. In addition, jail inmates who were veterans were also significantly more likely to have depression, bipolar disorder, or anxiety disorder compared to their non-veteran counterparts.

Among prison inmates, no significant difference was found between veterans and non-veterans on serious psychological distress in the past 30 days (14% veterans, 15% non-veterans). However, veterans were more likely to be in treatment for a mental health problem (18% vs. 15% for non-veterans), taking prescription medication (14% vs. 12%), and receiving counseling or therapy from a trained professional (13% vs. 11%). Veterans were also more likely than non-veterans to have ever stayed in a hospital overnight for a mental health problem (26% vs. 22%). Among jailed inmates, there was similarly no significant difference between veterans and non-veterans on past 30-day serious psychological distress (28% veterans, 26% non-veterans). However, jailed veterans were more likely than jailed non-veterans to be in treatment for a mental health problem (26% vs. 19%), taking prescription medication (22% vs. 17%), and receiving counseling or therapy from a trained professional (11% vs. 8%). Jailed veterans were more likely than jailed non-veterans to have ever stayed in a hospital overnight for a mental health problem (35% vs. 28%).

The BJS data also provide associations between combat experience and mental health indicators among incarcerated veterans (Bronson et al., 2015). Understanding the relation between combat trauma and criminal offending is a unique concern among justice-involved veterans (Blodgett et al., 2015). Among veterans in prison, those with combat experience were more likely to have an indicator of a mental health problem (64% vs. 49% of those without combat experience), equally likely to have had serious psychological distress in the past 30 days (16% vs. 13%), and more likely to have been told by a mental health professional that they had a mental health disorder (59% vs. 44%). Among veterans in jail, those with combat experience were also less likely to have no indicator of a mental health problem (27% vs. 41% of those without combat experience), more likely to have past 30-day serious psychological distress (31% vs. 27%), and more likely to have been told by a mental health professional that they had a mental health disorder (67% vs. 49%).

Associations between mental health problems and criminal involvement among veterans is further informed by a systematic review of the research literature in this area (Blodgett et al., 2015). Across 13 samples that reported a general rate of mental health problems among justice-involved veterans in either incarcerated or community settings, 13–62% of veterans had some mental health problems. Rates at the lower end of the range were based on more strict measures (e.g., a veteran reporting that a mental health professional had diagnosed them with a specific condition), whereas rates on the higher end were based on less strict measures (e.g., a veteran reporting any symptoms of a mental health disorder). Regarding specific mental health disorders, the prevalence of PTSD in general ranged from 4% to 39% (five samples), and the prevalence of combat-related PTSD ranged from 5% to 27% (four samples). Diagnostic rates at the higher end of these ranges were based on a formal assessment instrument. The prevalence of anxiety ranged from 10% to 51% (six samples), depression from 14% to 51% (nine samples), bipolar disorder from 3% to 11% (four samples), and adjustment disorder from 8% to 61% (four samples). In contrast to the wide ranges for these specific disorders, the diagnostic rate for the broader category of mood disorders ranged from 19% to 29% (5 samples). The prevalence of psychotic disorders ranged from 4% to 14% (four samples). Only two samples examined suicide in justice-involved veterans, with reports of 7% and 16% prevalence of suicidal ideation, and 0% and 1% prevalence of suicide attempts. In terms of substance use disorders , 21–71% of justice-involved veterans had an alcohol use disorder (15 samples), 26–65% had a drug use disorder (11 samples), and 57–61% had an alcohol use disorder, a drug use disorder, or both (two samples). Co-occurring mental health and substance use disorders were found for 23% and 53% of justice-involved veterans, respectively, in the latter two samples.

Four studies specifically compared justice-involved veterans to other justice-involved adults (i.e., non-veterans) on rates of mental health and substance use problems (Blodgett et al., 2015). Findings across these studies were mixed: in contrast to differences in the rate of mental health problems that were observed in the BJS data (Bronson et al., 2015), few significant differences were found in other studies identified in the systematic review. In addition, five studies compared justice-involved veterans to other veterans (Blodgett et al., 2015). Despite differences in how these studies were designed and how they defined the justice-involved and comparison groups, justice-involved veterans were consistently found to have more mental health problems, including substance misuse and co-occurring mental health and substance use disorders, than other veterans. For example, currently incarcerated veterans were more likely than veterans living in the community to report having had prior treatment for, or diagnosis of, a mental health disorder (13% vs. 8%; Greenberg & Rosenheck, 2009). Regarding alcohol use disorders , comparisons between justice-involved veterans and other veterans found rates of 44% versus 13% (Erickson, Rosenheck, Trestman, Ford, & Desai, 2008), 29% versus 13% (Black et al., 2005), and 48% versus 42% (McGuire, Rosenheck, & Kasprow, 2003), respectively. Regarding drug use disorders , comparisons yielded rates of 49% versus 7% (Erickson et al., 2008), and 62% versus 39% (McGuire et al., 2003) for justice-involved veterans versus other veterans, respectively.

Yet another approach to estimating the prevalence of mental health problems among justice-involved veterans has been to examine data for veterans served by VHA ’s Veterans Justice Programs. One of these programs—Health Care for Reentry Veterans—links veterans to VHA and community services upon reentry from state and federal prisons. Among veterans with an outreach visit from a Reentry Specialist who subsequently received VHA care, 69% were diagnosed with at least one mental health or substance use disorder (57% with at least one mental health disorder, and 49% with at least one substance use disorder; Finlay et al., 2017). The most common mental health disorders were depression, PTSD , and anxiety . The most common substance use disorders were alcohol, other drug, and cocaine use disorders. Thirty-five percent of justice-involved veterans seen in VHA were diagnosed with co-occurring mental health and substance use disorders.

The prevalence of mental health and substance use disorders has also been reported for veterans in treatment courts or jails who received an outreach visit from a Specialist from VHA’s Veterans Justice Outreach program (Finlay et al., 2015). This study also provides insight into gender differences. Among women, the prevalence of mental health and substance use disorders was 88% and 58%, respectively, compared to 76% and 72% among men. Women had higher odds than men of being diagnosed with any mental health disorder, depression, PTSD, anxiety, bipolar disorder, and personality disorders. Conversely, women had lower odds than men of being diagnosed with any substance use disorder, and with an alcohol, cocaine, cannabis, or other drug use disorder, specifically. Women also had lower odds than men of being diagnosed with co-occurring mental health and substance use disorders.

Gender differences in mental health indicators among justice-involved veterans have also been examined using data from the Jail Diversion and Trauma Recovery Program , which was initiated in 2008 by the Substance Abuse and Mental Health Services Administration (Stainbrook, Hartwell, & James, 2016). The goal of this program was to support the implementation of jail diversion in 13 states for persons with PTSD and other trauma-related disorders, with a priority emphasis on veterans. The rates of mental health and substance use problems were high for both women and men in this sample, but female veterans reported higher rates of mental health problems, whereas male veterans reported higher rates of alcohol use problems. Specifically, more female veterans had a history of mental health treatment, met criteria for PTSD, and had moderate or extreme difficulty related to global mental health and depression. Female veterans also had higher PTSD symptom severity scores. In contrast, male veterans were more likely than women to report heavy alcohol use in the past 30 days. There were no differences between women and men for illegal drug use, which was reported by slightly over one-third of the full sample. Overall, these findings are similar to those obtained with a different sample and definition of justice involvement (Finlay et al., 2015).

The studies reviewed thus far in this section have examined rates of mental health disorders among justice-involved veterans. Another approach to establishing links between mental health and criminal involvement among veterans is to examine the criminal histories of veterans in treatment for mental health problems. For instance, one study investigated the prevalence of specific types of criminal arrests among a large nationally representative sample of male patients in VHA addiction treatment programs, all of whom served in the military before September 11, 2001 (Weaver, Trafton, Kimerling, Timko, & Moos, 2013). Among these patients, 85% had at least one lifetime criminal charge, and 58% had at least three such charges. These charges were categorized as specific to drugs (25% of patients), driving under the influence (52%), not related to drugs or alcohol and non-violent (69%), or violent (25%). In addition, 46% of patients had at least one lifetime conviction, and 17% had at least three such convictions. Several comparisons were made between patients with comorbid alcohol and drug use disorders versus those with discrete alcohol or drug use disorders. The former group had greater odds of lifetime criminal charges (any and repeated), and of being on parole or probation at the time of treatment admission, than those with discrete alcohol- or drug-related diagnoses. Similarly, those with comorbid alcohol and drug use disorders had more than double the odds of convictions relative to those with just alcohol use disorders, whereas there was no reliable difference in odds in comparison to those with drug use disorders only. Patients with comorbid alcohol and drug use disorders had the greatest odds of reporting prior arrests for nonviolent crimes compared to both substance-specific groups. They also evidenced increased odds of having any prior violent charge relative to patients with discrete alcohol use disorders.

Another study of criminal involvement among veterans in VHA addiction treatment found that patients clustered into three profiles based on their criminal history, which was assessed by type of offense, number of convictions, and number of months incarcerated (Schultz, Blonigen, Finlay, & Timko, 2015). The three types of criminal history profiles identified were mild (low numbers of criminal offenses, convictions, and months incarcerated; 79% of patients); moderate (high number of public order offenses, repeated convictions, and >3 years incarcerated; 14% of patients); and severe (violent criminal offenses, repeated convictions, and >10 years incarcerated; 7% of patients). Patients with mild criminal histories had more severe alcohol problems than patients with severe criminal histories, whereas patients with moderate and severe criminal histories were more likely to report having had trouble controlling violent behavior in the 30 days before treatment. However, all groups improved during treatment such that they did not differ on alcohol or drug use severity or violent behavior one year after entering treatment.

A third study investigated veterans in addiction or mental health treatment who reported any lifetime history of military or non-military trauma exposure (Bennett, Morris, Sexton, Bonar, & Chermack, 2017). Overall, 46% reported a history of any violent or nonviolent legal charge. More specifically, 22% endorsed a history of any violent offense, most commonly for assault (20%), and 39% endorsed a history of any non-violent offense, most commonly forgery (34%) or violation of probation or parole (24%). Fifteen percent of the sample endorsed having had both a violent and nonviolent legal charge.

In summary, across both incarcerated and community settings, justice-involved veterans appear to have higher rates of mental health disorders than other veterans, with substance use disorders and PTSD among the most prevalent conditions. Based on several indicators, mental health problems appear to be more common for incarcerated veterans than non-veterans. PTSD and personality disorders, in particular, are consistently more prevalent among veterans in jails and prisons. Similar to gender differences in the general population, male justice-involved veterans tend to be at higher risk for substance use disorders , whereas female justice-involved veterans are at higher risk for other mental health conditions. Among veterans in treatment for substance use or mental health problems, a history of criminal justice involvement is the norm rather than the exception. A history of criminal justice involvement is also more common among those with co-occurring mental health and substance use disorders and those with polysubstance use disorders.

Criminal Recidivism Among Justice-Involved Veterans

Critical to the rehabilitation of justice-involved veterans and other adults is knowledge of the factors that drive risk for criminal behavior. Knowledge of these factors can guide assessment and treatment planning efforts with justice-involved adults to minimize risk for future run-ins with the legal system. At the heart of this issue is managing risk for recidivism, defined as rearrest, reconviction, or reincarceration for a new offense or violation of the terms of conditional release . For justice-involved adults generally, recidivism is the norm rather than the exception. For example, based on data compiled by the BJS on adult prisoners released from 30 states in 2005, 68% were rearrested within three years of release, and 77% were rearrested within five years of release (Durose, Cooper, & Snyder, 2014).

Comparisons Between Veterans and Non-Veterans

Whether the rate of recidivism is similar or different for incarcerated veterans and non-veterans following correctional release has, to date, not been directly estimated. However, some insights into this issue can be gleaned from the 2011–2012 National Inmate Survey, which examined differences between incarcerated veterans and non-veterans in their criminal justice histories (Bronson et al., 2015). In both jails and prisons, veterans are more likely than non-veterans to report being first-time offenders. Specifically, after adjusting for differences in age and race/ethnicity, 32% of veterans in jails (vs. 25% of non-veterans) and 27% of veterans in prisons (compared to 23% of non-veterans) had no prior history of incarceration. Further, 43% of veterans in prison (vs. 55% of non-veterans) and 62% of veterans in jails (vs. 68% of non-veterans) reported four or more prior arrests. Although firm conclusions cannot be drawn from these figures alone, they suggest that the rate of recidivism may be lower for incarcerated veterans than non-veterans. Nonetheless, these figures also illustrate that recidivism is still the norm among justice-involved veterans, given that 68% of veterans in jails and 73% of veterans in prisons had at least one prior episode of incarceration. Consistent with this, data from VHA’s Veterans Justice Program indicates that veterans served by these programs have an average of eight arrests in their lifetime (Department of Veterans Affairs, 2012). Together, these data suggest that many justice-involved veterans, like their non-veteran counterparts, are caught in a cycle of contact with the criminal justice system.

The Risk-Need-Responsivity Model: Application to Research on Recidivism Risk Among Justice-Involved Veterans

The Risk-Need-Responsivity (RNR) model is one of the leading frameworks for effective correctional rehabilitation (Andrews & Bonta, 2010a, 2010b). This model outlines the concepts and practices that have the strongest empirical support for reducing risk for recidivism among offenders. As reflected in its name, risk, need, and responsivity represent the model’s core principles. Several studies and meta-analyses have shown that interventions and services that adhere to these principles have the most evidence for reducing criminal recidivism among offenders (Andrews & Bonta, 2010a, 2010b; Lipsey & Cullen, 2007). This is particularly true for services that attend to program integrity in terms of selecting skilled staff, providing appropriate training and ongoing supervision, and using structured and manualized approaches (Bonta & Andrews, 2017). In the following sections, we review these core principles and discuss their potential application and relevance to efforts to study recidivism risk among justice-involved veterans (see Table 2.1 for a summary of this literature).

Table 2.1 Risk-Need-Responsivity (RNR) model principles and crime prevention initiatives for veterans
Full size table

The Risk Principle and Veterans

The risk principle refers to who should be treated. Specifically, this principle emphasizes the importance of matching service intensity to offenders’ level of risk for recidivism, and prioritizing resources and intensive services for offenders who are at moderate to high risk of criminal recidivism. Accordingly, offenders who are estimated to be at low risk for recidivism would not be referred to more intensive programming and would not be assigned to programs with high-risk offenders. Indeed, some evidence suggests that low-risk offenders who are assigned to intensive programs or monitored frequently may have a higher likelihood of negative outcomes (Andrews & Dowden, 2006). This may be due to affiliation with high-risk offenders in such programs or disruption of low-risk offenders’ prosocial networks and resources (Andrews & Dowden, 2006). At present, research to determine whether services for justice-involved veterans are adhering to these aspects of the risk principle is limited. For example, there is some evidence that veteran treatment courts are effective in reducing recidivism in this population (e.g., Hartley & Baldwin, 2016). However, it has also been suggested that higher-risk cases are systematically screened out of admission to these courts based on veterans’ offense severity or eligibility for VHA services (Erickson, 2016).

Adherence to the risk principle of the RNR model also goes along with the assessment principle , which directs the use of structured recidivism risk assessments that validly differentiate low-risk from high-risk cases (Bonta & Andrews, 2007). Actuarial and structured professional judgment approaches to recidivism risk assessment have consistently shown predictive superiority to unstructured professional judgment (Fazel, Singh, Doll, & Grann, 2012; Grove, Zald, Lebow, Snitz, & Nelson, 2000; Hanson & Morton-Bourgon, 2009). Structured assessments of recidivism risk, particularly fourth-generation tools that facilitate comprehensive RNR treatment planning and services delivery (e.g., Level of Service/Case Management Inventory; Andrews, Bonta, & Wormith, 2004) are not typically employed in diversion and rehabilitation services for veterans (Rodriguez et al., 2017). Thus, the implementation of tools that facilitate adherence to the risk principle may not be widely implemented in crime prevention services for justice-involved veterans.

In terms of the development and validation of criminogenic risk assessment tools for the veteran population, such work to date has been limited to the prediction of violence risk. A five-item violence risk screening tool, the “Violence Screening and Assessment of Needs” (VIO-SCAN) , was derived using two veteran samples—a random national sample of post-9/11 veterans, and a self-selected regional sample of veterans from the same service era (Elbogen et al., 2014). The violence risk factors measured by the VIO-SCAN are financial instability, combat experience, alcohol misuse, history of violence or arrests, and comorbid anger and probable PTSD. Items are scored dichotomously and added together to yield a total score ranging from 0 to 5. In the samples noted above, the VIO-SCAN total score was modestly to strongly predictive of any violence or aggression over the course of 1 year. Although these findings provide preliminary support for use of the VIO-SCAN to predict risk for violence among veterans, the screener is not regarded as an actuarial tool per se. Consequently, total scores on the VIO-SCAN should not be used to assign veterans into probabilistic risk categories (e.g., low, moderate, or high risk). Rather, the screener should be used to identify which veterans should be referred for a more comprehensive risk assessment. Although no structured violence risk assessment tools exist for such a follow-up assessment, evidence-based guidance is available (Elbogen et al., 2010).

The Need Principle and Veterans

The need principle refers to what should be treated. Specifically, this principle emphasizes that rehabilitation efforts should primarily target “criminogenic needs”—i.e., factors that are robust predictors of criminal recidivism and are changeable. Research to identify the criminogenic needs of justice-involved adults has been led by Andrews and Bonta (2010b), who highlighted the “Central Eight” risk factors for recidivism. These risk factors, derived from multiple meta-analyses (Andrews & Bonta, 2006), consist of the following: (1) history of antisocial behavior —particularly early, frequent, and varied antisocial activities; (2) antisocial personality pattern —traits such as impulsivity, hostility, sensation-seeking, and callous disregard for others; (3) antisocial cognitions —attitudes and beliefs that support a criminal identity and rationalization of criminal acts; (4) antisocial associates —close relationships with individuals who engage in or are supportive of criminal behavior; (5) family or marital dysfunction; (6) lack of positive involvement in school or work; (7) lack of positive involvement in prosocial activities, such as leisure and recreation; and (8) substance use.

The Central Eight represent intermediate targets for rehabilitation due to their theorized functional relationship to criminal behavior. By contrast, factors such as low self-esteem, low intelligence, emotional distress, or diagnoses of major depression or serious mental illness are generally weak predictors of recidivism and are thus categorized as non-criminogenic needs (Andrews & Bonta, 2006). Targeting such needs alone—which may need to be addressed for humanitarian, motivational, or other reasons—without a substantial focus on concomitant criminogenic needs is likely to be insufficient in reducing risk of recidivism among most criminal offenders. Further, the need principle also goes along with a breadth principle for treatment planning, which involves targeting multiple criminogenic needs when working with high-risk individuals. With a higher risk level comes a greater number of criminogenic needs. Thus, when working with high-risk individuals it is best to target the full range of criminogenic needs rather than focus treatment planning solely or predominantly on one or two risk factors or on non-criminogenic needs (e.g., a mental illness that does not have a case-specific connection to an individual’s offending behavior).

The Need Principle and Veterans: The Central Eight

The validity of the Central Eight in the prediction of criminal recidivism among civilians has been well established; however, the extent to which these risk factors apply to justice-involved veterans has received less empirical attention. The importance of conducting this research is underscored by the fact there are differences between justice-involved veterans and non-veterans in various demographics linked to recidivism (e.g., age, marital status, education, and employment), as well as differences between these groups in the prevalence of mental health conditions such as substance use disorders and PTSD that may suggest a relationship to criminal involvement (Bronson et al., 2015).

In response to this gap, Blonigen et al. (2016) reviewed the literature to identify studies examining one or more of the Central Eight risk factors for criminal justice involvement or criminal recidivism in samples that were exclusively or predominantly veterans. Thirteen studies were identified; however, due to the relatively small number of studies, Blonigen and colleagues conducted a narrative review rather than synthesizing the data through meta-analytic techniques. Notably, no studies were identified that systematically tested the Central Eight risk factors as predictors of criminal recidivism in veterans, which represents a significant gap in the extant literature.

Regarding antisocial history, cognitions, peers, and personality, each risk factor had at least one study that found a significant link with criminal justice involvement among veterans. As for the other Central Eight risk factors, substance use was consistently linked to higher risk for criminal justice involvement in veteran samples, with findings robust across different measurements of substance use and across different service eras (e.g., Vietnam, OEF/OIF/OND) . Among all of the Central Eight risk factors, substance use was most commonly examined as a correlate or predictor of criminal involvement among veterans. Evidence for an association between the remaining three risk factors of the Central Eight and criminal involvement in veterans was either mixed (family/marital dysfunction; school/work involvement) or no studies were identified (prosocial activities). Not included in the review by Blonigen et al. (2016) was a more recent longitudinal study, which found that family problems were a significant predictor of future legal problems among veterans in mental health treatment (Timko, Finlay, Schultz, & Blonigen, 2016).

The Need Principle and Veterans: Beyond the Central Eight

In addition to the Central Eight, there may be other criminogenic needs that are more common among, or unique to, veterans. Specifically, trauma exposure and PTSD, traumatic brain injury (TBI), and homelessness are all more prevalent among veterans than non-veterans, particularly those with a history of criminal involvement (Tsai, Rosenheck, Kasprow, & McGuire, 2014). We turn our attention to these additional needs in the sections below.

Trauma and PTSD

In the review by Blonigen et al. (2016), several studies were identified that observed a significant link between combat exposure and PTSD and risk for violent behavior. For example, multiple studies have found significant associations between combat exposure with PTSD and both general aggression and intimate partner violence (Elbogen et al., 2010). Other work suggests that links between PTSD and violence in the veteran population may not be direct and may depend on other intervening variables. For instance, in a large sample of OEF/OIF/OND veterans, PTSD was found to be a significant predictor of post-deployment arrests among individuals reporting high levels of anger or irritability (Elbogen, Johnson, Newton, et al., 2012). Notably, this combination of PTSD and anger/irritability was significant after controlling for history of prior arrests and substance use problems. PTSD has also been found to be more strongly linked to violent behavior among individuals with comorbid substance use disorders (Greenberg & Rosenheck, 2009). For example, recent work examining predictors of criminal justice involvement among substance-using veterans seeking specialty mental health care in VHA found that, after controlling for various demographic factors and cocaine use, PTSD symptom severity was associated with violent, but not non-violent, criminal charges (Bennett et al., 2017). Additional analyses suggested that this effect may have been particularly driven by intrusive symptoms (e.g., recurrent, involuntary distressing memories of the trauma). Collectively, these studies suggest that PTSD may be a specific pathway to criminal justice involvement, but that (a) the significance of this risk factor may be augmented by the presence of other criminogenic needs, and (b) the link between PTSD and criminal involvement in this population is more strongly related to violent offending .

TBI

The relationship between TBI and criminal justice involvement is often described as a function of the behavioral changes associated with TBI, such as increased impulsivity, aggression, and low frustration tolerance. Accordingly, TBI has been linked to increased risk of violent criminal offenses in the general population (Farrer, Frost, & Hedges, 2012), particularly in the presence of co-occurring mental health problems (Trudel, Nidiffer, & Barth, 2007). TBI has been highlighted as a specific mental health concern among incarcerated veterans (Pinals, 2010; Rosenthal & McGuire, 2013), and is linked to a higher risk of violent behavior in this population (Elbogen et al., 2010). The relevance of this issue is often raised for veterans who served in combat during the Iraq and Afghanistan eras, due to the common use of intermittent explosive devices in these conflicts (Hoge et al., 2008). However, whether TBI is uniquely related to criminal justice involvement in veterans above and beyond PTSD is difficult to determine, given that the two are marked by similar symptoms, behavioral changes, and often caused by the same event. In a study of Iraq and Afghanistan returnees, TBI in combination with anger/irritability was not significantly linked to an increased risk of post-deployment arrests after accounting for the effects of PTSD and substance use problems (Elbogen, Johnson, Newton, et al., 2012). Thus, the extent to which TBI is linked to criminal involvement among veterans may be a function of other comorbid mental health problems (Sreenivasan et al., 2013).

Homelessness

Some evidence suggests that the link with criminal involvement may be more common among veterans than non-veterans . For example, a BJS report indicated that among state prisoners, a higher proportion of veteran than non-veteran inmates were homeless prior to incarceration (12% vs. 10%; Mumola, 2000). Other data from VHA indicates that 30% of incarcerated veterans have a history of homelessness (Tsai et al., 2013), compared to a rate of 18% among a nationally representative sample of adults in the US with a history of incarceration (Greenberg & Rosenheck, 2013).

Outside of studies reporting differences in prevalence, there is limited research that has directly examined whether homelessness is a unique risk factor for criminal involvement in the veteran population. However, the link between financial stability and post-deployment adjustment has been examined in a large sample of OEF/OIF/OND veterans (Elbogen, Johnson, Wagner, Newton, & Beckham, 2012). In this study, across mental health diagnoses, greater financial stability was significantly associated with lower likelihood of arrests or aggression. Considered in reverse, financial instability, which may be conceptualized as a correlate of or precursor to homelessness, may be linked to criminal involvement among OEF/OIF/OND veterans.

The Need Principle and Veterans: Synthesizing the Literature

Subsequent to Blonigen et al.’ (2016) narrative review of RNR and veteran-specific risk factors for criminal justice involvement among veterans, King and Wade (2017) conducted a meta-analysis of the same studies. They broadly conceptualized the outcomes for 12 of the studies as justice involvement, and the outcome for the 13th study as violence. The 13 studies were coded for all indicators of the Central Eight as well as for PTSD, TBI, and homelessness/severe financial problems. Although more studies are being added to the meta-analysis, preliminary results (see Table 2.2) are that the number of studies per risk factor ranged from none (leisure–recreation) to 9 (substance use), with a mode of two studies (antisocial cognitions, homelessness or severe financial problems, and TBI). Most of the Central Eight risk factors—all but education and employment problems—were reliably associated with justice involvement among veterans. The magnitude of these effects was small and roughly consistent with that which is observed with offenders in general (Bonta & Andrews, 2017). Evidence in support of Blonigen et al.’s (2016) suggested veterans-specific risk factors was strongest for TBI.

Table 2.2 Preliminary meta-analysis of central eight risk factors and veteran-specific risk factors for justice involvement among veterans
Full size table

Although King and Wade’s (2017) findings are notable, it must be acknowledged that there was substantial heterogeneity for most risk factors examined in this meta-analysis, which was likely due to the high degree of measurement variability for both predictors and outcomes in the observed studies. Nonetheless, these preliminary meta-analytic results concur with the conclusions of the narrative review of this literature (Blonigen et al., 2016) and justify more research on the Central Eight risk factors in veterans. A useful future direction would be prospective, multi-wave validation studies of either general offender criminogenic risk assessment tools, or veteran-specific criminogenic risk assessment tools, in samples of justice-involved veterans.

The Responsivity Principle and Veterans

The responsivity principle refers to how justice-involved individuals should be treated and consists of two components—general responsivity and specific responsivity. General responsivity emphasizes the importance of using structured treatments with a cognitive-behavioral orientation to reduce risk for recidivism among offenders. In support of this principle, meta-analyses of cognitive-behavioral treatments for reducing recidivism, which generally focus on restructuring maladaptive cognitions and behaviors, indicate significant reductions in the rate of recidivism (ranging from 8% to 25%) relative to comparison treatments (Aos, Miller, & Drake, 2006; Landenberger & Lipsey, 2005; Wilson, Bouffard, & MacKenzie, 2005). These reductions have been observed across a range of offender types.

The cognitive-behavioral interventions that have shown the most promise in reducing risk of recidivism are ones that focus on restructuring antisocial thinking (Blodgett, Fuh, Maisel, & Midboe, 2013). The most commonly studied treatments of this kind are Moral Reconation Therapy (MRT; Little & Robinson, 1988), Thinking for a Change (T4C; Bush, Glick, & Taymans, 2011), and Reasoning and Rehabilitation (R&R; Ross, Fabiano, & Ross, 1986). Each intervention is manualized, delivered in a group format, and uses exercises and homework assignments to modify antisocial personality patterns, cognitions, and affiliations. In so doing, these interventions were designed to directly address the criminogenic needs from the RNR model that are associated with the highest risk of recidivism (Andrews & Bonta, 2010b). In terms of their evidence, meta-analyses of MRT (Aos et al., 2006; Ferguson & Wormith, 2013; Little, 2005) and R&R (Aos et al., 2006; Tong & Farrington, 2006; Wilson et al., 2005) have found that justice-involved adults receiving these treatments have significantly lower rates of recidivism relative to participants receiving other interventions or no treatment. For example, compared to participants from control conditions, the rate of recidivism is reduced by one-third among justice-involved adults receiving MRT (Ferguson & Wormith, 2013). However, none of these studies was a randomized controlled trial. The research base for T4C is much less extensive than MRT or R&R ; however, multiple studies have also reported positive effects on recidivism (Lee et al., 2012; Lipsey & Cullen, 2007) and social and interpersonal functioning (Golden, 2002).

For justice-involved veterans, particularly those who are eligible for and linked to VHA services (Blue-Howells, Clark, van den Berk-Clark, & McGuire, 2013), adherence to the principle of general responsivity is demonstrated by VHA’s commitment to providing evidence-based, cognitive-behavioral interventions for veterans with substance use and other mental health disorders (Karlin & Cross, 2014). Such interventions often include relapse prevention and social skills training, which do not directly target antisocial thinking but have nonetheless been found to be effective for improving outcomes among justice-involved adults (Milkman & Wanberg, 2007). Further, among veterans with co-occurring substance use and mental health disorders, one study found a 33% decrease in criminal recidivism among veterans who received services in VHA, but a 48% increase among veterans who received services from the state (Pandiani, Ochs, & Pomerantz, 2010).

Cognitive-behavioral interventions that more directly target antisocial thinking (MRT, T4C , and R&R ) have not yet been implemented systematically within VHA or other offender rehabilitation settings for veterans. Interviews with Specialists from VHA’s Veterans Justice Programs found that, by and large, justice-involved veterans have access to services that address most risk factors of the Central Eight, but access to services and cognitive-behavioral interventions that directly target antisocial thinking are limited (Blonigen et al., 2016). Indeed, no trials to date have examined the efficacy or effectiveness of MRT , T4C , or R&R with justice-involved veterans. However, a multisite randomized controlled trial of MRT for justice-involved veterans in VHA mental health residential treatment programs is currently underway (Blonigen et al., 2018). The results will provide the first test of the effectiveness and potential widespread implementation in VHA of a cognitive-behavioral treatment for antisocial thinking to reduce recidivism and improve health outcomes among justice-involved veterans.

The specific responsivity principle emphasizes the importance of tailoring or adapting services or interventions to recipients’ unique strengths and characteristics (e.g., learning styles, intellectual abilities; demographic or cultural factors) to facilitate their full engagement and participation in treatment. A review of evidence-based treatments for criminal recidivism, including MRT, T4C, and R&R, noted that during reintegration (i.e., entering civilian life after military service), veterans often feel isolated and disconnected from loved ones and other civilians and, in general, struggle with interpersonal relationships (Timko et al., 2016). Such struggles can lower tolerance for frustration, increase suspicion of others’ intentions, and ultimately increase risk for criminal involvement (Brown, 2008; Halvorson, 2010). Thus, treatments for recidivism for veterans may need to incorporate more trust-building activities and emphasize the formation of healthy interpersonal relationships. Self-stigma regarding substance use, mental health issues, and associated problems such as criminal justice involvement is also common among veterans (Glynn et al., 2014). To mitigate this stigma, treatments for recidivism could be adapted for veterans by framing the treatment as a “class” or “education” rather than therapy per se.

Specific responsivity often includes consideration of cultural factors. In this vein, treatments for recidivism could incorporate veteran culture into their curricula, similar to how military service is emphasized as a relevant demographic factor in veteran treatment courts (Douds, Ahlin, Howard, & Stigerwalt, 2017; Vaughan, Holleran, & Brooks, 2016). Among the three main treatments for antisocial thinking, only the materials for MRT have been adapted for veterans (Little & Robinson, 2013). These adaptations entailed revisions to the workbook, titled “Winning the Invisible War,” which includes veteran-centric examples and stories rather than changes to the content of the MRT steps or exercises. The ongoing trial of MRT in VHA (Blonigen et al., 2018) is using the new veteran-specific curriculum.

Impact of Criminal Records on Employment and Housing Among Veterans

Critical to breaking the cycle of recidivism and promoting long-term recovery among justice-involved veterans is identifying and addressing the barriers to employment and stable housing for these veterans following release from correctional settings. Those with a criminal history may face a range of barriers to finding employment or housing (Pager, 2003; Western, Kling, & Weiman, 2001). In this section, we review extant research examining the impact of a criminal record on the employment and housing statuses of formerly incarcerated veterans.

Employment

As for many justice-involved adults , unemployment is a significant issue among justice-involved veterans. The most recent BJS data indicate that approximately one-quarter of veterans incarcerated in jails and prisons were unemployed in the month prior to their arrest (Mumola, 2002). For these and other justice-involved veterans, a key question concerns the major barriers to employment in this population. This question was the focus of a narrative review that sought to identify the most salient barriers to employment faced by justice-involved adults and examine their generalizability to justice-involved veterans (McDonough, Blodgett, Midboe, & Blonigen, 2015). Thirty-two studies were reviewed and eight barriers were identified using qualitative methods. The study concluded that most of the employment barriers that justice-involved veterans likely face—i.e., lack of education and vocational skills; lack of job-readiness skills and criminogenic thinking; competing needs (e.g., mental health and other substance use problems); homelessness; legal restrictions; and employer stigma and criminal background checks—are barriers faced by justice-involved adults more generally. However, some important nuances to these findings were highlighted. For example, regarding lack of education and vocational skills, incarcerated veterans tend be more highly educated than their non-veteran counterparts (Bronson et al., 2015). In addition, veteran inmates are more likely to have held a job in the month prior to arrest (e.g., 78% vs. 67% in state prisons; 72% vs. 63% in jails; Mumola, 2000). Thus, a lack of education or vocational skills may not be as prominent of a barrier to employment for justice-involved veterans as other justice-involved adults.

In terms of competing needs and homelessness, compared to non-veterans in the criminal justice system, veterans have higher rates of substance use and other mental health disorders (Bronson et al., 2015) and homelessness (Fargo et al., 2012). Thus, it may be critical for employment interventions for justice-involved veterans to be integrated with treatment for substance use, other mental health problems, and housing assistance. Finally, in terms of legal restrictions, denying employment to applicants solely because of their criminal records may violate Title VII of the Civil Rights Act of 1964 (Equality Employment Opportunity Commission, 2012). However, such applicants can, in certain circumstances, be disqualified from employment in jobs that would put them in contact with vulnerable groups such as children or the elderly (Equality Employment Opportunity Commission, 2012). The offense profile of justice-involved veterans is marked by greater prevalence of violent criminal charges, particularly sexual offenses, than non-veterans, which may preclude more veterans from employment in positions that put them in contact with vulnerable groups.

Two unique barriers to employment for justice-involved veterans have also been identified (McDonough et al., 2015). First, employers can receive information on the military discharge status of applicants, and they may be reluctant to hire individuals with a less than honorable discharge. Less than a quarter of incarcerated veterans in jails and prisons received such a discharge from the military (Bronson et al., 2015); however, those who did tended to have lengthier and more serious criminal histories, as well as higher levels of prior substance use problems (Noonan & Mumola, 2007). Another unique barrier to employment for justice-involved veterans is entitlements and financial disincentives (e.g., VA disability compensation), which may reduce motivation to find formal or full-time employment (Tsai & Rosenheck, 2013, 2016). For example, most justice-involved adults face a host of financial obligations that may subject them to wage garnishment following release from prison (i.e., having a certain amount of one’s paycheck automatically withheld and sent directly to another institution or individual to pay off a debt; Visher, LaVigne, & Travis, 2004). Among service-connected veterans, a concern that employment would reduce receipt of benefits has been linked to an increased willingness to turn down a job (Meshberg-Cohen, Reid-Quinones, Black, & Rosen, 2014). Similarly, a national study of Veterans Treatment Court participants observed that those receiving VA or non-VA benefits were less likely to be employed (Tsai, Finlay, Flatley, Kasprow, & Clark, 2018). Thus, through efforts such as expanding vocational rehabilitation services, or offering benefits counseling at the time of benefits application to address potential misconceptions (Tsai & Rosenheck, 2013), there may be value in helping justice-involved veterans overcome disincentives to seeking or obtaining employment.

Few studies have directly examined the impact of a criminal record on employment among veterans (see McDonough et al., 2015). A recent study examined whether history of criminal justice involvement and other factors were associated with employment among homeless veterans across 19 sites in the Housing and Urban Development-Veterans Affairs Supportive Housing (HUD-VASH) program from 1992 to 2003 (Tsai & Rosenheck, 2016). A history of criminal justice involvement was not significantly associated with job attainment or earnings, whereas a diagnosis of a psychotic disorder and reliance on public-support income were negatively associated with these outcomes. In contrast, a secondary analysis of data from a randomized trial comparing supported employment with treatment as usual among job-seeking veterans with spinal cord injuries found that participants with felony convictions were generally less likely to find employment regardless of study condition (LePage, Ottomanelli, Barnett, & Njoh, 2014).

Taken together, the reviewed findings are mixed in terms of whether a criminal record directly impacts the employment prospects of justice-involved veterans. Nonetheless, barriers to work such as financial disincentives and competing needs (e.g., psychiatric problems; McDonough et al., 2015) may serve as indirect pathways that impact homeless veterans with a history of justice involvement. Such effects may also hamper the impact of evidence-based work-related interventions, such as supported employment. Further, as previously discussed, employment deficits have been associated with justice involvement among veterans, and thus may constitute a criminogenic risk factor for justice-related outcomes.

Housing

Homelessness and criminal justice involvement are closely intertwined (Greenberg & Rosenheck, 2008), and arrest history predicts longer duration of homelessness (Caton et al., 2005). It may be asked whether a history of incarceration has an impact on veterans’ housing status following reentry from a correctional setting. Release from a correctional setting is a high-risk period for many individuals and can lead to returns to homelessness for those who lack sufficient assistance or support (Metraux, Byrne, & Culhane, 2009). Studies have examined the role of criminal history on the housing status of veterans participating in permanent supportive housing programs, which combine permanent housing subsidies with supportive services such as case management to assist veterans with obtaining and maintaining their housing. The HUD-VASH program is the largest supported housing program in the country for homeless veterans. Among veterans in HUD-VASH from 1992 to 2003, neither having a substance use problem nor a more extensive criminal history were associated with how quickly veterans became housed (Tsai, O’Connell, Kasprow, & Rosenheck, 2011). Another study compared the outcomes of veterans enrolled in HUD-VASH with different criminal histories over a one-year period (Tsai & Rosenheck, 2013). At time of entry into HUD-VASH, most participants had at least one criminal charge (79%) and those with more extensive criminal histories had poorer housing status. However, after enrollment into the housing program, extent of criminal history was not associated with housing status and all groups showed substantial improvement on housing.

More recent research on veteran participants from the HUD-VASH program has examined factors affecting premature exit from this program, including the impact of criminal justice history. Among veterans who enrolled in HUD-VASH, several psychosocial factors have been linked to premature program exits (i.e., before being placed in permanent community housing), including a diagnosis of SUD and criminal justice involvement (Gabrielian et al., 2016). Data from a multisite study of the HUD-VASH program has also been used to identify factors associated with exiting the program due to incarceration and returning to homelessness (Cusack & Montgomery, 2017). While only 6.6% of exits were due to incarceration, veterans with a previous incarceration were 13 times more likely to exit the program because of reincarceration. A drug use disorder diagnosis also increased risk of this outcome more than two-fold, and a decrease in outpatient visits for substance use prior to exit increased risk for the outcome nearly four-fold. Finally, a history of incarceration, either prior to program entry or at the time of exit, was a significant predictor of experiencing another period of homelessness after program exit. Collectively, these findings highlight the vicious cycle of homelessness and incarceration that occurs among many veterans, and the need for housing and reentry services (i.e., services to assist persons with reintegration into the community from correctional custody) to assess for this risk and intervene to break the cycle. It is noteworthy that the receipt of service-connected income reduced the risk of exiting the HUD-VASH program due to incarceration by half (Cusack & Montgomery, 2017). Thus, services that provide employment training to assist veterans with obtaining a stable income during reentry from jail or prison may be beneficial in helping veterans break the cycle of homelessness and incarceration.

Directions for Future Research

The available literature on justice-involved veterans highlights the unique pathways to criminal justice involvement in this population and the range of characteristics and risk factors that characterize this distinct yet heterogeneous group. Importantly, while veterans make up a significant sub-population of adults incarcerated in jails and prisons in the US, most veterans never become involved in the criminal justice system. However, for those who do, substance use and mental health problems are common, with the available evidence suggesting that these issues are more common among veterans than non-veterans in correctional settings.

Although veterans who become involved in the criminal justice system appear to have lower rates of criminal recidivism than non-veterans, reoffending is still the norm among justice-involved veterans. Nevertheless, there remains an absence of studies directly estimating the rate of criminal recidivism among justice-involved veterans, relative to that of their non-veteran counterparts. Further, research focused on the development and validation of criminogenic risk assessment tools for the veteran population may be needed. The development of a violence risk screening instrument for use with veterans is promising, but there are as of yet no validated comprehensive risk assessment tools for general, violent, or sexual recidivism among veterans. As such veteran-specific tools are developed, it will be critical to verify whether they provide more accurate predictions of recidivism risk among justice-involved veterans than other established risk assessment tools for the general offender population.

The RNR model can serve as a useful framework for studying risk assessment and management with justice-involved veterans. In accordance with the need principle, prospective, multi-wave studies are needed to systematically identify the dynamic risk factors that predict recidivism in this population. Ideally, such studies would determine whether there are different risk factors, or differences in the effect size for these factors in the prediction of criminal recidivism, between veterans and non-veterans, and whether promoting change in these risk factors reduces risk for future criminal justice involvement in veterans. In this vein, the extent to which trauma/PTSD, TBI, and homelessness/financial instability may be veteran-specific risk factors for recidivism—exerting effects beyond the generally applicable Central Eight criminogenic needs from the RNR model—should be clarified. Future research exploring factors that moderate and mediate associations between risk factors and criminal involvement and recidivism among veterans will be beneficial. More research is also needed to understand the types of interventions that are most effective for reducing risk for criminal recidivism among veterans, and if and how treatments for criminal recidivism that were developed and validated with non-veterans need to be adapted to meet the needs of justice-involved veterans.

In terms of employment status among formerly incarcerated veterans, future studies should directly assess which barriers to employment for justice-involved adults are most relevant to the veteran population. At present, the available literature suggests that incorporating substance use or mental health treatment into employment training programs may be critical for justice-involved veterans. There is also value in further investigating how best to overcome the financial disincentives to seeking or obtaining employment among veterans who receive public benefits. Regarding housing status, formerly incarcerated veterans appear to benefit from supportive housing programs as much as veterans without a criminal history; however, formerly incarcerated veterans are at increased risk for dropout, particularly if they have a more extensive criminal history and do not stay engaged in addiction treatment services. Research examining how best to maintain veterans’ engagement in employment training services may help to mitigate this criminogenic risk.

In conclusion, the available literature highlights a number of key differences between justice-involved veterans and their civilian counterparts and illustrates how consideration of an individual’s military history can provide important context to understanding their pathway to criminal involvement. More research is needed to fully understand the unique criminogenic risks and needs of this group of justice-involved adults. A richer understanding of these issues will aid in the development and implementation of the highest quality correctional services for the men and women who served their country but are caught in a cycle of involvement with the criminal justice system.