Diversion of Veterans With Criminal Justice Involvement to Treatment Courts: Participant Characteristics and Outcomes

Abstract

Objective

This study compared characteristics and outcomes between veterans who participated in veterans treatment courts (VTCs) and veterans involved in criminal justice who participated in other treatment courts (TCs) or who participated in neither VTCs or TCs.

Methods

Data from 22,708 veterans (N=8,083 VTC participants, 680 participants in other TCs [other-TC participants], and 13,945 participants in neither VTCs nor TCs [non-TC participants]) in the Veterans Justice Outreach (VJO) program were analyzed by using multilevel regression models.

Results

VTC participants were more likely than other VJO participants to have served in Iraq or Afghanistan, but there were no sociodemographic disparities in access to VTCs. VTC participants were more likely than non-TC participants to have drug or public-order offenses, and they were more likely than other-TC participants to have DUI offenses. VTC participants had better independent housing outcomes than other VJO participants, and they had better employment outcomes than non-TC participants. However, VTC and other-TC participants were also more likely to have jail sanctions and new incarcerations compared with non-TC participants.

Conclusions

VTCs are a growing service model that serves a broad group of veterans with a range of criminal offenses. Although VTCs show moderate benefits in housing and employment, specialized services are needed to reduce recidivism and maximize these benefits.

The U.S. criminal justice system is facing a major crisis with overcrowding in prisons, inadequate mental health care for prisoners, and high recidivism rates (1–3). Treatment courts (TCs) offer a potential solution to some of these issues by diverting people to mental health and social services instead of incarceration. There are various types of TCs, including those dedicated to mental health, drug treatment, and community reentry. Veterans treatment court (VTC) is a new type of TC. Veterans are considered a population with unique needs because of their military experience (4). The U.S. Bureau of Justice Statistics estimates that 8% of all inmates in correctional facilities are veterans (5). Recent national concern for Iraq and Afghanistan veterans who are involved in the criminal justice system (6,7) and expansion of the Veterans Justice Outreach (VJO) (8) program in the U.S. Department of Veterans Affairs (VA) health care system have led to the development of VTCs throughout the country.

VTCs are a hybrid of mental health and drug courts and serve to address the needs of veterans who have been charged with criminal offenses (4,9). Eligibility requirements for admission to VTCs vary across jurisdictions (10); however, all courts follow a similar framework: veterans facing criminal charges who meet court admission requirements are provided the opportunity to avoid incarceration by receiving a reduced sentence or having the charges dropped once they successfully complete an individualized treatment program (9). Similar to practices at other types of TCs, participants are supervised by judges, and operations are managed by an interdisciplinary court team including representatives from the District Attorney and public defender’s offices, a probation officer, a treatment provider, and a court administrator. Unlike teams in other TCs, a VTC team also includes a VA representative, most often a VJO specialist, and a mentor coordinator, who matches the participant with a volunteer veteran mentor.

By the end of 2014, there were 351 veteran-focused courts in the country, which include separately designated VTCs and veterans’ dockets or tracks in mental health, drug, or criminal courts (11). Studies of VTCs so far have been limited to single-site studies (12) or have focused on VTC structure and admission criteria (13). This study contributes to the extant research by examining a national sample of veterans involved in criminal justice, comparing the characteristics and outcomes between veterans who were enrolled in VTCs nationally and those who received other specialized services.

Using national data from the VJO program, we conducted a retrospective study to first examine how VTC participants differed from participants in other TCs (other-TC participants) and from veterans who participated in neither VTCs nor TCs (non-TC participants) on sociodemographic, military, general medical, mental health, psychosocial, and legal characteristics at program admission. Other TCs included mental health courts, drug courts, and other specialized courts. We first compared background and health-related characteristics of the groups to examine whether there were disparities in access to TCs. Second, we examined how these groups fared on housing, employment, and legal outcomes at program exit, after the analyses were adjusted for differences among the groups at admission. We hypothesized that participants in VTCs would have better outcomes at program exit compared with participants of other TCs and with non-TC participants.

METHODS

National program data from the VJO program were extracted from the VA’s Homeless Operations Management and Evaluation System (HOMES). The aim of the VJO program is to avoid the unnecessary criminalization of mental illness and extended incarceration by ensuring that veterans involved in criminal justice have timely access to VA services, as clinically indicated. In most cases, to participate in the VJO program, veterans must be eligible for VA services. The VJO program consists of specialists who are responsible for direct outreach, assessment, and case management for justice-involved veterans in local courts and jails and who serve as a liaison with local justice system partners. A large component of the specialists’ work involves helping veterans enter VTC and other TCs (8).

The current study used data on 37,869 veterans across 142 VA-associated sites who were enrolled in the VJO program from July 2010 to November 2015. Veterans (N=15,161, 40%) who were missing program exit data (for example, because they were still in the program, they were lost to follow-up, or documentation was missing) were excluded, leaving a study sample of 22,708 veterans. There was nearly no difference in background characteristics between veterans who were included in the study and those who were excluded because of missing program exit data, except veterans who were included in the study were less likely to be in jail at program admission.

At program admission, information about whether veterans were enrolled in any treatment or specialty court and, if so, the type of court was recorded by VJO specialists. In cases in which veterans entered a treatment or specialty court after program admission, VJO specialists returned to the veterans’ admission form and updated information about the treatment or specialty court. In this study, VTCs were defined as specialty veterans courts or veterans dockets. Veterans were divided into three groups: those who entered VTCs (N=8,083; 35.6%); those who entered other TCs, such as drug treatment courts, problem-solving courts, and other specialty courts (N=680; 3.0%); and those who received VJO services but did not enter any TCs (N=13,945; 61.4%). Among those who entered other TCs, 262 (38.5%) entered drug treatment courts, 208 (41.2%) entered mental health courts, 47 (6.9%) entered domestic violence courts, 23 (3.4%) entered problem-solving courts, and 140 (20.6%) entered other specialty courts.

Measures

VJO specialists conducted in-person assessment interviews with veterans enrolled in the VJO program at admission and exit by using structured forms. At program admission, information on sociodemographic characteristics, military service history, general medical and mental health, and psychosocial and legal status was collected. At program exit, information on length of time in the program and housing, employment, income, benefits, treatment, legal, and program outcomes was documented.

Military history

Veterans were asked whether they had ever served in any major theaters of operations. Combat exposure was assessed by asking veterans whether they ever received hostile or friendly fire in a combat zone. Veterans were also asked whether they received any VA service-connected disability compensation for a psychiatric or other condition.

Psychosocial status

Veterans were asked where they spent each of the past 30 days, with responses collapsed into five categories: own place (owned or rented an apartment or house), someone else’s place (family or friend’s house or apartment), residential treatment or transitional housing (VA or non-VA residential treatment, domiciliary, transitional housing, or hotel), institution (hospital, prison, or jail), or homeless (shelter, outdoors, or automobile). Veterans who met the federal definition of chronic homelessness were also categorized as being chronically homeless (14). Employment history in the past three years was assessed and coded as either employed full-time or part-time, enrolled in a vocational rehabilitation program, not employed, or disabled or retired.

Legal status

VJO specialists documented whether the veteran was in jail at program entry and the type of offense the veteran was currently facing, including violent offense (for example, manslaughter, sexual assault, and robbery), property offense (for example, burglary, motor vehicle theft, and vandalism), drug offense (for example, possession and trafficking), public-order offense (for example, weapons offense, public intoxication, and disorderly conduct), probation or parole violation, or other offense. VJO specialists also recorded whether the veteran was involved in a driving-under-the-influence (DUI) offense or domestic dispute or had arrearage or delinquency problems involving child support orders.

Incarceration history was assessed by asking veterans about the total amount of time they had spent in jail or prison during their lifetime, which was categorized as none, one year or less, or more than one year.

Health status

Veterans rated their general health in the past month on a 5-point scale, which was categorized as excellent or very good, good or fair, and poor. Medical history was assessed by presenting veterans with a list of ten conditions, such as heart disease, and asking whether a doctor or nurse had ever told them that they had any of the conditions listed. Responses were summed for a total score. Psychiatric diagnoses were based on clinical impressions of VJO specialists, who are mostly social workers trained in conducting mental health assessments. VJO specialists determined psychiatric diagnoses on the basis of interviews with veterans and on review of any existing VA medical records. Veterans were also asked whether they had ever been psychiatrically hospitalized.

Outcomes

Veterans’ duration of involvement in the VJO program was recorded by VJO specialists. Veterans exited the program because of positive, negative, and neutral reasons. Housing and employment were assessed in the same manner at program exit and program admission. VJO specialists also documented the number of jail sanctions (incarceration ordered by a judge as a sanction for noncompliance with the treatment program or other infraction) as well as arrests and incarcerations for new offenses during a veteran’s time in the program.

Data Analysis

First, characteristics of VTC, other-TC, and non-TC participants at program admission were compared by using analysis of variance (ANOVA) and chi-square tests. Log transformations were conducted on income and housing variables because of nonnormal distributions. Because our hypotheses were focused on VTCs, the VTC participants served as the reference group in all group comparisons. Given large sample sizes, effect sizes (indicated by Cohen’s d score or change in percentage) were relied on more than statistical significance as indicators of group differences. Second, notable differences (d>±.5 or change in percentage of ±5 points) between groups were further tested with stepwise logistic regressions. Site was entered into a first block, and participant admission characteristics were entered into a second block by using a backward-elimination procedure. Third, the groups were compared on outcomes at program exit by using ANOVAs and chi-square tests. Fourth, multivariable analyses of outcomes were conducted by controlling for site and for differences between the groups at program admission by using generalized linear mixed modeling; site was entered as a random factor. Finally, two-block logistic regressions that adjusted for site effects were conducted to examine associations between number of jail sanctions, new arrests, and new incarcerations and other outcomes at program exit.

RESULTS

VTC Versus Non-TC Participants

Table 1 and Table 2 show bivariate comparisons between characteristics of VTC and non-TC participants at program admission. Notable differences were entered into a stepwise logistic regression, which revealed that VTC participants were more likely than non-TC participants to have served in Iraq and Afghanistan (odds ratio [OR]=1.14, 99% confidence interval [CI]=.98–1.33), to report combat exposure (OR=1.24, CI=1.07–1.44), to have a drug offense (OR=1.85, CI=1.59–2.16), and to have a public-order offense (OR=1.34, CI=1.17–1.54). VTC participants were less likely than non-TC participants to be in jail at program admission (OR=.14, CI=.12–.17), to be chronically homeless (OR=.83, CI=.70–.98), to have a probation offense (OR=.82, CI=.68–.99), to have any prior psychiatric hospitalizations (OR=.89, CI=.79–1.01), to have an affective disorder diagnosis other than bipolar disorder (OR=.89, CI=.79–1.02), and to report having spent fewer days in an institution in the past month (OR=.64, CI=.56–.73).

TABLE 1.

Characteristics of 22,708 participants in veterans treatment courts (VTCs), other treatment courts (TCs), or neither VTCs nor TCs (non-TC participants)

Characteristic	Group 1: VTC participants (N=8,083)		Group 2: other-TC participants (N=680)		Group 3: non-TC participants (N=13,945)		Test of difference	df	Effect sizea
	N	%	N	%	N	%			1 vs. 2	1 vs. 3
Sociodemographic
Age (M±SD)	43.7±13.9		45.8±12.9		45.0±13.2		F=30.0**	2, 22,667	d=−.16	d=−.10
Gender							χ2=7.0	2
Male	7,665	94.8	646	95.0	13,332	95.6			Δ%=−.2	Δ%=−.8
Female	418	5.2	34	5.0	613	4.4			Δ%=.2	Δ%=.8
Race							χ2 = 27.5**	4
White	5,296	68.6	387	60.5	9,127	68.0			Δ%=8.1	Δ%=.6
Black	2,109	27.3	231	36.1	3,836	28.6			Δ%=−8.8	Δ%=−1.3
Other	315	4.1	22	3.4	466	3.5			Δ%=.7	Δ%=.6
Ethnicity							χ2 = 145.4**	2
Hispanic	903	11.2	54	7.9	912	6.5			Δ%=3.3	Δ%=4.7
Not Hispanic	7,180	88.8	626	92.1	13,033	93.5			Δ%=−3.3	Δ%=−4.7
Marital status							χ2 = 54.6**	4
Married or coupled	1,958	24.8	134	20.5	2,827	20.9			Δ%=4.3	Δ%=3.9
Divorced, separated, or widowed	3,734	47.3	323	49.4	6,968	51.6			Δ%=−2.1	Δ%=−4.3
Single	2,195	27.8	197	30.1	3,708	27.5			Δ%=−2.3	Δ%=.3
Education (M±SD years)	13.2±1.8		13.0±1.8		12.9±1.7		F=39.4**	2, 21,728	d=.11	d=−.17
Monthly income (M±SD $)b	1,393.6±2,239.5		1,073.3±1,350.5		895.3±1,786.2		F=266.9**	2, 22,705	d=.11	d=−.32
Military history
Theater of operations							χ2=36.1**	10
WWII or Korean War	32	.6	1	.1	59	.4			Δ%=.5	Δ%=.2
Vietnam War	922	11.4	77	11.3	1,368	9.8			Δ%=.1	Δ%=1.6
Persian Gulf War	632	7.8	54	7.9	1,128	8.1			Δ%=−.1	Δ%=−.3
Afghanistan or Iraq	2,848	35.2	167	24.6	3,810	27.3			Δ%=10.6	Δ%=7.9
Other	678	8.4	47	6.9	926	6.6			Δ%=1.5	Δ%=1.8
None	2,971	36.8	334	49.1	6,654	47.7			Δ%=−12.3	Δ%=−10.9
Combat exposure	3,827	48.6	248	38.5	5,197	39.1	χ2 = 189.5**	2	Δ%=10.1	Δ%=9.5
VA service-connected disability
Psychiatric condition	1,422	17.6	134	19.7	2,083	14.9	χ2=33.9**	2	Δ%=−2.1	Δ%=2.7
Other condition	1,505	18.6	89	13.1	1,982	14.2	χ2=78.6**	2	Δ%=5.5	Δ%=4.4
Psychosocial status
Housing in past month (M±SD nights)c
Own place	14.8±14.7		10.4±13.7		8.0±12.2		F=478.3**	2, 22,705	d=.28	d=−.44
Someone else’s place	7.1±12.3		6.8±11.6		5.6±10.6		F=12.9**	2, 22,705	d=−.03	d=−.06
Residential or transitional	2.1±7.1		2.3±7.3		2.2±7.1		F=5.7*	2, 22,705	d=−.07	d=−.05
Institution	3.6±8.9		7.5±11.3		11.1±12.6		F=1,766.1**	2, 22,705	d=−.49	d=−.87
Homeless	1.4±5.8		1.5±5.6		1.6±5.8		F=8.2**	2, 22,705	d=−.06	d=−.06
Employment history in past 3 years							χ2 = 51.3**	6
Employed	3,386	42.8	235	35.7	5,358	39.7			Δ%=7.1	Δ%=3.1
Unemployed	1,999	25.3	174	26.4	3,816	28.2			Δ%=−1.1	Δ%=−2.9
Vocational rehabilitation	30	.4	8	1.2	96	.7			Δ%=−.8	Δ%=−.3
Disabled or retired	2,487	31.5	241	36.6	4,242	31.4			Δ%=−5.1	Δ%=.1
History of chronic homelessness	996	12.7	97	15.0	2,380	17.9	χ2 = 99.3**	2	Δ%=−2.3	Δ%=−5.2
Legal status
Incarceration history							χ2=439.3**	4
None	1,159	14.9	54	8.3	1,350	10.2			Δ%=6.6	Δ%=4.7
≤1 year	5,277	67.7	419	64.6	7,979	60.1			Δ%=3.1	Δ%=7.6
>1 year	1,362	17.5	176	27.1	3,954	29.8			Δ%=−9.6	Δ%=−12.3
In jail at program admission	688	8.5	184	27.1	7,064	50.7	χ2=4,016.9**	2	Δ%=−18.6	Δ%=−42.2
Legal offenses
Violent	1,734	21.5	124	18.2	3,606	25.9	χ2 = 67.4**	2	Δ%=3.3	Δ%=−4.4
Property	1,160	14.4	104	15.3	2,297	16.5	χ2 = 17.5**	2	Δ%=−.9	Δ%=−2.1
Drug	1,855	22.9	235	34.6	2,611	2,432	χ2 = 18.7	2	Δ%=−11.7	Δ%=4.2
Public order	2,830	35.0	185	27.2	3,886	27.9	χ2 = 126.9**	2	Δ%=7.8	Δ%=7.1
Probation or parole violation	534	6.6	83	12.2	2,432	17.4	χ2 = 517.5**	2	Δ%=−5.6	Δ%=−10.8
Other	689	8.5	62	9.1	1,669	12.0	χ2 = 65.5**	2	Δ%=−.6	Δ%=−3.5
DUI	2,539	31.4	130	19.1	2,913	20.9	χ2=316.9**	2	Δ%=12.3	Δ%=10.5
Domestic dispute	1,265	15.7	99	14.6	2,479	17.8	χ2 = 19.3**	2	Δ%=1.1	Δ%=−2.1
Child support issues	444	5.5	44	6.5	1,087	7.8	χ2=42.2**	2	Δ%=−1.0	Δ%=−2.3

^aA Cohen’s d score of .±.5 or a change in percentage (Δ%) of ±5 points indicates a notable difference.

^bA log transformation was conducted on monthly income for tests of difference.

^cA log transformation was conducted on all housing variables for tests of difference.

^*p<.01

^**p<.001

TABLE 2.

Health characteristics of 22,708 participants in veterans treatment courts (VTCs), other treatment courts (TCs), or neither VTCs nor TCs (non-TC participants)

Characteristic	Group 1: VTC participants (N=8,083)		Group 2: other-TC participants (N=680)		Group 3: non-TC participants (N=13,945)		Test of difference	df	Effect sizea
	N	%	N	%	N	%			1 vs. 2	1 vs. 3
General health							χ2=94.5**	4
Poor	904	11.6	81	12.5	1,896	14.2			Δ% = −.9	Δ% = −2.6
Fair or good	5,344	68.5	448	69.0	9,410	70.6			Δ% = −.5	Δ% = −2.1
Very good or excellent	1,552	19.9	120	18.5	2,024	15.2			Δ%=1.4	Δ%=4.7
N of medical conditions (M±SD)b	.7±.9		.9±1.0		.8±1.0		F=28.6**	2, 22,216	d=−.21	d = −.11
Psychiatric diagnosis
Alcohol use disorder	4,449	55.0	389	57.2	7,681	55.1	χ2=122	2	Δ% = −2.2	Δ% = −.1
Drug use disorder	3,051	37.7	370	54.4	5,823	41.8	χ2=88.6**	2	Δ% = −16.7	Δ% = −4.1
Psychotic disorder	378	4.7	73	10.7	933	6.7	χ2=62.6**	2	Δ%=−6.0	Δ% = −2.0
Bipolar disorder	633	7.8	80	11.8	1,369	9.8	χ2=29.9**	2	Δ% = −4.0	Δ% = −2.0
Other affective disorder	2,618	32.4	240	35.3	5,550	39.8	χ2=121.4**	2	Δ% = −2.9	Δ% = −7.4
PTSD	3,033	37.5	237	34.9	4,738	34.0	χ2=28.25**	2	Δ%=2.6	Δ%=3.5
Other anxiety disorder	1,732	21.4	130	19.1	3,194	22.9	χ2=10.5*	2	Δ%=2.3	Δ% = −1.5
Any psychiatric hospitalizations	2,733	34.0	283	41.7	5,676	40.9	χ2=104.2**	2	Δ% = −7.7	Δ% = −6.9

^aA Cohen’s d score of .±.5 or a change in percentage (Δ%) of ±5 points indicates a notable difference.

^bChosen by the veteran from a list of ten conditions

^*p<.01

^**p<.001

Table 3 shows bivariate comparisons between outcomes of VTC and non-TC participants. As Table 4 shows, after the analyses controlled for site and differences between the groups at program admission, VTC participants were in the VJO program longer than non-TC participants, and they were more likely to be housed in their own places and to be employed at program exit. VTC participants were also more likely than non-TC participants to have experienced any jail sanctions, new arrests, and new incarcerations during the program. These results remained the same after the analyses were adjusted for program tenure.

TABLE 3.

Outcomes at program exit of 22,708 participants in veterans treatment courts (VTCs), other treatment courts (TCs), or neither VTCs nor TCs (non-TC participants)^a

Characteristic	Group 1: VTC participants (N=8,083)		Group 2: other-TC participants (N=680)		Group 3: non-TC participants (N=13,945)		Test of difference	df	Effect sizeb
	N	%	N	%	N	%			1 vs. 2	1 vs. 3
Days in program	326.76221.5		268.66230.1		191.96232.4		F=895.54**	2 and 22,705	d=.26	d=.59
Housing							χ2=1,417.33**	8
Own place	4,624	66.7	280	52.4	4,375	41.0			Δ% = 14.3	Δ%=25.7
Someone else’s place	1,182	17.1	108	20.2	1,878	17.6			Δ% = −3.1	Δ% = −.5
Residential or transitional	457	6.6	82	15.4	2,120	19.9			Δ% = −8.8	Δ% = −13.3
Institution	604	8.7	56	10.5	2,028	19.0			Δ% = −1.8	Δ% = −10.3
Homeless	63	.9	8	1.5	269	2.5			Δ% = −.6	Δ% = −1.6
Employment							χ2=780.97**	6
Employed	2,241	32.8	108	20.7	1,580	15.9			Δ% = 12.1	Δ%=16.9
Unemployed	1,848	27.1	142	27.3	4,034	40.7			Δ% = −.2	Δ% = −13.6
Vocational rehabilitation	153	2.2	29	5.6	376	3.8			Δ% = −3.4	Δ% = −1.6
Disabled or retired	2,584	37.9	242	46.4	3,917	39.5			Δ% = −8.5	Δ% = −1.6
VA benefits							χ2=67.28**	4
Currently receiving benefits	4,019	64.0	289	57.8	5,334	57.9			Δ%=6.2	Δ%=6.1
Pending application	366	5.8	48	9.6	619	6.7			Δ% = −3.8	Δ% = −.9
No benefits	1,896	30.2	163	32.6	3,267	35.4			Δ% = −2.3	Δ% = −5.1
Non-VA benefits							χ2=42.63**	4
Currently receiving benefits	1,398	29.3	148	40.3	2,041	31.8			Δ% = −11.0	Δ% = −2.5
Pending application	153	3.2	21	5.7	282	4.4			Δ% = −2.5	Δ% = −1.2
No benefits	3,228	67.5	198	54.0	4,086	63.8			Δ% = 13.5	Δ%=3.7
Monthly income	952.9±1,838.8		663.0±1,264.8		469.0±1,441.3		F=424.97**	2, 22,705	d=.19	d=.40
(M±SD $)c
Jail sanctions							χ2=490.32**	4
0	6,445	79.7	514	75.6	12,477	89.5			Δ%=4.1	Δ% = −9.8
1–5	1,592	19.7	155	22.8	1,455	10.4			Δ% = −3.1	Δ%=9.3
≥6	46	.6	11	1.6	13	.1			Δ% = −1.0	Δ% = .5
New arrests							χ2=111.09**	4
0	7,027	86.9	588	86.5	12,709	91.1			Δ% = .4	Δ% = −4.2
1	813	10.1	75	11.0	1,008	7.2			Δ% = −.9	Δ%=2.9
≥2	243	3.0	17	2.5	228	1.6			.5	Δ%=1.4
New							χ2=187.16**	4
incarcerations
0	6,950	86.0	569	83.7	12,610	90.4			Δ% = 2.3	Δ% = −4.4
1	810	10.0	81	11.9	1,132	8.1			Δ% = −1.9	Δ%=1.9
≥2	323	4.0	30	4.4	203	1.5			Δ% = −.4	Δ%=2.5

^aAll participants were enrolled in the Veterans Justice Outreach program.

^bA Cohen’s d score of >±.5 or a change in percentage (Δ%) of ±5 points indicates a notable difference.

^cA log transformation was conducted on monthly income to normalize the data for tests of difference.

^**p<.001

TABLE 4.

Multivariable analyses of outcomes at program exit among 22,708 participants in veterans treatment courts (VTCs), other treatment courts (TCs), or neither VTCs nor TCs (non-TC participants)^a

Characteristic	VTC vs. other-TC participants		VTC vs. non-TC participants		Other-TC vs. non-TC participants
	OR	99% CI	OR	99% CI	OR	99% CI
Days in program	.19b**		.56b**		.23b**
Housed in own place	1.51**	1.16–1.95	2.23**	1.97–2.48	1.28	.99–1.67
Employed	1.40	.98–1.99	2.18**	1.88–2.51	1.40*	1.02–1.95
Receiving VA benefits	1.23	.93–1.63	1.40**	1.26–1.58	1.04	.82–1.32
Receiving non-VA benefits	.79	.57–1.09	1.27**	1.09–1.49	1.51**	1.12–2.01
Monthly incomec	.09b		.25b**		.11b*
Any jail sanctions	1.08	.79–1.48	4.14**	3.49–4.90	3.86**	2.83–5.26
Any new arrests	1.30	.90–1.86	2.48**	2.08–2.97	1.88**	1.32–2.66
Any new incarcerations	1.21	.85–1.70	2.44**	2.05–2.89	1.99**	1.43–2.77

^aAll participants were enrolled in the Veterans Justice Outreach program. The analyses controlled for site (N=142) and notable differences in characteristics of the groups at program admission.

^bCohen’s d score; d>±.5 indicated a notable difference.

^cA log transformation was conducted on monthly income to normalize the data.

^*p<.01

^**p<.001

VTC Participants Versus Other-TC Participants

Table 1 and Table 2 also show bivariate comparisons between VTC and other-TC participants at program admission. A stepwise logistic regression revealed that VTC participants were more likely than other-TC participants to have served in Iraq or Afghanistan (OR=1.35, CI=1.00–1.82), to have a DUI offense (OR=1.33, CI=.94–1.87), and to have a VA service-connected disability for a nonpsychiatric condition (OR=1.49, CI=1.02–2.17). VTC participants were less likely than other-TC participants to be in jail at program admission (OR=.48, CI=.34–.68) and to have a drug offense (OR=.76, CI=.55–1.04), a drug use disorder (OR=.71, CI=.53–.96), and a psychotic disorder (OR=.53, CI=.34–.85).

Table 3 shows bivariate comparisons between outcomes for VTC and other-TC participants. As Table 4 shows, after the analyses controlled for site and differences between the groups at program admission, VTC participants were more likely than other-TC participants to be housed at program exit.

Jail Sanctions, New Arrests, New Incarcerations, and Other Outcomes

Logistic regressions that adjusted for site showed that number of new arrests and incarcerations were each negatively associated with being housed in one’s own place versus not being housed in one’s own place at program exit (OR=.78, CI=.67–.92, and OR=.60, CI=.51–.71, respectively). However, number of jail sanctions was not associated with being housed in one’s own place versus not being housed in one’s own place at program exit. Number of new arrests (OR=.70, CI=.55–.88), new incarcerations (OR=.62, CI=.49–.78), and jail sanctions (OR=.84, CI=.71–.99) were each negatively associated with employment versus no employment at program exit.

DISCUSSION

In a national sample of over 20,000 veterans in the VJO program, over one-third (38.6%) were in some type of TC. A large majority of veterans who entered a TC were enrolled in a VTC, which likely reflects the VJO program’s focus on staffing VTCs. Notably over one-third of VTC participants served in Iraq or Afghanistan, and Iraq and Afghanistan veterans were more likely than veterans of other theatres of operations to enter VTCs. This finding may reflect public concern for the health and well-being of veterans of recent wars and the priorities of the VA, which has dedicated itself to providing these veterans with accessible treatment (15). There were no major sociodemographic disparities in access to VTCs, which is important given long-standing concerns about racial and ethnic bias in the criminal justice system and the disproportionate number of people from racial-ethnic minority groups who are imprisoned in the United States (16).

VTCs were originally conceived for combat veterans (17) and for veterans with nonviolent offenses (18). However, our findings clearly show that VTCs have broadened the population of veterans they serve to include noncombat veterans and violent offenders. Whereas VTC participants reported higher rates of combat exposure compared with non-TC participants, less than half of VTC participants reported combat exposure and barely over one-third were diagnosed as having posttraumatic stress disorder. Eligibility and procedures for VTCs vary by jurisdiction (10) and are affected by state legislation. For example, both Nevada and Texas have passed legislation requiring VTCs to serve only veterans who have brain injury, mental illness, or substance use disorders broadly related to military service; interestingly, this criterion has allowed veterans charged with a broader range of offenses to enter VTCs (18).

The most common offenses with which VTC participants were charged were DUI and public-order offenses; VTC participants were more likely than other-TC participants to have DUI offenses and more likely than non-TC participants to have a public-order offense. Although there was no difference between rates of violent offenses among VTC participants and other VJO participants, 22% of VTC participants were facing a violent offense at program entry. The U.S. Bureau of Justice Statistics has reported for over a decade that inmates who are veterans are more likely than other inmates to have been convicted of a violent crime (5,19). Our finding demonstrates that VTCs do not exclude veterans charged with violent crimes, which addresses concerns that VTCs “fence out many of the veterans whose crimes are most tied to their combat trauma” (18).

The outcome analyses partly supported our study hypotheses by finding that VTC participants had better independent housing outcomes than other-TC and non-TC participants. VTC participants also had better employment outcomes than non-TC participants, although there was no employment difference between VTC participants and other-TC participants. VTCs as well as other TCs offer the opportunity for charges to be dismissed or reduced so that veterans are less likely to have criminal records that can impede access to housing and employment opportunities (20,21). VTCs use a team approach in which VA providers and volunteer veteran mentors are involved in the rehabilitation process, and this unique approach may have contributed to these improved outcomes. The benefits of peer support and a recovery-oriented approach have been documented in various client populations (22–24) and may be relevant to veterans in VTCs as well. Many VTCs also have judges and legal providers who have developed expertise on veterans’ issues (18). Anecdotally, some VTCs have been described as having a strong sense of community for veterans and a cultural respect for their military service (10), which may also affect outcomes, although further research is needed to empirically examine these aspects of VTCs.

However, both VTC and other-TC participants were also more likely to have jail sanctions and new incarcerations compared with veterans who did not enter a TC. That outcome is likely due to the fact that TCs carefully monitor participants, which leads to more opportunities for jail sanctions and discovery of new offenses. TC participants also stayed longer in the program than non-TC participants, which was expected because TC participants often agree to accept supervision that lasts longer than a conventional sentence (18). Jail sanctions are an essential part of TCs, and previous studies have shown that jail sanctions may help discourage counterproductive behaviors, especially among individuals (25,26) with fewer instances of recidivism, so long-term evaluation of the impact of jail sanctions on outcomes is needed. Finally, these findings underscore the high recidivism rates among some veterans involved in criminal justice (27) and suggest that specialized interventions, such as moral reconation therapy (28), are needed for some VTC participants.

Several study limitations are worth noting. Participants were not randomly assigned to TCs, so no inferences about causation can be made. Data were based on VA administrative records and did not include measures of social support, psychiatric symptoms, and other psychosocial constructs that we could not control for in the analyses. Only veterans in the VJO program who had program outcome data (60% of total sample) were included in the study, so the results may not be generalizable to veterans involved in criminal justice who are not engaged in VA care or who dropped out of the program. This study provides data on a national level, but there is considerable variability in the operations of VTCs in different jurisdiction at the local level.

CONCLUSIONS

Notwithstanding its limitations, this study provides important insights about which veterans are being served by various TCs, the outcomes of veterans in various criminal-justice settings, and the value of a rehabilitative, therapeutic jurisprudence model in our court systems.