Family Factors Associated With Delinquency Outcomes in Court-Involved Youth in Mental Health Treatment

Abstract

Objective

To examine the impact of baseline family functioning and parental monitoring on engagement in and severity of delinquent acts of court-involved youth (CIY) after 6 months of mental health treatment.

Method

Adolescent (mean age =15.16 years) CIY (N = 165) recruited from 2 US cities completed questionnaires at baseline and at 6 months during their court-mandated mental health treatment with a 71% (n =117) completion rate. Youth were mostly male (61.5%) and White (64.1%). Baseline demographics and psychosocial variables that were significantly associated with 6-month delinquency engagement and severity in initial analyses were entered into regressions.

Results

In initial analyses, baseline alcohol use, cannabis use, parental monitoring, and family functioning were associated with 6-month delinquency engagement and severity (ps < .05). Regressions demonstrated that after controlling for baseline psychiatric symptoms and demographics, baseline alcohol use, cannabis use, and parental monitoring had small to medium effects on 6-month delinquency engagement and severity.

Conclusion

In CIY enrolled in mental health treatment, youth with substance use and less parental monitoring at baseline were more likely to have higher engagement in and severity of delinquency at 6 months. This suggests that clinical interventions that target these factors could reduce delinquency. Future directions for this research include improving our understanding of the biopsychosocial factors in this population and better tailoring of existing family-based interventions that target substance use for CIY in mental health treatment.

Clinical Trial Registration Information

Integrated Mental Health Treatment & HIV Prevention for Court-Involved Youth (ITP); https://clinicaltrials.gov/study/NCT01421485.

Diversity & Inclusion Statement

We worked to ensure sex and gender balance in the recruitment of human participants. We worked to ensure that the study questionnaires were prepared in an inclusive way. We worked to ensure race, ethnic, and/or other types of diversity in the recruitment of human participants. One or more of the authors of this paper self-identifies as a member of one or more historically underrepresented sexual and/or gender groups in science.

Plain language summary

This study utilized the longitudinal data of 117 court-involved youth who were mandated to mental health treatment in two Eastern US cities. The results showed that substance use and less parental monitoring at the start of treatment predicted engagement in and severity of delinquency after six months of treatment. Findings suggest that interventions tailored toward specifically targeting these factors may help to reduce delinquency.

Between 2018 and 2020, approximately 2 million youths younger than age 18 received dispositions by the US court systems. Of these youths, approximately 74.4% were not detained.^1,2 One potential disposition for this large group of court-involved youth (CIY) is referral to outpatient mental health services. This pathway reflects the high rates of psychiatric illness within CIY populations, including disorders of attention, anxiety, mood, and substance use.3, 4, 5 Psychiatric symptoms^6,7 and disorders^8,9 as well as substance use^7,8,10 have also been associated with delinquent behaviors. Therefore, diversion to mental health treatment could be a potential means of reducing further involvement in the legal system and the negative health outcomes associated with detention. To date, studies have found that receiving mental health treatment reduces time-to-reoffending and recidivism rates in CIY and emerging adults.11, 12, 13, 14

Further, therapies that involve the family have also been helpful for CIY. Families can play a crucial role in youth behaviors as illustrated by the social personal framework (SPF), which postulates that adolescent risk-taking does not occur in isolation, but rather in the setting of personal attributes, family context, peer relationships, and environmental circumstances.¹⁵ Consistent with this theoretical framework, research has demonstrated associations between family functioning,^16,17 parental mental health,¹⁶ parenting practices,^18,19 and behaviors in CIY. Interventions for juvenile offenders that involve caregivers have shown promise in targeting unsafe behaviors, such as substance use^20,21 and illicit behaviors that result in legal system re-entry.21, 22, 23 Effective family-based interventions for CIY include multisystemic therapy^13,14 and functional family therapy^22,23 as well as interventions that largely focus on affect management.²⁰ Family-based interventions may be of particular benefit to CIY who are not detained because families can have a direct role in supporting and managing behavior of youth. However, these programs, particularly multisystemic therapy, can be intensive in terms of commitments of patient time, staff support, and resources. Intensive family programs that are separate from mental health treatment of youth also may not be feasible for family systems already under significant stress and in instances where these services are not covered by insurance providers.

Despite evidence that individual mental health treatment and family involvement are important factors in the delinquency outcomes of CIY, few studies have examined the impact of specific family factors on delinquency outcomes during the course of mental health treatment. This is a considerable gap as, in keeping with the SPF, symptoms and behaviors of CIY cannot be viewed in isolation from the family system. Therefore, establishing which specific, modifiable, psychosocial factors impact youth delinquency can allow for the creation of feasible and tailored interventions that effectively address the needs of this vulnerable population, prevent negative long-term outcomes, and be easily incorporated into mental health services. With these aims, we previously investigated the effects of family functioning and parental monitoring on the severity and variety of delinquent acts by CIY at the start of their court-mandated mental health treatment. This investigation demonstrated significant associations between parental monitoring and youth delinquent behaviors.²⁴ Therefore, for this current investigation, our objective was to examine if baseline family functioning and parental monitoring continue to have an impact on delinquency in CIY after they have received 6 months of mental health treatment. We hypothesized that baseline family functioning and parental monitoring would significantly predict later engagement of CIY in delinquency as well as the severity of acts of delinquency due to the notable influence family factors can have on youth behaviors as postulated by the SPF.

Method

Participants

Adolescents (N = 598) from 2 cities in the Eastern United States were referred for study participation by court officials (eg, intake worker, probation officer, magistrate, or judge) after adjudication to mental health treatment between November 2011 and April 2015. Youth were eligible if they were able to speak and read English, between the ages of 11 and 17 years, had an open petition with the partnering family court at the time of referral, and residing with a legal guardian who was able to attend weekly therapy sessions. Referred adolescents were excluded if they were currently enrolled in mental health treatment; had prior psychotic symptoms that necessitated intensive treatment at a higher level of care; or required evidence-based care not included in the current intervention, including youth charged with sexual offenses and/or with diagnoses of obsessive-compulsive disorder or pervasive developmental disorder. Approximately 53% (n = 317) of referred adolescents were eligible to participate, among whom 54% (n = 170) provided consent and 52% (n =165) enrolled. Of the enrolled participants, 71% (n = 117) completed assessments at baseline and at 6 months.

Procedures

Adolescents were recruited for a longitudinal randomized controlled trial (NCT01421485) that compared a novel intervention, termed integrated treatment program (ITP) (n = 64), with standard mental health counseling in the community (n = 53). ITP incorporated mental health and substance abuse treatment based in cognitive-behavioral therapy and motivational interviewing with an HIV prevention program. It also included parent and family sessions that centered on family communication regarding sexual and HIV risk reduction. Youth in the standard treatment arm received counseling services at the discretion of their provider, which generally consisted of a mental health evaluation and consideration for psychotherapy. Both treatment arms also received psychiatric medication management when indicated and case management services. Participants being diverted to the community by the courts in place of detention were referred by court officials when mental health services and evaluation were warranted. Consent was obtained before participation. The institutional review boards of the two participating sites approved all study protocols. Adolescents completed measures on laptop computers using an audio computer-assisted self-interview program. For the purposes of this study, the survey responses of youth at study intake (baseline) and at 6-month follow-up were analyzed. Study methods have also been previously reported.^24,25

Measures

Independent Variables

Self-reported demographic data were obtained from all participants at intake, including age, gender identity, race, and ethnicity. The demographic data are reported in Table 1.

Table 1.

Baseline Demographic Data for Adolescent Participants

	N	Mean (SD) / %
Age	117	15.16 (1.37)
Gender
Female	45	38.5
Male	72	61.5
Race
African American/Black/Haitian	8	6.8
American Indian/Alaska Native	2	1.7
Asian	3	2.6
Multiple/other	25	21.4
White	75	64.1
Ethnicity
Hispanic or Latinx
Yes	26	22.4
No	90	77.6
Substance use
Cannabis
Yes	44	37.6
No	73	62.4
Alcohol
Yes	64	54.7
No	53	45.3
Other drugs
Yes	11	9.3
No	106	90.6
Delinquency engagement
Yes	71	60.7
No	46	39.3
Delinquency severity
Serious	40	34.2
Moderate	29	24.8
Minor	23	19.7
None	25	21.4

The Customary Drinking and Drug Use Record (CDDR) is an interviewer-administered questionnaire that assesses drug and alcohol use over the past 3 months as well as lifetime use.²⁶ It examines the quantity and frequency of substance use, age of initiation, use progression, consequences of use, withdrawal symptoms, and psychological dependence. The CDDR has been shown to be both valid and reliable (r = 0.83 for alcohol and 0.92 for drugs) for use with youth from substance-using (α = .89 for alcohol and .72 for drugs) and community (α = .78 for alcohol and .85 for drugs) samples.²⁷ For the purposes of this study, the CDDR was administered at baseline and assessed for alcohol, cannabis, and other drug use in the past 3 months.

The Family Assessment Device (FAD) is a 60-item self-report measure that assesses adolescents’ current perceptions of family functioning.²⁸ Participants rated their families in areas such as, “We resolve most everyday problems around the house” and “There are rules about dangerous situations.” Higher scores on a 4-point Likert scale indicate poorer family functioning. The FAD was developed and normed on clinical and nonclinical samples, and adequate reliability (α = .72-.92, r = 0.66-0.76) and validity have been established.²⁸ In this study, the subscale of general family functioning was used to assess the overall climate and behavior of the families at baseline. It has been shown to highly correlate with the other subscales (r = 0.85-0.88) and can be used as a single measure to represent overall family functioning.²⁹ This subscale has been shown to be valid and reliable in clinical and nonclinical samples (α = .83-.86).²⁹

The Parental Monitoring Questionnaire (PMQ) is a 24-item self-report measure that assesses adolescents’ current perceptions of parental monitoring and sources of parental knowledge.³⁰ Participants were asked to use a 5-point Likert scale, ranging from no/never (1) to yes/always (5), to rate how often certain parenting practices take place. Examples include, “How often do you need to have your parent’s permission to stay out late on a weekday evening” and “In the last month, how often have your parents talked with the parents of your friends?” Higher scores indicate elevated levels of parental monitoring. For the purposes of this study, the parental monitoring subscale completed at baseline was used in the analyses. This subscale has demonstrated good reliability (α = .85, r = 0.83) and correlates with adolescent internalizing and externalizing maladjustment, deviant peer relationships, and family discord.³⁰

The Symptom Checklist-90-Revised (SCL-90-R) is a 90-item self-report measure that assesses the presence and severity of mental health symptoms over the past 7 days.³¹ It uses 9 primary symptom dimensions, namely, interpersonal sensitivity, depression, anxiety, phobic anxiety, obsessive-compulsive, somatization, hostility, paranoid ideation, and psychostimulation. Participants are asked to use a 5-point Likert scale, ranging from not at all (1) to extremely (5), to rate their level of distress related to specific symptoms over the past week. Each symptom examined is prefaced by the phrase, “How distressed were you by ….” Examples of these symptoms include, “Feeling hopeless about the future” and “Feelings of panic or anxiety.” The sum of these 9 subscales and any additional items included are then divided by the total number of responses to generate the Global Severity Index (GSI). The GSI has been demonstrated to be reliable and valid (α = .95).³² For the purposes of this study, baseline GSI from the SCL-90-R was used when analyzing adolescent self-reported psychiatric symptoms in relation to delinquency outcomes. The 6-month GSI responses were used when comparing treatment groups.

Dependent Variables

The National Youth Survey (NYS) of Self-Reported Delinquency is a 40-item self-report measure designed to assess the frequency at which adolescents commit delinquent acts, including stealing, carrying weapons, engaging in violence, using and selling drugs, and public misconduct.³³ It also assesses how often drugs and alcohol were involved in these delinquent acts. Youth were asked if and which delinquent acts they had engaged in over the past 3 months. The NYS has been shown to have acceptable reliability and validity with correlations equaling r = 0.75 for frequency and r = 0.84 for variety of acts reported with a mean coefficient of r = 0.74.³⁴ NYS responses of CIY obtained at the 6-month-follow-up visit were used to generate the primary outcome variables: engagement in delinquency and severity.

The engagement in delinquent acts variable was generated from responses at the 6-month follow-up visit to the well-validated (α = .93) and widely used general delinquency subscale of the NYS, which is a summary measure that examines a full range of delinquent acts.^33,35, 36, 37 For the purposes of this study, the 23-item version of the general delinquency subscale was used and converted into a binary measure (“yes/no”) to indicate if youth had engaged in any delinquent acts. This modification was made to better assess the clinically meaningful difference between delinquent and nondelinquent youth. It was anticipated that many youth would not engage in delinquent acts after 6 months of mental health treatment.

Youth were divided into 4 delinquency severity categories based on their responses to the NYS at the 6-month follow-up visit: serious, moderate, minor, and none. These categories have been previously used and validated in the literature on the CIY population^6,24 and are fully outlined in Table 2.

Table 2.

Delinquency Severity Categories Using Items From the National Youth Survey of Self-Reported Delinquency^6,24,33

Serious o Have you stolen (or tried to steal) a motor vehicle, such as a car or motorcycle? o Have you stolen (or tried to steal) something worth more than $50? o Have you attacked someone with the idea of seriously hurting or killing him/her? o Have you been involved in gang fights? o Have you sold hard drugs such as heroin, cocaine, or LSD? o Have you had (or tried to have) sexual relations with someone against their will? o Have you used force (strong-arm methods) to get money or things from other students? o Have you used force (strong-arm methods) to get money or things from a teacher or other adults at school? o Have you used force (strong-arm methods) to get money or things from people (not students or teachers)? o Have you broken into a building or vehicle (or tried to break in) to steal something or just look around? o Have you stolen money or things from your parents or other members of your family?

Moderate o No endorsement of serious category items o Have you knowingly bought, stole, or held stolen goods (or tried to do any of these things)? o Have you carried a hidden weapon other than a plain pocketknife? o Have you stolen (or tried to steal) something worth less than $5? o Have you sold marijuana or hashish (pot, grass, hash)? o Have you hit (or threatened to hit) a teacher or other adult at school? o Have you hit (or threatened to hit) one of your parents? o Have you hit (or threatened to hit) other students? o Have you been loud, rowdy, or unruly in a public place (disorderly conduct)? o Have you taken a vehicle for a ride (drive) without the owner’s permission? o Have you stolen (or tried to steal) things worth between $5 and $50?

Minor o No endorsement of serious or moderate category items o Have you been paid for having sexual relations with someone? o Have you begged for money or things from strangers? o Have you run away from home? o Have you purposely damaged or destroyed property belonging to your parents or other family members? o Have you purposely damaged or destroyed property belonging to a school? o Have you purposely damaged or destroyed other property that did not belong to you (not counting family or school property)? o Have you thrown objects (such as rocks, snowballs, or bottles) at cars or people? o Have you lied about your age to gain entrance or purchase something; for example, lying about your age to buy liquor or get a movie? o Have you cheated on school tests? o Have you hitchhiked where it was illegal to do so? o Have you bought or provided liquor for a minor? o Have you avoided paying for things such as movies, bus or subway rides, and food? o Have you been drunk in a public place? o Have you stolen or tried to steal something at school such as someone’s coat from a classroom, locker, cafeteria, or a book from the library? o Have you skipped classes without an excuse? o Have you failed to return extra change that a cashier gave you by mistake? o Have you been suspended from school? o Have you made obscene phone calls, such as calling someone and saying dirty things?

No endorsed delinquency (none) o No endorsement of serious, moderate, or minor category items

Note: LSD = lysergic acid diethylamide.

Statistical Analysis

Means and standard deviations were calculated for scale scores and categorical variables. The t test and χ² test of independence were used to compare participants who completed baseline and 6-month assessments with youth who provided only baseline data. We also conducted multivariate tests to assess differences at baseline and 6 months in psychiatric symptoms and delinquency outcomes between the 2 treatment arms. To examine the impact of self-reported baseline psychosocial factors on delinquency engagement and severity at 6-month follow-up, initial analyses, including t tests, χ² tests of independence, correlations, and analyses of variance and covariance were conducted to compare baseline demographic and psychosocial variables with 6-month delinquency outcomes.

After examining these descriptive analyses, the race category was recoded into 2 groups with White being compared with all other groups due to the small proportions in all other categories. Also, a psychosocial factor, the FAD score, was excluded from the regression analyses because it was significantly correlated with GSI scores (r² = 0.36, p < .001) and PMQ scores (r² = −0.43, p < .001), whereas PMQ and GSI scores were not (r² = −0.16, p = .095), which was concerning in terms of model stability in the setting of a small sample size with limited power to determine clinically meaningful differences between 2 measures that are both centered on a family construct (eg, FAD and PMQ). Furthermore, as parental monitoring (PMQ) is a modifiable skill that can be taught and practiced, while family functioning (FAD) is an indicator of overall family climate, it was thought that monitoring would be more clinically useful to investigate. Two scales were transformed so that the results from the regressions would be more understandable and relevant. First, the GSI scores were divided at t scores greater than or equal to 63 because that is the clinical cutoff suggesting clinical psychiatric symptoms.³¹ Second, because there are no criteria for adequate parental monitoring, PMQ scores were standardized. The PMQ was normally distributed as evidenced by a Shapiro-Wilk test (W = 0.97, p = .020) with a skew of −0.15 and kurtosis of −0.95.

Variables were next entered into regressions to examine the extent to which they predicted delinquency engagement (logistic regression) severity (multinomial logistic regression) 6 months after baseline. In addition to demographic variables (eg, gender, race, and age), we included psychiatric symptoms and substance use in our analyses because of the high prevalence of psychiatric symptoms and substance use in CIY populations, the potential impacts of these factors on youth behavior, and the fact that participants received mental health treatment. As this was an exploratory investigation, adjusted odds ratio (AOR) and 95% CI were used to assess effect sizes. All analyses were done with IBM SPSS Statistics for Windows, version 26 (IBM Corp, Armonk, New York).

Results

Demographics

Adolescent participants ranged in age from 12 to 17 years old with an mean (SD) age of 15.16 (1.37) years. The majority were White (64.1%) and male (61.5%). Full demographic data are presented in Table 1. Regarding delinquency, most participants endorsed engagement in delinquency (60.7%) as well as serious (34.2%) and moderate (24.8%) levels of severity. Cannabis (37.6%) and alcohol (54.7%) were the primary substances used with only 9.3% reporting use of other drugs. Participants were most likely to endorse use of both substances (31.6%) or neither of them (39.3%), rather than using only one (χ²₁ = 24.58, p < .001).

Analyses comparing the participants who completed assessments at baseline and 6 months with participants who completed only baseline measures found that there were no significant differences in age, gender, race, alcohol use, cannabis use, psychiatric symptoms, family functioning, parental monitoring, delinquency engagement, and delinquency severity. Psychiatric symptoms improved over 6 months for the entire sample (F_1,113 = 11.38, p = .001), and there were no significant differences in symptom improvement between the 2 groups (F_1,113 = 2.87, p = .093). At 6-month follow-up, there were no significant differences between the ITP and standard treatment arms in delinquency engagement (χ² = 1.45, p = .229) and severity (χ² = 3.64, p = .303). Therefore, as delinquency engagement and severity were the primary outcomes of the current study, both treatment groups were combined for the remaining analyses.

Primary Analyses

The results of the initial analyses are shown in Tables 3 and 4. There were no significant age or racial differences in delinquency engagement and severity. Gender was associated with delinquency severity at 6 months, with male participants being overrepresented in the serious category (χ²₃ = 10.09, p = .018).

Table 3.

Initial Analyses Between Baseline Predictor Variables and Delinquency Engagement at 6 Months

	Delinquency engagement				Test statistic	p
	Yes (n = 71)		No (n = 46)
	Mean	(SD)	Mean	(SD)	t
Age	15.21	(1.33)	15.09	(1.44)	-0.48	.634
GSI	1.69	(0.71)	1.45	(0.56)	-1.90	.060
FAD	29.42	(6.52)	26.17	(6.09)	-2.70	.008∗
PMQ	73.63	(19.22)	82.30	(19.07)	2.39	.018∗
	n	(%)	n	(%)	χ2
Gender
Female	28	(39.4)	17	(37.0)	0.07	.788
Male	43	(60.6)	29	(63.0)
Race
BIPOC	25	(35.2)	20	(43.5)	0.81	.369
White	46	(64.8)	26	(56.5)
Alcohol use
Yes	47	(66.2)	17	(37.0)	9.63	.002∗
No	24	(33.8)	29	(63.0)
Cannabis use
Yes	34	(47.9)	10	(21.7)	8.13	.004∗
No	37	(52.1)	36	(78.3)

Note: BIPOC = Black, Indigenous, and People of Color; FAD = Family Assessment Device; GSI = Global Severity Index; PMQ = Parental Monitoring Questionnaire.

^∗p < .05.

Table 4.

Initial Associations Between Baseline Variables and Delinquency Severity at 6 Months

	Delinquency severity										Test statistic	p
	Total (N = 117)		Serious (n = 40)		Moderate (n = 29)		Minor (n = 23)		None (n = 25)
	n	(%)	n	(%)	n	(%)	n	(%)	n	(%)	χ2
Gender, male	72	(61.5)	30	(75.0)	11	(37.9)	15	(65.2)	16	(64.0)	10.09	.018∗
Race, BIPOC	45	(38.5)	14	(35.0)	10	(34.5)	9	(39.1)	12	(48.0)	1.36	.715
Alcohol use	64	(54.7)	30	(75.0)	17	(58.6)	7	(30.4)	10	(40.0)	14.48	.002∗
Cannabis use	44	(37.6)	23	(57.5)	11	(37.9)	6	(26.1)	4	(16.0)	13.02	.005∗
	Mean	(SD)	Mean	(SD)	Mean	(SD)	Mean	(SD)	Mean	(SD)	F
Age	15.16	(1.37)	15.33	(1.33)	15.07	(1.32)	15.57	(1.04)	14.64	(1.60)	2.17	.096
GSI	1.59	(0.66)	1.58	(0.59)	1.82	(0.85)	1.50	(0.54)	1.43	(0.59)	1.78	.155
FAD	28.15	(6.53)	30.35	(7.27)	28.21	(5.47)	26.57	(5.21)	26.0	(6.72)	3.02	.033∗
PMQ	30.09	(9.18)	25.38	(8.67)	32.56	(7.50)	32.0	(9.06)	33.04	(9.36)	6.12	<.001∗

Note: BIPOC = Black, Indigenous, and People of Color; FAD = Family Assessment Device; GSI = Global Severity Index; PMQ = Parental Monitoring Questionnaire.

^∗p < .05.

Delinquency Engagement

In initial comparisons, engagement in delinquent acts at 6 months was significantly associated with baseline family functioning (t₁₁₅ = −2.70, p = .008), parental monitoring (t₁₁₅ = 2.39, p = .018), alcohol use (χ²₁ = 9.63, p = .002), and cannabis use (χ²₁ = 8.13, p = .004) (Table 3). Logistic regression demonstrated that after controlling for baseline psychiatric symptoms and demographic factors, baseline cannabis use (AOR 2.87, 95% CI 1.05-7.84), alcohol use (AOR 2.28, 95% CI 0.86-6.03), and parental monitoring (AOR 1.86, 95% CI 0.75-4.66) had small to medium effects on delinquency engagement at 6 months (χ²₇ = 19.15, p = .008). Table 5 outlines the full model.

Table5.

Logistic Regressions of Baseline Predictor Variables for Delinquency Engagement and Severity at 6 Months

Variable	B	SE	AOR	95% CI	p
Delinquency engagement
Age	-0.15	0.17	0.86	0.62-1.20	.378
Gender	0.15	0.45	1.61	0.48-2.82	.741
Race	0.30	0.43	1.35	0.58-3.13	.484
Alcohol use	0.82	0.50	2.28	0.86-6.03	.098
Cannabis use	1.06	0.51	2.87	1.05-7.84	.040∗
GSI	0.83	0.49	2.29	0.88-5.99	.090
PMQ	0.62	0.47	1.86	0.75-4.66	.183
Serious delinquency
Age	0.15	0.23	1.61	0.75-1.81	.507
Gender	-0.87	0.67	0.42	0.11-1.56	.195
Race	-0.42	0.59	0.66	0.21-2.10	.477
Alcohol use	-0.24	0.71	0.79	0.20-3.15	.736
Cannabis use	1.83	0.73	6.24	1.49-26.14	.012∗
GSI	0.55	0.56	1.73	0.58-5.14	.325
PMQ	0.75	0.32	2.11	1.13-3.96	.020∗
Moderate delinquency
Age	0.12	0.24	1.12	0.71-1.78	.620
Gender	0.98	0.63	2.65	0.78-9.05	.119
Race	-0.45	0.61	0.64	0.19-2.08	.453
Alcohol use	-0.06	0.75	0.95	0.22-4.13	.941
Cannabis use	1.37	0.79	3.94	0.84-18.41	.082
GSI	0.82	0.55	2.28	0.78-6.64	.132
PMQ	-0.13	0.33	0.88	0.46-1.69	.692
Minor delinquency
Age	0.72	0.28	2.06	1.20-3.54	.009∗
Gender	0.18	0.66	1.20	0.33-4.41	.782
Race	-0.26	0.63	0.77	0.23-2.65	.684
Alcohol use	1.66	0.80	5.28	1.10-25.28	.038∗
Cannabis use	0.98	0.84	2.66	0.51-13.87	.244
GSI	0.27	0.60	1.31	0.40-4.25	.652
PMQ	0.13	0.33	1.14	0.59-2.18	.698

Note: AOR = adjusted odds ratio; GSI = Global Severity Index; PMQ = Parental Monitoring Questionnaire.

^∗p < .05.

Delinquency Severity

Youth with higher delinquency severity at 6 months had greater baseline alcohol use (χ²₃ = 14.48, p = .002), greater baseline cannabis use (χ²₃ = 13.02, p = .005), less parental monitoring (F_3,117 = 6.12, p < .001), and poorer family functioning (F_3,117 = 3.02, p = .033) (Table 4). Multinomial logistic regression, with no delinquency at 6 months as the reference group (Table 5), demonstrated that after controlling for psychiatric symptoms and demographic factors, baseline cannabis use (AOR 6.24, 95% CI 1.49-26.17) and parental monitoring (AOR 2.11, 95% CI 1.13-3.96) continued to predict serious delinquency, while alcohol use (AOR = 5.28, 95% CI 1.10-25.28) predicted minor delinquency (χ²₂₁ = 54.29, p < .001).

Discussion

Adolescence is an important developmental period in which risky behaviors can have serious long-term consequences, and families can intervene as a significant protective factor. CIY diverted to mental health treatment are a unique population of adolescents because they have a significant psychiatric symptom burden3, 4, 5 and are completing their court-mandated sentences in the community where they may have exposure to both positive (eg, family support, outpatient services) and negative (eg, substance use, neighborhood violence) psychosocial influences. In this investigation involving CIY who received mental health treatment, youth with a history of substance use and less parental monitoring at baseline were more likely to have engagement in and higher severity of delinquency at 6 months. To our knowledge, this is one of the first investigations to examine the predictors of delinquency as youth progress through mental health treatment. These associations between parental monitoring, substance use, and recurrent delinquency are particularly notable as their influence remained even though the sample had overall improvement in psychiatric symptoms during the course of treatment. While substance use and parenting practices are directly assessable treatment targets that have been successfully addressed in previous interventions for CIY,^13,20,38,39 these predictors of delinquent behaviors may not always be directly and effectively addressed in mandated mental health treatment. Our results support the importance of targeting parenting practices and substance use as CIY engage in mental health treatment, as these factors continued to predict delinquency outcomes even after controlling for psychiatric symptoms. Therefore, while the literature supports the utility of mental health treatment for CIY,^11,12,21 in keeping with our results and the SPF,¹⁵ it cannot be provided in isolation without consideration for substance use and family factors, as CIY with difficulties in these areas may be particularly vulnerable and could stand to greatly benefit from more effective tailored and targeted interventions.

While effective evidence-based treatments that involve the family system do exist, including multisystemic therapy and functional family therapy, they may not be feasible for many families and communities due to issues with extensive time commitment, staff support, insurance coverage, and other resources required. Therefore, a targeted focus on improving parenting skills, such as parental monitoring related to delinquent behaviors, may allow for more feasible and effective programs to be incorporated into existing mental health treatment. Our investigation supports this assertion and makes a unique contribution to the literature by examining CIY who have engaged in mental health treatment, identifying notable contributory factors for further delinquency. Furthermore, our finding that baseline parental monitoring had a significant impact on delinquency outcomes even after accounting for other factors is distinctive. Previous studies among community youth and CIY not receiving mental health treatment have also found associations between a variety of family measures (eg, family functioning and parenting skills) and youth behavior (eg, substance use and delinquency), which places them at increased risk for legal contact and poor health outcomes.^16,17,19,38, 39, 40, 41, 42

The concepts of family functioning and parenting skills, such as parental monitoring, are likely closely related. In fact, our study found that the scales measuring family functioning and parental monitoring were significantly correlated. Furthermore, the relation between youth behavior and these family factors is intuitive because, in general terms, family functioning reflects overall family climate, and parental monitoring describes specific actions to know the adolescent’s activities. Ultimately, it may be easier to teach the skills and behaviors of effective parental monitoring than it is to change the overall family climate. Therefore, our findings along with previous literature on family factors in CIY indicate that parenting practices are a critical buffer against engagement in risky youth behaviors and thus warrant incorporation into prevention and diversion efforts in a manner that is directly targeted, measurable, and feasible for families.

Similar to the need for family-based care, CIY also require substance use treatment as a part of their diversion plans. Our results indicate that baseline substance use at the start of mental health treatment is predictive of youth delinquency 6 months later, even as psychiatric symptoms improve. This aligns with previous literature involving CIY that has demonstrated similar associations between substance use and delinquency.^7,10,43,44 Other investigations in samples not referred for mental health treatment have found a stronger association between delinquency and mental health symptoms rather than between delinquency and substance use.⁴⁵ Overall, there is significant evidence that a connection exists between substance use and mental health symptoms^46,47 and that high rates of co-occurring diagnoses are present in CIY and contribute to recidivism rates.^8,9 Therefore, dual diagnosis care is warranted as a part of court-mandated mental health diversion programs.

While the results of this study are consistent with previous literature as indicated above, there are some notable differences. First, among this group of CIY, baseline demographic variables, including age and race, were not predictive of delinquency outcomes at 6 months. Furthermore, gender, while significant in initial analyses examining delinquency severity, lost significance at the multivariate level. This in not in keeping with previous investigations, such as the study by Tolou-Shams et al.,⁷ in which male participants, older participants, and participants who identified as non-Latinx Black, Latinx, or non-Latinx multiracial were more likely to recidivate over a 24-month follow-up period. Furthermore, Dawkins and Dawkins⁴³ found that some significant associations between alcohol use and delinquent behaviors of various severity emerged only after analyzing the participants by racial group. Similarly, McReynolds et al.⁸ demonstrated that gender influences the connection between psychiatric diagnoses and recidivism. The discrepancies between these results and ours may be reflective of potential underlying differences between youth referred to mental health treatment as in this study, youth in first-time legal system contact as in the study by Tolou-Shams et al.,⁷ incarcerated youth as in the study by Dawkins and Dawkins,⁴³ and youth referred to probation as in the study by McReynolds et al.⁸ Furthermore, as our sample was predominantly White and male, there might not have been enough power to detect differences in delinquency outcomes across gender and racial identities. Also, we might not have detected any racial or ethnic differences because we measured self-reported delinquency rather than recidivism or contact with the legal system. Therefore, it is crucial to thoughtfully consider these factors in CIY populations due to major inequities seen in involvement in the court system, particularly in terms of discrepant rates of detention vs diversion, for racial⁴⁸ as well as gender and sexual minoritized youth.^5,49

This study has several limitations, including limited generalizability, as only CIY diverted to mental health treatment in 2 cities in the Eastern United States were recruited. As these participants were enrolled in mental health treatment, psychiatric symptoms were elevated, and the range was restricted compared with projects with a broader sample. Therefore, our power to detect associations with symptoms and generalizability to samples not in care were limited. Further limiting generalizability is that these findings were generated from participants who enrolled in an intervention study and not a broader epidemiological sample. This sample also predominantly identified as White and male, and thus the results do not capture the nuances of racial and gender inequities among individuals involved in the court system, especially as there are racial inequities between youth who are diverted vs youth are detained for similar behaviors.⁴⁸ Another limitation is that all the study measures used self-report, including measures related to our independent and dependent variables. This characteristic may have introduced common rater effects, such as recall and social desirability bias across these variables, which could have contributed to underreporting, such as with substance use (an independent variable) and/or engagement in delinquent acts (a dependent variable).

This investigation had a number of strengths, including having no significant differences between youth who completed the study and youth who did not complete it. This study also used scales that have been validated in psychiatric populations and/or justice-involved youth^35,36 and employed categorizations of delinquency severity that have previously been used in the literature.^6,24 Notably, this study assessed adolescents’ perspectives via self-report, which are important to consider when examining the predictors of youth delinquent behaviors. Previous research has demonstrated that youth reports of parenting behaviors alone are highly predictive of delinquent behaviors¹⁸; thus using adolescent perspectives is valuable. Critically, this study addresses a gap in the literature regarding which psychosocial factors predict delinquency outcomes in CIY who meet criteria for and are diverted to mental health treatment. Examination of these factors and their associations with delinquent behaviors in this group of CIY is valuable because it can contribute to the creation of more tailored, transdiagnostic family-based interventions.

Future directions for this research include continuing to use longitudinal data in this sample to assess how predictors of delinquency may shift with continued court-mandated mental health treatment beyond 6 months. In addition, further investigation is warranted to assess the mechanisms that may underlie changes in parental behavior, such as emotional management and parental stress, as it could better inform clinical interventions that address parenting practices and allow for feasible incorporation into mental health treatment. Similarly, in keeping with the SPF, examination of parental and peer views on substance use may also help to inform future interventions, as previous studies have shown links between these opinions and youth substance use.^17,38,46 Future investigations should seek to recruit diverse populations from a variety of geographical settings to better capture and promote improved understanding of the biopsychosocial factors and disparities faced by these youth and their families. In total, these efforts may allow for better tailoring of existing interventions to ensure that they are relevant to the needs of this population, especially youth referred to mental health treatment.

In conclusion, mental health treatment alone is not enough to promote optimal behavioral outcomes for CIY mandated to it, especially for CIY with substance use and less parental monitoring at the start of treatment. Therefore, clinical interventions should consider specifically and effectively targeting parental monitoring and youth substance use to reduce delinquency. Future investigations should seek to recruit diverse populations from a variety of geographical settings to improve our understanding of the biopsychosocial factors in this population and better tailor existing family-based interventions for CIY in mental health treatment.

CRediT authorship contribution statement

Elizabeth M. Olsen: Writing – review & editing, Writing – original draft, Investigation, Formal analysis, Conceptualization. Laura B. Whiteley: Writing – original draft, Supervision, Investigation, Formal analysis, Conceptualization. Marina Tolou-Shams: Writing – review & editing, Supervision, Investigation, Funding acquisition, Conceptualization. Christianne Esposito-Smythers: Writing – review & editing, Supervision, Investigation, Funding acquisition, Conceptualization. Larry K. Brown: Writing – review & editing, Supervision, Investigation, Funding acquisition, Conceptualization.