Prevalence, Comorbidity, and Continuity of Psychiatric Disorders in a 15-Year Longitudinal Study of Youths Involved in the Juvenile Justice System

Key Points

Question

What is the prevalence, continuity, and comorbidity of psychiatric disorders among youths during the 15 years after detention in a juvenile justice facility, and do outcomes vary by sex and race/ethnicity?

Findings

In this cohort study of 1829 youths who were detained in a juvenile justice facility, 64% of males and 35% of females with a psychiatric disorder during detention had a disorder 15 years later. Substance use and behavioral disorders were more common among non-Hispanic White youths than Hispanic and Black youths.

Meaning

This study’s findings suggest that persistent psychiatric disorders may complicate the transition from adolescence to adulthood among youths who have been detained in a juvenile justice facility; the pediatric health community should advocate for early identification and treatment of disorders among this population.

Abstract

Importance

Previous studies have found that one-half to three-quarters of youths detained in juvenile justice facilities have 1 or more psychiatric disorders. Little is known about the course of their disorders as they age.

Objective

To examine the prevalence, comorbidity, and continuity of 13 psychiatric disorders among youths detained in a juvenile justice facility during the 15 years after detention up to a median age of 31 years, with a focus on sex and racial/ethnic differences.

Design, Setting, and Participants

The Northwestern Juvenile Project is a longitudinal cohort study of health needs and outcomes of 1829 randomly selected youths in a temporary juvenile detention center in Cook County, Illinois. Youths aged 10 to 18 years were interviewed in detention from November 20, 1995, through June 14, 1998. Participants were reinterviewed up to 12 times during the 15-year study period through February 2015, for a total of 16 372 interviews. The sample was stratified by sex, race/ethnicity (Black, Hispanic, and non-Hispanic White), age (10-13 years or 14-18 years), and legal status (processed in juvenile or adult court). Data analysis was conducted from February 2014, when data preparation began, to March 2020.

Exposures

Detention in a juvenile justice facility.

Main Outcomes and Measures

Psychiatric disorders, assessed by the Diagnostic Interview Schedule for Children, version 2.3 at the baseline interviews. Follow-up interviews were conducted using the Diagnostic Interview Schedule for Children, version IV; the Diagnostic Interview Schedule, version IV; and the World Mental Health Composite International Diagnostic Interview (beginning at the 6-year follow-up interview).

Results

The study included 1829 youths sampled at baseline (1172 males and 657 females; mean [SD] age, 14.9 [1.4] years). Although prevalence and comorbidity of psychiatric disorders decreased as the 1829 participants aged, 52.3% of males and 30.9% of females had at least 1 or more psychiatric disorders 15 years postdetention. Among participants with a disorder at baseline, 64.3% of males and 34.8% of females had a disorder 15 years later. Compared with females, males had 3.37 times the odds of persisting with a psychiatric disorder 15 years after baseline (95% CI, 1.79-6.35). Compared with Black participants and Hispanic participants, non-Hispanic White participants had 1.6 times the odds of behavioral disorders (odds ratio, 1.56; 95% CI, 1.27-1.91 and odds ratio, 1.59; 95% CI, 1.23-2.05, respectively) and greater than 1.3 times the odds of substance use disorders (odds ratio, 1.90; 95% CI, 1.55-2.33 and odds ratio, 1.39; 95% CI, 1.11-1.73, respectively) throughout the follow-up period. Behavioral disorders and substance use disorders were the most prevalent 15 years after detention.

Conclusions and Relevance

This study’s findings suggest that persistent psychiatric disorders may complicate the transition from adolescence to adulthood, which is already challenging for youths involved in the juvenile justice system, many of whom are from racial/ethnic minority groups and low-income backgrounds. The pediatric health community should advocate for early identification and treatment of disorders among youths in the justice system.

This cohort study examines the prevalence, comorbidity, and continuity of psychiatric disorders among youths detained in a juvenile justice facility during the 15 years after detention, with a focus on sex and racial/ethnic differences.

Introduction

Youths involved in the juvenile justice system have a substantially higher prevalence of psychiatric disorders compared with those in the general population,^{1,2,3,4,5,6,7,8,9} with 45% to 66% of males and 45% to 73% of females in the system meeting the criteria for 1 or more psychiatric disorders.^2,10,11,12 Anxiety, mood, and behavioral disorders are prevalent.^{1,2,3,4,6,10,11,13,14} Substance use disorders (SUDs), which are the most common, affect up to 50% of males and 22% to 46% of females.^{1,2,6,10,11,12,13} Comorbid psychiatric disorders are also common^5,8,13,15; 34% of males and 60% of females in juvenile detention facilities have 3 or more disorders.¹⁶

Less is known about the course of psychiatric disorders after youths leave detention facilities. Findings have been inconsistent. A 2015 study of 75 first-time juvenile offenders found that the prevalence of disorders had decreased 2 years later from more than 70% to approximately 46%.⁹ A longitudinal study of 97 males found that 2 years after juvenile detention, mental health problems had persisted; 88%, 73%, and 51% of the sample met criteria for conduct disorder, drug use disorder, and alcohol use disorder, respectively.¹⁷ These studies have limitations, including small sample sizes, brief follow-up periods, the inclusion of few females, and the inability to study the Hispanic population, which is now the largest minority group in the US.¹⁸ Moreover, the study by Harrington et al¹⁷ was conducted in the UK, limiting the generalizability of its findings to populations in the US, which has greater income inequality¹⁹ and racial/ethnic diversity.^20,21

To date, the largest investigation of youths detained in juvenile justice facilities is the Northwestern Juvenile Project, a study of 1829 youths aged 10 to 18 years.¹⁰ This study found that at the time of detention, two-thirds of males and almost three-quarters of females had 1 or more psychiatric disorders.¹⁰ Approximately one-half of the youths had comorbid psychiatric disorders.¹⁵ Five years later, at a median age of 20 years, more than 50% of males and 40% of females had 1 or more psychiatric disorders.²² Comorbidity remained common, occurring in 27% of males and 14% of females.²³

Many psychiatric disorders, such as substance use^24,25,26,27 and mood disorders,^24,25,28 may begin during young adulthood. However, to our knowledge, no large longitudinal study has explored the prevalence, comorbidity, and continuity of psychiatric disorders among youths detained in a juvenile justice facility as they age into adulthood. This omission is critical. Involvement with the juvenile justice system may impede developmental trajectories by limiting youths’ opportunities to attain normative adult milestones,^29,30,31 such as establishing careers, gaining economic and residential independence, and starting families. For youths involved in the juvenile justice system, many of whom are from racial/ethnic minority groups and low-income backgrounds,^32,33,34 transitioning from adolescence to young adulthood may be further complicated by psychiatric disorders, which are associated with negative psychosocial outcomes, especially if untreated.^35,36,37

The current investigation presents new epidemiologic data from participants in the Northwestern Juvenile Project who were followed up to a median age of 31 years. The study examined 13 psychiatric disorders: major depression, mania, dysthymia, hypomania, generalized anxiety disorder (GAD), panic disorder, posttraumatic stress disorder (PTSD), attention-deficit/hyperactivity disorder (ADHD), conduct disorder (CD), oppositional defiant disorder (ODD), antisocial personality disorder (ASPD), alcohol use disorder, and drug use disorder. Focusing on sex and racial/ethnic differences, we examined the prevalence of psychiatric disorders, including patterns of comorbidity, and the continuity of psychiatric disorders over time.

Methods

For this cohort study, the institutional review boards of Northwestern University, the US Centers for Disease Control and Prevention, the Illinois Department of Child and Family Services, Cermak Health Services, the Illinois Department of Corrections, and the Federal Bureau of Prisons approved relevant study procedures. Participants signed either an assent form (if younger than 18 years) or a consent form (if 18 years or older). The Northwestern University institutional review board, Cermak Health Services, and the Illinois Department of Corrections also waived parental consent when participants were younger than 18 years, consistent with federal regulations regarding research with minimal risk.³⁸ This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies as applicable.

Sample, Procedures, and Measures

We recruited and interviewed a stratified random sample of 1829 youths at the time of intake to the Cook County Juvenile Temporary Detention Center in Chicago, Illinois, between November 20, 1995, and June 14, 1998. The Cook County Juvenile Temporary Detention Center is used for pretrial detention and detention of juvenile offenders sentenced to less than 30 days. To ensure adequate representation of important subgroups, we stratified the sample by sex, race/ethnicity (Black, Hispanic, non-Hispanic White, and other), age (10-13 years or 14-18 years), and legal status (processed in juvenile court or adult court). Face-to-face structured interviews were conducted at the detention center, with most interviews occurring within 2 days of intake. The stratified random sample included 1172 males and 657 females; 1005 were Black (55%), 524 were Hispanic (29%), 296 were non-Hispanic White (16%), and 4 identified as other race/ethnicity (0%). The median age was 15 years (mean [SD], 14.9 [1.4] years).

All participants were reinterviewed up to 12 times during the 15-year study period through February 2015 at approximately 3, 5, 6, 8, 12, 14, and 15 years after the baseline interviews. A random subsample of 997 participants received additional interviews at 3.5 years and 4.0 years after baseline. We received funding to interview the last 800 participants who were enrolled at baseline at 10, 11, and 13 years after baseline. Participants were interviewed regardless of whether they lived in the community or a correctional facility. We conducted 16 372 interviews overall. Additional details are available in eMethods 1 in the Supplement and are published elsewhere.^10,15,22,39 For each follow-up interview, we chose the most reliable and valid diagnostic measures that were available in English and Spanish at the time.

Baseline

At baseline, we administered the Diagnostic Interview Schedule for Children, version 2.3 (DISC-2.3),^40,41 which is based on the Diagnostic and Statistical Manual of Mental Disorders (Third Edition Revised) (DSM-III-R), to assess all disorders excluding PTSD. Thirteen months after the study began, we added the PTSD module from the DISC, version IV (DISC-IV), which is based on the Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition) (DSM-IV), when available.³⁹ Additional details on baseline diagnostic decisions can be found elsewhere.^10,15,39

Follow-up Interviews

At follow-up interviews, we administered the DISC-IV^42,43 (child and young adult versions) to assess past-year prevalence for all disorders excluding SUDs; the DISC-IV was used until the 6-year follow-up interview, when most participants had reached age 18 years. Beginning with the 6-year follow-up interview, we administered the World Mental Health Composite International Diagnostic Interview,^44,45 which is based on the DSM-IV, to assess mood and anxiety disorders during the previous year. For ASPD (age ≥18 years only) and SUDs, we administered the Diagnostic Interview Schedule, version IV (DIS-IV),^46,47 which is based on the DSM-IV. Sensitivity analyses indicated that changes in prevalence over time were not associated with changes in measurements (eMethods 1 and eMethods 2 in the Supplement).

Statistical Analysis

We used Stata software, version 12 (StataCorp LLC)⁴⁸ and StatTag, version 5⁴⁹ for reproducibility. Data analysis was conducted from February 2014, when data preparation began, to March 2020. To generate prevalence estimates and inferential statistics that reflect the Cook County Juvenile Temporary Detention Center population, participants were assigned sampling weights, augmented with nonresponse adjustments to account for missing data.⁵⁰ We used Taylor series linearization to estimate standard errors.^51,52 We used α = .05 to assess statistical significance with 2-sided tests as applicable.

Because some participants were, by design, interviewed more frequently, we summarize prevalence at 8 time points: baseline and follow-up interviews corresponding to approximately 3, 5, 6, 8, and 12, 14, and 15 years after baseline. eTable 1 in the Supplement summarizes demographic and retention data; 77% of participants had a 15-year interview. Race/ethnicity (Black, Hispanic, non-Hispanic White, and other) was self-identified; analyses exclude 4 participants who identified as other. eMethods 1 in the Supplement details management of missing data, including multiple imputations for item nonresponse and assessing effects of attrition.

Changes in Prevalence as Youths Age

As in previous analyses of the Northwestern Juvenile Project data,^22,53 we used generalized estimating equations⁵⁴ to fit marginal models examining differences in prevalence of psychiatric disorders by sex and race/ethnicity and changes in prevalence of disorders as youths aged. Models used all available interviews, an average of 9 interviews per person (range, 1-13 interviews per person). Unless otherwise noted, odds ratios (ORs) indicate sex and racial/ethnic differences over time.

Generalized estimating equations models included covariates for sex, race/ethnicity, aging (time since baseline), age at baseline (10–18 years), and legal status at detention. Details of model formulation are in eMethods 1 in the Supplement. For models with significant interactions between sex and aging (Wald P values), we report model-based ORs for sex differences at 3, 5, 6, 8, and 12, 14, and 15 years postdetention. There were no significant interactions between sex and race/ethnicity. We estimated generalized estimating equations models with sampling weights to account for study design.

Models for Continuity of Disorders Over Time

As in previous analyses of the Northwestern Juvenile Project data,²³ we examined continuity of disorders over time: did youths with a disorder at baseline have greater odds of having the same (or different) disorders 15 years later compared with youths without the disorder? We used logistic regression: disorder at baseline was the independent variable, and disorder at the 15-year interview was the dependent variable. We estimated separate models for males and females. All models assessing SUDs at follow-up were adjusted for time in corrections because access to substances is restricted in correctional settings.

Results

Prevalence of Disorder as Youths Age

Sex and racial/ethnic differences in prevalence over time are shown in Figure 1 (any disorder), eFigure 1 in the Supplement (comorbid disorder), Figure 2 (mood, anxiety, and behavioral disorders), and Figure 3 (any substance, alcohol, and drug disorders). eTables 2-4 in the Supplement provide specific prevalence estimates. eTables 5-9 in the Supplement show adjusted ORs examining changes in prevalence and sex and racial/ethnic differences as youths age.

Figure 1. Sex and Racial/Ethnic Differences in Prevalence of Any Disorder From Detention (Baseline) to 15 Years After Baseline.

A, A total of 1167 males and 655 females were included. B, A total of 207 non-Hispanic White males, 386 Hispanic males, and 574 Black males were included. C, A total of 89 non-Hispanic White females, 136 Hispanic females, and 430 Black females were included.

Figure 2. Sex and Racial/Ethnic Differences in Prevalence of Mood, Anxiety, and Behavioral Disorders From Detention (Baseline) to 15 Years After Baseline.

A, A total of 1167 males and 655 females were included. B, A total of 207 non-Hispanic White males, 386 Hispanic males, and 574 Black males were included. C, A total of 89 non-Hispanic White females, 136 Hispanic females, and 430 Black females were included.

Figure 3. Sex and Racial/Ethnic Differences in Prevalence of Any Substance, Alcohol, and Drug Use Disorders From Detention (Baseline) to 15 Years After Baseline.

A, A total of 1167 males and 655 females were included. B, A total of 207 non-Hispanic White males, 386 Hispanic males, and 574 Black males were included. C, A total of 89 non-Hispanic White females, 136 Hispanic females, and 430 Black females were included.

Any Disorder

Sex differences

Prevalence decreased as participants aged (Figure 1), although only modestly for males: nearly two-thirds of males (64.5%) had any disorder at baseline, and more than half of males (52.3%) had any disorder 15 years later. In contrast, prevalence among females decreased from 68.4% to 30.9%. At most time points, males had greater odds of disorder than females (eg, 3 years after baseline: adjusted OR, 1.41; 95% CI, 1.13-1.76; 12 years after baseline: adjusted OR, 1.56; 95% CI, 1.24-1.97; and 15 years after baseline: adjusted OR, 2.66; 95% CI, 2.08-3.40) (eTable 5 in the Supplement). However, excluding behavioral disorders, patterns over time for males and females were similar (eFigure 2 in the Supplement).

Racial/ethnic differences

Throughout the follow-up period, non-Hispanic White participants had nearly twice the odds of any disorder compared with Black participants (adjusted OR, 1.91; 95% CI, 1.56-2.34) and 1.69 times the odds of Hispanic participants (95% CI, 1.28-2.22). Differences remained even after excluding behavioral disorders (non-Hispanic White vs Black participants: adjusted OR, 1.96; 95% CI, 1.60-2.40; non-Hispanic White vs Hispanic participants: adjusted OR, 1.42; 95% CI, 1.11-1.81; and Hispanic vs Black participants: adjusted OR, 1.38; 95% CI, 1.09-1.75) (eTable 5 in the Supplement).

Comorbid Disorders

Sex differences

Although prevalence decreased as participants aged (eFigure 1 in the Supplement), 15 years after baseline, nearly 1 in 6 males (16.0%) and 1 in 9 females (11.6%) still had comorbid disorders. The rate of decrease was associated with sex (eTable 5 in the Supplement). At 3, 5, and 6 years postdetention, males had greater odds of comorbid disorders than females (respectively, adjusted OR, 1.77, 95% CI, 1.40-2.24; adjusted OR, 1.81; 95% CI, 1.47-2.22; and adjusted OR, 1.55; 95% CI, 1.24-1.94).

Racial/ethnic differences

Prevalence of comorbid disorders was highest among non-Hispanic White participants, followed by Hispanic participants and then Black participants (non-Hispanic White vs Black participants: adjusted OR, 1.77; 95% CI, 1.44-2.17; non-Hispanic White vs Hispanic participants: adjusted OR, 1.39; 95% CI, 1.12-1.74; and Hispanic vs Black participants: adjusted OR, 1.27; 95% CI, 1.02-1.58).

Mood Disorders

Sex differences

Mood disorders generally decreased from baseline to 15 years: among males, from 15.8% to 9.9%; among females, from 21.4% to 7.9% (Figure 2). Overall, females had 1.30 times the odds of any mood disorder (95% CI, 1.06-1.59) and 1.54 times the odds of major depression (95% CI, 1.24-1.92) compared with males (eTable 6 in the Supplement).

Anxiety Disorders

Sex differences

Compared with males, females had 1.66 times the odds of any anxiety disorder (95% CI, 1.33-2.06) and 1.75 times the odds of panic disorder (95% CI, 1.16-2.64) over time. Throughout most follow-up interviews, females also had greater odds of PTSD (the most prevalent anxiety disorder) compared with males (eg, 8 years after baseline, adjusted OR, 3.21; 95% CI, 1.87-5.53) (eTable 7 in the Supplement). PTSD decreased over time, but the rate of decrease was associated with sex.

Racial/ethnic differences

Non-Hispanic White participants were more likely than Black participants to have panic disorder (adjusted OR, 3.00; 95% CI, 1.71-5.27) but less likely to have GAD (adjusted OR, 0.54; 95% CI, 0.31-0.93). Hispanic participants were more likely than Black participants and non-Hispanic White participants to have PTSD (respectively, adjusted OR, 1.53; 95% CI, 1.07-2.19 and adjusted OR, 1.96; 95% CI, 1.32-2.90). Compared with Black participants, Hispanic participants were also more likely to have any anxiety disorder and panic disorder (respectively, adjusted OR, 1.44: 95% CI, 1.06-1.95 and adjusted OR, 2.05; 95% CI, 1.18-3.55).

Behavioral Disorders

Sex differences

Behavioral disorders (including ASPD after age 18 years) were among the most common. For example, 15 years after baseline, 38.6% of males and 14.8% of females had a behavioral disorder. Compared with females, males had more than twice the odds of ASPD (adjusted OR, 2.29; 95% CI, 1.89-2.78) (eTable 8 in the Supplement). Approximately 40% of males had ASPD both 5 and 15 years postdetention. For females, prevalence decreased from 26.8% to 14.8% during this period.

Racial/ethnic differences

Compared with Hispanic participants and Black participants, non-Hispanic White participants had 1.6 times the odds of behavioral disorders (respectively, adjusted OR, 1.59; 95% CI, 1.23-2.05 and adjusted OR, 1.56; 95% CI, 1.27-1.91).

Substance Use Disorders

Sex differences

Although prevalence decreased over time, 15 years after baseline, 16.1% of males and 9.9% of females had an SUD. Among males, prevalence of alcohol use and drug use disorders were each approximately 10%; among females, prevalence of each was approximately 5% (Figure 3). After baseline, males had higher prevalence of SUDs than females. For example, 15 years after baseline, males had 2.69 times the odds of alcohol use disorder compared with females (95% CI, 1.71-4.22).

Racial/ethnic differences

Compared with Black participants and Hispanic participants, non-Hispanic White participants had greater than 1.3 times the odds of any SUD (respectively, adjusted OR, 1.90; 95% CI, 1.55-2.33 and adjusted OR, 1.39; 95% CI, 1.11-1.73), alcohol use disorder (respectively, adjusted OR, 1.72; 95% CI, 1.35-2.18 and adjusted OR, 1.42; 95% CI, 1.12-1.79), and any drug use disorder (respectively, adjusted OR, 1.71; 95% CI, 1.39-2.10 and adjusted OR, 1.26; 95% CI, 1.01-1.57) (eTable 9 in the Supplement). Hispanic participants also had higher prevalence of any SUD and any drug use disorder compared with Black participants (respectively, adjusted OR, 1.37; 95% CI, 1.11-1.70 and adjusted OR, 1.36; 95% CI, 1.10-1.69).

Continuity of Disorders Over Time

Among participants who had a disorder at baseline, 64.3% of males and 34.8% of females had a disorder 15 years later. Compared with females, males had 3.37 times the odds of persisting with disorders 15 years after baseline (95% CI, 1.79-6.35). Even after excluding behavioral disorders, males had more than twice the odds of persisting with a disorder (OR, 2.02; 95% CI, 1.06-3.83).

For males and females respectively, Figure 4 and Figure 5 illustrate associations between having a disorder at detention and 15 years later (either the same or different disorders). eTables 10 and 11 in the Supplement provide specific prevalence estimates and ORs, with and without adjusting for comorbidity at baseline.

Figure 4. Odds Ratios for the Association Between Baseline and 15-Year Diagnoses Among Males.

A total of 837 males were assessed for a disorder at both baseline and at the 15-year follow-up. The disorder diagnosed at baseline could be the same as or different than the disorder diagnosed 15 years later. Only odds ratios with a significance of P < .05 are shown. Filled circles illustrate that participants with the disorder at baseline had higher odds of being diagnosed with the disorder 15 years later. Open circles illustrate that participants with the disorder at baseline had lower odds of being diagnosed with the disorder 15 years later. Shaded boxes represent disorders in the same diagnostic category. ADHD indicates attention-deficit/hyperactivity disorder; ASPD, antisocial personality disorder; GAD, generalized anxiety disorder; ODD, oppositional defiant disorder; PTSD, posttraumatic stress disorder; and SUD, substance use disorder.

Figure 5. Odds Ratios for the Association Between Baseline and 15-Year Diagnoses Among Females.

A total of 530 females were assessed for a disorder at both baseline and at the 15-year follow-up. The disorder diagnosed at baseline could be the same as or different than the disorder diagnosed 15 years later. Only odds ratios with a significance of P < .05 are shown. Filled circles illustrate that participants with the disorder at baseline had higher odds of being diagnosed with the disorder 15 years later. Open circles illustrate that participants with the disorder at baseline had lower odds of being diagnosed with the disorder 15 years later. Shaded boxes represent disorders in the same diagnostic category. ADHD indicates attention-deficit/hyperactivity disorder; ASPD, antisocial personality disorder; GAD, generalized anxiety disorder; ODD, oppositional defiant disorder; PTSD, posttraumatic stress disorder; and SUD, substance use disorder.

Males

Males with major depression at baseline were more likely to have any mood (OR, 2.88; 95% CI, 1.02-8.16), any anxiety (OR, 5.41; 95% CI, 1.32-22.21), PTSD (OR, 5.07; 95% CI, 1.01-25.46), and alcohol use disorder (OR, 3.49; 95% CI, 1.30-9.37) 15 years later. Males with dysthymia at baseline were more likely to have any anxiety (OR, 4.71; 95% CI, 1.01-21.97), any SUD (OR, 3.36; 95% CI, 1.29-8.71), and alcohol use disorder (OR, 3.98; 95% CI, 1.43-11.06). Baseline ADHD was associated with all disorders 15 years later except for major depression and ASPD. For example, males with ADHD at baseline had 2.78 times the odds of any SUD 15 years later (95% CI, 1.23-6.27). Baseline CD was associated with all externalizing disorders (ASPD and SUDs). For example, males with CD at baseline had 1.96 (95% CI, 1.12-3.42) times the odds of ASPD and 2.44 (95% CI, 1.09-5.45) times the odds of drug use disorder compared with males without CD at baseline. GAD at baseline was associated with lower prevalence of ASPD 15 years later (OR, 0.27 95% CI, 0.09-0.82).

Females

Major depression and dysthymia were associated with many disorders 15 years later (eg, major depression with any anxiety: OR, 2.39; 95% CI, 1.20-4.74; major depression with any SUD: OR, 2.23; 95% CI, 1.18-4.23). Females with GAD at baseline were more likely to have any anxiety and alcohol use disorders 15 years later (respectively, OR, 3.17; 95% CI, 1.22-8.25 and OR, 3.62; 95% CI, 1.12-11.75). Conduct disorder, ODD, and ADHD at baseline were associated with ASPD 15 years later (respectively, OR, 3.58; 95% CI, 2,14-6.00; OR, 2.03; 95% CI, 1.14-3.62; and OR, 2.11, 95% CI, 1.19-3.74) but were inconsistently associated with other disorders. Alcohol use disorder at baseline was associated with several disorders at follow-up (eg, with ASPD: OR, 1.99; 95% CI, 1.17-3.36); drug use disorder was not.

Discussion

Many psychiatric disorders extend well past adolescence. Among participants who had a disorder at baseline, nearly two-thirds of males and more than one-third of females had a disorder 15 years later. As was true for youths in detention¹⁰ and 5 years later,²² behavioral disorders and SUDs were the most prevalent, followed by mood and anxiety disorders. Prevalence of several disorders, such as PTSD,⁵⁵ drug use disorders,⁵⁵ and ASPD,^55,56,57 was notably higher than the general population. Comorbidity was less prevalent 15 years after detention than at baseline and 5 years²³ but nevertheless problematic.

There were significant sex and racial/ethnic differences. At detention, approximately two-thirds of males and females had psychiatric disorders. As they aged, males fared worse than females: 15 years after detention, more than half of males had a disorder compared with less than one-third of females. Males had more than 3 times the odds of persisting with disorders 15 years after detention. In addition, males had higher prevalence of ASPD and SUDs as they aged. Females may fare better than males because they are more likely to have adolescent-limited trajectories of offending.^58,59,60 Moreover, as they age, females are more likely to engage in prosocial activities and relationships⁶¹ and to seek⁶² and receive⁶³ mental health services.

Racial/ethnic differences persisted as youths aged. Significantly more non-Hispanic White participants had substance use and behavioral disorders than Hispanic participants and Black participants at baseline and in adulthood, similar to the general population.^64,65,66,67 Because individuals who are from racial/ethnic minority groups are disproportionately incarcerated,^68,69,70 we had anticipated finding differences in the opposite direction, that is, greater prevalence in racial/ethnic minority groups. Our findings may indicate the selective enforcement of laws in minority communities.^71,72,73

We found substantial continuity of disorders over time. We also found multiple pathways to addiction: several disorders at baseline (median age, 15 years), including depressive and behavioral disorders, were associated with SUD 15 years later. In addition, depression during childhood was a powerful predictor linked to many disorders in adulthood. Males with depressive disorder at baseline were more likely than those without to have mood, anxiety, and alcohol use disorders 15 years later. Females with depressive disorders at baseline were more likely than those without to have nearly all disorders.

Limitations

This study has limitations. We changed some diagnostic measures as participants aged and as the DSM was updated. Our data are subject to the limitations of self-reporting. Generalizability may be limited to detained youths in urban centers with similar demographic characteristics. Participants may have denied symptoms to hasten the interview; actual prevalence rates may be higher. Although retention rates were high, participants who missed interviews may have been more likely to have disorders. Findings do not take into account mental health services that have been provided.

Implications

First, educate pediatricians that disruptive behavior disorders (such as oppositional defiant disorder and conduct disorder) are serious but treatable. Up to 25% of children and adolescents seen in primary care have a significant behavioral or emotional health problem.^{8,74,75,76,77} By facilitating early detection and referral, pediatricians can play a critical role in reducing the likelihood of youths entering the juvenile justice system. Several promising evidence-based interventions are available. For young children, treatments include Parent Management Training and Parent–Child Interaction Therapy.⁷⁸ For older children, including those in the juvenile justice system, the Coping Power Program⁷⁹ and Multisystemic Therapy⁸⁰ can reduce aggressive behaviors. Treatments are most effective when they include youth, parents or caregivers, and other key stakeholders (eg, schools). The most effective pediatric integrated mental health care models include population-based care (identifying, treating, and tracking outcomes of all patients), measurement-based care (validated assessment tools to identify patients and evaluate outcomes), and evidence-based mental health services.⁸¹ Juvenile justice is the system of last resort.

Second, researchers should investigate variables that predict the desistance of psychiatric disorders, focusing on high-risk populations. Most of the participants in this study were at risk for continued psychiatric disorders. Yet, risk does not convey certainty. Why do some high-risk youths improve despite limited access to quality care,⁸² meager social networks,⁶¹ and few environmental resources?⁸³ Which characteristics of mental health services, family systems, schools, and communities increase resilience? Studies of the highest-risk populations (eg, youths in the juvenile justice and child welfare systems) will inform intervention for those most in need.

Third, the public health system must expand mental health services for youths while they are involved in the juvenile justice system and when they return to the community. Trauma, loss, and maltreatment, which are common precipitants of depression and other disorders, are prevalent in juvenile justice youth: more than 90% report experiencing at least 1 traumatic event,^39,84 including loss of intimates,⁸⁵ childhood maltreatment,^84,86 and witnessing violence.^39,84 Treating disorders during childhood may reduce psychiatric morbidity as youths age. Yet, despite decades of research documenting the prevalence of psychiatric disorders,^{2,10,22,87,88} only approximately one-third of youths are referred for or receive any mental health service during or after detention.^9,89,90,91 Logistic challenges abound: short stays,⁹² insufficient resources^93,94 and time to provide evidence-based assessments and treatments,⁹⁵ lack of coordination among service agencies,^96,97 and engagement of families.^98,99

Federal law prohibits Medicaid from funding services in public institutions, such as correctional facilities, except for off-site hospitalizations of more than 24 hours.^100,101 Quality of services thus vary by an institution’s budget. Moreover, the Children’s Health Insurance Program, which insures nearly 10 million of the 45 million children who receive medical coverage through public programs,¹⁰² terminates benefits for incarcerated youth,^100,103 as do qualified health plans through the health insurance marketplaces.¹⁰⁴ Consequently, families must endure the bureaucratic delay of re-enrollment when youths return home.¹⁰⁵ The SUPPORT Act (Public Law No. 115-271), passed in 2018, may improve this situation. It prohibits termination of Medicaid eligibility for juveniles “who are inmates of public institutions.”¹⁰⁶ We must continue to advocate for the positions set forth by the American Medical Association and the American Academy of Pediatrics that state and county medical societies help provide necessary medical services to incarcerated youth¹⁰⁷; that Medicaid, the Children’s Health Insurance Program, and private insurance be available for all incarcerated youths without suspension or termination¹⁰⁸; and that youths without insurance coverage before incarceration be automatically eligible for Medicaid if incarcerated.¹⁰⁹

The Patient Protection and Affordable Care Act (still in force as of this writing) mandates that mental health and substance abuse services be included in insurance plans and considered central components of medical intervention.^110,111,112 Nevertheless, the most significant barrier remains: access to high-quality treatment and health care professionals at all stages of the juvenile justice system. The challenge for the pediatric health community is to implement needed changes. We must increase the number of pediatric health care professionals who can implement evidence-based treatments and are willing to provide services to low-income families. Integrated mental health care models are among the most promising, especially if they include the components now known to be most effective.⁸¹ We must continue to expand the availability of innovative treatments (eg, telehealth options) useful for patients who are unable to access services because of illness, lack of mobility, or distance from available care.^113,114 In addition, we must increase funding to expand services that are accessible to at-risk youth, thereby reducing the likelihood that children will fall between the cracks of inadequate mental health service systems into the juvenile justice net.

Conclusions

Persistent psychiatric disorders may complicate the transition from adolescence to adulthood, which is already challenging for youths in the juvenile justice system, many of whom are from racial/ethnic minority groups and low-income backgrounds. The pediatric health community must advocate for early identification and treatment of disorders among youths involved in the juvenile justice system.

Supplement.

eMethods 1. Expanded Notes on Study Methods

eMethods 2. Maximizing Comparability Among Measures: Changes to Scoring Algorithms for Testing Results

eTable 1. Demographic Characteristics and Retention of the Sample Recruited from the Cook County Juvenile Detention Center Between 1995 and 1998

eTable 2. Prevalence of Psychiatric Disorders Over Time: Sex Differences

eTable 3. Prevalence of Psychiatric Disorders Over Time: Racial/Ethnic Differences Among Males

eTable 4. Prevalence of Psychiatric Disorders Over Time: Racial/Ethnic Differences Among Females

eTable 5. Odds Ratios Describing Demographic Differences in the Prevalence of Selected Psychiatric Disorders From Detention (Baseline) Through the 15-Year Time Point

eTable 6. Odds Ratios Describing Demographic Differences in the Prevalence of Selected Psychiatric Disorders from Detention (Baseline) Through the 15-Year Time Point

eTable 7. Odds Ratios Describing Demographic Differences in the Prevalence of Selected Psychiatric Disorders From Detention (Baseline) Through the 15-Year Time Point

eTable 8. Odds Ratios Describing Demographic Differences in the Prevalence of Selected Psychiatric Disorders From Detention (Baseline) Through the 15-Year Time Point

eTable 9. Odds Ratios Describing Demographic Differences in the Prevalence of Selected Psychiatric Disorders From Detention (Baseline) Through the 15-Year Time Point

eTable 10. 15-Year DSM-IV Diagnoses Predicted From Baseline Diagnoses Among Males

eTable 11. 15-Year DSM-IV Diagnoses Predicted From Baseline Diagnoses Among Females

eFigure 1. Prevalence of Comorbid Disorder in Delinquent Youth From Detention (Baseline) to 15 Years Postbaseline: Sex and Racial/Ethnic Differences

eFigure 2. Prevalence of Any Disorder (Excluding Behavioral Disorders) in Delinquent Youth From Detention (Baseline) to 15 Years Postbaseline: Sex and Racial/Ethnic Differences

eReferences