An Updated Systematic Review and Meta-regression Analysis: Mental Disorders Among Adolescents in Juvenile Detention and Correctional Facilities

Objective To synthesize evidence on the prevalence of mental disorders in adolescents in juvenile detention and correctional facilities and examine sources of heterogeneity between studies. Method Electronic databases and relevant reference lists were searched to identify surveys published from January 1966 to October 2019 that reported on the prevalence of mental disorders in unselected populations of detained adolescents. Data on the prevalence of a range of mental disorders (psychotic illnesses, major depression, attention-deficit/hyperactivity disorder [ADHD], conduct disorder, and posttraumatic stress disorder [PTSD]) along with predetermined study characteristics were extracted from the eligible studies. Analyses were reported separately for male and female adolescents, and findings were synthesized using random-effects models. Potential sources of heterogeneity were examined by meta-regression and subgroup analyses. Results Forty-seven studies from 19 countries comprising 28,033 male and 4,754 female adolescents were identified. The mean age of adolescents assessed was 16 years (range, 10–19 years). In male adolescents, 2.7% (95% CI 2.0%–3.4%) had a diagnosis of psychotic illness; 10.1% (95% CI 8.1%–12.2%) major depression; 17.3% (95% CI 13.9%–20.7%) ADHD; 61.7% (95% CI 55.4%–67.9%) conduct disorder; and 8.6% (95% CI 6.4%–10.7%) PTSD. In female adolescents, 2.9% (95% CI 2.4%–3.5%) had a psychotic illness; 25.8% (95% CI 20.3%–31.3%) major depression; 17.5% (95% CI 12.1%–22.9%) ADHD; 59.0% (95% CI 44.9%–73.1%) conduct disorder; and 18.2% (95% CI 13.1%–23.2%) PTSD. Meta-regression found higher prevalences of ADHD and conduct disorder in investigations published after 2006. Female adolescents had higher prevalences of major depression and PTSD than male adolescents. Conclusion Consideration should be given to reviewing whether health care services in juvenile detention can address these levels of psychiatric morbidity.

dolescents account for approximately 5% of the custodial population in Western countries, and on any given day in the United States, 53,000 young people are detained in various correctional facilities. 1 Psychiatric disorders are known to be prevalent in juvenile offenders. 2 Furthermore, a number of studies indicate that psychiatric disorders in this population are linked to a wide range of negative outcomes, including elevated risk of repeat offenses, 3,4 poor prognosis of mental health problems, high rates of substance misuse, 5,6 increased likelihood to experience or perpetrate violence in intimate relationships, and psychosocial difficulties in adulthood. 7 A previous systematic review and meta-analysis synthesized evidence up to 2006 on the prevalence of mental disorders in detained adolescents. The findings highlighted considerable mental health needs. 8 Since then, a significant body of new primary research has been published. However, recent systematic reviews have been limited by their scope (eg, by including only Englishlanguage reports or not searching the gray literature), a lack of quantitative methods (including heterogeneity analyses), and the use of inconsistent time frames for psychiatric diagnoses (eg, in past month, past year, and lifetime). [9][10][11] This article presents an updated systematic review and meta-analysis on the prevalence of mental disorders in detained adolescents, including posttraumatic stress disorder (PTSD), 12 which has become increasingly researched in this population over the last decade. The findings should inform service provision, planning, and future research.

Protocol and Registration
We conducted this systematic review following the Preferred Reporting Items for Systematic Review and Meta-Analyses 13 and the Meta-analysis of Observational Studies in Epidemiology guidelines (see Table S1, available online). 14 The study protocol was also registered with the PROSPERO International Prospective Register of Systematic Reviews (CRD42019117111).

Search Strategy
We identified studies published between January 1966 and October 2019 reporting the prevalence of mental disorders in adolescents aged between 10 and 19 years in juvenile detention and correctional facilities. For the period January 1966 to May 2006, the methods were described in a previous review conducted by two of the authors (S.F. and N.L.). 8 For this update, we searched the following electronic databases: EMBASE, PsycINFO, Medline, US National Criminal Justice Reference System Abstract Database, Global Health, and Google Scholar. Our search strategy featured terms related to adolescents (juvenile*, adol*, young*, youth*, boy*, or girl*) and custody (prison*, jail*, incarcerat*, custod*, imprison*, or detain*), which was identical to the previous review. For psychotic illnesses, major depression, attention-deficit/hyperactivity disorder (ADHD), and conduct disorder, new search dates ranged from December 2005 to October 2019. However, for PTSD, searches began in January 1980 to coincide with the addition of this disorder to DSM-III. 15 Reference lists were hand-searched. No language restriction was set, and non-English surveys were translated ( Figure 1).

Study Eligibility
We included studies reporting diagnoses of psychotic illnesses, major depression, ADHD, conduct disorder, and PTSD among adolescents in juvenile detention and correctional facilities based on clinical examination or  interviews conducted with semistructured diagnostic instruments. 16 We defined adolescence from the age of 10 to 19 years, 17 comparable with the previous review and consistent with research. 18 We excluded studies that did not report the prevalence rates of mental disorders separately for male and female adolescents (with the exception of samples including <10% girls), surveys featuring enriched or selected samples of juveniles in custody, and studies that employed exclusively self-report instruments to diagnose individuals (but did include the Diagnostic Interview Schedule for Children [DISC], as it was typically administered in a semistructured way). Furthermore, included studies reported current prevalence of psychotic illnesses, major depression, ADHD, and PTSD or lifetime prevalence of conduct disorder that adhered to international classifications (ICD and DSM). Thus, one study 19 was partially excluded because the prevalences of psychotic illnesses, major depression, and ADHD were reported for the past year rather than the past 6 months. Another reason to include PTSD was correspondence from the original review that recommended its inclusion to expand the clinical scope. 20 For psychosis, we excluded one small study 21 (n ¼ 173) owing to being an outlier (11.0%).

Data Extraction
One reviewer (G.B.) extracted data from the newly identified studies according to the protocol used in the previous review. In the case of any uncertainty in data extraction, R.Y. and S.F. were consulted. Gender-specific information was collected in regard to prespecified characteristics: geographic location, year of interview, sampling method (consecutive admissions, total population, random, stratified random, or some combination thereof), participation rate, number of interviewed adolescents, diagnostic instrument and criteria (ICD or DSM), type of interviewer (psychiatrist versus other), proportion of individuals diagnosed with each disorder, mean age and age range, mean duration of incarceration at the interview, and proportion with violent offenses. 8 Authors of primary studies were contacted when further information was required (Table 1).

Quality Assessment
Study quality was assessed in the included surveys using a modified version of the Newcastle-Ottawa Scale, which appraises sample representativeness and size, participation rate, statistical quality, and ascertainment of diagnosis. 22,23 We employed the same version of the checklist used in a recent study of the prevalence of PTSD in prisoners. 24 The potential total score ranged from 0 to 6 points. Studies with a score of 0 to 2 points were considered low quality, studies with scores of 3 to 4 points were considered medium quality, and studies with scores of 5 to 6 points were high quality (see Supplement 1 and Table S2, available online).

Data Analysis
A random-effects meta-analysis was conducted to calculate pooled prevalence of each disorder, given that heterogeneity among studies was high. 25 We aggregated smaller studies, for which the sample size was <100 individuals. For these small studies, prevalences reported in the text were from the nonaggregated data, whereas the figures were generated using results from the aggregated data. The Poisson distribution was used to obtain 95% confidence intervals when events were rare. 26 Two studies 27,28 for which the prevalence of psychotic illnesses was zero were imputed according to standard methods (ie, confidence intervals were calculated using "3" as the numerator and the real population size as the denominator). 29 We reported the I 2 statistic and Cochran's Q to indicate the degree of heterogeneity between studies. In line with guidelines, heterogeneity was considered to be low when I 2 ranged from 0 to 40%; moderate, from 30% to 60%; substantial, from 50% to 90%; and considerable, from 75% to 100%. 30 We conducted subgroup and meta-regression analyses to explore source of heterogeneity on a range of study characteristics: year of publication ( 2006 versus >2006), gender (male versus female), mean age (both as a continuous and as a dichotomous variable; 15 or >15 years), sample size (both as a continuous and as a dichotomous variable; 250 versus >250 adolescents), study origin (United States versus elsewhere), instrument (DISC versus other instruments), diagnostic criteria (ICD versus DSM), interviewer (psychiatrist versus nonpsychiatrist), sampling strategy (stratified/nonstratified random versus consecutive/complete) and study quality score (both as a continuous and as a dichotomous variable; high-quality studies versus low-and medium-quality studies)). We first conducted univariate meta-regression, followed by multivariable analysis including factors that reached statistical significance (set at p < .05) in the univariate models. To test group differences, subgroup analyses were conducted on all dichotomous variables. All analyses were performed using STATA statistical software package, version 13.0 using metan and metareg commands. 31

DISCUSSION
In this updated systematic review of the prevalence of mental disorders among adolescents in juvenile detention and correctional facilities, we identified 47 studies with 32,787 adolescents from 19 different countries. We doubled the number of primary studies compared with a 2008 systematic review. 8 Moreover, we broadened our scope of search by adding a new psychiatric diagnosis (PTSD) and more carefully analyzed heterogeneity. The prevalence estimates confirm high levels of mental disorders in detained adolescents. The two commonest treatable disorders in male adolescents were depression (present in about 1 in 10) and ADHD (prevalent in 1 in 5). In female adolescents, approximately one in four had depression, and one in five had PTSD. We found higher prevalences of depression and PTSD in girls in custody compared with boys.
Our review suggests that mental disorders are substantially more common among detained adolescents compared with general population counterparts. Approximately 3% of detained adolescents were diagnosed with a current psychotic illness, a 10-fold increase compared with ageequivalent individuals in the general population. 75,76 Higher prevalences of current major depression were found in both male (10%) and female (26%) adolescents compared with the general adolescent population (5% and 11%, respectively). 77 About 1 out of 5 detained adolescents had ADHD compared to 1 out of 10 adolescents in the general population. 78 Nearly two-thirds of detained adolescents were diagnosed with any lifetime conduct disorder, whereas the estimated lifetime rate of conduct disorder in US adolescents is approximately 10%. 79 In addition, adolescents in detention also had higher rates of PTSD than those in the general population, 9% versus 2% in male adolescents and 18% versus 8% in female adolescents. 80 These differences underscore the large burden of psychiatric morbidity in detained adolescents.
Apart from higher prevalence than the general population, prevalence estimates in adolescent juvenile detention and correctional facilities were also different from those found in adult prison populations. Psychotic illnesses and major depression appear to be more prevalent in adult prisoners than in adolescent custodial populations. 81 However, the prevalence estimates for PTSD are similar in both groups. 24 These comparisons suggest that the mental health needs of detained adolescents could be different from those of adult prisoners and may require separate and specifically targeted programs to meet these needs.
The prevalences for ADHD and conduct disorder are higher than in the previous 2008 review. Regarding ADHD, this upward trend may be specific to detained adolescents, as ADHD diagnoses in youths in the general population have not increased when standardized diagnostic methods are used. 82 There are two possible explanations for this finding. First, increased prevalence could result from increased awareness of ADHD symptoms among health professionals working in custodial services. That is, the true prevalence of these disorders remains unchanged, but clinicians might be identifying them more accurately. Second, higher prevalence may result from improved identification of adolescents at high risk of reoffending over time. Some individuals with ADHD and conduct disorders who previously might not have been identified may be more likely to be selected for placement in custodial correctional facilities due to improved identification of these disorders. Another main finding was the higher prevalence of major depression and PTSD in detained female adolescents compared with their male counterparts. These results are consistent with results from adult prison samples 24,81,83 as well as the general population, military personnel, and terror attack survivors. [84][85][86][87] However, the explanations for this specific to incarcerated youths are not clear. Criminality in female adolescents may be more strongly associated with internalizing mental disorders than crime in male adolescents, or girls might be more vulnerable to adverse and traumatic experiences related to an antisocial lifestyle either within or outside the detention centers.
Finally, the funnel plot results suggest publication bias in male adolescents toward lower prevalence for conduct disorder and toward higher prevalence for PTSD. This could be due to the increased attention that trauma theory has received as a putative causal mechanism for juvenile criminality. In contrast, a highly prevalent descriptive diagnosis such as conduct disorder might be perceived as less useful for etiologic understanding, treatment planning, and primary prevention regarding juvenile delinquency.
One implication of this updated review is that there is no pressing need for conducting more primary prevalence studies, especially in high-income countries, considering that the evidence base is quite large and with most prevalence estimates remaining stable over time. Hence, future research could move toward treatment and interventions in custodial settings and investigate modifiable risk factors for adverse outcomes within custody such as self-harm and violence that may be associated with mental health problems. Effective treatment will likely improve prognosis and reduce suicidality, violence, and reoffending risk. 88 Some limitations should be noted. First, owing to discrepancies in how substance use disorder and other mental disorders were classified between studies, it was not possible to reliably examine comorbidity. As adolescents who have comorbid disorders generally present an elevated criminogenic risk, future research on comorbidity is needed. 45,69,89 Second, there were insufficient data on the type of facilities (pretrial versus sentenced; short-term versus long-term) where youths were detained. Therefore, we could not explore whether this variable was associated with heterogeneity. Future studies should consider reporting this information on juvenile justice facilities. Third, our analyses were solely based on formal diagnoses of mental disorders according to DSM and ICD, which provide standard ways of communication between mental health professionals. However, we did not report on subthreshold psychiatric symptoms, which future work could examine, as these individuals could benefit from preventive programs. An additional limitation from this review is that the quality appraisal scale was not specifically designed for the purpose of prison prevalence studies, and therefore some of the scoring made assumptions that need further examination (including a lower score for interviews conducted by laypersons using standardized measures versus unstructured clinical interviews conducted by psychiatrists or psychologists, although most of the latter also used standardized tools). Further, there were high levels of between-study heterogeneity. This is expected due to the differences in jurisdictions regarding whom they detain, availability and effectiveness of health care services, and prison environments. Therefore, further work could examine prevalence rates longitudinally in the same individuals to study trends over time. Moreover, we primarily used data from the US general population as a point of comparison for the calculated pooled prevalences because of similar diagnostic instruments, age ranges, and prevalence periods. [77][78][79][80] Nevertheless, as worldwide rates differ, including for ADHD between high-income countries, prevalences should be interpreted in relation to national or regional general population prevalences. Finally, it is notable that all included studies were conducted in high-and upper middle-income countries despite the global search. Determining whether new research in other countries is required will need to be balanced by information in this review, local needs, and whether such research can be linked to improved services.
In conclusion, our updated systematic review has reported high rates of treatable mental disorders in detained adolescents. The findings underscore the importance of access to mental health services and effective treatment. Such treatment will likely improve prognosis of this population, almost all of whom will reenter the community, and decrease risk of repeat offending, reducing the substantial social and financial costs related to imprisonment. 90