;
Prevalence of interpersonal trauma exposure and posttraumatic stress disorder among juvenile offenders: A three-level meta-analysis
Masterscriptie Forensische Orthopedagogiek Graduate School of Development and Education Universiteit van Amsterdam.
R. Riad, 11975075 mw. R. op den Kelder mw. dr. M. Hoeve Amsterdam, april, 2020
Abstract
High prevalence rates of trauma exposure and posttraumatic stress disorder among children and adolescents have been reported in research. Studies on trauma exposure and posttraumatic stress disorder among adult offenders demonstrate even higher prevalence rates compared to the general population. The question arises whether these high rates are also found in juvenile offenders, given the vulnerability of this population. The current study is the first meta-analysis on the prevalence of interpersonal trauma exposure and posttraumatic stress disorder among juvenile offenders. A three-level meta-analysis was conducted on 64 studies and 160 effect sizes published between 1993 and 2019 that were retrieved from MEDLINE, Embase and PsycINFO. The 159.458 participants of the selected studies were aged between 9 to 27 years old. Moderator analyses were conducted to examine possible moderators (age, gender and type of trauma). From the included studies, small effect sizes were found for interpersonal trauma exposure (OR = 1.43) and posttraumatic stress disorder (OR = 0.85), indicating an overall prevalence rate of 63.1% of interpersonal trauma exposure and 18.7% of posttraumatic stress disorder among juvenile offenders. Moderator-analysis showed that age and gender were not significant moderators. Type of trauma was only found to be a significant moderator for interpersonal trauma exposure. Results suggest that
interpersonal trauma exposure and posttraumatic stress disorder are prevalent in juvenile offenders, which emphasizes the importance of a trauma-sensitive approach in juvenile detention facilities.
Background
Trauma can have serious effects on well-being and development throughout the human life span (Bussey & Wise, 2007; Zoellner & Maercker, 2006). According to the Diagnostic and Statistical Manual of Mental Disorders, trauma refers to an actual or threatened death, serious injury, or sexual violence (American Psychiatric Association, 2013). Many children and adolescents experience at least one traumatic event in their lifetime, with prevalence rates varying between 15%–82.5% (Breslau, Lucia, Alvarado, & 2006; Copeland, Keeler, Angold, & Costello, 2007; Landolt, Schnyder, Maier, Schoenbucher, & Mohler-Kuo, 2013;
McLaughlin et al., 2013). Exposure to traumatic events in childhood has been linked to a variety of negative outcomes (Chartier, Walker, & Naimark, 2010), which include anxiety symptoms, poor academic performances, depression, suicidal attempts, eating disorders, behavioural and psychological problems, conduct/antisocial behaviour, psychosis and substance abuse (Dyregrov & Yule, 2006).
In particular, children and adolescents are more susceptible to develop posttraumatic stress disorder (PTSD) after having experienced a traumatic event (Caffo, Forresi, & Lievers, 2005; Ford, Grasso, Hawke, & Chapman, 2013). PTSD is a mental health disorder, which is caused by exposure to a traumatic event (de Vries, 2019). PTSD is the most studied
psychological consequence of traumatic exposure in children (Alisic, Jongmans, van Wesel, & Kleber, 2011). PTSD symptoms are intrusive re-experiences (e.g. intrusive thoughts and nightmares), persistent avoidance (e.g. avoidance of thoughts/feelings related to traumatic events), negative alterations in cognition and mood (e.g. feelings of detachment), and alterations in arousal and reactivity (e.g. hypervigilance, sleep problems; American Psychiatric Association, 2013). Approximately 1.4%–8.3% of children and adolescents develop PTSD (Breslau et al., 2006; Copeland et al., 2007; McLaughlin et al., 2013), with a mean prevalence rate for children of 15.9% (Alisic et al., 2014). PTSD is associated with
functional impairments, which may lead to development of other psychiatric disorders
(Rodriguez, Holowka, & Marx, 2012). Moreover, youth who have been diagnosed with PTSD experience developmental problems in academic, social, emotional and physical domains (Fairbank & Fairbank, 2009; Pynoos et al., 2009; Seng, Graham-Bermann, Clark, McCarthy, & Ronis, 2005). All in all, it seems that PTSD among youth poses a problem, especially given the negative effects on the various developmental domains.
Besides being related to PTSD, trauma exposure is also related to delinquency. According to the theory of the cycle of violence, childhood maltreatment increases the possibility of offending (DeLisi et al., 2009; Widom, 1989). The theory of the cycle of violence suggests that the adoption of antisocial beliefs and attitudes that excuse violence as an acceptable way of solving problems is the causal process in the association between childhood maltreatment and criminal behaviour (Sunday et al., 2011). The theory of the cycle of violence is supported by a large amount of recent studies that showed a positive link between maltreatment and offending in adolescence and adulthood (Cicchetti & Toth, 2005; Maas, Herrenkohl, & Sousa, 2008; Malvaso, Delfabbro, & Day, 2015; Norman, Byambaa, De Butchart, Scott, & Vos, 2012). For example, a meta-analysis of 18 studies showed a positive link between childhood violence exposure and antisocial behaviour in juvenile offenders (Wilson, Stover, & Berkowitz, 2009). Thus, children who have experienced maltreatment are not only at risk for developing PTSD, but also risk having an increased likelihood of juvenile offending (Widom, 1989).
Compared to the general population, prevalence rates of traumatic exposure and PTSD are higher among offenders. Adult offenders have extensive histories of trauma prior to incarceration (Wolff & Shi, 2012), including childhood trauma with prevalence estimates of more than 60% of childhood maltreatment (Dudeck et al., 2011; Messina & Grella, 2006;
the general population (Fazel & Seewald, 2012; James & Glaze, 2006; Trestman, Ford, Zhang, & Wiesbrock, 2007). In sum, there is a growing body of evidence that indicate that offenders experience much higher rates of trauma exposure and PTSD compared to the community sample.
Given the negative outcomes of trauma exposure and PTSD, the question that arises is whether these high rates of trauma exposure and PTSD are also found in juvenile offenders (i.e. a child or young person who has been found guilty of some criminal offence before a juvenile court; Collins English Dictionary, z.j.). Juvenile offenders are already a vulnerable population. However, those who are traumatized or show symptoms of PTSD show higher prevalence rates of school dropout, comorbidity, recidivism and suicide attempts, which makes them even more vulnerable (Cauffman, Monahan, & Thomas, 2015; Haynie, Petts, Maimon, & Piquero, 2009; Marmorstein, 2007; McReynolds, Wasserman, DeComo, Keating, & Nolen, 2008; Wasserman & McReynolds, 2011; Wolff, Baglivio, & Piquero, 2015). On top of that, juvenile offenders are in prison during a critical developmental stage, which is a restrictive environment (West & Sabol, 2008). Prison itself places youth at risk for additional exposure to trauma and may also trigger those with a history of trauma in response to coercive practices that could provoke psychological distress (Dierkhising, Lane, & Natsuaki, 2014; Geller, Fagan, Tyler, & Link, 2014). Altogether, trauma exposure and PTSD influence the development of children and adolescents negatively. With juvenile offenders already being a vulnerable population, trauma exposure and PTSD increases the susceptibility to
developmental problems even more (Benedini & Fagan, 2018; McLaughlin et al., 2013). On top of that, previous research shows that there is a discrepancy between the prevalence rates of trauma exposure (15%–82.5%) and PTSD (1.4%–8.3%) among children and adolescents. This discrepancy between the prevalence rates of trauma exposure and PTSD may also be found in the sample of juvenile offenders. A reason to assume this could be the
misinterpretation of PTSD symptoms as symptoms of behavioural disorders (Bertram & Dartt, 2009), as youth are often diagnosed with attention-deficit hyperactivity disorder or conduct disorder (Bernhard, Martinelli, Ackermann, Saure, & Freitag, 2018; Nock, Joiner, Gordon, Lloyd-Richardson, & Prinstein, 2006). Being a victim or witness of stressful life events, such as community violence, has been associated with maladaptive emotional
responses (Chauhan, Reppucci, Burnette, & Scott Reiner, 2014). As a result, youth exposed to chronic violence behave aggressively and are more likely to be angry, distrustful, depressed and fearful (Goguen, 2008). Social service workers often interpret these symptoms as maladaptive and evidence of behavioural disorders, while these could also be symptoms of PTSD (Bertram & Dartt, 2009). Altogether, it could be possible that there is also a
discrepancy between the prevalence rates of trauma exposure and PTSD among juvenile offenders due to misinterpretation of PTSD symptoms.
Gender differences
Through research, gender differences have been found in trauma exposure and PTSD. In the sample of adults, gender differences are reported in specific types of trauma (Wamser-Nanney & Cherry, 2018). For example, women are more likely to have experienced domestic violence and sexual abuse compared to men, whereas men are more likely to be exposed to nonsexual assaults, motor vehicle accidents, other accidents and natural disaster (Tolin & Foa, 2006). Furthermore, women who have experienced trauma exposure demonstrate higher rates of PTSD and other trauma-related symptoms, such as internalizing disorders (e.g., depression and anxiety) than men who experience more externalizing disorders (e.g., conduct disorder and substance abuse; Kimerling, Weitlauf, Iverson, Karpenko, & Jain, 2014; Tolin & Foa, 2006). As for children and adolescents, previous studies have reported higher prevalence rates of trauma exposure in boys than in girls (Breslau et al., 2006), whereas a meta-analysis
PTSD compared to boys (32.9%). Boys who experienced non-interpersonal trauma were at the least risk of developing PTSD (8.4%; Alisic et al., 2014). In sum, gender differences in trauma exposure and PTSD have been reported, where it seems that females experience more interpersonal trauma and are more likely to develop PTSD and other trauma-related
symptoms than males.
High PTSD prevalence rates in women could be a result of women being more likely to experience interpersonal trauma such as sexual abuse and rape (Chivers-Wilson, 2006), Sexual abuse has the potential for various physical and mental health consequences (Gilbert et al., 2009; Jina & Thomas, 2013), especially given the tendency of disclosure of sexual abuse due to shame and threats made by the perpetrator (McElvaney, 2013; McElvaney, Greene, & Hogan, 2014). In a cross-sectional study, with a sample of 17.970 women across 11 countries, 70.4% reported lifetime trauma exposure. Furthermore, 12.1% of the respondents reported sexual assaults, with an average of 2.4 incidents per respondent. With the majority of
children, adolescents and adults developing PTSD after experiencing sexual abuse (Johnson, Maxwell, & Galea, 2009), sexual assault is the strongest predictor of developing PTSD (Frans, Rimmo, Aberg, & Fredrikson, 2005).
Age
Life-course theories point out adolescence as the critical period of development (Benedini & Fagan, 2018). This life stage is marked by significant and rapid cognitive, physical and social growth, including biological changes which are related to brain development, puberty and development of peer and romantic relationships (Shlafer et al., 2014). Adolescence is also characterized by important milestones, such as forming personal identity and independence from parents (Sawyer et al., 2012; Steinberg, 2001). Therefore, adolescence is an especially vulnerable stage for trauma exposure (Klein & Romeo, 2013). In conclusion, adolescence is a vulnerable stage for trauma exposure in which traumatized youth
are especially susceptible to experience the negative outcomes of trauma exposure.
Given that the adolescence is already a vulnerable stage, juvenile offenders are even more susceptible to experiencing traumatic stress and developing PTSD, as older juvenile offenders have experienced more time in prison. Imprisonment causes traumatic stress given that juvenile offenders face an increased risk of exposure to verbal or physical abuse from peers or staff, which increases the prevalence of trauma exposure. Detention staff intimidate juvenile offenders using verbal and physical confrontation, as they believe that juvenile offenders are too aggressive and defiant to respond to any approach except intimidation (Ford & Blaustein, 2012). As a result, the older juvenile offenders get, the more traumatic events and traumatic stress they experience. This increases the risk of developing PTSD. Therefore, older juvenile offenders could experience more traumatic events than younger juvenile offenders, which increases their likelihood to develop PTSD.
Type of trauma
The current study focusses on interpersonal trauma exposure which is divided into: (1) domestic violence, (2) physical abuse, (3) emotional abuse, (4) sexual abuse, and (5) neglect (Leeb, Paulozzi, Melanson, Simon, & Arias, 2008). In a study with a sample of children in the child welfare system, respondents reported an average of 4.36 different types of trauma experiences. The most frequent type of trauma was neglect (96.4%), followed by domestic violence (77.5%), emotional abuse (53.5%), physical abuse (44.4%), and sexual abuse (44.3%; Whitt-Woosley, Sprang, & Royse, 2018). Moreover, domestic violence, emotional abuse, physical abuse, sexual abuse and neglect are frequently experienced by juvenile offenders (Dixon, Howie & Franzcp, 2005). Therefore, the focus is on these five types of interpersonal trauma as they are most frequently experienced by children and juvenile offenders. To the best of the authors’ knowledge, to date, no meta-analysis on prevalence of
The present study
This study is the first meta-analysis on the prevalence of interpersonal trauma
exposure and PTSD among juvenile offenders. Valid data on the prevalence of interpersonal trauma exposure and PTSD is important for several reasons. Firstly, a meta-analysis on the prevalence of interpersonal trauma exposure and PTSD contributes to the existing body of scientific literature. Secondly, a meta-analysis on the prevalence of interpersonal trauma exposure and PTSD offers insight into the needs of traumatized juvenile offenders, which could improve treatment. Thirdly, investigation of interpersonal trauma exposure and PTSD in juvenile offenders may broaden our understanding of how juvenile offenders process traumatic events. Finally, PTSD among juvenile offenders seriously affects the rehabilitation trajectory, resulting in recidivism, or continuous involvement with the juvenile justice system (Ford, Kerig, Desai, & Feierman, 2016). More insight into the development of PTSD could improve the rehabilitation trajectory. Clinically, more insight into interpersonal trauma exposure and PTSD among juvenile offenders may result in a decrease of recidivism.
Economically and politically, untreated PTSD, incarceration and recidivism (Johnson, Arditti, & McGregor, 2018; Kessler, Chiu, Demler, & Walters, 2005) can come with huge societal costs. Accurate prevalence rates of interpersonal trauma exposure and PTSD enables to improve treatment and balance the cost of prevention.
The current study aims to answer the following research questions: (1) What is the prevalence of interpersonal trauma exposure among juvenile offenders?; (2) What is the prevalence of PTSD among juvenile offenders?; and (3) Do age, gender and type of trauma influence the prevalence of interpersonal trauma exposure and PTSD in juvenile offenders? Prior research reported high prevalence rates of trauma exposure among children, adolescents and adult offenders. As for the prevalence of PTSD, relatively higher PTSD rates are reported in adult offenders than in the general population. Research discussed above on trauma
exposure and PTSD among children, adolescents and adult offenders give us reason to draw the following hypotheses: (1) prevalence rates of interpersonal trauma exposure and PTSD among juvenile offenders are at least as high as, or even higher than in the community sample; (2) there is a discrepancy between prevalence rates of interpersonal trauma exposure and PTSD among juvenile offenders; (3) girls report higher prevalence rates of interpersonal trauma exposure and PTSD than boys; (4) prevalence rates of interpersonal trauma exposure and PTSD are higher in late adolescence than in early adolescence among juvenile offenders; and (5) prevalence rates of interpersonal trauma exposure and PTSD differ between the examined types of trauma.
Method Selection of studies
Studies were included in the meta-analysis if they met four criteria. Firstly, studies had to use a sample of juvenile offenders with a mean age between 10 to 24 years old
(adolescents). Secondly, studies had to report information about the prevalence of
interpersonal trauma exposure and/or PTSD according to DSM (any version) and/or ICD-10. Thirdly, studies had to report in English or Dutch language. Finally, studies had to be peer-reviewed journal articles. Studies were excluded when they included participants with: (1) medical conditions (seizures, medication toxicity, delirium, metabolic or subtle co-occurring neurodevelopmental disorders); (2) psychiatric disorders due to a general medical condition; or (3) psychiatric disorders due to substance intoxication.
This meta-analysis was performed by following the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (Vrabel, 2015). The search strategy, criteria for inclusion and exclusion along with data to be extracted was defined in advance. A systematic literature search was conducted in the electronic databases MEDLINE,
OR exp child abuse/ OR physical abuse/ OR emotional abuse/ OR childhood trauma survivor/ OR exp violence etc. The procedure of article identification included two steps. First, all duplicates were removed. Then, all remaining articles were checked in a first (title, abstract, keywords) and second (full text) screening phase. The full search was conducted until October 2019. For the full electronic search strategy, see Appendix A.
Study selection
The study selection was performed by one reviewer in a standardized manner. Normally, the study selection is conducted by two independent reviewers to ensure an objective and transparent approach for research synthesis and to reduce bias (Borenstein, 2009). Due to time constraints, this was not possible. Moreover, this study was done for thesis purposes which means there was no budget available to hire a second reviewer. The search yielded 2218 papers (after duplicate removal). These papers were screened of which 33 papers on interpersonal trauma exposure and 31 papers on PTSD met the selection criteria. The eligibility assessment is displayed in Figure 1.
Data items
For the data collection process a data-extraction sheet was developed (see Appendix B). The following information was derived from the included studies: (1) study characteristics (publication status and study quality); (2) characteristics of participants (age and gender); (3) interpersonal trauma exposure (prevalence of interpersonal trauma exposure and type of trauma); (4) PTSD (PTSD prevalence and PTSD diagnosis); and (5) type of outcome measure (e.g., case files, Posttraumatic Stress Disorder Reaction Index).
The Effective Public Health Practice Project (EPHPP) created a quality assessment tool for the assessment of study quality of quantitative studies (Thomas, Ciliska, Dobbins, & Micucci, 2004). In the current study, this tool was used in order to evaluate the findings of research on the prevalence of interpersonal trauma exposure and PTSD. This assessment does
not only provide an overview of the quality of the results reported by previous studies, it also provides the opportunity to make recommendations for future research. An overall quality rating of weak, moderate or strong was appointed by one reviewer. Although study quality assessment by two reviewers is preferred (similar to the screening phase), this was beyond the scope of this thesis due to time and budget limitations.
Strategy of analysis
In 58% of the included papers about interpersonal trauma exposure and 52% of the papers about PTSD, multiple relevant effect sizes were reported. Papers reported on effect sizes for two reasons: (1) data on interpersonal trauma exposure and/or PTSD prevalence among girls and boys was reported separately; and (2) different types of trauma were reported. The log odds ratio (OR) has been used as measure to describe the effect size, as scholars argue this to be statistically preferable for meta-analyses (Fleiss & Berlin, 2009). The OR was calculated using logit and the variance of logit (Barendregt, Doi, Lee, Norman, & Vos, 2013). To interpret the size of OR, the method of Chen, Cohen, and Chen (2010) was used. This method indicates that OR = 1.68, 3.47, and 6.71 are respectively interpreted as small, medium and large. The effect size calculation (in the SPSS syntax) was double-checked by the thesis supervisor.
A three-level meta-analytic random effects model was used so that maximum
statistical power could be achieved (Assink & Wibbelink, 2016). A three-level meta-analytic random effects model provides more information given the fact that effect sizes are not eliminated or averaged (Assink & Wibbelink, 2016; Cheung, 2014). Three levels of variance were modelled; (1) sampling variance; (2) variance between the effect sizes from the same study; and (3) variance between studies. A multi-level approach allows for examining within-study heterogeneity as well as between-within-study heterogeneity (Assink & Wibbelink, 2016).
explain within- or between-study differences in effect sizes (Borenstein, Hedges, Higgins, & Rothstein, 2010). For the moderator analyses, as categorical variable, dummy variables were created, and a reference category was chosen. Physical abuse was chosen as the reference category as physical abuse has shown to be the most conspicuous precursor to delinquency (Platt, 2016). The statistical analyses for the three-level meta-analytic random effects model were executed with the software R (using the metafor package; Assink & Wibbelink, 2016). Publication bias
Publication bias is a prevalent issue in meta-analytic research (Borenstein et al., 2010), meaning that studies that report significant results are more probable to be published than studies without significant results (Lin et al., 2018). This could result in an overestimation of the true effect size in a meta-analysis (Borenstein, 2009). In order to investigate publication selection bias, the PET-PEESE approach was used. This approach has been shown to be a better fit than the Trim & Fill strategy and Fail Safe N analysis (Stanley & Doucouliagos, 2014). The PET-PEESE approach consists of two steps. First, the precision-effect-test (PET; based on results on the Egger test). For the PET analysis the standard error is used as a moderator, in which a significant moderator implicates possible publication bias. When the intercept is significantly different from zero, the next step is executed: precision-effect estimate (PEESE). For this analysis, the variance is included as a moderator. Publication bias is implicated when the effect size varies significantly with the standard error. However, it should be noted that the PET-PEESE approach is designed for a fixed effects model. In the current study it was used in a random model. Furthermore, it is important to note that all publication bias analyses have a low power to detect bias (Borenstein et al., 2010; Stanley & Doucouliagos, 2014).
Figure 1
Assessment for eligibility
Papers identified through search MEDLINE, Embase, PsycINFO (n =
16.271)
Papers screened after duplicate removal (n = 2.218)
Papers assessed for full text eligibility (n = 253)
Papers excluded (n = 2.492) for following reasons:
• No data on interpersonal trauma exposure and PTSD
• Population not fitting inclusion criteria
• Other language than English or Dutch
• Insufficient information
Papers included for interpersonal trauma exposure (n = 33)
Sample
For the prevalence of interpersonal trauma exposure, 33 samples and 100 effect sizes were examined, reporting data on 141.148 juvenile offenders of whom 38.137 were girls. The ages in the samples ranged from 9 to 27 years. For the prevalence of PTSD, 31 samples and 60 effect sizes were examined, reporting data on 18.310 juvenile offenders of whom 4.678 were girls. The ages in the samples ranged from 9 to 27 years. Of the included samples, 15% and 13% of respectively the studies on interpersonal trauma exposure and PTSD, consisted of girls only. Additionally, of the included samples, 9% and 23% of respectively the studies on interpersonal trauma exposure and PTSD, consisted of boys only. Lastly, 76% and 61% of the included samples on respectively interpersonal trauma exposure and PTSD consisted of both girls and boys.
Constructs
In order to be able to perform unambiguous empirical observations during data collection, a translation must be made from theoretical concepts to measurable "operational" concepts (Vennix, 2010). In the present study interpersonal trauma exposure, PTSD and type of trauma are operationalized by the prevalence rates of these concepts.
In addition, constructs were defined prior to the screening in order to give direction to the search. The concept of ‘juvenile offender’ was defined as an adolescent who has been found guilty of some criminal offence before a juvenile court (Collins English Dictionary, z.j.). The concept of ‘adolescence’ was defined as youth aged 10 to 24 years as increasing numbers of studies with relevance to adolescent health consider the 10 to 24-year age group as adolescence (Patton et al., 2009; Sawyer et al., 2012; Patton et al., 2013; Sawyer,
Azzopardi, Wickremarathne, & Patton, 2018).
The concept of ‘type of trauma’ is broad. In this study, the focus was on five types of interpersonal trauma, namely (1) domestic violence, (2) emotional abuse, (3) physical abuse,
(4) sexual abuse, and (5) neglect. Firstly, the concept of ‘domestic violence’ was defined as abusive behavioural acts by all members in the household and ‘house friends’ (people that are close to the household; Atria, 2019). Secondly, the concept of ‘emotional abuse’ was defined as non-accidental verbal or symbolic abusive acts (excluding physical and sexual abusive acts) such as degradation and humiliation (Smith Slep, Heymana, & Snarr, 2011). Thirdly, ‘physical abuse’ was defined as any non-accidental act or behaviour causing injury, trauma, or other physical suffering (American Humane Association, z.j.; Mash & Wolfe, 2012). As for ‘sexual abuse’, this concept was defined as forced or coerced sexual activity (Levine et al., 2018). Lastly, ‘neglect’ was divided in two dimensions, namely ‘emotional neglect’ and ‘physical neglect.’ Emotional neglect was defined as failure to provide a child with emotional support, security and encouragement. Physical neglect was defined as failure to provide a child with basic necessities of life, such as food, clothes and shelter (Miller-Perrin & Perrin, 2012).
Results Interpersonal trauma exposure
In the present study two separate multilevel meta-analyses were performed. Overall effect sizes are displayed in Table 1. For the prevalence of interpersonal trauma exposure, 33 samples and 100 effect sizes were examined. Figure 2 displays a forest plot presenting the effect sizes and their confidence intervals.
The overall mean effect size for the prevalence rate of interpersonal trauma exposure among juvenile offenders was significant (OR = 1.43, p = <0.001, k = 100), demonstrating that the prevalence of interpersonal trauma exposure among juvenile offenders differs significantly from zero. The analyses yielded a small effect size, according to Chen, Cohen, and Chen (2010). When transforming the overall mean effect size back to percentages, it
corresponds with an overall effect prevalence rate of interpersonal trauma exposure of 63.1% among juvenile offenders.
Table 1
Effect sizes (ES) and confidence intervals (CI) for meta-analyses on the prevalence of interpersonal trauma exposure and PTSD
K ES n OR 95% CI p Interpersonal trauma exposure 33 100 141.148 1.43 1.18 ; 1.68 <0.001 PTSD 31 60 18.310 0.85 0.71 ; 0.99 <0.001 K = number of samples
In order to examine whether moderator analyses were necessary, an analysis was conducted to examine whether variety in effect sizes could be assigned to random sampling error, within-study variance, or between-study variance. Effect sizes were heterogeneous for interpersonal trauma exposure; within-study variance (σ2 = 0.06, X2 (1) = 10671.00, p < 0.001) and between-study variance were found to be significant (σ2 = 0.34, X2 (1) = 102.03, p < 0.001). The results of the analysis indicated that 15.6% of the total variance was assignable to within-study differences and 84.3% to between-study differences. This suggests that study outcomes are likely influenced by study characteristics which justifies the investigation of potential moderators. Therefore, moderator analyses were conducted for interpersonal trauma exposure.
Moderator-analyses were conducted to examine the moderators age, gender and type of trauma. Table 2 presents the results of the analyses for interpersonal trauma exposure. From Table 2, it can be derived that the moderators age (F(1, 98) = 0.04, p = 0.847), gender (F(1, 98) = 1.74, p = 0.190), and study quality (F(2, 97) = 1.08, p = 0.343) were not
significant moderators, indicating that age, gender and study quality did not differ. This demonstrated that the prevalence rate of interpersonal trauma exposure did not differ in the
age group of 10 to 24-year-old juvenile offenders. Furthermore, this indicated that the prevalence rate of interpersonal trauma exposure did not differ between female and male juvenile offenders and that study quality did not influence the prevalence rate of interpersonal trauma exposure.
Type of trauma was also examined as a possible moderator. A significant overall moderator effect was found, indicating that the prevalence of interpersonal trauma exposure differs significantly between types of trauma (F(4, 78) = 3.57, p = 0.01). The mean effect size of physical abuse (t = 8.63, p = <0.001) was significantly higher than the effect sizes of sexual abuse (t = -2.55, p = 0.013) and neglect (t = -2.62, p = 0.011), corresponding with a
prevalence rate of 45.6%. Additionally, emotional abuse (t = 8.06, p = <0.001) has shown to have a significant higher prevalence rate than neglect (t = -2.46, p = 0.016). Domestic violence, sexual abuse, emotional abuse and neglect corresponded respectively with a prevalence rate of 50.6%, 30.5%, 44.7% and 23.4%.
PTSD
For the prevalence of PTSD, 31 samples and 60 effect sizes were examined. Figure 3 displays a forest plot presenting the effect sizes and their confidence intervals.
The overall mean effect size of the prevalence of PTSD among juvenile offenders was significant (OR = 0.85, p = <0.001, k = 60 effect sizes), indicating that the prevalence of PTSD among juvenile offenders differs significantly from zero. The analyses yielded a small effect size, according to Chen, Cohen, and Chen (2010). When transforming the overall mean effect size back to percentages, it corresponds with an overall effect prevalence rate of PTSD of 18.7% among juvenile offenders.
As well as for interpersonal trauma exposure, an analysis to examine whether
to be significant, but between-study variance was, in contrast, found to be significant (σ2 = 0.16), X2 (1) = 3.88, p = 0.049). The analysis indicated that 10.3% was assignable to within-study differences and 86.2% to between-within-study differences. Because of the significant between-study variance, moderator analyses were conducted for PTSD.
From Table 3, it can be derived that the moderators age (F(1, 58) = 0.18, p = 0.678), gender (F(1, 58) = 0.05, p = 0.831), study quality (F(2, 57) = 0.30, p = 0.740), and type of trauma (F(1, 9) = 0.89, p = 0.508) were, contrary to the hypotheses of this study, not significant moderators. This indicated that the prevalence rate of PTSD did not differ in the age group of 10 to 24-year-old juvenile offenders. Furthermore, this indicated that the
prevalence rate of PTSD did not differ between female and male juvenile offenders, that study quality did not influence the prevalence rate of PTSD, and that the prevalence of PTSD did not differ significantly between types of trauma.
Publication bias
Possible publication bias was examined by applying the PET-PEESE approach. The PET analysis demonstrated that the effect sizes varied significantly with the standard error for interpersonal trauma exposure and PTSD (p < 0.001), which implicated a file-drawer
problem. Figure 4 displays the funnel plot of the effect sizes of interpersonal trauma exposure and Figure 5 presents the effect sizes of PTSD. After examining the funnel plots of
interpersonal trauma exposure and PTSD, most studies were distributed in the upper half of the funnel plots and spread asymmetrically. This indicated that researchers most likely dropped the imprecisely estimated effects that were close to zero, but not those far from zero. This means that there is bias in published research on the prevalence of interpersonal trauma exposure and PTSD among juvenile offenders. If there was no publication bias, the plots would have been symmetric (Bryman, 2016). Altogether, the funnel plots of interpersonal
trauma exposure and PTSD indicated the presence of a file-drawer problem in research on prevalence of interpersonal trauma exposure and PTSD among juvenile offenders.
Figure 2
Forest plot of the meta-analysis on the prevalence of interpersonal trauma exposure. RE, random effects
Table 2
Moderator analyses for the prevalence rate of interpersonal trauma exposure
Variable K ES β0 (SE) t0 β1 (SE) t1 F (df1,df2)
Study characteristics Age (mean centred) 27 80 1.43(0.13) 11.07 0.01(0.07) 0.03(1,98) Gender (mean centred) 31 95 1.42(0.12) 11.44 -0.00 (0.00) 1.17(1,98) Trauma characteristics Type of trauma 3.57(4,78)* Physical abuse (RC) 12 24 1.25(0.15) 8.63*** Sexual abuse 13 25 1.05(0.14) 7.52*** -0.20 (0.078) -2.55* Neglect 5 7 0.93(0.17) 5.41*** 0.31 (0.12) -2.62* Emotional abuse 7 13 1.24(0.15) 8.06*** -0.01 (0.09) -0.10 Domestic violence 7 13 1.30(0.17) 7.48*** 0.05(0.17) 0.31 Study Quality 1.08(2,97)
Strong (RC) 1 1 1.61(0.76) 2.11 Moderate 4 8 1.92(0.37) 5.25 0.31(0.85) Weak 28 91 1.81(0.19) 9.34 -0.03 (0.18)
K = number of samples; ES = number of effect sizes; β0 = mean effects size (Odds Ratio); t0 = test statistic for difference mean effect with zero; β1 = regression coefficient; t1 = test statistic of difference of mean effect size with the reference category (RC); F (df1,df2) = test statistic for testing significance of moderator (p < 0.05).
Figure 3
Table 3
Moderator analyses for the prevalence rate of PTSD
Variable K ES β0 (SE) t0 β1 (SE) t1 F (df1,df2)
Study characteristics Age (mean centred) 26 50 0.86 (0.07) 12.10 -0.02 (0.04) 0.17(1,58) Gender (mean centred) 31 60 0.85 (0.07) 12.07 0.00(0.00) 0.05(1,58) Trauma characteristics Type of trauma 0.51(4,9) Physical abuse (RC) 3 3 0.97(0.22) 4.45** Sexual abuse 4 4 0.84(0.21) 4.10** -0.13 (0.18) -0.72 Neglect 1 1 0.86(0.29) 3.01* -0.11 (0.28) -0.40 Emotional abuse 1 1 1.34(0.31) 4.30** 0.37 (0.27) 1.37 Domestic violence 3 5 1.05(0.27) 4.25** 0.08(0.33) 0.24
Strong (RC) 2 2 1.61(0.76) 2.11 Moderate 9 12 1.92(0.37) 5.25 0.31(0.85) Weak 20 46 1.81(0.19) 9.34 -0.03 (0.18)
K = number of samples; ES = number of effect sizes; β0 = mean effects size (Odds Ratio); t0 = test statistic for difference mean effect with zero; β1 = regression coefficient; t1 = test statistic of difference of mean effect size with the reference category (RC); F (df1,df2) = test statistic for testing significance of moderator (p < 0.05).
Figure 4
Figure 5
Discussion
To the authors’ knowledge, this is the first meta-analysis on the prevalence of interpersonal trauma exposure and PTSD among juvenile offenders. This study included 33 studies on the prevalence of interpersonal trauma exposure with a sample of 141.148 juvenile offenders and 31 studies on the prevalence of PTSD with a sample of 18.310 juvenile
offenders. The aim was to determine prevalence rates of interpersonal trauma exposure and PTSD among juvenile offenders and examine whether age, gender and type of trauma affect the prevalence rates of interpersonal trauma exposure and PTSD.
Results demonstrated high prevalence rates of interpersonal trauma exposure (63.1%) and PTSD (18.7%) among juvenile offenders. Results of this meta-analysis are comparable to findings on the prevalence of trauma exposure and PTSD among the community sample. These results also correspond with this study’s hypothesis, expecting a discrepancy between the prevalence rates of interpersonal trauma exposure and PTSD among juvenile offenders. Additionally, for interpersonal trauma exposure, the moderator-analysis showed a significant effect size for type of trauma. Age and gender were not significant moderators. As for PTSD, moderator-analysis showed that age, gender and type of trauma were not significant
moderators.
Interpersonal trauma exposure
Results demonstrated a small but significant effect size of interpersonal trauma exposure (OR = 1.43). The overall effect size of interpersonal trauma exposure indicated that approximately 63.1% of juvenile offenders have experienced at least one traumatic event in their lives, which is a substantive number of interpersonal trauma exposure. Furthermore, type of trauma was a significant moderator, but contrary to the hypotheses of this study, age and gender were not. Thus, the prevalence of interpersonal trauma exposure differs between types
female and male juvenile offenders. The mean effect size of physical abuse was significantly higher than the effect sizes of sexual abuse and neglect. Additionally, the results of the moderator-analysis showed that the prevalence of emotional abuse was also significantly higher than the prevalence of neglect.
Based on research about interpersonal trauma exposure, a prevalence rate of interpersonal trauma exposure at least as high as, or even higher than in the community sample was expected. With approximately 63% of juvenile offenders with a history of interpersonal trauma exposure, the results showed that the prevalence rate of interpersonal trauma exposure among juvenile offenders is as high as in the community sample. This emphasizes the importance of a trauma-sensitive approach within the criminal justice system as research has showed numerous negative outcomes related to trauma exposure (Dyregrov & Yule, 2006).
Although the prevalence rate of interpersonal trauma exposure has found to be as high as in the community sample, it is expected that this is just the tip of the iceberg, since only five types of interpersonal trauma were examined. Prior research on trauma exposure have examined more than the five types of interpersonal trauma examined in the current study. For example, a longitudinal study examined rape/sexual assault, torture, kidnapping, being shot/stabbed, being mugged/getting threatened with a weapon, serious accident, natural disaster, diagnosis with life-threatening illness, witnessing killing, etc. This study reported 76% trauma exposure among 17-year-old adolescents who have been followed since they were 6 years old (Breslau et al., 2006). Another study examined 15 types of trauma with results indicating a prevalence rate of 68% of trauma exposure among children and
adolescents (Copeland et al., 2007). In sum, only examining five types of interpersonal trauma may have resulted in missing data on the prevalence of interpersonal trauma exposure among juvenile offenders.
Another reason to assume that the prevalence rate of interpersonal trauma exposure is higher in reality, is that the majority of the data is based on self-reports (88%). Disclosure of interpersonal trauma has been studied extensively and research shows that underreporting is an important problem. A possible underlying mechanism of this underreporting could be shame (Sable, Danis, Mauzy, & Gallagher, 2006; Thompson, Sitterle, Clay, & Kingree, 2007). For example, sexually abused victims tend to have more adversity with disclosure due to stigma. Research suggests that shame and reduced disclosure may be partially caused by the fear of negative social reactions (Budden, 2009). Additionally, sexually abused victims may fear the perpetrator, as they are often threatened to keep the sexual abuse a secret
(McElvaney, Greene, & Hogan, 2014). In conclusion, it is a possibility that the respondents of this study have underreported due to shame, fear and stigma.
As for the moderator analysis, gender was, contrary to the hypothesis of this study, not found to be a significant moderator. The hypothesis was that girls would report significantly higher rates of interpersonal trauma exposure than boys, as prior research has shown gender differences in trauma exposure, with females being more likely to develop PTSD and trauma-related symptoms (Kimerling et al., 2014; Tolin & Foa, 2006). The results of this meta-analysis did not show gender differences in the prevalence rate of interpersonal trauma exposure due to the majority of this study’s sample consisting of boys (73.1%). The author of this study could not have foreseen this, given that there are more males in the criminal justice system than females (Wagner, 2012).
Additionally, age was not a significant moderator either. It was expected that older juvenile offenders would report more interpersonal trauma exposure than younger juvenile offenders due to the possibility of experiencing traumatic events and traumatic stress in prison as juvenile offenders get older. Perhaps another factor related to age could have had more
majority has experienced cumulative trauma (i.e., the amount of different types of trauma experienced; Breslau, Reboussin, Anthony, & Storr, 2005; Finkelhor, Turner, Shattuck, & Hamby, 2013). Youth who have been exposed to cumulative trauma experience more chronic and complex psychological issues than those who have only experienced a single traumatic event, and are at increased risk for delinquency (Cloitre et al., 2009; Ford, Elhai, Connor, & Frueh, 2010; Ford et al., 2013; Richmond, Elliott, Pierce, Aspelmeier, & Alexander, 2009). In sum, cumulative trauma is a risk factor that can threaten successful development and
contributes to unsuccessful adaptations like delinquency (Benedini & Fagan, 2018). Therefore, cumulative trauma could influence the prevalence rate of interpersonal trauma exposure among juvenile offenders.
Lastly, type of trauma, the prevalence of interpersonal trauma exposure differs
significantly between types of trauma. This result supports the hypothesis of this study. In the current study, domestic violence has the highest prevalence rate (50.6%). This could possibly be explained as follows: children who witness domestic violence are at great risk for
developing externalizing behaviour, such as physical aggression and behavioural problems (Boeckel, Wagner, & Grassi-Oliveira, 2015; Evans, Davies, & DiLillo, 2008). The theory of the cycle of violence suggests that children who, for example witness domestic violence, adopt antisocial beliefs and attitudes. Violence is then seen as an acceptable way of solving problems, which could lead to criminal behaviour (Widom, 1989). Following this theory, it could be possible that juvenile offenders who have witnessed or experienced domestic violence copied antisocial and aggressive behaviour that led to criminal behaviour, which explains the high prevalence of domestic violence among juvenile offenders.
The second most reported type of trauma exposure is physical abuse (45%). This corresponds with the prevalence rate of physical abuse among adult offenders (49.8% of male adult offenders). As physically abused victims show the highest rate of arrests for violent
criminal behaviour (Halsey, 2018), this could be an explanation for the high rate of physical abuse among juvenile offenders.
PTSD
The results demonstrated a small but significant effect size for PTSD (OR = 0.85), corresponding with approximately 18.7% of juvenile offenders that are diagnosed with PTSD. As expected, the prevalence of PTSD among juvenile offenders is higher than in the
community sample (1.4%–8.3%; Breslau et al., 2006; Copeland et al., 2007; McLaughlin et al., 2013). This result is important, as research suggests that PTSD is a predictor of
delinquency; PTSD rates of adult offenders are 1.5-4 times higher compared to the general population (Fazel & Seewald, 2012; James & Glaze, 2006; Trestman et al., 2007). This finding is also important given that juvenile offenders often face an increased risk of trauma exposure in prison. Children with PTSD show alterations in arousal and reactivity (American Psychiatric Association, 2013), which are expressed in behavioural dyscontrol such as
aggression (Ostrov & Godleski, 2010). As a result, detention staff verbally and physically intimidate juvenile offenders due to their belief that this is the only way to discipline (Ford & Blaustein, 2012). In conclusion, prison is by definition an unsafe environment. This forms an immense problem as incarcerating youth with PTSD increases the risk of additional trauma exposure and traumatic stress (Ford & Blaustein, 2012).
Findings of this study also showed a discrepancy between the prevalence rates of interpersonal trauma exposure and PTSD among juvenile offenders, which is in line with the hypothesis of this study. As discussed in the introduction of this study, misinterpretation of PTSD symptoms as symptoms of behavioural disorders occurs frequently. Furthermore, men who have experienced trauma exposure show high rates of externalizing disorders (e.g., conduct disorder and substance abuse; Kimerling et al., 2014; Tolin & Foa, 2006). As this
the explanation of the discrepancy between the prevalence rates of interpersonal trauma exposure and PTSD among juvenile offenders. In conclusion, it seems that PTSD symptoms are not always recognised as such, which could explain the discrepancy between the
prevalence rates of interpersonal trauma exposure and PTSD among juvenile offenders. Another possible explanation for the discrepancy between the prevalence rates of interpersonal trauma exposure and PTSD may be the comorbidity of PTSD and conduct disorder (Bernhard et al., 2018). Twin studies have reported shared risk factors for
externalizing and internalizing problems among children and adolescents (Rhee et al., 2015), which also could underlie comorbidity of PTSD and conduct disorder. Moreover, research have found a comorbid relationship between PTSD and conduct disorder in samples of veterans and the general population (Allwood, Dyl, Hunt, & Spirito, 2008; Connor, Ford, & Albert, 2007; Koenen et al., 2005; Nock, Kazdin, Hiripi, & Kessler, 2006). In sum, symptoms of conduct disorders are recognised more frequently than the symptoms of PTSD (Bertram & Dartt, 2008). However, reality shows that there could actually be comorbidity of PTSD and conduct disorder, which should result in higher prevalence rates of PTSD than currently reported.
In contrast to the hypothesis on the moderators age, gender and type of trauma, analyses showed that these three moderators were not significant. As for gender, the same explanation that has been discussed before in the section of interpersonal trauma exposure applies to PTSD as well. The majority of the sample consisted of boys. As a result, the
moderator-analysis did not show gender differences in the prevalence rate of PTSD due to the low number of girls in the sample.
For the moderator age, it was expected that older adolescents would report more PTSD than younger adolescents. This finding could be explained by coping. In a study of adult offenders, researchers have found that their shared identity grew stronger in prison. They
sensed social support by their fellow incarcerated offenders and they coped more effectively with situational stressors (Haslam & Reicher, 2006). In another study, with a sample of juvenile offenders who have experienced cumulative trauma, greater levels of cumulative trauma were associated with greater levels of spiritual coping resources. These findings suggest that juvenile offenders may already have some internal tools to deal with the negative sequelae of trauma exposure (Maschi, MacMillan, Morgen, Gibson, & Stimmel, 2010). Therefore, experiencing a traumatic event does not have to lead to PTSD. This could explain why older juvenile offenders did not report higher prevalence rates of PTSD than younger juvenile offenders.
Lastly, for the moderator type of trauma, the hypothesis was that the prevalence rate of PTSD would differ between the five types of trauma examined in this study. This moderator has shown not to be significant. It could be possible that the experience of a traumatic event in itself raises the susceptibility to developing PTSD, regardless of what type of trauma is
experienced. Children experience traumatic events differently compared to adults. They have far less experience in the world. Children could experience some events as extremely
stressful, whereas an adult would not (Verlinden, 2014). As a result, most traumatic events could lead to PTSD among children as they may experience traumatic events as extremely stressful.
Strengths and limitations
This study was the first meta-analysis to examine the prevalence rate of interpersonal trauma exposure and PTSD among juvenile offenders with a three-level meta-analysis
approach. A three-level meta-analytic model is a relatively new, advanced and strong method for managing dependency of effect sizes. Applying a multi-level model made it possible to include all potential effect sizes and correct for statistical dependency. Thus, the findings of
trauma exposure and PTSD and adds scientific and clinical value.
Despite these strengths, this meta-analysis has several limitations. Firstly, the analysis showed a presence of publication bias. This means that possibly only studies that reported significant results were published than studies without significant results (Lin et al., 2018). Due to the publication bias, results of this meta-analysis should be interpreted carefully, as the effects found in this study may be smaller in reality.
Secondly, PTSD and comorbid disorders were not taken into account in this meta-analysis. Comorbidity is high in PTSD patients. Of individuals with PTSD, 80–90% have at least one comorbid condition and around 60% has two or more additional diagnoses (Foa, Keane, Friedman, & Cohen, 2009). Although the author of the present study was not able to assess for comorbid disorders, to provide and improve evidence-based interventions for juvenile offenders, it is necessary that valid and reliable prevalence estimates of comorbid PTSD and other disorders are available.
Lastly, the quality of the measurement instrument used could have had an effect on the study outcomes. The possible influence of the quality of the measurement on the prevalence rates of interpersonal trauma exposure and PTSD was not tested, as this was beyond the scope of this thesis. As a result, the reliability of the outcomes of this study cannot be indicated. Additionally, it is difficult to determine the reliability of case files. Processing data consistently from case files is challenging. The information about maltreatment is often vague, incomplete and inconsistent, which results in difficulty coding the information for statistical analysis (Huffhines et al., 2016). Interestingly, a study on interpersonal trauma exposure among juvenile offenders found roughly 95% interpersonal trauma exposure using a questionnaire (Modrowski, Bennett, Chaplo, & Kerig, 2017), whereas another study found around 5% examining case files (Wylie & Rufino, 2018). This shows that there is still a lot of
confusion in processing data from case files. Controlling quality of measurement could have made interpretation of the results more reliable.
Future research
The rate of comorbidity is high in individuals with PTSD (Foa et al., 2009). Prior research indicates that comorbidity appears due to the causal relationship between certain symptoms (Borsboom, Cramer, Schmittmann, Epskamp, & Waldorp, 2011). Future research should therefore attempt to examine the shared variance between PTSD symptoms and symptoms of other mental health and substance abuse disorders. Such analyses could result in more clarification of the specific aspects of the comorbid relationship between PTSD and other disorders, which could benefit treatment. In addition, the predictors of PTSD and other psychopathy should be examined in future research. Understanding the association between predictors of PTSD would substantially enhance knowledge regarding the impact of these predictors, which could further benefit treatment. Furthermore, clinicians are urged to implement differential assessment at the beginning of the trajectory of juvenile offenders to classify possible PTSD and perhaps comorbid disorders. This could benefit the trajectory of juvenile offenders as differential assessment benefits treatment.
Secondly, the examined moderators could not explain the high level of heterogeneity between samples. Future research should therefore examine other moderators (aside from the moderators used in this meta-analysis), such as other types of interpersonal trauma,
cumulative trauma and age of onset of trauma. It is possible that examining other moderators generates more insight into interpersonal trauma exposure and PTSD among juvenile
offenders, which could provide guidelines for improving their rehabilitation. Examining the age of onset of trauma could benefit preventive care as literature on delinquency has indicated that the age of onset trauma is the most robust predictor of delinquency (Dierkhising et al.,
measurements used to assess interpersonal trauma exposure and PTSD, were not examined. It could be interesting for future studies to include this in research as it could explain part of the heterogeneity. Additionally, examining study characteristics results in more accurate study findings.
Lastly, over 90% of the studies included in this meta-analysis were coded as low quality, which underlines the importance for researchers to further increase the quality of their studies. The author of this study urges researchers to systematically report selection bias, study design, confounders, blinding and data collection methods. They should be aware of the role of blinding and possible confounders when drawing conclusions (Armijo-Olivo, Stiles, Hagen, Biondo, & Cummings, 2010).
Conclusion
This meta-analysis draws special attention to the prevalence rates of interpersonal trauma exposure and PTSD among juvenile offenders. The results indicate high prevalence rates of interpersonal trauma exposure and PTSD, with a significant higher prevalence rate of interpersonal trauma exposure compared to PTSD. This emphasizes the importance of a trauma-sensitive approach for juvenile offenders as research has shown various negative outcomes of interpersonal trauma exposure and PTSD. Understanding behaviours through a trauma lens could benefit treatment and rehabilitation. The findings of this study imply that juvenile detention facilities should incorporate a trauma-sensitive approach by accurately assessing and identifying trauma and PTSD (Smithgall, Cusick, & Griffin, 2013). A trauma-sensitive approach would help to perceive juvenile offenders as more than their behaviours and move beyond a system of punishment.
Appendix A Table A1
Search strategy for the Embase database
1. exp child abuse/ OR child neglect/ OR emotional abuse/ OR physical abuse/ OR incest/ OR institutionalization/ OR institutionalized adolescent/ OR institutionalized child/ OR exp child welfare/ OR abandoned child/ OR institutionalized child/ OR (childhood trauma OR trauma exposure OR emotional trauma OR psychological trauma OR neglect* OR abuse OR maltreat* OR sexual abuse OR child mistreatment OR institutionaliz* OR early adversity).ti,ab,kw.
2. posttraumatic stress disorder/ OR early life stress/ OR psychotrauma/ OR psychotrauma assessment/ OR (ptsd* OR ptss* OR post-traumatic stress* OR posttraumatic stress* OR traumatic stress* OR psychotrauma* OR psycho-trauma* OR stress disorder* OR ((stressor OR trauma) ADJ2 disorder*) OR traumatic experience* OR traumatised youth).ti,ab,kw. 3. juvenile delinquency/ OR juvenile justice/ OR juvenile offenders/ OR juvenile OR
(juvenile* OR delinq* OR resident* care) OR ((child* OR adolesc* OR youth* OR juvenile) ADJ3 (delinq* OR devian* OR crim* OR offend*)).ti,ab,kw.
4. correctional institutions/ OR juvenile justice/ OR prisons/ OR (correctional OR prison* OR justice OR institutional care).ti,ab,kw.
(1 or 2) and 3 and 4
Table A2
Search strategy for the PsycINFO database
1. child abuse/ OR child abuse reporting/ OR child neglect/ OR domestic violence/ OR emotional abuse/ OR incest/ OR physical abuse/ OR rape/ OR sexual abuse/ OR sexual harassment/ OR verbal abuse/ OR abandonment/ OR (child abus* OR child maltreat* OR child mistreatment OR child* welfare recidiv* OR institutionaliz* OR neglect OR out-of-home placement* OR psychological abus* OR sexual abus* OR verbal abus* OR physical abuse OR early adversity OR emotional trauma OR psychological trauma OR human traffic* OR traumatic experience*).ti,ab,id.
2. posttraumatic stress disorder/ OR complex PTSD/ OR DESNOS/ OR post-traumatic stress/ OR (posttrauma* OR post-trauma* OR psychotrauma* OR psycho-trauma* OR traumatic stress* OR PTSD* OR PTSS* OR stress disorder* OR ((stressor OR trauma) ADJ2 disorder*) OR traumatised youth).ti,ab,id.
3. juvenile justice/ OR juvenile delinquency/ OR (((child* OR adolesc* OR youth* OR juvenile) ADJ3 (delinq* OR crim*)) OR (adolesc* ADJ3 offender*) OR justice involved youth).ti,ab,id.
4. correctional institutions/ OR juvenile justice/ OR prisons/ OR residential care institutions/ OR (residential OR correctional OR prison* OR justice OR institutional care).ti,ab,id. (1 or 2) and 3 and 4
Table A3
Search strategy for the Medline database
1. child abuse/ OR child abuse reporting/ OR child neglect/ OR domestic violence/ OR emotional abuse/ OR incest/ OR physical abuse/ OR rape/ OR sexual abuse/ OR sexual harassment/ OR verbal abuse/ OR abandonment/ OR (child abus* OR child maltreat* OR child mistreatment OR child* welfare recidiv* OR institutionaliz* OR neglect OR out-of-home placement* OR psychological abus* OR sexual abus* OR verbal abus* OR physical abuse OR early adversity OR emotional trauma OR psychological trauma OR human traffic* OR traumatic experience*).ti,ab,kf.
2. posttraumatic stress disorder/ OR early life stress/ OR psychotrauma/ OR psychotrauma assessment/ OR (ptsd* OR ptss* OR post-traumatic stress* OR posttraumatic stress* OR traumatic stress* OR psychotrauma* OR psycho-trauma* OR stress disorder* OR ((stressor OR trauma) ADJ2 disorder*) OR traumatic experience* OR traumatised youth).ti,ab,kf. 3. juvenile delinquency/ OR juvenile justice/ OR juvenile offenders/ OR (juvenile OR
(juvenile* OR delinq* OR resident* care) OR ((child* OR adolesc* OR youth* OR juvenile) ADJ3 (delinq* OR devian* OR crim* OR offend*))).ti,ab,kf.
4. correctional institutions/ OR juvenile justice/ OR prisons/ OR (correctional OR prison* OR justice OR institutional care).ti,ab,kf.
(1 or 2) and 3 and 4
Appendix B Table B1
Coding scheme
Variable Variable labels
Study characteristics
StudyID Study identification number (001, 002, 003, etc.) ESID Effect size identification number (001, 002, 003, etc.)
Authors Author names
Year Publication year
Publication status 1 = published, 0 = not published
N Number of participants in total sample
Age mean Mean age, total sample
AgeMin Minimal age, total sample
AgeMax Maximum age, total sample
AgeSD Standard deviation age, total sample Gender Percentage girls in total sample Interpersonal trauma exposure and PTSD characteristics
Trauma_exposure Percentage interpersonal trauma exposure in total sample Type_trauma 0 = unknown 1 = physical abuse, 2 = neglect, 3 = sexual
abuse, 4 = domestic violence, 5 = verbal abuse, 6 = emotional abuse
Trauma_reported 1 = parent, 2 = juvenile delinquent, 3 = clinical, 4 = case files
PTSD Percentage PTSD in total sample Measurement characteristics
Trauma_measurement 1 = The BRFSS ACEs module, 2 = JVQ, 3 = The ACE study, 4 = MAYSI-2, 5 = CTQ, 6 = PADDI, 7 = AEIII, 8 = case files, 9 = PTSD-RI, 10 = TLEQ, 11 = K-SADS-PL, 12 = TESI, 13 = WSJJCA, 14 = PACT, 15 = DISC-V, 16, = PTSD-I
PTSD_measurement 1= CRIES 2 = CAPS-CA 3 = ADIS 4 = TSCYC, 5= SCID-1, 6= interview, 7= PTSD-RI, 8= self-report, 9= K-SADS, 10= CPTS‐RI, 11= MINI-KID 8, 12= semi-structured interview, 13= DISC-IV
References
Aherns, C. E., Campbell, R., Ternier-Thames, N. K., Wasco, S. M., & Sefl, T. (2007).
Deciding whom to tell: Expectations and outcomes of rape survivors’ first disclosures. Psychology of Women Quarterly, 31, 38-49. doi:10.1111/j.1471- 6402.2007.00329.x Alisic, E., Jongmans, M. J., Van Wesel, F., & Kleber, R. J. (2011). Building child trauma
theory from longitudinal studies: A meta-analysis. Clinical Psychology Review, 31(5), 736-47. doi: 10.1016/j.cpr.2011.03.001
Alisic, E., Zalta, A. K., Van Wesel, F., Larsen, S. E., Hafstad, G. S., Hassanpour, K., & Smid, G. E. (2014). Rates of post-traumatic stress disorder in trauma-exposed children and adolescents: Meta-analysis. The British Journal of Psychiatry, 204, 335–340. doi: 10.1192/bjp.bp.113.131227
Allwood, M. A., Dyl, J., Hunt, J. I., & Spirito, A. (2008). Comorbidity and service utlization among psychiatrically hospitalized adolescents with posttraumatic stress disorder. Journal of Psychological Trauma,7, 104-121.
https://doi.org/10.1080/19322880802231791
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th edition). Washington DC: American Psychiatric Association. Armijo-Olivo, S., Stiles, C. R., Hagen, N. A., Biondo, P. D., & Cummings, G. G. (2010).
Assessment of study quality for systematic reviews: A comparison of the Cochrane Collaboration Risk of Bias Tool and the Effective Public Health Practice Project Quality Assessment Tool: Methodological research. Journal of Evaluation in Clinical Practice, 18(1), 12-18. doi: 10.1111/j.1365-2753.2010.01516.x
Assink, M., & Wibbelink, C. (2016). Fitting three-level meta-analytic models in R: A step-by-step tutorial. The Quantitative Methods for Psychology, 12, 154–174.
Atria. (2019). Comments on ‘The prevalence of domestic violence and child abuse in the Netherlands’. Retrieved, february 2, 2020, from https://rm.coe.int/atria-report-for-grevio-2019/16809386dd
Barendregt, J. J., Doi, S. A., Lee, Y. Y., Norman, R. E., & Vos, T. (2013). Meta-Analysis of Prevalence. Epidemiology and Community Health, 67, 974–978. doi: 10.1136/jech-2013-203104
Bedard-Gilligan, M., Jaeger, J., Echiverri-Cohen, A., Zoellner, L. A. 2012. Individual differences in trauma disclosure. Journal of Behavior Therapy and Experimental Psychiatry, 43(2), 716-723. https://doi.org/10.1016/j.jbtep.2011.10.005
Benedini, K. M., & Fagan, A. A. (2018). A Life-Course Developmental Analysis of the Cycle of Violence. Journal of Developmental and Life-Course Criminology, 4, 1–23.
https://doi.org/10.1007/s40865-017-0073-6
Bernhard, A., Martinelli, A., Ackermann, K., Saure, D., & Freitag, C. M. (2018). Association of trauma, Posttraumatic Stress Disorder and Conduct Disorder: A systematic review and meta-analysis. Neuroscience and Biobehavioral Reviews, 91, 153–169.
https://doi.org/10.1016/j.neubiorev.2016.12.019
Bertram, R. M., & Dartt, J. L. (2009). Post traumatic stress disorder: A diagnosis for youth from violent, impoverished communities. Journal of Child and Family Studies, 18, 294–302. doi:10.1007/s10826-008-9229-7
Borenstein, M. (2009). Introduction to Meta-Analysis. Hoboken, New Jersey: Wiley. Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H.R. (2010). A basic
introduction to fixed-effect and random-effects models for meta-analysis. Research Synthesis Methods, 1, 97–111. doi:10.1002/jrsm.12
Borsboom, D., Cramer, A. O. J., Schmittmann, V. D., Epskamp, S., & Waldorp, L. J. (2011). The Small World of Psychopathology. PLoS ONE, 6(11), e27407.
https://doi.org/10.1371/journal.pone.0027407
Bussey, M. C., & Wise, J. B. (2007). Trauma transformed: An empowerment response. New York: Columbia University Press.
Breslau, N., Lucia, V. C., & Alvarado, G. F. (2006). Intelligence and other predisposing factors in exposure to trauma and posttraumatic stress disorder: A follow-up study at age 17 years. Archives of General Psychiatry, 63(11), 1238–1245. doi:
10.1001/archpsyc.63.11.1238
Breslau, N., Reboussin, B. A., Anthony, J. C., Storr, C. L. (2005). The structure of
posttraumatic stress disorder: Latent class analysis in 2 community samples. Archives of General Psychiatry, 62(12), 1343-51. doi:10.1001/archpsyc.62.12.1343
Bryman, A. (2016). Social Research Methods (5th edition). Oxford: Oxford University Press. Budden, A. (2009). The role of shame in posttraumatic stress disorder: A proposal for a
socio-emotional model for DSM-5. Social Science & Medicine, 69, 1032-1039. doi:10.1016/j.socscimed.2009.07.032
Caffo, E., Forresi, B., & Lievers, L. S. (2005). Impact, psychological sequelae and management of trauma affecting children and adolescents. Current Opinion in Psychiatry, 18, 422–428. doi: 10.1097/01.yco.0000172062.01520.ac
Cauffman, A., Monahan, K. C., & Thomas, A. (2015). Pathways to Persistence: Female Offending From 14 to 25. Journal of Developmental and Life-Course Criminology, 1, 236–268. doi: 10.1007/s40865-015-0016-z
Chartier, M. J., Walker, J. R., & Naimark, B. (2010). Separate and cumulative effects of adverse childhood experiences in predicting adult health and health care utilization. Child Abuse Neglect, 34, 454–464. http://dx.doi.org/10.1016/j.chiabu.2009.09.020
Chauhan, P., Reppucci, N. D., Burnette, M., & Reiner, S. (2014). Race, neighborhood disadvantage, and antisocial behavior among female juvenile offenders. Community Psychology, 38(4), 532-540. https://doi.org/10.1002/jcop.20377
Chen, H., Cohen, P., & Chen, S. (2010). How Big is a Big Odds Ratio? Interpreting the Magnitudes of Odds Ratios in Epidemiological Studies. Communications in Statistics - Simulation and Computation, 39(4), 860-864. doi: 10.1080/03610911003650383 Cheung, M. W. L. (2014). Modeling dependent effect sizes with three-level meta-analyses: A
structural equation modeling approach. Psychological Methods, 19, 211. doi:10.1037/a0032968
Chivers-Wilson, K. A. (2006). Sexual assault and posttraumatic stress disorder: A review of the biological, psychological and sociological factors and treatments. McGill Journal of Medicine, 9(2), 11-118.
Cicchetti, D., & Toth, S. L. (2005). Child maltreatment. Annual Review of Clinical Psychology, 1, 409-438. doi: 10.1146/annurev.clinpsy.1.102803.144029
Cloitre, M., Stolbach, B. C., Herman, J. L., Van der Kolk, B., Pynoos, R., Wang, J., Petkova, E. (2009). A developmental approach to complex PTSD: Childhood and adult
cumulative trauma as predictors of symptom complexity. Journal of Traumatic Stress, 22(5), 399-408. doi: 10.1002/jts.20444
Collins English Dictionary. (z.j.). Dictionary. Retrieved, January 12, 2020, from https://www.collinsdictionary.com/us/dictionary/english/juvenile-offender
Connor, D. F., Ford, J. D., & Albert, D. B. (2007). Conduct disorder subtype and comorbidity. Annals of Clnical Psychiatry, 19, 161-168. doi: 10.1080/10401230701465269
Copeland, W. E., Keeler, G., Angold, A., & Costello, E. J. (2007). Traumatic events and posttraumatic stress in childhood. Archives of General Psychiatry, 64, 577–584. doi:
