Participants in all seven studies were enrolled after 1995. Combined, the study authors observed 94,103 detained youth in the juvenile justice system. Sample size in individual studies ranged from 264 to 64,329. Latinx youth represented 9% to 30% of the samples. Of the seven studies that met inclusion criteria, youth represented three different intercepts of the juvenile justice system: (a) pretrial detention (n = 5), incarcerated (n = 1), and repeat offenders (n = 1). All studies included a mix of ethnicities/races (i.e., Black, White, and Latinx) and genders; males were a majority in each study. Table 1 is a summary of all seven studies. (Insert Table 1).
Measurement of Trauma Experience
There were differences in the number and types of traumatic experiences that the youth were asked about, which precluded allowing direct data comparisons across studies. At the lower end, four studies queried for eight or fewer traumatic experiences (Abram et al., 2004; Ford et al., 2008; Lau et al., 2015; Rawal et al., 2004) versus three that queried for 10 or more (Baglivio & Epps, 2016; Ford et al., 2013; King et al., 2011; min = 3 [Rawal et al., 2004], max = 21 [Ford et al., 2013]). Even when two articles (Ford et al., 2008; Ford et al., 2013) used the same measure (i.e., the Traumatic Events Screening Inventory) to query trauma exposure, the difference between the number of traumatic exposures queried was 14. A traumatic exposure that was not commonly assessed for was traumatic separation or loss due to immigration (Abram et al., 2004; Baglivio & Epps, 2016; Ford et al., 2008; Lau et al., 2015; Rawal et al., 2004).
Measurement of PTSD
Most of the studies varied in their measurement of PTSD. Only two studies compared PTSD symptom severity across races/ethnicities (Abram et al., 2004; Ford et al., 2008), and only one study compared PTSD severity by Latinx gender (Abram et al., 2004). Both studies used a standardized and highly accepted measure of PTSD that elicited Diagnostic and Statistical Manual of Mental Disorders (DSM) criteria.
The responses to trauma were collected in three ways. In all but one study (Rawal et al., 2004), interviewers were used to elicit self-report from youth; unfortunately, there is a reporting bias inherent in self-report, In four studies (Abram et al., 2004; Ford et al., 2008; Ford et al., 2013; King et al., 2011), measures were delivered by individuals with counseling degrees, who presumably would be more sensitive to assessing traumatic experiences from youth than the officers from the court used to assess trauma exposure in two other studies (Baglivio & Epps, 2016; Lau et al., 2015). Baglivio and Epps (2016) used trauma screening as part of a risk assessment done by an officer of the court to predict the likelihood that the youth would reoffend, and there was no mention of whether the youth were blinded to this, which would further create response bias because youth are less likely to report traumatic history when doing so puts them at risk to reoffend in the rater’s perspective (van Batenburg-Eddes et al., 2012). Lastly, two studies reviewed prior court documentation on the youth (e.g., Child Protective Services reports) to assess past abuse/maltreatment; King et al. (2011) used this as a supplement to court documentation, and Rawal et al. (2004) used it by itself. King et al. (2011) noted that if there was a discrepancy between the court report and the self-report, they would keep the response that indicated trauma exposure. Allowing the researcher to choose the desired response is also highly problematic.
Type of Studies
I evaluated the studies in this systematic review to determine the strength of the findings. The quality of the study designs was mixed. Of the seven studies, three (Abram et al., 2004; Lau et al., 2015; Rawal et al., 2004) stratified their samples to be representative of the state’s juvenile justice sample where the research took place in terms of age, sex, race/ethnicity, and legal status. Baglivio and Epps (2016) and Lau et al. (2015) used secondary databases containing samples of youth in detention throughout the entire state, allowing for much larger sample sizes. However, these studies are problematic in that Latinx youth were consistently underrepresented in the study samples. In the same two studies, the authors also reported attrition; 22% in Baglivio and Epps and 5% in Lau et al. King et al. (2011) reported that after reviewing the characteristics of the participants they lost, they determined that these participants were more likely to be male or Latinx than the participants who were sampled.
Most of the studies used count data to summarize cumulative traumatic responses. The number of possible count outcomes ranged from 0 to 19; only Abram et al. (2004) used a count regression model to analyze count data. Ford et al. (2013) used an exploratory latent class analysis to determine the number of latent classes of polyvictimization and subsequent PTSD, while two studies used logistic regression to compare rates of abuse by demographic characteristics (Ford et al., 2008; King et al., 2011). Rawal et al. (2004) used analysis of variance to detect and report differences chi-square tests to report differences, and Baglivio and Epps used prevalence rates.
I examined seven studies that assessed either the number of traumatic experiences or PTSD severity by ethnicity/race or Latinx gender. I also investigated traumatic loss by ethnicity/race and sex abuse by Latinx gender. The studies varied in that not all allowed comparisons by trauma experience and PTSD severity. Table 2 shows the effect sizes of traumatic exposure, PTSD severity, traumatic loss, and sex abuse with odds ratios and 95% confidence intervals. (Insert Table 2).
Trauma Exposure by Race/Ethnicity
There were mixed findings in terms of rates of trauma exposure by race/ethnicity. In Abram et al. (2004), Latinx youth reported a mean number of traumatic experiences that was higher than that of White youth and comparable to that of Black youth. In Ford et al. (2013), polyvictimization and subsequent PTSD were combined using exploratory latent class analysis, which created a high polyvictimization class (M = 11.37, SD = 1.11; trauma types), relatively moderate adversity (M = 8.9, SD = 0.34; trauma types), and low (M = 7.39, SD = 0.42; trauma types). Latinx youth (n = 18, 3%) were the lowest proportion of youth in the highest polyvictimization class (White = 5.9%, Black = 5.4%); however, in the same study, the proportion of Latinx youth that made up the other two subcategories was similar to White and Black youth at the moderate and low categories of polyvictimization. Baglivio and Epps (2016) and Lau et al. (2015) found that Latinx youth reported the lowest number of cumulative trauma exposures. King et al. (2011) reported that Latinx youth had trauma rates similar to those of Black youth but less than those of White youth. One study with a low N was inconclusive (Rawal et al., 2004). Lastly, Ford et al. (2008) indicated that Latinx youth, when compared to Black and White youth, were more likely to report both traumatic loss and community violence.
Trauma Exposure by Latinx Gender
Trauma exposure by Latinx gender was also mixed. Abram et al. (2004) found that a higher proportion of Latinx males than females reported experiencing trauma. King et al. (2011) found that, on every construct of types of abuse (i.e., physical and sexual abuse), a higher proportion of Latinx females disclosed every type compared to Latinx males. Abram et al. allowed analysis by trauma type and reported that a higher proportion of Latinx males than Latinx females disclosed two items: “witnessing or experiencing someone get hurt badly or killed” and “being threatened by a weapon.” Abram et al. also found that a higher proportion of Latinx females than males disclosed “forced to do something sexual.”
PTSD by Race/Ethnicity
Abram et al. (2004) and Ford et al. (2008) reported PTSD symptoms by ethnicity/race. Abram et al. found that Latinx youth disclosed clinically significant PTSD symptoms at a statistically significantly higher rate than did White and Black youth. Ford et al. reported a similar trend, while not statistically significant, with Latinx youth reporting more severe PTSD symptoms.
PTSD by Latinx Gender
Abram et al. (2004) were the only researchers to compare PTSD symptoms by Latinx gender. They found that a higher proportion of Latinx males were diagnosed with clinically significant PTSD than were Latinx females.