Trauma informed interventions: A systematic review

Background Health inequities remain a public health concern. Chronic adversity such as discrimination or racism as trauma may perpetuate health inequities in marginalized populations. There is a growing body of the literature on trauma informed and culturally competent care as essential elements of promoting health equity, yet no prior review has systematically addressed trauma informed interventions. The purpose of this study was to appraise the types, setting, scope, and delivery of trauma informed interventions and associated outcomes. Methods We performed database searches— PubMed, Embase, CINAHL, SCOPUS and PsycINFO—to identify quantitative studies published in English before June 2019. Thirty-two unique studies with one companion article met the eligibility criteria. Results More than half of the 32 studies were randomized controlled trials (n = 19). Thirteen studies were conducted in the United States. Child abuse, domestic violence, or sexual assault were the most common types of trauma addressed (n = 16). While the interventions were largely focused on reducing symptoms of post-traumatic stress disorder (PTSD) (n = 23), depression (n = 16), or anxiety (n = 10), trauma informed interventions were mostly delivered in an outpatient setting (n = 20) by medical professionals (n = 21). Two most frequently used interventions were eye movement desensitization and reprocessing (n = 6) and cognitive behavioral therapy (n = 5). Intervention fidelity was addressed in 16 studies. Trauma informed interventions significantly reduced PTSD symptoms in 11 of 23 studies. Fifteen studies found improvements in three main psychological outcomes including PTSD symptoms (11 of 23), depression (9 of 16), and anxiety (5 of 10). Cognitive behavioral therapy consistently improved a wide range of outcomes including depression, anxiety, emotional dysregulation, interpersonal problems, and risky behaviors (n = 5). Conclusions There is inconsistent evidence to support trauma informed interventions as an effective approach for psychological outcomes. Future trauma informed intervention should be expanded in scope to address a wide range of trauma types such as racism and discrimination. Additionally, a wider range of trauma outcomes should be studied.

consistently improved a wide range of outcomes including depression, anxiety, emotional dysregulation, interpersonal problems, and risky behaviors (n = 5).

Conclusions
There is inconsistent evidence to support trauma informed interventions as an effective approach for psychological outcomes. Future trauma informed intervention should be expanded in scope to address a wide range of trauma types such as racism and discrimination. Additionally, a wider range of trauma outcomes should be studied.

Background
Despite the United States' commitment to health equity, health inequities remain a pressing concern among some of the nation's marginalized populations, such as racial/ethnic or gender minority populations. For example, according to the 2016 National Health and Nutrition Examination Survey (NHANES), 29.1% of Mexican Americans and 24.3% of African Americans with diabetes had hemoglobin A1C greater than 9% (the gold standard of glucose control with levels � 7% deemed adequate), compared to 11% in non-Hispanic whites [1]. The 2016 survey also revealed that 40.9% and 41.5% of Mexican Americans and African Americans with hypertension, respectively, had their blood pressure under control, compared to 51.7% in non-Hispanic whites. In 2014, 83% of all new diagnoses of HIV infection in the United States occurred among gay, bisexual, and other men who have sex with men, with African American men having the highest rates [2].
Several factors have been discussed as root causes of health inequities. For example, Farmer et al. [3] noted structural violence-the disadvantage and suffering that stems from the creation and perpetuation of structures, policies and institutional practices that are innately unjust -as a major determinant of health inequities. According to Farmer et al., because systemic exclusion and disadvantage are built into everyday social patterns and institutional processes, structural violence creates the conditions which sustain the proliferation of health and social inequities. For example, a recent analysis [4] using a sample including 4,515 National Health and Nutrition Examination Survey participants between 35 and 64 years of age revealed that black men and women had fewer years of education, were less likely to have health insurance, and had higher allostatic load (i.e., accumulation of physiological perturbations as a result of repeated or chronic stressors such as daily racial discrimination) compared to white men (2.5 vs 2.1, p<.01) and women (2.6 vs 1.9, p<.01). In the analysis, allostatic load burden was associated with higher cardiovascular and diabetes-related mortality among blacks, independent of socioeconomic status and health behaviors.
Browne et al. [5] identified essential elements of promoting health equity in marginalized populations such as trauma-informed and culturally competent care. In particular, traumainformed care is increasingly getting closer attention and has been studied in a variety of contexts such as addiction treatment [6][7][8] and inpatient psychiatric care [9]. While there is a growing body of the literature on trauma-informed care, no prior review has systematically addressed trauma-informed interventions; one published review of literature [10] limited its scope to trauma survivors in physical healthcare settings. As such, the purpose of this paper is to conduct a systematic review and synthesize evidence on trauma-informed interventions.
For the purpose of this paper, we defined trauma as physical and psychological experiences that are distressing, emotionally painful, and stressful and can result from "an event, series of events, or set of circumstances" such as a natural disaster, physical or sexual abuse, or chronic articles and 8,864 studies were forwarded to the title and abstract screening stage. Eight thousand five hundred and twenty-one studies were excluded because they were irrelevant. Three hundred and forty-three abstracts were identified to be read fully. Following this, 311 articles were excluded for focusing on other psychological conditions (n = 120), were non-experimental studies (n = 78) and were in inpatient or incarcerated populations (n = 46). One additional companion article was identified during full text review. Therefore, thirty-three articles met the inclusion criteria and are reported in this review. Fig 1 provides details of the selection process and identifies the reasons why articles were excluded at each stage.

Characteristics of trauma-informed interventions
Type of intervention. Table 5 details the trauma informed intervention characteristics included in this review. The two most frequently used interventions were eye movement desensitization and reprocessing (EMDR) [28,30,31,33,36,38]-a multi-phase intervention using bilateral stimulation, such as left-to-right eyes movements or hand tapping, to desensitize individuals to a traumatic memory or image-and trauma-focused cognitive behavioral therapy or cognitive behavioral therapy (CBT) [26,27,32,46,48]-a psychological approach to introduce emotional regulation and coping strategies (e.g., deep muscle relaxation, yoga, thought discovery and breathing techniques) to deal with negative feelings and behaviors surrounding a trauma of interest [32,48]. The implementation of CBT varied on the trauma of interest. Other studies implemented interventions using general trauma focused therapy [22,43], emotion focused therapy [23,25], stress reduction programs [17], cognitive processing therapy [24], brief electric psychotherapy [31], present focused group therapy [26], compassion focused therapy [44], prolonged exposure [45], stress inoculation training [45], psychodynamic therapy [45], and visual schema displacement therapy [30]. A number of studies included more than one of these therapies [13,26,30,31,33,36,45].

PLOS ONE
or therapists rate therapy sessions [26,34,45]. Finally, one study had quality assurance worksheets completed after each session that were later reviewed by the study coordinator [34].
A select number of the studies found associations between trauma-informed interventions and other psychological outcomes such as attachment anxiety, attachment avoidance, psychiatric symptoms or dental distress. For example, the trauma-informed mindfulness-based reduction program implemented by Kelly and colleagues was associated with a greater decrease in anxious attachment, measured by the Relationship Structures Questionnaire, compared to the waitlist group [17]. Similarly, Masin-Moyer and colleagues found that TREM and an attachment-informed TREM (ATREM) were associated with significant reductions in group attachment anxiety, group attachment avoidance, and psychological distress in women  Brief post-session survey completed at end of every session; One randomly selected session for each group rated on the post-session questionnaire by 2 raters who were kept blind to condition PFGT had greater advantage than TFGT in total HIV risk reduction (p = .05); but all three groups had significant reduction in total HIV risk scores overtime. Both TFGT and PFGT had an advantage on PTSD severity compared to waitlist condition (p<.05); but all three groups showed significant reduction in PTSD severity over time. TFGT had a significantly greater reduction in anger/irritability compared with PFGT (p<.01).

Trauma informed interventions
Dalton ( Therapists rated their overall treatments using 9 points scales from 1 (extremely uncharacteristic) to 9 (extremely characteristic) via the Psychotherapy Process Q Set Reduced subjective PTSD symptoms but showed no change in subjective dissociation, depression, anxiety, or interpersonal sensitivity symptoms after 12 weeks. Greater presence of PDT process was significantly associated with greater reductions in PTSD and depression symptoms (p<.05 for both). Greater presence of SIT process was related to greater reduction in PTSD symptoms (p<.05). Clients showed significant improvement in trauma-related attention bias (p<.01) and anxiety-related attention bias (p<.05) but not in attention bias for neutral words. Greater PDT and greater SIT process were both marginally related to reduced implicit memory for anxiety cues (p<.1 for both). PE process levels were unrelated to any significant change after 12 weeks of treatment.
Decker (2017) Those who received violence related discussion and/or safety resources felt more confident in their providers concern for their safety and ability to respond appropriately to violence. Treatment increased knowledge of violencerelated resources. Close to two thirds (65%) of women reported receiving at last one element of the intervention on their exit survey and reported that clinic base Interpersonal violence assessment was helpful irrespective of past violence history.
Decker ( Therapist received specialized EMDR supervision for the treatment of dental phobia All sessions videotaped. One randomly selected session rated by five different raters The intervention group improved on all outcome variables except for depression. Dental anxiety total score pretreatment to 12 months (d = 3.28) was significant (p<.001). There was continuing decrease of dental anxiety up to 3 months after treatment and plateaued. Significant reduction of PTSD symptoms between baseline and 3 months follow up (at 12 months, difference was no longer significant).
Dutton (2016) [41]/Imagery exposure 8 trauma focused sessions and 2 neutral session. Sessions were 30 min each and included 5-min baseline exposure and five 5-minute exposure trials. The imagery exposure was conducted with standardized imagery scenes and cued the participants to focus on their active responses (e.g., did your breathing or heart rate change?)

Not addressed
Mean responses to script-driven imagery scale scores following the first exposure trial were > zero (p<0.001), and symptom ratings decreased significantly across exposure trials (p = 0.001). Past month CAPS score significantly predicted responses to the first trauma script presentation (p<0.001).

Ford (2018)[27]/CBT and Trauma Affect Regulation: Guide
for Education and Therapy (TARGET) 8 sessions of manualized internet-supported CBT for problem drinking with or without trauma-focused emotional regulation skills University of Connecticut counseling center PhD clinical psychology students received training (10 hours) to conduct both therapies and were randomly assigned to participants. Each therapist conducted at least 5 cases of each therapy modality First author reviewed therapist's first two cases and 33% of the sessions following (randomly chosen). Fidelity was achieved on 100% of all items in all sessions in both therapies Both treatments showed significant reduction in days of alcohol use in the past month (p = .006); days of impairment due to alcohol use were reduced at post-treatment and follow-up only for the CBT+TARGET group but the base rate was very low (approximately 1.25 days in the past month) and the change for both groups was not statistically significant.
Gawande ( Outpatient treatment unit 2 female group leaders did all group sessions in all 10 groups together Not addressed No group differences in psychological and PTSD symptoms and sense of coherence. Significant reductions for the study group in the total symptom score and in 8 of 9 scales of Global Severity Index (p<.05); reductions for the short-term group in 4 of 9 subscales (p<.05); and no differences for the wait-list group. A PTSD reduction for the study group, from 87% to 40% (p<.01) but not for the waiting-list group. An increase in sense of coherence for both groups (10-point and 7-point, respectively; p<.05).
(Continued ) Outpatient treatment unit 2 female group leaders did all group sessions in all 10 groups together Not addressed No group differences in psychological and PTSD symptoms and sense of coherence. Significant reductions for the study group in the total symptom score and in 8 of 9 scales of Global Severity Index (p<.05); reductions for the short-term group in 4 of 9 subscales (p<.05); and no differences for the wait-list group. A PTSD reduction for the study group, from 87% to 40% (p<.01) but not for the waiting-list group. An increase in sense of coherence for both groups (10-point and 7-point, respectively; p<.05).
Lundqvist, (2009) [43]/Trauma focused group therapy 2-year long trauma focused group therapy. 46 sessions total with phase 1 containing 22 weekly sessions over 5 months, phase 2 containing 15 weekly sessions over 4 months, and phase 3 comprising 9 sessions over 1 year. Sessions were designed to help women tell their childhood sexual-abuse narratives and to discuss relationships within the family. Each participant was the central narrator in 3 sessions during which she could tell the others about sexual details in abuse, feelings of shame, and feelings of guilt.
Outpatient treatment setting First author was group leader (faculty of Heath and Society at Malmo University) but second group leader was not specified. Both leaders were female Not addressed Levels of social interaction significantly improved, with most evident improvements in total score and adequacy of social integration. The effect size values were .55 and .64, respectively. Social adjustment was significantly improved particularly in subscale of work/studies and homework.
Effect size values were .53 and .56, respectively. No significant changes in family climate except for the expressed emotion subscale perceived criticism in relation to the partner that showed a reduction.
MacIntosh There was significant reduction in Inventory of Interpersonal Problems scores from pre to post treatment, suggesting lower levels of interpersonal problems (p = .002).
There was significant reduction in the mean levels of trauma symptoms reported by participants from pre to post treatment (p = .004).
ATREM had the same 16 topics as TREM but also had 3 open weeks to add new attachment information involving imagery, arts, fables, group meditation, transitional objects, body tapping and written and verbal feedback. Open weeks were to integrate more processing by pausing content and initiating in-the-moment exploration of relational dynamics and facilitating dyadic and group connections.
Therapy took place at: an outpatient behavioral health facility, a residential substance use treatment, and an outpatient victim services agency First-author was TREM trained and trained all other clinicians. Each group had at least 1 licensed masters level social worker or counselor. All facilitators participated in training prior to intervention implementation A facilitator report fidelity checklist was created by 1st author to ensure weekly discussion questions and activities in the TREM manual were addressed Pre and post intervention results showed statistically significant reductions in individual and group attachment anxiety (p = .03), group attachment avoidance (p<.001), perceived social support (p = .002), emotional regulation capacities (p<.001), psychological distress, depression, anxiety, and PTSD symptom severity (p<.001) for ATREM and TREM. ATREM associated with statistically significant reductions in individual attachment avoidance.
Lundqvist, (2009) [43]/Trauma focused group therapy 2-year long trauma focused group therapy. 46 sessions total with phase 1 containing 22 weekly sessions over 5 months, phase 2 containing 15 weekly sessions over 4 months, and phase 3 comprising 9 sessions over 1 year. Sessions were designed to help women tell their childhood sexual-abuse narratives and to discuss relationships within the family. Each participant was the central narrator in 3 sessions during which she could tell the others about sexual details in abuse, feelings of shame, and feelings of guilt.
Outpatient treatment setting First author was group leader (faculty of Heath and Society at Malmo University) but second group leader was not specified. Both leaders were female Not addressed Levels of social interaction significantly improved, with most evident improvements in total score and adequacy of social integration. The effect size values were .55 and .64, respectively. Social adjustment was significantly improved particularly in subscale of work/studies and homework.
Effect size values were .53 and .56, respectively. No significant changes in family climate except for the expressed emotion subscale perceived criticism in relation to the partner that showed a reduction.
(Continued )  There was significant reduction in Inventory of Interpersonal Problems scores from pre to post treatment, suggesting lower levels of interpersonal problems (p = .002).
There was significant reduction in the mean levels of trauma symptoms reported by participants from pre to post treatment (p = .004).
ATREM had the same 16 topics as TREM but also had 3 open weeks to add new attachment information involving imagery, arts, fables, group meditation, transitional objects, body tapping and written and verbal feedback. Open weeks were to integrate more processing by pausing content and initiating in-the-moment exploration of relational dynamics and facilitating dyadic and group connections.
Therapy took place at: an outpatient behavioral health facility, a residential substance use treatment, and an outpatient victim services agency First-author was TREM trained and trained all other clinicians. Each group had at least 1 licensed masters level social worker or counselor. All facilitators participated in training prior to intervention implementation A facilitator report fidelity checklist was created by 1st author to ensure weekly discussion questions and activities in the TREM manual were addressed Pre and post intervention results showed statistically significant reductions in individual and group attachment anxiety (p = .03), group attachment avoidance (p<.001), perceived social support (p = . There was significant reduction in Inventory of Interpersonal Problems scores from pre to post treatment, suggesting lower levels of interpersonal problems (p = .002).
There was significant reduction in the mean levels of trauma symptoms reported by participants from pre to post treatment (p = .004). Fidelity checks were based upon video recordings that were carried out on a pilot sample to ensure the procedure was carried out properly In experiment 1, VSDT emotionality scores were higher than EMDR (p<.001) and the control (p<.001), VSDT and EMDR vividness scores were no different (p = 1.00), VSDT vividness scores were higher than the control (p = .02) and EMDR emotionality and vividness scores were higher than the control (p = .02, p = .01). In experiment 2, VSDT emotionality scores were higher than EMDR and the control (p = .001, p<.001). There was no difference in emotionality score between EMDR and the control (p = .08). There were no differences in vividness score between EMDR and the control (p = .83) and between VSDT and EMDR (p = 1.00). The VSDT vividness score was higher than the control (p = .01).

Masin
Nijdam (2012) [33]/Brief eclectic psychotherapy and EMDR Weekly sessions were applied according to the Dutch treatment manual. Sessions lasted 1.5 hrs. EMDR therapy consisted of identification and processing of distressing images of the traumatic events. After patient focused on image with corresponding negative cognition, the patient was asked to follow the therapist's finger making saccadic movements in alternation with the patient's own associations. Distress was measured every 5-10 minutes, until distress level was 0 or 1 and then more positive cognition was introduced as it related to target image. Procedure was repeated for other distressing images and treatment sessions were terminated when trauma memory felt neutral.
Outpatient setting 38 psychiatry residents or master's level clinical psychologists. Therapists received a 3-day level-I training for EMDR and for brief electric psychotherapy. Therapists received biweekly group supervision. All sessions audiotaped. Treatment adherence protocols developed to rate 6 brief eclectic psychotherapy sessions and three EMDR sessions using an EMDR Fidelity Scale adapted for use with the Dutch EMDR protocol. Significant, small-to medium-sized improvements in verbal memory, information processing speed, and executive functioning were found after trauma-focused psychotherapy (Cohen's d = 0.16-0.68). No differences emerged between treatment conditions. Greater PTSD symptom decrease was related to better post-treatment neurocognitive performance (all p<.005). Patients with comorbid depression improved more than patients with PTSD alone on interference tasks (p<.01).
(Continued ) with a history of interpersonal trauma [47]. Additionally, individuals in an outpatient substance abuse treatment program, consisting of psychoeducational seminars and traumainformed addiction treatment, experienced significantly better outcomes of psychiatric severity, measured by the Global Appraisal of Individual Needs scale, compared to a control treatment group [19]. Doering and colleagues found that EMDR, compared to the control group, was associated with significantly greater improvement in dental stress, anxiety and fear in patients with dental-phobia [28].
There was a series of interpersonal, emotional and behavioral outcomes assessed in the included studies. For example, adult females that were sexually abused in childhood experienced a significant improvement in social interaction and social adjustment after receiving trauma focused group therapy [43]. Similarly, Dalton and colleagues found that couples that received emotion focused therapy experienced a significant reduction in relationship distress [23] and MacIntosh and colleagues found that individuals that received CBT reported lower interpersonal problems post-treatment [46]. Trauma-based interventions were also associated with emotional outcomes. Visual schema displacement therapy and EMDR both were superior to the control treatment in reducing emotional disturbance and vividness of negative memories [30]. In a separate study, CBT was found to reduce levels of emotional dysregulation in individuals that experienced childhood sexual abuse [46]. Lastly, trauma-informed interventions were associated with behavioral outcomes, including HIV risk reduction [26], decreased days of alcohol use [27], and improvements in avoidance of client condom negotiations, frequency of sex trade under influence of drugs or alcohol, and use of intimate partner violence support [40]. Interventions that were associated with these behavioral outcomes included trauma focused and present focused group therapy [26], CBT [27], and a trauma-informed support, validation, and safety-promotion dialogue intervention [40].

Publication bias
We analyzed three sets of outcome variables for publication bias: PTSD, depression, and anxiety. Based on Begg and Mazumdar test, there was no evidence of publication bias for PTSD (z = 1.55, p = 0.121) and anxiety (z = 0.29, p = 0.769). However, there was some evidence of publication bias for depression (z = 5.19, p<.001). The statistically significant publication bias for depression appears to be mainly due to large effect sizes in Nixon [24] and Bowland [35].

Discussion
According to our database search, this is the first systematic review to critically appraise trauma-informed interventions using a comprehensive definition of trauma. In particular, our definition encompassed both physical and psychological experiences resulting from various circumstances including chronic adversity. Overall, there was inconsistent evidence to suggest trauma informed interventions in addressing psychological outcomes. We found that traumainformed interventions were effective in improving PTSD [17,20,21,24,26,28,34,42,[45][46][47] and anxiety [21,28,35,47,48] in less than half of the studies where these outcomes were included. We also found that depression was improved in less than about two thirds of the studies where the outcome was included [17,18,20,21,24,32,35,47,48]. Although limited in the number of published studies included this review, available evidence consistently supported traumainformed interventions in addressing interpersonal [23,43,46], emotional [30,46], and behavioral outcomes [26,27,40].
In several studies, therapists took on both traditional treatment and research responsibilities (e.g., delivery of the intervention) [20,25,29,32,33,36,40,46,47], yet blinding of those delivering treatment was discussed clearly in only one study [25]. This dual role is likely to have led to the disclosure of group allocation, hence, threatening the internal validity of the results. Future studies should address these issues by calculating proper sample size a priori, using a comparison group, and concealing group assignments.

Review limitations
Several limitations of this review should be noted. First, by using narrowly defined search terms, it is possible that we did not extract all relevant articles in the existing literature. However, to avoid this, we conducted a systematic electronic search using a comprehensive list of MeSH terms, as well as similar keywords, with consultation from an experienced health science librarian. Additionally, we hand searched our reference collections, Second, the trauma informed interventions included in this review were implemented to predominantly address trauma related to sexual or physical abuse among women. Thus, our findings may not be applicable to trauma related to other types of incidence such as chronic adversity (e.g., racism or discrimination). Likewise, there were insufficient studies addressing a wider range of trauma impacts such as emotion regulation, dissociation, revictimization, non-suicidal selfinjury or suicidal attempts, or post-traumatic growth. Future research is warranted to address these broader impacts of trauma. We included only articles written in English; therefore, we limited the generalizability of the findings concerning studies published in non-English languages. Finally, we used arbitrary cutoff scores to categorize studies as low, medium, and high quality (quality ratings of 0-4, 5-8, and 9+ for RCTs and 0-3, 4-6, 7+ for quasi-experimental studies, respectively). Using this approach, each quality-rating item was equally weighted. However, certain factors (e.g., randomization method) may contribute to the study quality more so than others.

Conclusions
Our review of 33 articles shows that there is inconsistent evidence to support trauma informed interventions as an effective intervention approach for psychological outcomes (e.g., PTSD, depression, and anxiety). With growing evidence in health disparities, adopting trauma informed approaches is a growing trend. Our findings suggest the need for more rigorous and continued evaluations of the trauma informed intervention approach and for a wide range of trauma types and populations.