Predictors of Treatment Completion in a Sample of Youth Who Have Experienced Physical or Sexual Trauma

Abstract

Despite significant advances in knowledge and availability of evidence-based models for child traumatic stress, many children simply do not complete treatment. There remain notable gaps in the services research literature about treatment completion among youth, particularly those who have experienced trauma and related sequelae. This study investigated the linkages among child physical and sexual trauma, posttraumatic stress disorder (PTSD) symptomatology, and treatment completion utilizing a clinical sample drawn from a large database from community treatment centers across the United States specializing in childhood trauma. Results from regression analyses indicated that neither the experience of sexual nor physical trauma directly predicted successful treatment completion. The links between sexual trauma and treatment completion, however, were mediated by PTSD avoidance symptoms. Children and youth experiencing sexual trauma reported higher levels of avoidance symptoms that were, in turn, significantly associated with a lower likelihood of completing trauma-focused mental health treatment. Practice implications are discussed and include strategies for clinicians to intervene during pivotal points of treatment to improve rates of service utilization and treatment completion.

Keywords

posttraumatic stress psychotherapy children and adolescents attrition

Despite multiple years in which physical and sexual maltreatment and assault of children have decreased in the United States, national survey data have continued to indicate notable 1-year prevalence rates of 4.2% for physical abuse, 49.6% for any type of physical assault, 4.2% for sexual abuse, and 3.2% for sexual assault (Turner, Finkelhor, & Ormrod, 2010; U.S. Department of Health and Human Services Administration for Children and Families, 2010). Maltreatment portends a markedly heightened risk of mental health problems that often endure well into adulthood (Briere & Elliott, 2003; Stroud & Petersen, 2012) with more than one third of maltreated youth experiencing clinical symptoms of posttraumatic stress disorder (PTSD; Kerig, Ward, Vanderzee, & Arnzen Moeddel, 2009). Relative to other children requiring mental health services, those who have experienced trauma receive fewer referrals and less treatment (Cooper, Masi, Debabnah, Aratani, & Knitzer, 2007; Palusci & Ondersma, 2012). Consistent with the definition of a traumatic event as causing an overwhelming threat or fear for oneself or a loved one, it may be the experience of posttraumatic symptoms (e.g., hyperarousal, avoidance) rather than the event, per se, that impedes treatment completion (Kerig et al., 2009).

The proliferation of effective trauma-focused treatments (Cohen, Mannarino, & Deblinger, 2006; Forbes et al., 2010; Lieberman & Van Horn, 2008; Pynoos et al., 2008; Runyon, Deblinger, & Schroeder, 2009; Saunders, Berliner, & Hanson, 2004) notwithstanding, service utilization studies generally delineate a pattern of low rates of treatment completion and seldom have focused on child traumatic stress. An estimated 40% to 60% of children who begin psychotherapy discontinue care prior to completion and likely receive diminished benefit from a partial course of an evidence-based treatment (EBT; Kazdin, 1996, 2011). Low rates of service utilization have been reported for children with disruptive behavior disorders (Kazdin, Mazurick, & Siegel, 1994), mixed internalizing and externalizing child psychiatric disorders (Gonzalez, Weersing, Warnick, Scahill, & Woolston, 2011; Warnick, Gonzalez, Weersing, Scahill, & Woolston, 2012), and psychosocial stressors and impairment (Miller, Southam-Gerow, & Allin, 2008), as well as among community samples of youth and adults experiencing different forms of adversity (Amaya-Jackson et al., 1999; Harpaz-Rotem, Murphy, Rosenheck, Berkowitz, & Marans, 2007; Miller et al., 2008; Murphy, Kasprow, & Rosenheck, 2005; Tsai, Edens, & Rosenheck, 2011).

Several factors have been highlighted in the empirical literature that affect whether children and families utilize a full course of mental health services. Although not consistent across all studies, perhaps due in part to the heterogeneity within and across samples, factors that have been identified include child age, sex, race and ethnicity, socioeconomic status, extent of mental health and medical problems, comorbid diagnoses, single-parent status, and quality of parental discipline (Arnow et al., 2007; Kazdin, 1996; Saxe, Ellis, & Kaplow, 2007). Several studies, albeit inconsistently, have reported a connection between PTSD symptom severity and treatment attrition and outcome (Lewis et al., 2010; Zayfert et al., 2005).

Youth affected by trauma experience many of the same risks and challenges that interfere with service participation and completion for youth with other psychiatric concerns (Costello, Erkanli, Fairbank, & Angold, 2002). Cohen and colleagues, in a study of childhood traumatic grief, observed that only 65% of children referred for trauma treatment completed all sessions (Cohen, Mannarino, & Knudsen, 2004). Less is known, however, about the barriers to treatment completion that may be unique to this population. Little is known about how abuse histories influence treatment completion (Garcia-Rea & LePage, 2010), despite the strong empirical link between abuse and mental health problems. Although research is notably limited on predictors of treatment completion among youth who have experienced trauma, demographic characteristics (Costello et al., 2002; Last & Perrin, 1993; Tolin & Foa, 2006) and the nature of traumatic events, such as repeated exposure, acts that involve severe interpersonal aggression, or caregivers as perpetrators (Luthra et al., 2009), are known to predict PTSD symptom presentation and severity. We hypothesize that these factors may in turn, relate to treatment completion.

In sum, in community care, as well as in well-controlled research studies, low rates of service utilization and treatment completion pose significant barriers to achieving desired outcomes. Youth likely derive greatest benefit from a completed course of treatment, whereas incomplete treatment might dilute benefits or exacerbate symptomatology. Children who do not complete treatment may not sufficiently learn critical skills for coping with intrusive and hyperarousal symptoms, potentially leaving them in a heightened state of distress or avoidance. Limited research in this area suggests that demographic characteristics, the type of traumatic experience, and PTSD symptomatology may directly and indirectly affect treatment completion. No known study has heretofore examined these issues simultaneously among youth who have experienced trauma, yet understanding these issues is critical to identifying youth who are less likely to complete a full course of well-validated EBTs. Such knowledge has the potential to inform mental health care, including targeted engagement strategies, to enhance treatment completion.

The Current Study and Hypothesis

This study examines the direct and indirect associations among physical and sexual trauma, child PTSD symptomatology (i.e., overall severity, as well as re-experiencing, avoidance, and hyperarousal symptom clusters), and treatment completion, controlling for demographic variables (i.e., gender, race, ethnicity) and treatment site, in a large, clinical sample of youth receiving evidence-based trauma treatment in community settings. The primary hypothesis was that PTSD symptomatology would mediate the well-known relationship of physical and sexual trauma with lower rates of treatment completion.

Method

Participants

Participants were selected from the National Child Traumatic Stress Network (NCTSN) Core Data Set (CDS) which includes youth from birth to 21 years of age, who received assessment and treatment services between 2004 and 2010 at one of 56 NCTSN community-based sites funded by the Substance Abuse and Mental Health Services Administration (SAMHSA) to serve youth affected by a range of traumas, including maltreatment and assault (Briggs et al., 2012). The centers are located across diverse urban, suburban, and rural areas of the United States. The following inclusion criteria were used in this study: (a) Youth in the data set must have received care from a site that submitted 80% or more of expected follow-up data with at least one youth completing treatment as planned (n = 5,028); (b) youth were 7 years or older and had received assessment that included the well-validated University of California Posttraumatic Stress Disorder Reaction Index (PTSD-RI; n = 1,202); and (c) youth had experienced at least one traumatic event (n = 4,692) and participated in a trauma-related treatment program (n = 1,202). Across the CDS, treatment services primarily consisted of cognitive behavioral or attachment-oriented interventions. Sample heterogeneity resulted from the reporting options in the CDS for 15 non-mutually exclusive treatments. Given our focus on service delivery, as opposed to the efficacy of a specific EBT, we selected primary and secondary modalities that would be sufficiently consistent with a trauma-oriented approach (i.e., trauma-focused cognitive behavioral therapy [TF-CBT], non-specific cognitive behavioral therapy [CBT], parent training [excluding Parent–Child Interaction Therapy], relaxation training, narrative therapy, phase-oriented therapy, family therapy, attachment-focused treatment, trauma psychoeducation, and care coordination). The final sample for this study consisted of 928 youth.

Procedures

The University of California Los Angeles (UCLA)–Duke University National Center for Child Traumatic Stress (NCCTS) developed a quality improvement initiative to collect assessment and treatment data from participating NCTSN centers through a standardized protocol. The CDS includes information on client sociodemographics, trauma histories, service utilization, treatment types, functional impairments, symptomatology, and comorbid conditions. Data were collected at clinical intake prior to beginning treatment services and at treatment completion by NCTSN site staff members who received instruction about administering and recording data from clinical protocols and standardized measures.

Measures

Demographics

Data on youth age, sex, race, and ethnicity were drawn from an initial or intake interview prior to beginning any treatment. For the purposes of this study, race and ethnicity variables were categorized as follows: White, Black, and Other or Multiracial, and Hispanic/Latin American or non-Hispanic.

Trauma history and type

Youth history of physical or sexual trauma was drawn from the trauma history profile (THP), which was derived from the trauma history component of the PTSD-RI; (Pynoos, Rodriguez, & Steinberg, 2000). The THP includes a total of 20 types of trauma based on the National Child Abuse and Neglect Data System Glossary (U.S. Department of Health and Human Services Administration for Children and Families, 2000). Providers queried parents/caregivers, school-age and adolescent youth, and other collateral respondents (e.g., caseworkers) to determine whether each type of trauma had occurred. For the current sample, physical trauma was operationalized as either physical abuse by a caregiver or physical assault by a non-caregiver; similarly, sexual trauma consisted of sexual abuse by a caregiver or sexual assault by a non-caregiver. We opted to combine abuse and assault into the composite variables of physical trauma and sexual trauma based on the substantial overlap between youth identified through child protection and legal systems, as well as the ubiquity of multiple forms of victimization. For the set of analyses in this study, physical and sexual trauma variables were coded dichotomously as present or absent (i.e., 1, 0).

PTSD symptoms

Child PTSD symptoms were assessed with the PTSD-RI (Pynoos et al., 2000). In addition to the trauma history items noted earlier, other items address the extent of PTSD symptoms during the past month. Frequency is rated on a Likert-type scale ranging from “none of the time” to “most of the time.” The items map directly onto the Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; American Psychiatric Association [APA], 1994) PTSD symptom cluster criteria of intrusion/re-experiencing (Criterion B), avoidance (Criterion C), and hyperarousal (Criterion D). Twenty items directly assess DSM-IV PTSD symptoms and are reflected in the cluster scores; the total PTSD score includes two additional items that assess associated features of fear of recurrence and trauma-related guilt. The PTSD-RI is a widely disseminated measure that has robust psychometric properties with strong evidence for reliability and validity (Steinberg, Brymer, Decker, & Pynoos, 2004).

Treatment completion

Treatment completion was represented as a binary variable of treatment completed as planned versus incomplete (1, 0) based on follow-up forms completed by clinicians at the NCTSN sites. Participants who for any reason, known or unknown, dropped out, lacked follow-up data, or did not complete treatment were defined as not completing treatment. Owing to the nature of data submission requirements, continuous variables for session number and treatment duration were not available.

Data Analysis

At the outset, the relations between demographics and NCTSN site to mediation outcomes were tested separately. A set of demographic variables was selected for inclusion and represented those with a significant bivariate relationship with at least one of the mediation model outcomes, either the mediator (PTSD symptoms) or treatment completion outcome. Once selected, the set remained constant through the course of the models. Adjustment for potential effects of site involved a random effects model based on site which allowed us to generalize inferences to the entire population rather than restricting inferences to the sites included in our sample. The mean number of prior exposures to traumatic events did not differ between the complete and incomplete treatment groups, so they were not retained in subsequent analyses.

Direct and indirect associations were tested using a series of multiple regression analyses. Direct associations were examined between trauma type (sexual, physical) and treatment completion through logistic regressions, controlling for demographic variables (i.e., age, sex, race, ethnicity). The site variable was not predictive of any of the mediator variables in the bivariate tests and was not retained for the analyses. Next, the roles of overall PTSD symptoms and those reflecting the three symptom clusters (i.e., re-experiencing, avoidance, and hyperarousal) were examined separately as mediators of these associations. As a robust test of mediation, we used the product of the coefficients approach with tests of empirical asymmetric confidence intervals (CIs; MacKinnon, 2008; MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002). This method allows detection of significant mediation when the data do not demonstrate a direct association between the independent and dependent variables. Path coefficients are generated by the regression models to provide a joint estimation of mediating variable effects on the outcome, where Path a is the path between the predictors (sexual and physical trauma) and the mediator (PTSD symptoms), and Path b is the path between the mediator and the outcome (treatment completion). As the design contains a continuous mediator and a dichotomous outcome, the coefficients were adjusted to be on a comparable scale. The test statistic was then computed from the product of these terms and their standard error. The distribution of the test statistic is not expected to be normal, so we computed asymmetric confidence intervals (ACIs) for each test statistic based on the distribution of the product of the two variables. Mediated effects are significant when the ACIs do not cross zero.

Results

Sample Characteristics and Preliminary Analyses

Participants (N = 928, 57% female) had a mean age of 12.1 years and a racial distribution as follows: 58% White, 19% Black, and 24% Other minority or multiracial background. In total, 41% were of Hispanic or Latin American heritage. Table 1 provides descriptive information for key demographic, independent, and dependent variables.

Table 1.

Treatment Completion Status by Demographic and Trauma Exposure Variables.

		No	Yes
Treatment completed as planned	n	M (SD) or n (%)	M (SD) orn (%)	Total M (SD) or n (%)
		513 (56.7%)	415 (43.3%)	928
Age	928	11.9 (3.0)	12.3 (3.1)	12.1 (3.1)
Gender	928
Male		236 (46.0%)	159 (38.3%)	395 (42.6%)
Female		277 (54.0%)	256 (61.7%)	533 (57.4%)
Race	928
White		267 (52.1%)	267 (64.3%)	534 (57.5%)
Black		106 (20.7%)	68 (16.4%)	174 (18.8%)
Other minority or multiracial		140 (27.3%)	80 (19.3%)	220 (23.7%)
Ethnicity	920
Hispanic or Latin American		183 (36.2%)	195 (47.0%)	378 (41.1%)
Non-Hispanic		303 (60.0%)	206 (49.6%)	509 (55.3%)
Unknown		19 (3.8%)	14 (3.4%)	33 (3.6%)
Trauma exposures		4.1 (2.6)	4.0 (2.3)	4.1 (2.5)
Sexual trauma	902
No		272 (54.6%)	226 (55.9%)	498 (55.2%)
Yes		226 (45.4%)	178 (44.1%)	404 (44.8%)
Physical trauma	909
No		273 (54.5%)	233 (57.1%)	506 (55.7%)
Yes		228 (45.5%)	175 (42.9%)	403 (44.3%)
PTSD	910	n = 495	n = 415	910
Re-experiencing (Criterion B)		8.4 (5.5)	8.2 (5.2)	8.3 (5.4)
Avoidance (Criterion C)		10.6 (6.5)	10.0 (6.1)	10.4 (6.3)
Arousal (Criterion D)		10.0 (4.8)	9.8 (4.7)	9.9 (4.8)
PTSD overall	928	n = 513 29.0 (14.9)	n = 415 28.0 (14.2)	928 28.5 (14.6)

Note. PTSD = posttraumatic stress disorder.

Results of independent logistic regressions to test the direct associations between trauma type (physical and sexual) and treatment completion, after controlling for demographics and site, resulted in neither physical nor sexual trauma being directly related to the probability of treatment completion (physical trauma: B = −0.13, SE = −.04, p = .36; sexual trauma: B = −0.12, SE = −.03, p = .42). The independent variables for the mediation analyses, which also controlled for demographics and site, consisted of the two trauma types (physical and sexual trauma) and four possible mediators (total PTSD symptom score and the three cluster scores) in separate equations to avoid issues of multicollinearity. Thus, eight sets of mediation models were constructed (linear regression for Path a, logistic regression for Path b). Regarding indirect associations, physical trauma was associated with hyperarousal (Criterion D), but hyperarousal did not predict treatment completion (see Table 2). As such, the products of the coefficients tests were not examined. Physical trauma was not significantly associated with Criterion B (re-experiencing), C (avoidance), or overall scores. Therefore, we did not proceed with further testing of these potential mediating models.

Table 2.

Summary of Regression Results—Associations Among Physical Trauma, PTSD Hyperarousal Cluster (Criterion D) Score, and Treatment Completion (N = 853).

	Path a (gen. linear regression): Physical trauma → PTSD Criterion D score		Path b (logistic regression): PTSD Criterion D score → Treatment completion
	B (SE)	B	B (SE)	β
Black	0.51 (0.46)	0.04	−0.23 (0.20)	−0.05
Other minority or multiracial	−0.47 (0.43)	−0.04	−0.71 (0.19)*	−0.16
Male	−0.56 (0.33)	−0.06	−0.32 (0.14)*	−0.09
Hispanic or Latin American	−0.12 (0.37)	−0.01	0.57 (0.16)*	0.15
Physical trauma	0.83 (0.33)*	0.09	−0.15 (0.14)	−0.04
PTSD Criterion D score			−0.01 (0.01)	−0.02

Note. PTSD = posttraumatic stress disorder.

p < .05.

For sexual trauma, the product of the coefficients test demonstrated significant mediation (ab = −0.03, CI = [−0.075, −0.003]), such that sexual trauma was significantly associated with higher avoidance symptoms, which in turn, were associated with lower likelihood of treatment completion (see Table 3). Sexual trauma also was associated significantly with overall PTSD symptoms, which in turn were associated with treatment completion at a marginally significant level (see Table 4). The product of the coefficients test demonstrated no significant mediation (ab = −0.02, CI = [−0.101, 0.003]). Sexual trauma was not significantly associated with Criterion B (re-experiencing) or D (hyperarousal). Thus, we did not proceed with further testing of these models.

Table 3.

Summary of Regression Results—Associations Among Sexual Trauma, PTSD Avoidance Cluster (Criterion C) Score, and Treatment Completion (N = 849).

	Path a (gen. linear regression): Sexual trauma → PTSD Criterion C score		Path b (logistic regression): PTSD Criterion C score → Treatment completion
	B (SE)	B	B (SE)	β
Black	0.10 (0.60)	0.01	−0.28 (0.20)	−0.06
Other minority or multiracial	0.31 (0.56)	0.02	−0.71 (0.19)*	−0.16
Male	−0.74 (0.45)	−0.06	−0.38 (0.15)*	−0.10
Hispanic or Latin American	0.65 (0.48)	0.05	0.62 (0.16)*	0.17
Sexual trauma	1.45 (0.45)*	0.11	−0.07 (0.15)	−0.02
PTSD Criterion C score			−0.02 (0.01)*	−0.09

Note. PTSD = posttraumatic stress disorder.

p < .05.

Table 4.

Summary of Regression Results—Associations Among Sexual Trauma, PTSD Overall Score, and Treatment Completion (N = 849).

	Path a (gen. linear regression): Sexual trauma → PTSD overall score		Path b (logistic regression): PTSD overall score → Treatment completion
	B (SE)	B	B (SE)	β
Black	0.47 (1.38)	0.01	−0.32 (0.20)	−0.07
Other minority or multiracial	−0.44 (1.27)	−0.01	−0.80 (0.19)*	−0.19
Male	−2.33 (1.03)*	−0.08	−0.39 (0.15)*	−0.11
Hispanic or Latin American	0.85 (1.10)	0.03	0.65 (0.16)*	0.18
Sexual trauma	2.20 (1.03)*	0.08	−0.11 (0.15)	−0.03
PTSD overall score			−0.01 (0.01)^†	−0.07

Note. PTSD = posttraumatic stress disorder.

†

p < 0.10. *p < .05.

Discussion

Notwithstanding the ubiquity of physical and sexual trauma during childhood and the growing availability of EBTs for children with PTSD symptoms, little is known about the factors that affect treatment completion. Drawing from a data set specifically focused on characteristics of and services provided to youth who have experienced trauma, this study is one of the first to examine the links among child sexual and physical trauma, PTSD symptomatology, and treatment completion. The study is particularly relevant to community mental health providers, given the large representation of community treatment centers specializing in treating symptoms secondary to childhood trauma.

As expected, findings of this study supported the well-established link between trauma and PTSD symptoms. This study found that trauma type (i.e., physical, sexual) was associated differently with PTSD symptom clusters. Children who have experienced physical trauma were more likely to experience hyperarousal symptoms (PTSD Criterion D), whereas those who had experienced sexual trauma were more likely to experience avoidance symptoms (PTSD Criterion C). Neither physical nor sexual trauma directly affected treatment completion.

The primary finding of this study indicated that avoidance symptoms of PTSD mediated the relations between sexual trauma and treatment completion. Youth who have experienced sexual trauma in the form of abuse or assault are prone to avoidance symptoms (e.g., not wanting to think about, talk about, or have feelings about the trauma). They are significantly less likely to complete mental health treatment focused on traumatic stress. A similar pattern approached significance for a mediating effect of total PTSD symptoms; thus, any inference is no more than suppositional. These findings provided partial support for the hypothesis that sexual and physical trauma would be associated with the severity of PTSD symptoms, which in turn would be associated with lower likelihood of treatment completion.

Practice Implications

Results from this study reinforce clinical wisdom that previously has not been examined empirically. Namely, an experience of a potentially traumatic event alone does not account entirely for the problem of low rates of treatment completion. For youth who have experienced trauma and whose symptom pattern is typified by the avoidance of distress and reminders about their sexual trauma, they may find treatment more stressful when the process of gradual exposure brings the trauma and associated symptoms to the fore. This may, in turn, result in premature treatment termination. This pattern seems especially relevant for mental health clinicians treating youth who have experienced sexual trauma and, perhaps, other types of trauma, as they strive to retain youth in treatment to derive maximal benefit. With a fuller understanding of the interplay between abuse and assault, PTSD symptoms, and treatment completion, clinicians may anticipate barriers to retention and respond proactively. Combined with the wider dissemination of EBTs for child traumatic stress, managing treatment engagement and participation may yield higher completion rates and, presumably, better outcomes.

When children’s symptom presentations are consistent with the current findings, clinicians may incorporate an early and continuing emphasis on treatment engagement (e.g., session reminders, follow-up contact), the therapeutic alliance, or use of motivational interviewing approaches (McKay & Bannon, 2004) that can improve treatment completion. Efforts to calibrate exposure and support coping skills may warrant continuous focused attention along with specific psychoeducation about sexual trauma and avoidance, barriers to treatment completion, and effective coping during the intervals between in-person sessions.

Limitations

Several limitations are relevant when interpreting results of this study. Clinicians were asked to determine whether treatment was completed as planned. Although they were trained in the expected data collection and measures administration, the CDS does not include assessment of inter-rater reliability. One could readily envision more precise operational definitions of treatment completion (e.g., number of sessions relative to those indicated in a manualized treatment), yet these were not possible due to variable time frames in follow-up reporting. The same applies to the definition of incomplete treatment where reasons for premature termination are rarely known. The opportunity of this large data set likewise set constraints on our sample selection. The larger NCTSN data set allowed selection of multiple types of treatment that were not mutually exclusive. We opted for a definition that drew in categories that likely reflected relevant EBTs, albeit at the expense of precision about any particular treatment model or its attendant techniques.

With a sample restricted to youth who have experienced physical or sexual trauma, generalization to children affected by other types of trauma would be speculative. Although the approach used herein is consistent with approaches used throughout the research literature on service delivery, it is important to keep in mind that these results do not allow inferences about treatment dosage, fidelity, or outcome. Rather, it represents a descriptive study of a topic that has received little attention in studies related to child traumatic stress. Future research should address these clinically relevant topics.

Sample heterogeneity represented another challenge, as it does for many studies that rely on administrative data sets. To some extent, inclusion and exclusion criteria and statistical control for covariates reduced heterogeneity, yet the study covariates did not account for the full range of contributing factors that has been posited in the extant literature. The method of testing the hypothesized model utilized separate regression analyses. Future research might evaluate relations among all variables of interest simultaneously via structural equation modeling or path analyses to better understand the simultaneous direct and indirect associations of these types of trauma with PTSD symptoms and treatment completion.

Whereas the focus of the current study was on physical and sexual trauma, future research could expand to other trauma types (e.g., domestic violence, neglect) to investigate how the nature of trauma exposure affects symptom presentation and the likelihood of treatment completion, providing clinicians with further opportunities to tailor treatments accordingly. Moreover, future examinations of other presenting symptomatology (e.g., depression, oppositional behavior) may further illuminate knowledge regarding treatment completion among youth who have experienced trauma.

Conclusions

The current study utilized a large, clinical data set representing diverse U.S. community treatment centers that specialize in childhood traumatic stress to highlight the relations among physical and sexual trauma, PTSD symptomatology, and treatment completion. Results of the study yield important implications for reducing alarmingly high rates of treatment attrition, especially for youth affected by sexual trauma. Future studies could address a broader range of traumatic experiences, other clinical presentations, alternate methods of analysis, and the associations of traumatic event, symptoms, and treatment completion with treatment outcome.

Footnotes

Acknowledgements

The authors wish to thank Robert Lee and Carrie Purbeck-Trunzo for their assistance and advice in preparing this manuscript.

Authors’ Note

The views and opinions reflect those of the authors and not necessarily those of Substance Abuse and Mental Health Services Administration (SAMHSA).

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by a grant (5U79SM059467-05 and 3U79SM054284-10S) from the Substance Abuse and Mental Health Services Administration (SAMHSA).

Author Biographies

Robert A. Murphy, PhD, is an associate professor in the Department of Psychiatry and Behavioral Sciences at Duke University School of Medicine. His research interests include services evaluation and intervention for traumatized, violence exposed, and HIV/AIDS-affected youth.

Holli E. Sink, PhD, is a clinical psychologist at the Southmountain Center of Excellence in Evidence-Based Treatment. Her research interests include childhood trauma, youth aggression, bullying, psychopathic traits, and school-based prevention programs.

George S. Ake III, PhD, is an assistant professor in the Department of Psychiatry and Behavioral Sciences at Duke University School of Medicine. His research interests include children and families coping with trauma, domestic violence, and the role of religion.

Karen Appleyard Carmody, PhD, is a clinical associate in the Department of Psychiatry and Behavioral Sciences at Duke University School of Medicine. Her research addresses the correlates and consequences of parenting, developmental processes underlying adaptation to adversity, and empirically based interventions relating to trauma and attachment.

Lisa M. Amaya-Jackson, MD, MPH, is an associate professor in the Department of Psychiatry and Behavioral Sciences at Duke University School of Medicine. Her research interests include trauma and exposure to violence in children and adolescents, epidemiological research on risk factors, protective factors, treatment effects, and service utilization.

Ernestine C. Briggs, PhD, is an assistant professor in the Department of Psychiatry and Behavioral Sciences at Duke University School of Medicine. Her research interests include child traumatic stress, family violence, PTSD, resiliency, mental health disparities, and implementation science.

References

Amaya-Jackson

Davidson

J. R.

Hughes

D. C.

Swartz

Reynolds

George

L. K.

Blazer

D. G.

(1999). Functional impairment and utilization of services associated with posttraumatic stress in the community. Journal of Traumatic Stress, 12, 709-724.

American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author.

Arnow

B. A.

Blase

Manber

Constantino

M. J.

Markowitz

J. C.

Klein

D. N.

Rush

A. J.

(2007). Dropouts versus completers among chronically depressed patients. Journal of Affective Disorders, 97, 197-202.

Briere

Elliott

D. M.

(2003). Prevalence and psychological sequelae of self-reported childhood physical and sexual abuse in a general population sample of men and women. Child Abuse & Neglect, 27, 1205-1222.

Briggs

E. C.

Fairbank

J. A.

Greeson

J. K. P.

Layne

C. M.

Steinberg

A. M.

Amaya-Jackson

L. M.

Pynoos

R. S.

(2013). Links between child and adolescent trauma exposure and service use histories in a national clinic-referred sample. Psychological Trauma: Theory, Research, Practice, and Policy, 5(2), 101-109. doi:10.1037/a0027312

Cohen

J. A.

Mannarino

A. P.

Deblinger

(2006). Treating trauma and traumatic grief in children and adolescents. New York, NY: Guilford Press.

Cohen

J. A.

Mannarino

A. R.

Knudsen

(2004). Treating childhood traumatic grief: A pilot study. Journal of the American Academy of Child & Adolescent Psychiatry, 43, 1225-1233.

Cooper

J. L.

Masi

Debabnah

Aratani

Knitzer

(2007). Strengthening policies to support children, youth, and families who experience trauma. New York, NY: National Center for Children in Poverty, Columbia University.

Costello

E. J.

Erkanli

Fairbank

J. A.

Angold

(2002). The prevalence of potentially traumatic events in childhood and adolescence. Journal of Traumatic Stress, 15, 99-112.

10.

Forbes

Creamer

Bisson

J. I.

Cohen

J. A.

Crow

B. E.

Foa

E. B.

. . .Ursano

R. J.

(2010). A guide to guidelines for the treatment of PTSD and related conditions. Journal of Traumatic Stress, 23, 537-552. doi:10.1002/jts.20565

11.

Garcia-Rea

E. A.

LePage

J. P.

(2010). Effects of childhood abuse on relapse in a recently homeless substance dependent veteran population. Psychological Trauma: Theory, Research, Practice, and Policy, 2, 201-205.

12.

Gonzalez

Weersing

V. R.

Warnick

E. M.

Scahill

L. D.

Woolston

J. L.

(2011). Predictors of treatment attrition among an outpatient clinic sample of youths with clinically signfiicant anxiety. Administration and Policy in Mental Health and Mental Health Services Research, 38, 356-367.

13.

Harpaz-Rotem

Murphy

R. A.

Rosenheck

R. A.

Berkowitz

Marans

(2007). Clinical epidemiology of urban violence: Responding to children exposed to violence in ten communities. Journal of Interpersonal Violence, 22, 1479-1490.

14.

Kazdin

A. E.

(1996). Dropping out of child psychotherapy: Issues for research and implications for practice. Clinical Child Psychology and Psychiatry, 1, 133-156.

15.

Kazdin

A. E.

(2011). Evidence-based treatment research: Advances, limitations, and next steps. American Psychologist, 66, 685-698.

16.

Kazdin

A. E.

Mazurick

J. L.

Siegel

T. C.

(1994). Treatment outcome among children with externalizing disorders who terminate prematurely versus those who complete psychotherapy. Journal of the American Academy of Child & Adolescent Psychiatry, 33, 549-557.

17.

Kerig

P. K.

Ward

R. M.

Vanderzee

K. L.

Arnzen Moeddel

(2009). Posttraumatic stress as a mediator of the relationship between trauma and mental health problems among juvenile delinquents. Journal of Youth and Adolescence, 28, 1214-1225.

18.

Last

C. G.

Perrin

(1993). Anxiety disorders in African-American and white children. Journal of Abnormal Child Psychology, 21, 153-164.

19.

Lewis

C. C.

Simons

A. D.

Nguyen

L. J.

Murakami

J. L.

Reid

M. W.

Silva

S. G.

March

J. S

. (2010). Impact of childhood trauma on treatment outcomes in the Treatment of Adolescents with Depression Study (TADS). Journal of the American Academy of Child and Adolescent Psychiatry, 49, 120-132.

20.

Lieberman

A. F.

Van Horn

(2008). Psychotherapy with infants and young children: Repairing the effects of stress and trauma on early attachment. New York, NY: Guilford Press.

21.

Luthra

Abromovitz

Greenberg

Schoor

Newcorn

Scheidler

. . .Chemtob

(2009). Relationship between type of trauma exposure and postraumatic stress disorder among urban children and adolescents. Journal of Interpersonal Violence, 24, 1919-1927.

22.

MacKinnon

D. P.

(2008). Introduction to statiscal mediation analysis. New York, NY: Erlbaum.

23.

MacKinnon

D. P.

Lockwood

C. M.

Hoffman

J. M.

West

S. G.

Sheets

(2002). A comparison of methods to test mediation and other intervening variable effects. Psychological Methods, 7, 83-104. doi:10.1037/1082-989X.7.1.83

24.

McKay

Bannon

(2004). Evidence update: Engaging families in child mental health services. Child and Adolescent Psychiatric Clinics of North America, 40, 1-17.

25.

Miller

L. M.

Southam-Gerow

M. A.

Allin

R. B.

Jr. (2008). Who stays in treatment? Child and family predictors of youth client retention in a public mental health agency. Child Youth Care Forum, 37, 153-170.

26.

Murphy

R. A.

Kasprow

W. J.

Rosenheck

R. A.

(2005). Predictors of childlren’s involvement in parents’ treatment among homeless veterans in community residential care. Psychiatric Services, 56, 1147-1149.

27.

Palusci

V. J.

Ondersma

S. J.

(2012). Services and recurrence after psychological maltreatment confirmed by child protective services. Child Maltreatment, 17, 153-163.

28.

Pynoos

R. S.

Fairbank

J. A.

Steinberg

A. M.

Amaya-Jackson

Gerrity

Mount

M. L.

Maze

(2008). The national child traumatic stress network: Collaborating to improve the standard of care. Professional Psychology: Research and Practice, 39(4), 389-395. doi:10.1037/a0012551

29.

Pynoos

R. S.

Rodriguez

Steinberg

(2000). PTSD Index for DSM-IV. Los Angeles: University of California Los Angeles.

30.

Runyon

M. K.

Deblinger

Schroeder

C. M.

(2009). Pilot evaluation of outcomes of combined parent-child cognitive behavioral group therapy for children at risk for child physical abuse. Cognitive and Behavioral Practice, 16, 101-118.

31.

Saunders

B. E.

Berliner

Hanson

R. F.

(Eds.). (2004). Child physical and sexual abuse: Guidelines for treatment (Revised Report). Charleston, SC: National Crime Victims Research and Treatment Center.

32.

Saxe

G. N.

Ellis

B. H.

Kaplow

J. B.

(2007). Collaborative treatment of traumatized children and teens: The trauma systems therapy approach. New York, NY: Guilford Press.

33.

Steinberg

A. M.

Brymer

M. J.

Decker

K. B.

Pynoos

R. S.

(2004). The University of California at Los Angeles Post-traumatic Stress Disorder Reaction Index. Current Psychiatry Reports, 6, 96-100.

34.

Stroud

Petersen

A. C.

(2012). Child maltreatment research, policy, and practice for the next decade. Archives of Pediatric & Adolescent Medicine, 166, 506-508.

35.

Tolin

D. F.

Foa

E. B.

(2006). Sex differences in trauma and posttraumatic stress disorder: A quantitative review of 25 years of research. Psychological Bulletin, 132, 959-992.

36.

Tsai

Edens

E. L.

Rosenheck

R. A.

(2011). A typology of childhood problems among chronically homeless adults and its association with housing and clinical outcomes. Journal of Health Care for the Poor and Underserved, 22, 853-870. doi:10.1353/hpu.2011.0081

37.

Turner

H. A.

Finkelhor

Ormrod

(2010). Poly-victimization in a national sample of children and youth. American Journal of Preventive Medicine, 38, 323-330. doi:10.1016/j.amepre.2009.11.012

38.

U.S. Department of Health and Human Services Administration for Children and Families. (2000). National Child Abuse and Neglect Data System (NCANDS) Glossary (SDC and DCDC Combined). Washington, DC: Author.

39.

U.S. Department of Health and Human Services Administration for Children and Families. (2010). Child maltreatment 2009. Washington, DC: Author.

40.

Warnick

E. M.

Gonzalez

Weersing

V. R.

Scahill

L. D.

Woolston

J. L.

(2012). Defining dropout from youth psychotherapy: How definitions shape the prevalence and predictors of attrition. Child and Adolescent Health, 17(2), 76-85.

41.

Zayfert

DeViva

J. C.

Becker

C. B.

Pike

J. L.

Gillock

K. L.

Hayes

S. A.

(2005). Exposure utilization and completion of cognitive behavior therapy for PTSD in “real world” clinical practice. Journal of Traumatic Stress, 18, 637-645.