Predicting Treatment Attrition for Child Sexual Abuse Victims

Abstract

Sexually abused children drop out of treatment more frequently than children receiving services for other issues. While researchers suggest that chaotic family dynamics may lead to inflated attrition rates in this population, other factors that potentially contribute to treatment attrition are virtually unknown. Therefore, the purpose of this study was to investigate the relationship between child and parent characteristics with attrition for sexual abuse victims (N = 132) and their nonoffending caregivers. Results indicate that children with parents who confirmed past or current intimate partner violence were 2.5 times more likely to prematurely terminate from treatment.

Keywords

child sexual abuse treatment attrition intimate partner violence concurrent family violence child advocacy centers

For nearly 30 years, researchers have discovered a number of long-term benefits associated with mental health treatment for sexually abused children. These benefits include reduced depression, anxiety, shame, posttraumatic stress, and fear (Mannarino, Cohen, Deblinger, Runyon, & Steer, 2012), and enhanced interpersonal relationships, stress management skills, self-esteem, and child safety skills (Briere, 1992; Cohen, Mannarino, Berliner, & Deblinger, 2000; Ducharme, Atkinson, & Poulton, 2000; Lau & Weisz, 2003; Mannarino et al., 2012). In empirically supported therapies for child sexual abuse (CSA) victims, over 80% of children experience some improvement in these areas within 12–16 sessions occurring once per week for 60–90 min (Cohen, Mannarino, et al., 2000; Deblinger, Lippmann, & Steer, 1996; Mannarino et al., 2012).

However, high therapy drop-out rates prevent abused children from getting the full treatment benefits they need (Kazdin & Wassell, 1998; Lau & Weisz, 2003; Phillips et al., 2000; Risser & Schewe, 2013; Weed, 2007). Treatment attrition can be defined as missing regular sessions, dropping out of treatment before attaining therapeutic goals, and resistance toward applying therapist suggestions to one’s everyday life (Drieschner, Lammers, & van der Staak, 2004). Relatively few studies have focused on the attrition patterns of maltreated children. Out of the studies on abused children and attrition that do exist, researchers tend to either study a single subtype of abuse or lump children with all types of abuse together in one sample (e.g., Koverola, Murtaugh, Connors, Reeves, & Papas, 2007; Lau & Weisz, 2003). In a longitudinal study on abused children and treatment outcomes, Lau and Weisz (2003) identified that abused children (N = 161) displayed poorer treatment persistence and earlier termination than nonabused children. Other treatment attrition studies reported that up to 60% of abused children in need of counseling prematurely terminate from treatment before receiving any therapeutic benefits (Kazdin, Holland, & Crowley, 1997; Schottenbauer, Glass, Arnkoff, Tendick, & Gray, 2008). Because children are dependent on caregivers in order to access, attend, and/or engage in treatment, researchers would benefit from including caregiver variables in attrition studies (Kinniburgh, Blaustein, Spinazzola, & van der Kolk, 2005).

Concurrent Family Violence and Attrition

Child abuse and family violence is a considerable health problem and is the most prevalent form of violence in the United States (Tolan, Gorman-Smith, & Henry, 2006). In a recent survey of over 4,500 children aged 0–17, 60% of children were exposed to violence in the year 2009 alone (Finkelhor, Turner, Ormrod, & Hamby, 2009). Families with one type of violence (i.e., CSA) oftentimes have co-occurring child or elder abuse (Dixon, Browne, Hamilton-Giachritsis, & Ostapuik, 2010). Furthermore, a substantial link exists between caregiver intimate partner violence (IPV) and co-occurring child abuse, both of which are capable of producing a variety of negative emotional, social, and psychological outcomes in the lives of both battered women and their abused children (Taylor, Guteman, Lee, & Rathouz, 2009). Researchers document the co-occurrence of CSA, child physical abuse, and/or neglect as occurring in over 50% of IPV cases (Herrenkohl, Sousa, Tajima, Herrenkohl, & Moylan, 2008), bringing estimates of co-occurring abuse to 500,000 children each year (Koverola et al., 2007). Therefore, chaotic family dynamics may contribute to poor treatment adherence among CSA victims.

Families with concurrent forms of violence face several barriers to treatment. Unlike single-victim crimes, families with both IPV and child abuse have two or more victims, typically including a mother (e.g., nonoffending caregiver) and one or more of her children (Dong, Anda, Dube, Giles, & Felitti, 2003). As previously established in IPV literature (Rosenbaum & Leisring, 2003), a perpetrator of IPV and/or CSA may create a frightening and manipulative atmosphere of power and control over the victim(s), which can undermine the ability of the nonoffending caretaker to shield the children from harm. Moreover, because families with co-occurring IPV and CSA are typically disorganized, chaotic, and secretive, child victims of abuse are more frequently referred to treatment after it is discovered by a third party, such as law enforcement or school personnel (Renner & Slack, 2006).

Therefore, in CSA cases where IPV is present, there exists a reduced likelihood that the parent victim will come forward for help due to the risks it poses to the family, such as further physical or sexual harm to the nonoffending caregiver and/or the child. While there is some evidence that higher levels of parental stress are related to lower rates of attendance in therapy sessions (Andra & Thomas, 1998; Kazdin & Wassell, 1998; Koverola et al., 2007) and higher attrition rates (Kazdin, 1990), some findings suggest the opposite. Risser and Schewe (2013) found no difference in parental–child stress levels between those who completed treatment and those who did not. However, the authors speculated that these findings may be due to having a younger sample of children (N = 1,365, 1–11 years old) versus adolescents, who may have more autonomy in deciding to continue treatment.

In another study examining a sample of child victims of physical abuse, sexual abuse, or neglect who also witnessed IPV (N = 105), McNamara and Fields (2000) concluded that parental stress interferes with parent’s ability to keep their children in treatment. The authors also found that parent self-report of stress associated with their own parental functioning was the only predictor (across child symptoms, agreement of child and parent report, and parent stress) of child attrition. Other researchers hypothesized that reluctance to adhere to treatment may be due to the potential legal liabilities faced by a nonoffending caregiver who reports abuse, potentially leading to loss of child custody (Koverola et al., 2007). Thus, IPV is an important factor to consider when examining child attrition.

Rationale for Study

Identifying factors related to attrition may bolster outcomes research for CSA victims. Widespread treatment attrition among CSA victims could lead researchers to falsely conclude that therapeutic interventions cause the observed changes in outcomes, when greater outcome variance could be explained by unidentified attrition-related variables (Howard, Krause, & Orlinsky, 1986; Peters, Calam, & Harrington, 2005). Clinicians might benefit from highlighting caregiver factors (i.e., chaotic, disorganized family dynamics, IPV) that influence treatment adherence during the screening process so that they can increase support from the very first session (Cohen, Berliner, & Mannarino, 2000; Maniglio, 2009).

Researchers have already identified that a positive therapeutic alliance is related to treatment engagement. For example, Prinz and Miller (1994) examined treatment dropout among 147 families with an aggressive child (age 4–9 years) and found that more intense parent engagement among high-adversity families produced a significantly lower drop-out rate than standard family treatment. Similarly, Kazdin, Marciano, and Whitley (2005) reported that positive child–therapist and parent–therapist alliances were associated with increased therapeutic change, fewer perceived treatment barriers, and an increase in treatment acceptability. Therefore, looking at caregiver and child functioning at intake may help clinicians focus on nurturing the therapeutic alliance for high-risk families and potentially increase the likelihood of successful treatment completion.

Examining both child and family characteristics with attrition patterns may offer a more complete picture for the ways concurrent family violence influences treatment. Koverola, Murtaugh, Connors, Reeves, and Papas (2007) examined attrition patterns among children aged 4–17 (N = 118) receiving outpatient treatment, and identified that caregivers who reported a high level of child-related parental distress and high psychological distress had children who were less likely to engage in treatment. Specifically, they found that 64% of children completed treatment, whereas 20% did not engage in treatment, and 16% stopped treatment prior to reaching their counseling goals. However, this study focused on a heterogeneous sample of children, all of whom had a history of family violence, thus differences in treatment adherence between children from violent and nonviolent homes could not be compared. Additional studies need to be conducted to explore whether a relationship exists between caregiver IPV, child trauma, and attrition for CSA victims.

Horowitz, Putnam, and Noll (1997) conducted a longitudinal study on factors affecting utilization of treatment for 81 sexually abused girls (6–16 years old) receiving counseling at a child protection agency. The authors reported that age of abuse onset and child psychopathology (i.e., depression, externalizing behaviors) were predictors of total completed therapy sessions, while family functioning did not predict total sessions of therapy. However, this study sample was comprised of only female clients who completed their treatment, and the authors did not provide information about clients who dropped out of treatment prematurely. In another study of 1,365 children receiving community-based services for exposure to violence, Risser and Schewe (2013) found that child emotional and behavioral problems at intake, parent–child stress, and income did not differ by treatment engagement. However, while their sample was comprised of younger children exposed to some form of violence (e.g., IPV, community violence, or child abuse), no children were identified specifically as CSA victims.

Therefore, this study aims to determine whether a relationship exists between child trauma symptomology (e.g., trauma-related cognitive schemas and posttraumatic stress) and a CSA client’s graduation status (e.g., whether the child finished counseling or dropped out of counseling early). Additionally, the researchers examined whether a relationship exists between caregiver violence (e.g., past or current IPV) and whether their child victim of CSA graduated from treatment or prematurely terminated. The first research question examines whether a CSA victim’s level of posttraumatic stress and trauma-related cognitive schemas predicts whether a CSA victim completes treatment. The second research question examines whether IPV status of a CSA victim’s caregiver predicts whether a CSA victim completes treatment.

Method

Participants

The population of interest in this investigation was child victims of CSA. The researchers accessed sanitized reports from a database program managed by the National Children’s Alliance (NCA), known as “NCA Trak.” These reports contained information on clients (e.g., type of abuse, client and caretaker information, service usage, graduation status, referral source, law enforcement outcomes, and assessment administration dates) receiving services between January 2009 and December 2011 at a child advocacy center (CAC) in Central Florida.

The researchers employed purposive sampling to meet the specific inclusion criteria (Gall, Gall, & Borg, 2007). We selected participants based on a number of conditions, including the requirements that they (a) were sexual abuse victims, (b) were no longer in treatment, (c) completed both the Trauma and Attachment Belief Scale (TABS) and Trauma Symptom Checklist for Children (TSCC) assessments at intake, and (d) had caregivers who either confirmed or denied past or current IPV. Based on the inclusion criteria for this study, our final sample size included 132 electronic data packets. Participants in this study ranged in age from 8 years to 16 years, and the majority of participants were female (n = 108; 81.8%). The participants’ demographic data are presented in Table 1.

Table 1.

Participants’ Demographic Characteristics.

Characteristic	n	Total %
Gender
Female	108	81.8
Male	24	15.2
Ethnic background
Caucasian (non-Hispanic)	75	56.8
African American/African/Black	28	21.2
Hispanic/Latina/Latino	20	15.2
Biracial/multiracial	7	5.3
Asian/Asian American	2	1.5
Education level
Elementary school	29	22.0
Middle school	58	43.9
High school	45	19.4

Note. N = 132.

Procedure

Upon the Institutional Review Board’s approval, the researchers contacted the director of mental health services at the child abuse agency to establish on-site data collection. The collection of archival data occurred between July 2010 and January 2012. Since the researchers used existing data, no incentives were offered to any participants. The child abuse agency collected data on clients at various points during their treatment. Prior to the client beginning treatment, clients who were at least 8 years of age completed forms and questionnaires within the agency’s intake paperwork packet including the TABS and TSCC at intake, 12 sessions, 6 months, and posttreatment. However, for this study, only pretest scores were analyzed, which measured baseline levels of trauma and cognitive disturbance. The treating therapists collected information on service usage, attendance, and graduation status on a weekly basis and entered the information into the NCA Trak database. Regarding data analysis for the assessment instruments, the researchers analyzed T-scores for both the TABS and TSCC as opposed to mean scores. According to Pallant (2007), T-scores are a better indicator of how the sample compares to the norm groups for each assessment.

Instruments

The researchers gathered information from different client sources for measuring the dichotomous variable of graduation status and the continuous variables of posttraumatic stress and trauma-related cognitive schemas. These sources included archival data from (a) NCA Trak, (b) the TABS assessment, and (c) the TSCC assessment. We include a brief overview of each form or instrument below.

NCA database (NCA Trak)

The NCA database, otherwise known as NCA Trak, is a national data collection system developed specifically for CACs. Based on the information entered into the system, the CAC generates reports on child and caretaker demographics, child protective services and law enforcement outcomes, screening and referral information, staff caseloads, assessment information, the client’s presenting information, and client allegations. In this study, children were labeled as graduators if they fully completed their treatment goals as defined by the therapist. Children were labeled as premature terminators if they (a) were administratively discharged due to excessive absences or 30-day inactivity; (b) dropped out of counseling; (c) left after partially, but not fully, completing their treatment goals; (d) were referred by the agency to other types of services, such as residential treatment; or (e) left for special circumstances not listed above.

Caregiver IPV status was recorded into NCA Trak as “yes” or “no” if (a) the caretaker verbally confirmed past or current IPV in the home, (b) the caretaker indicated in writing the occurrence of past or present IPV in the home on intake paperwork, or (c) if a referring agency provided paperwork to the child abuse agency indicating the presence of past or current IPV in the home.

TABS

The TABS (Pearlman, 2003), formerly known as the Traumatic Stress Institute Belief Scale revision L (Pearlman, 2003), is an 84-item assessment designed to measure trauma-related cognitive schemas. Based on Constructivist Self-Development Theory (CSDT; McCann & Pearlman, 1990; Pearlman & Saakvitne, 1995; Saakvitne, Gamble, Pearlman, & Lev, 2000), which posits that survivors of trauma adopt maladaptive and dysfunctional beliefs about themselves and others after a traumatic event, the TABS incorporates CSDT and elements selected from existing theoretical and empirical literature on the associations among traumatic stress, cognitions, and needs. Specifically, the TABS measures disruptions in five psychological need areas, including safety (i.e., I believe I am safe), trust (i.e., You can’t trust anyone), esteem (i.e., I have bad feelings about myself), intimacy (i.e., I feel close to lots of people), and control (i.e., I feel like I can’t control myself). Evaluation on the TABS for reliability and validity (Pearlman, 2003; Varra, Pearlman, Brock, & Hodgson, 2008) yielded sound evidence of content, construct, and criterion validity. Reliability is also acceptable, with a Pearson r score of .75, and an internal consistency α score of .96.

TSCC

The TSCC is a 54-item self-report instrument used with children between the ages of 8 and 16 to assess how often an individual has thoughts, feelings, and behaviors related to traumatic events they have experienced (Briere, 1996). The measure includes six clinical scales, including items measuring anxiety (i.e., Getting scared all of a sudden and don’t know why), depression (i.e., Crying), posttraumatic stress (i.e., Scary ideas or pictures just pop into my head), sexual concerns (i.e., Not trusting people because they might want sex), dissociation (i.e., Feeling like I’m not in my body), and anger (i.e., Arguing too much). The TSCC also includes four subscales (e.g., Sexual Preoccupation, Sexual Distress, Fantasy, and Overt Dissociation). The TSCC is one of the most commonly used measures for exploring sexual abuse sequelae in children and adolescents (Elhai, Gray, Kashdan, & Franklin, 2005). The assessment has moderate to high convergent and discriminant validity (Boyle, 2003) and high construct validity (Sadowski & Friedrich, 2000), with a Pearson r score of .81 and a high internal consistency α score of .91 (Bal, De Bourdeaudhuij, Crombez, & Van Oost, 2005; Crouch, Smith, Ezzell, & Saunders, 1999; Lanktree & Briere, 1995).

Results

Data Entry, Handling, and Analysis

The researchers conducted power analyses a priori to determine the appropriate sample size for both χ² test for independence and logistic regression analyses. The researchers referred to Cohen’s (1992) rules regarding sample size in order to identify a median effect size for correlational analyses (power = .80) at the .05 level. A logistic regression with one dichotomous predictor variable (i.e., caregiver IPV) required a minimum of 55 participants to achieve adequate power. A χ² test for independence required a minimum of 107 participants to achieve adequate power. Therefore, we exceeded the minimum recommended sample size to ensure adequate power for both research questions posed in the current study.

The authors assessed the normality of the data (Tabachnik & Fidell, 2013). First, it is worthy to note that the percentage of missing scores replaced with norm scores for this study is unknown. However, both assessments have a built-in procedure for supplanting missing items with recommended replacement scores based on the norm sample. Therefore, the authors opted to not employ list-wise deletion (Hair, Black, Babin, Anderson, & Thatham, 2010).

Additionally, the authors assessed the fit between the distribution of the variables and the assumptions of the statistical analysis, such as normality, homogeneity of variance, linearity, and multicollinearity. Upon further assessment of multicollinearity using the Tolerance and variance inflation factor (VIF) statistics, all VIF scores were lower than the recommended threshold of 3 (Tabachnick & Fidell, 2013). No outliers were identified and assumptions were met for each analysis. Therefore, our data were deemed appropriate for χ² test of independence and logistic regression analyses.

Trauma-Related Cognitive Schemas and Client Graduation

We used logistic regression analysis to test the predictive relationship between the dependent, dichotomous variable of graduation status (e.g., graduated or prematurely terminated), and the continuous predictor variables of posttraumatic stress and trauma-related cognitive schemas (see Table 2). The full model containing all seven predictors (e.g., posttraumatic stress score of the TSCC and the self-safety, self-trust, self-intimacy, self-esteem, and self-control, other intimacy, and other control subscale scores of the TABS) was not statistically significant, χ²(8, N = 131) = 15.10, p > .05, indicating that the model was not able to distinguish between those who graduated or prematurely terminated from treatment based on trauma scores. The whole model correctly classified 63.6% of cases. As shown in Table 2, none of the independent variables made a unique statistically significant contribution to the model. Thus, scores related to trauma symptomology and disturbances in trauma-related cognitive schemas were not able to predict whether a CSA victim graduated or prematurely terminated from treatment, and therefore, the null hypothesis was retained.

Table 2.

Predicting Graduation Status by Trauma/Cognitive Schema Disturbance.

Sub scale	β	SE	Wald	ρ	Odds Ratio	95% CI for Odds Ratio
Sub scale	β	SE	Wald	ρ	Odds Ratio	Lower	Upper
Constant	−.386	1.19	0.103	.75	0.68
Posttraumatic stress	.024	0.024	1.00	.32	1.0	.98	1.07
Other safety	−.002	0.03	0.007	.93	1.0	.95	1.05
Other intimacy	−.013	0.04	0.001	.71	1.0	.92	1.06
Other control	.000	0.03	0.000	.99	1.0	.95	1.06
Self-control	−.020	0.04	0.311	.58	0.98	.92	1.05
Self-trust	.009	0.02	0.160	.69	1.0	.97	1.06
Self-esteem	−.001	0.03	0.001	.98	1.0	.94	1.06

Note. df = 1. SE = standard error; CI = confidence interval.

Caregiver IPV and Client Graduation

We utilized a χ² test for independence to examine the relationship between caregiver IPV status (e.g., yes or no) and a client’s graduation status (e.g., graduated or prematurely terminated). The χ² test for independence (with Yates continuity correction) indicated a statistically significant association between graduation status and caregiver IPV, χ²(1, N = 107) = 4.208, p < .05, ϕ = −2.18. The cross tabulation (see Table 3) indicated that children of caregivers who denied past or current IPV were just as likely to graduate from treatment as they were to prematurely terminate from treatment. However, among children of caregivers who confirmed past or current IPV, a higher proportion (71.4%) prematurely terminated from treatment. This seems to represent the fact that, based on the odds ratio, the odds of a CSA victim prematurely terminating are 2.5 times higher if their parents confirm past or current IPV than those children with parents who denied IPV.

Table 3.

Caregiver IPV and Graduation Status Cross Tabulation.

Caregiver IPV Status	Premature Termination	Graduated	Total
No caregiver IPV
Count	22	22	44
%	50	50	100
% of total	21	21	41
Yes caregiver IPV
Count	45*	18*	63
%	71	29	100
% of total	42	17	59

Note. N = 107. IPV = intimate partner violence.

*p ≤ .05.

Discussion

Previous studies on multiply victimized children and attrition tend to either examine a single subtype of abuse or lump children with all types of abuse together into one sample (e.g., Koverola et al., 2007; Lau & Weisz, 2003). Thus, researchers fail to account for the unique ways in which specific combinations of abuse influence outcomes and/or attrition rates. The literature commonly refers to these combined effects of multiple abuses as overlap. The purpose of this study was to examine both child and caregiver factors to learn more about how these factors are related to treatment attrition for CSA victims.

The findings in both research questions indicated that child factors, specifically traumatization and schema distortions, do not place children at increased risk for premature termination, while parent factors, specifically past or current IPV, place children at significant risk for early termination. This supports the notion that CSA treatment attrition is a function of caregiver distress, especially if the caregiver is also a victim of abuse. These findings are echoed in previous research from other fields that determined parent factors, such as single motherhood, low socioeconomic status, and high maternal stress, predicted child attrition (Armbruster & Fallon, 1994; Kazdin, Marciano, & Whitley, 2005; McCabe, 2002; Peters et al., 2005; Prinz & Miller, 1994). Furthermore, the trauma scores of individuals who graduated were statistically similar to those who prematurely terminated, which are similar to other findings identifying no differences in behavioral and emotional functioning among the nonengagers, attriters, and completers at intake (Risser & Schewe, 2013).

Regarding the second research question of this study, our data identified a relationship between caregiver IPV and a CSA victim’s graduation status. Specifically, CSA victims with confirmed caregiver IPV were 2.5 times more likely to prematurely terminate than to graduate. Exactly 50% of caregivers who denied past or present IPV had children who graduated from treatment, while the other 50% prematurely terminated. Thus, CSA victims without caregiver violence had an equal probability of finishing treatment as they did terminating early. However, of children with caregivers who reported past or current IPV, 71% left treatment before meeting their therapeutic goals and only 29% graduated.

The results of our study supported findings reported by other researchers. Koverola and colleagues (2007), for example, reported that (a) child trauma scores were statistically equivalent among those who completed treatment and those who left early and (b) caregiver distress was related to level of treatment engagement. The authors concluded, “This may reflect that children in this sample do not have a strong voice in determining whether they engage in or complete treatment, but rather they will be involved in treatment at the discretion of their caregivers” (p. 36).

Implications

Future Research

The findings of this research study highlight a number of implications for future researchers of attrition and overlapping family violence. First, there is a need for a valid and reliable assessment of concurrent family violence for both children and caretakers. Currently, no assessments exist that quantitatively measure detailed experiences of violence, as well as cognitive, emotional, or social disturbances associated with concurrent family violence. Furthermore, as is mentioned in recent literature (Thackeray, Scribano, & Rhoda, 2010), CACs would benefit from regularly administering assessments to caregivers in an effort to learn more about other potential caregiver factors influencing attrition. However, Saunders (2003) states:

[ …] it is unrealistic to think that evaluation protocols can be constructed to assess every important type of violence and trauma, all possible mediators and moderators, and the many potential outcomes. Such a protocol would simply be too long, too complex, and too taxing on participants, particularly children. Therefore, the idea comprehensive study on child violence may never be done. (p. 369)

As mentioned in previous studies on concurrent family violence (Kazdin, 1996; Saunders, 2003), there exists a need for researchers and practitioners from multiple specialties and expertise to share information across disciplines. Considering the estimated overlap of research samples is as high as 50% across studies (Saunders, 2003), researchers of trauma, child abuse, IPV, and attrition might benefit from creating a comprehensive task force to explore these complex issues together and not separately. Similarly, the results of this study demonstrate a need for a clinical multidisciplinary approach to investigate, manage, treat, and prosecute caregiver IPV in families with CSA.

Practice

The results from our study provide several implications regarding how counselors and agencies may better support families with multiple abuse victims. Currently, the majority of the focus of treatment for CSA victims is placed on the child victim of abuse, with the majority of assessments and therapeutic attention directed at the child. However, due to the relationship between caregiver well-being and their child’s successful completion of treatment, additional efforts should be made to support and encourage caregivers to participate in treatment. Specifically, the results of this study demonstrate a need for a multidisciplinary approach to investigate, manage, treat, and prosecute caregiver IPV in families with CSA. CSA victims and their families might benefit from the partnerships and collaborations of local domestic violence shelters, beyond a referral for services.

Currently, mental health counselors and agencies working with caregivers in violent relationships engage in advocacy for their clients to help connect the family to resources such as welfare assistance, free food programs, transportation programs, medical assistance, and family therapy to decrease the burdens faced by the family in order to better protect the child. Similar forms of advocacy could be used to increase the likelihood of treatment completion. For example, many perpetrators of abuse extol power and control tactics in order to dominate victims and discourage them from leaving the relationship. These power and control tactics include controlling finances, isolating the victim from friends and family, confiscating important documents such as drivers licenses, birth certificates, and immigration paperwork, and threatening to harm pets and/or children. A caregiver experiencing this type of control and manipulation would benefit from having resources, such as group therapy with other IPV victims, financial advisement, career planning, and safety planning to help the victim relinquish her autonomy. Moreover, alternative forms of therapeutic attention, potentially in the form of home-based visits, could be offered to the caregiver throughout the child’s treatment to help bolster treatment engagement. We believe these interventions would support a CSA victim’s appropriate completion of therapy.

Furthermore, systematic efforts to involve and retain caregivers whose children present with more severe trauma symptomology may be beneficial for the client’s treatment fidelity. However, in instances where the child is engaging in therapy to reduce trauma symptomology (e.g., Eye Movement Desensitization and Reprocessing), including the caregiver in treatment may complicate and even hinder progress. Clinical decisions must be thoroughly discussed in weekly staff meetings and carefully examined on a case-by-case basis, as complex family dynamics—such as parental traumatization, depression, reversed parent–child roles, and/or unsupportive parents—may damage the child’s positive growth and healing in treatment. In these cases, parents should be referred to individual counseling and extensively prepared before joining a therapeutic session with a child with acute traumatization.

Limitations

Prior to applying these findings to the research, practice, and counselor education fields, it is important to discuss study limitations. First, when compared to more complex studies on child abuse, it is clear that this study only “scratches the bivariate surface” (Saunders, 2003, p. 369). In his article on the challenges of researching overlapping family violence, Saunders (2003) reveals that most child victims of violence are victims of more than one type of violence. Therefore, research on the link between violence exposure and outcomes are often confounded by the child’s and/or caretaker’s exposure to other types of trauma. Considering most child abuse researchers examine either (a) one subtype of abuse or (b) all subtypes of abuse in a single sample, the current study selected children with a specific combination of family violence (e.g., CSA and IPV) because of the unique cognitive schema distortions particular to sexual abuse (vs. physical abuse or neglect). However, this does not constitute a comprehensive method of evaluation.

It is also necessary to note limitations associated with our sampling method. The current study selected participants based on several criteria. While making informed decisions in sampling can help reduce threats to validity faced by research studies that do not account for differences in abuse experiences, results are likely to be shaped by a number of unknown confounding variables. In some cases, these confounding variables can elevate findings on negative outcomes (i.e., high attrition) within simple bivariate research. While the vast majority of archival child abuse research is correlational, Saunders (2003) stated, “If these alternative explanatory factors are not measured, then criterion group equality cannot be assessed, the impact of other risk factors cannot be determined, and important potential confounds cannot be controlled” (p. 358). However, due to the complex and numerous differences in the families’ risk and protective factors, it is nearly impossible to account for every confounding factor.

Another limitation deals with research design. Within correlational analyses, etiological conclusions cannot be made about the observed relationships between caregiver IPV, graduation status, posttraumatic stress, and trauma-related cognitive schemas. Furthermore, while archival data offer a wealth of knowledge on clinical samples, it has several limitations. For example, the definition of abuse in archival research is a function of (a) how victim describes it to the therapist, and (b) how the therapist records it. In a sense, the subjective nature of reporting prevents the researcher from crafting a definition for different types of abuse that is informed by the literature. Without agreement across studies on the common definitions and measurements of child maltreatment and IPV, generalizability is greatly diminished. Moreover, researchers working with archival data are not able to monitor the accuracy of the data, which may threaten the study’s internal validity.

One of the greatest limitations of this study is the manner in which information was recorded by the CAC on caregiver IPV. As is typical in many CACs (Thackeray et al., 2010), the child abuse agency did not administer specific IPV assessments to caretakers and, therefore, the records contain little detail about the specific nature of past or current IPV, including the degree of IPV exposure experienced by the CSA victim. Therefore, it was not possible to account for important details of the violence (i.e., abuse severity, duration, and frequency) that might influence the findings of this study.

Conclusion

The goal of this study was to illuminate the relationships between caregiver IPV, child traumatization, trauma-related cognitive schemas, and treatment attrition. Our data failed to identify a relationship between level of traumatization, disturbance of trauma-related cognitive schemas, and whether a CSA victim graduated or prematurely terminated from treatment. Yet, children with confirmed caregiver IPV were 2.5 times more likely to prematurely terminate from treatment than those without caregiver IPV. Therefore, the findings from this study highlight a need for more detailed caregiver assessments in CACs and call for cross-field collaboration in both research and practice. Without these partnerships, the depth and breadth of research are sacrificed. Most importantly, these collaborations would better serve the millions of vulnerable children and families who desperately need healing and help.

Footnotes

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) received no financial support for the research, authorship, and/or publication of this article.

References

Andra

M. L.

Thomas

A. M.

(1998). The influence of parenting stress and socioeconomic disadvantage on therapy attendance among parents and their behavior disordered preschool children. Education and Treatment of Children, 21, 195–208.

Armbruster

Fallon

(1994). Clinical, sociodemographic, and systems risk factors for attrition in a children’s mental health clinic. American Journal of Orthopsychiatry, 64, 577–585. doi:10.1037/h0079571

Bal

De Bourdeaudhuij

Crombez

Van Oost

(2005). Predictors of trauma symptomology in sexually abused adolescents: A 6-month follow-up study. Journal of Interpersonal Violence, 20, 1390–1405. doi:10.1177/0886260505278720

Boyle

G. J.

(2003). Trauma symptom checklist for children. In Plake

B. S.

Impara

J. C.

Spies

R. A.

(Eds.), The fifteenth mental measurements yearbook (pp. 974−976). Lincoln: The University of Nebraska Press.

Briere

J. N.

(1992). Child abuse trauma: Theory and treatment of the lasting effects. Newbury Park, VA: Sage.

Briere

J. N.

(1996). Trauma Symptom Checklist for Children (TSCC): Professional manual. Odessa, FL: Psychological Assessment Resources, Inc.

Cohen

J. A.

(1992). A power primer. Quantitative Methods in Psychology, 112, 155–159. doi:10.1037/0033-2909.112.1.155

Cohen

J. A.

Berliner

Mannarino

A. P.

(2000). Treatment of traumatized children: A review and synthesis. Journal of Trauma, Violence, and Abuse, 1, 29–46. doi:10.1177/1524838000001001003

Cohen

J. A.

Mannarino

A. P.

Berliner

Deblinger

(2000). Trauma-focused cognitive behavioral therapy for children and adolescents: An empirical update. Journal of Interpersonal Violence, 15, 1202–1223. doi:10.1177/088626000015011007

10.

Crouch

J. L.

Smith

D. W.

Ezzell

C. E.

Saunders

B. E.

(1999). Measuring reactions to sexual trauma among children: Comparing the Children’s Impact of Traumatic Events Scale and the Trauma Symptom Checklist for Children. Child Maltreatment, 4, 255–263. doi:10.1177/1077559599004003006

11.

Deblinger

Lippmann

Steer

(1996). Sexually abused children suffering posttraumatic stress symptoms: Initial treatment outcome findings. Child Maltreatment, 1, 310–321. doi:10.1177/1077559596001004003

12.

Dixon

Browne

Hamilton-Giachritsis

Ostapuik

(2010). Differentiating patterns of aggression in the family. Journal of Aggression, Conflict and Peace Research, 2, 32–44. doi:10.5042/jacpr.2010.0003

13.

Dong

Anda

R. F.

Dube

S. R.

Giles

W. H.

Felitti

V. J.

(2003). The relationship of exposure to childhood sexual abuse to other forms of abuse, neglect and household dysfunction during childhood. Child Abuse and Neglect, 27, 625–639. doi:10.1016/s0145-2134(03)00105-4

14.

Drieschner

K. H.

Lammers

S. M.

van der Staak

C. P.

(2004). Treatment motivation: An attempt for clarification of an ambiguous concept. Clinical Psychology Review, 23, 1115–1137. doi:10.1016/j.cpr.2003.09.003

15.

Ducharme

J. M.

Atkinson

Poulton

(2000). Success-based noncoercive treatment of oppositional behavior in children from violent homes. Journal of the American Academy of Child & Adolescent Psychiatry, 39, 995–1004. doi:10.1097/00004583-200008000-00014

16.

Elhai

J. D.

Gray

M. J.

Kashdan

T. B.

Franklin

C. L.

(2005). Which instruments are most commonly used to assess traumatic event exposure and posttraumatic effects? A survey of traumatic stress professionals. Journal of Traumatic Stress, 18, 541–545. doi:10.1002/jts.20062

17.

Finkelhor

Turner

Ormrod

Hamby

S. L.

(2009). Violence, abuse, and crime exposure in a national sample of children and youth. Pediatrics, 124, 1411–1423.

18.

Gall

M. D.

Gall

J. P.

Borg

W. R.

(2007). Educational research: An introduction. Boston, MA: Pearson/Allyn & Bacon.

19.

Hair

J. F.

Black

W. C.

Babin

B. J.

Anderson

R. E.

Tatham

R. L.

(2010). Multivariate data analysis. Upper Saddle River, NJ: Pearson.

20.

Herrenkohl

T. I.

Sousa

Tajima

E. A.

Herrenkohl

R. C.

Moylan

C. A.

(2008). Intersection of child abuse and children’s exposure to domestic violence. Trauma, Violence, & Abuse, 9, 84–89. doi:10.1177/152438008314797

21.

Horowitz

L. A.

Putnam

F. W.

Noll

J. G.

(1997). Factors affecting utilization of treatment services by sexually abused girls. Child Abuse & Neglect, 21, 35–48. doi:10.1016/s0145-2134(96)00129-9

22.

Howard

K. I.

Krause

M. S.

Orlinsky

D. E.

(1986). The attrition dilemma: Toward a new strategy for psychotherapy research. Journal of Consulting and Clinical Psychology, 54, 106–110. doi:10.1037/0022-006x.54.1.106

23.

Kazdin

A. E.

(1990). Premature termination from treatment among children referred for antisocial behavior. Journal of Child Psychology and Psychiatry, 31, 415–425. doi:10.1111/j.1469-7610.1990.tb01578.x

24.

Kazdin

A. E.

(1996). Dropping out of child psychotherapy: Issues for research and implications for practice. Clinical Child Psychology and Psychiatry, 1, 133–156. doi:10.1177/1359104596011012

25.

Kazdin

A. E.

Holland

Crowley

(1997). Family experience of barriers to treatment and premature termination from child therapy. Journal of Consulting and Clinical Psychology, 65, 453–463. doi:10.1037//0022-006x.65.3.453

26.

Kazdin

A. E.

Marciano

P. L.

Whitley

M. K.

(2005). The therapeutic alliance in cognitive-behavioral treatment of children referred for oppositional, aggressive, and antisocial behavior. Journal of Consulting and Clinical Psychology, 73, 726–730. doi:10.1037/0022-006x.73.4.726

27.

Kazdin

A. E.

Wassell

(1998). Treatment completion and therapeutic change among children referred for outpatient therapy. Professional Psychology, 29, 332–340. doi:10.1037//0735-7028.29.4.332

28.

Kinniburgh

K. J.

Blaustein

Spinazzola

Van der Kolk

B. A.

(2005). ARC: Attachment, self-regulation, and competency: A comprehensive framework for intervention with complexly traumatized youth. (n.d.). PsycEXTRA Dataset. doi:10.1037/e732542011-00

29.

Koverola

Murtaugh

C. A.

Connors

K. M.

Reeves

Papas

M. A.

(2007). Children exposed to intra-familial violence: Predictors of attrition and retention in treatment. Journal of Aggression, Maltreatment & Trauma, 14, 19–42. doi:10.1300/j146v14n04_02

30.

Lanktree

C. B.

Briere

J. N.

(1995). Outcome therapy for sexually abused children: A repeated measures study. Child Abuse and Neglect, 19, 1145–1155. doi:10.1016/0145-2134(95)00075-j

31.

Lau

A. S.

Weisz

J. R.

(2003). Maltreatment among clinic-referred children: Implications for presenting problems, treatment attrition, and long term outcomes. Journal of American Academy of Child and Adolescent Psychiatry, 42, 1327–1334. doi:10.1097/01.chi.0000085754.71002.14

32.

Maniglio

(2009). The impact of child sexual abuse on health: A systemic review of reviews. Clinical Psychology Review, 29, 647–657. doi:10.1016/j.cpr.2009.08.003

33.

Mannarino

A. P.

Cohen

J. A.

Deblinger

Runyon

M. K.

Steer

R. A.

(2012). Trauma-focused cognitive-behavioral therapy for children sustained impact of treatment 6 and 12 months later. Child Maltreatment, 17, 231–241. doi:10.1177/1077559512451787

34.

McCabe

K. M.

(2002). Factors that predict premature termination among Mexican–American children in outpatient psychotherapy. Journal of Child and Family Studies, 11, 347–359. doi:10.1023/a:1016876224388

35.

McCann

I. L.

Pearlman

L. A.

(1990). Psychological trauma and the adult survivor: Theory, therapy, and transformation. New York, NY: Bruner/Mazel.

36.

McNamara

J. R.

Fields

S. A.

(2000). Differential functioning of outpatients and residents of a domestic violence shelter on the Abuse Disability Questionnaire. Psychological Reports, 86, 893–894. doi:10.2466/pr0.86.3.893-894

37.

Pallant

(2007). SPSS survival manual 3rd edition: A step-by-step guide to data analysis using SPSS. Buckingham, England: Open University Press.

38.

Pearlman

L. A.

(2003). Trauma and Attachment Belief Scale (TABS) manual. Los Angeles, CA: Western Psychological Services.

39.

Pearlman

L. A.

Saakvitne

K. W.

(1995). Trauma and the therapist: Countertransference and vicarious traumatization in psychotherapy with incest survivors. New York, NY: Norton.

40.

Peters

Calam

Harrington

(2005). Maternal attributions and expressed emotion as predictors of attendance at parent management training. Journal of Child Psychology and Psychiatry, and Allied Disciplines, 46, 436–448. doi:10.1111/j.1469-7610.2004.00365.x

41.

Phillips

S. D.

Hargis

M. D.

Kramer

T. L.

Lensing

S. Y.

Taylor

J. L.

Burns

B. J.

Robbins

J. M.

(2000). Toward a level playing field: Predictive factors for the outcomes of mental health treatment for adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 39, 1485–1495. doi:10.1097/00004583-200012000-00008

42.

Prinz

R. J.

Miller

G. E.

(1994). Family-based treatment for childhood antisocial behavior: Experimental influences on dropout and engagement. Journal of Consulting and Clinical Psychology, 62, 645–650. doi:10.1037/0022-006x.62.3.645

43.

Renner

L. M.

Slack

K. S.

(2006). Intimate partner violence and child maltreatment: Understanding intra-and intergenerational connections. Child Abuse & Neglect, 30, 599–617. doi:10.1016/j.chiabu.2005.12.005

44.

Risser

H. J.

Schewe

P. A.

(2013). Predictors of treatment engagement for young children exposed to violence. Journal of Aggression, Maltreatment & Trauma, 22, 87–108. doi:10.1080/10926771.2013.743946

45.

Rosenbaum

Leisring

P. A.

(2003). Beyond power and control: Towards an understanding of partner abusive men. Journal of Comparative Family Studies, 34, 7–22.

46.

Saakvitne

K. W.

Gamble

Pearlman

L. A.

Lev

B. T.

(2000). Risking connection: A training curriculum for working with survivors of childhood abuse. Baltimore, MD: The Sidran Press.

47.

Sadowski

C. M.

Friedrich

W. N.

(2000). Psychometric properties of the Trauma Symptom Checklist for Children (TSCC) with psychiatrically hospitalized adolescents. Child Maltreatment, 5, 364–372. doi:10.1177/1077559500005004008

48.

Saunders

B. E.

(2003). Understanding children exposed to violence: Toward an integration of overlapping fields. Journal of Interpersonal Violence, 18, 356–376. doi:10.1177/0886260502250840

49.

Schottenbauer

M. A.

Glass

C. R.

Arnkoff

D. B.

Tendick

Gray

S. H.

(2008). Nonresponse and dropout rates in outcome studies on PTSD: Review and methodological considerations. Psychiatry: Interpersonal and Biological Responses, 71, 134–168. doi:10.1521/psyc.2008.71.2.134

50.

Tabachnick

B. G.

Fidell

L. S.

(2013). Using multivariate statistics (6th ed.). Upper Saddle River, NJ: Pearson.

51.

Taylor

C. A.

Guterman

N. B.

Lee

S. J.

Rathouz

P. J.

(2009). Intimate partner violence, maternal stress, nativity, and risk for maternal maltreatment of young children. American Journal of Public Health, 99, 175–183. doi:10.2105/ajph.2007.126722

52.

Thackeray

J. D.

Scribano

P. V.

Rhoda

(2010). Domestic violence assessments in the child advocacy center. Child Abuse & Neglect, 34, 172–182. doi:10.1016/j.chiabu.2009.10.002

53.

Tolan

Gorman-Smith

Henry

(2006). Family violence. Annual. Review of Psychology, 57, 557–583. doi:10.1146/annurev.psych.57.102904.190110

54.

Varra

E. M.

Pearlman

L. A.

Brock

K. J.

Hodgson

S. T.

(2008). Factor analysis of the Trauma and Attachment Belief Scale: A measure of cognitive schema disruption related to traumatic stress. Journal of Psychological Trauma, 7, 185–196. doi:10.1080/19322880802266813

55.

Weed

K. B.

(2007). Predicting treatment completion with maltreated children. Doctoral dissertation. Retrieved from ProQuest Dissertations and Theses. (UMI No. 3258569).