Elsevier

Child Abuse & Neglect

Volume 67, May 2017, Pages 98-108
Child Abuse & Neglect

The blind men and the elephant: Identification of a latent maltreatment construct for youth in foster care

https://doi.org/10.1016/j.chiabu.2017.02.020Get rights and content

Abstract

Child maltreatment is a major public health concern due to its impact on developmental trajectories and consequences across mental and physical health outcomes. Operationalization of child maltreatment has been complicated, as research has used simple dichotomous counts to identification of latent class profiles. This study examines a latent measurement model assessed within foster youth inclusive of indicators of maltreatment chronicity and severity across four maltreatment types: physical, sexual, and psychological abuse, and neglect. Participants were 500 foster youth with a mean age of 12.99 years (SD = 2.95 years). Youth completed survey questions through a confidential audio computer-assisted self-interview program. A two-factor model with latent constructs of chronicity and severity of maltreatment revealed excellent fit across fit indices; however, the latent constructs were correlated 0.972. A one-factor model also demonstrated excellent model fit to the data (χ2 (16, n = 500) = 28.087, p = 0.031, RMSEA (0.012–0.062) = 0.039, TLI = 0.990, CFI = 0.994, SRMR = 0.025) with a nonsignificant chi-square difference test comparing the one- and two-factor models. Invariance tests across age, gender, and placement type also were conducted with recommendations provided. Results suggest a single-factor latent model of maltreatment severity and chronicity can be attained. Thus, the maltreatment experiences reported by foster youth, though varied and complex, were captured in a model that may prove useful in later predictions of outcome behaviors. Appropriate identification of both the chronicity and severity of maltreatment inclusive of the range of maltreatment types remains a high priority for future research.

Introduction

Child maltreatment is a risk factor associated with a range of negative outcomes across physical, social, and mental health domains (Cicchetti & Toth, 2005; Kaplan, Pelcovitz, & Labruna, 1999), and it remains a leading public health concern due to its financial and societal costs (Fang, Brown, Florence, & Mercy, 2012). A recent report released by the US Department of Health and Human Services (2016) revealed that approximately 3.5 million children are identified to child protective services each year as potential victims of maltreatment. Of these youth, approximately 20% were cases of substantiated child maltreatment and represent a population with high risk for adverse behavioral and emotional outcomes. The influence of exposure to maltreatment in childhood on later child development has been explored extensively in the literature, with effects related to various aspects of maltreatment, such as type (English, Upadhyaya et al., 2005), chronicity (English, Graham, Litrownik, Everson, & Bangdiwala, 2005), severity (Litrownik et al., 2005), or age of onset (Thornberry, Ireland, & Smith, 2001).

Despite the growth of research on child maltreatment, the construct of child maltreatment remains difficult to define given that each maltreatment experience is comprised of several dimensions (e.g., severity, perpetrator type, duration), and that there is a wide range of means of assessing these experiences and their dimensions (e.g., child self-report, caregiver report, case-file review). No consistent definition of maltreatment has been established across states for either legal purposes related to removal of a child from their home or for research purposes in establishing relations between maltreatment exposure and outcomes. Thus, maltreatment is a complicated construct that can be conceptualized in a variety of ways, creating the potential for divergent research findings based on how the construct is operationalized.

Consequently, little guidance is available for ways to account for multiple dimensions of maltreatment using more comprehensive measurement, and yet advances in statistical methodologies provide new alternatives for accounting for multidimensionality in constructs that have layered features, such as the type, severity, and chronicity of maltreatment. Our goal is to improve the dimensionality of maltreatment measurement. To this end, we examine a latent measurement model of foster youth self-report of maltreatment experiences, inclusive of indicators of maltreatment chronicity and severity (English, Upadhyaya et al., 2005) across four maltreatment types: physical abuse, sexual abuse, psychological abuse, and neglect, to determine if a multidimensional construct of maltreatment can be fit to the data. If so, later outcomes predictions might be facilitated by the use of a coherent and inclusive model, and the model could be used as an example for the creation of future maltreatment measurement models.

Prior research has consistently indicated that the type of maltreatment experienced may have implications for outcomes (e.g., Perez and Widom, 1994, Spinazzola et al., 2014). However, research in which variables representing maltreatment type are used to predict outcomes has led to some discrepant findings (e.g., Moran, Vuchinich, & Hall, 2004; Taussig, 2002, Wall and Kohl, 2007). These discrepancies may be due in part to researchers’ use of idiosyncratic strategies of determining which type of maltreatment a child has experienced, reducing the likelihood results will generalize. To illustrate the challenges in using maltreatment type to predict outcomes, the methods and findings from studies using maltreatment type to predict adolescent substance abuse and then academic outcomes are described.

Taussig (2002) explored risk behaviors for substance use longitudinally during adolescence following physical abuse, sexual abuse, and neglect in a sample of youth in foster care. In a multiple regression analysis, her research indicated that a history of neglect was a significant predictor of substance use in youth, while physical abuse and sexual abuse were not (maltreatment categories not mutually exclusive). Conversely, in a community sample, Moran et al. (2004) found that all types of maltreatment (physical, sexual, emotional, and physical + sexual, maltreatment categories mutually exclusive) were associated with increased odds for substance use behavior in youth when compared to non-abused peers. Associations were strongest in youth who experienced both physical and sexual abuse when compared to just singular forms of abuse. Finally, research by Wall and Kohl (2007) in a nationally-representative sample of adolescents revealed that odds for substance use were lower in youth experiencing neglect than youth experiencing physical abuse. Although maltreatment types were mutually exclusive in this study, when a child experienced multiple maltreatment types, they were categorized as sexually abused if sexual abuse was their most severe abuse experience, and so on. Thus, the role of maltreatment type in predicting substance abuse in adolescents remains largely unclear. Whether or not maltreatment type is a useful predictor of substance abuse seems to depend at least in part on other dimensions of maltreatment, such as severity (Wall & Kohl, 2007) or the potential co-occurrence of other maltreatment types (Taussig, 2002).

Similarly, in work on maltreatment and academic outcomes, specifying maltreatment by type also has important implications regarding outcomes. A community-based longitudinal study evaluated the impact of childhood physical and sexual abuse (non mutually exclusive categories) on later academic outcomes (e.g., years of education completed, graduation rates) and previous childhood school performance (e.g., teacher and parent report of academic performance; Tanaka, Georgiades, Boyle, & MacMillan, 2015). Physical abuse was associated with problematic childhood and later academic performances, whereas sexual abuse demonstrated no such relation (Tanaka et al., 2015). Perez and Widom (1994) found that compared to a matched control group, adults with a neglect history significantly differed on overall IQ and reading ability, with a physical abuse history significantly differed on overall IQ, and with a sexual abuse history did not differ on cognitive abilities. Lastly, another prospective, longitudinal study compared academic outcomes in a group of young adults with a neglect history to a matched control group (Nikulina, Widom, & Czaja, 2011). In a multiple regression analysis, it was determined that a history of neglect was uniquely predictive of negative academic outcomes, such as cognitive ability and number of years completed in school. However, other types of maltreatment were not included in the analyses. Similar to the substance abuse findings, the association between maltreatment type and academic outcomes seems to depend on factors such as which other maltreatment types are included, and whether maltreatment types are mutually exclusive.

Given the discrepancies in the findings reviewed above, it is unsurprising that research has identified significant overlap across maltreatment types (Herrenkohl & Herrenkohl, 2009). Use of latent profile analysis has suggested this co-occurrence of maltreatment types within child may be highly meaningful for outcomes (e.g., Pears, Kim, & Fisher, 2008). Indeed, a recent study by Spinazzola et al. (2014) identified that psychological maltreatment accounts for a significant proportion of the variance in mental health outcomes for youth, above and beyond effects of physical and sexual abuse and thus, should be considered along with measures of physical and sexual abuse. In sum, given the multiple dimensions of maltreatment, the frequent overlap between maltreatment types, and the varied findings regarding how maltreatment type relates to outcomes in youth, a broad, more-inclusive approach to conceptualization of maltreatment that accounts for other dimensions (such as severity and chronicity) to use when examining the path between maltreatment and outcomes is warranted (Butt, Chou, & Browne, 2011; English, Upadhyaya et al., 2005).

While it seems intuitive that the severity of a maltreatment experience may be relevant for outcomes, the ability to classify severity of experience is confounded by a number of individual (e.g., salience and appraisal of event, age of victim) and situational characteristics (e.g., perpetrator type, type of maltreatment). One cannot simply ask a child, a case manager, or a caregiver if the child has experienced mild, moderate, or severe maltreatment, as the individual appraisal of these levels could be quite varied. Nor can one use the severity of the child’s injury or behavioral response to indicate severity of maltreatment, as children’s responses can vary greatly. Severity, then, often remains an unmeasured aspect of maltreatment. Fortunately, some models of classification of severity across maltreatment types have been created and have growing empirical support for their validity and reliability. One such example is the Maltreatment Classification System (Barnett, Manly, & Cicchetti, 1993), which was then updated with the Modified Maltreatment Classification System (MMCS, English and the LONGSCAN Investigators, 1997). This classification system has been used across many studies (e.g., Bolger & Patterson, 2001; Manly, Kim, Rogosch, & Cicchetti, 2001), and results indicate that severity, classified in a number of different ways, accounts for a significant portion of variance in outcomes. However, it is also important to consider severity in light of other maltreatment dimensions.

Litrownik et al. (2005) examined various approaches to defining maltreatment severity across types of maltreatment and within types of maltreatment. Their findings revealed that accounting for maximum severity of maltreatment experience within type of maltreatment offered the greatest predictive power for youth outcomes. A study by Clemmons, Walsh, DiLillo, and Messman-Moore (2007) identified maltreatment severity as a stronger predictor than number of maltreatment types experienced for later trauma symptomology. Moreover, a study of the effects of frequency and severity of maltreatment on behavioral and emotional outcomes in foster youth revealed that only severity of maltreatment significantly predicted externalizing problem behavior and adaptive behavior (Jackson, Gabrielli, Fleming, Tunno, & Makanui, 2014). Thus, severity of maltreatment appears to be an important component of youths’ maltreatment experiences, and it has significant import for youth behavioral and emotional outcomes following maltreatment exposure. In sum, severity contributes meaningful information above and beyond what can be accounted for by other maltreatment dimensions, such as frequency (Jackson et al., 2014) and type (Clemmons et al., 2007, Litrownik et al., 2005), and at the same time, the ways severity affects outcomes is dependent upon the valence of other maltreatment dimensions.

Recognition that the chronicity of adverse life events may be significant for youth outcomes has also garnered increasing support. Chronicity is likewise a difficult construct to operationalize due to the varied ways it can be measured and its overlap with the construct of dosage of maltreatment experience. For example, within maltreatment, chronicity could be the duration of a specific maltreatment experience, the proportion of the child’s life in which a type of maltreatment occurred, and the number of different maltreatment events experienced, to identify a few ways of defining it. There is little debate, however, that ongoing and repeated child maltreatment, as compared to the experience of a singular event, likely has greater negative impact on a child’s successful developmental progression (Copeland, Keeler, Angold, & Costello, 2007; Egeland, Sroufe, & Erickson, 1983).

Relatedly, dosing (or experiencing multiple types of maltreatment) of maltreatment experiences appears to matter across outcomes, and its impact can extend beyond the effects associated with experiencing a single maltreatment type, or the effects attributable to severity (e.g., Bolger & Patterson, 2001; Bolger, Patterson, & Kupersmidt, 1998). Flaherty et al. (2009) noted that exposure to five or more types of adversity (such as maltreatment experiences) in early childhood was associated with increased odds for health complaints and illness at age twelve. However, simple counts of dosing across maltreatment types do not account for other important aspects of the maltreatment experience (such as severity and chronicity). For example, against expectations, some research has shown greater negative findings for youth with neglect alone in relation to cognitive functioning when compared to youth with neglect and physical abuse exposure (O’Hara et al., 2015). Because other maltreatment dimensions were unexplored in this study, it remains unknown whether, for example, some youth with multiple maltreatment-type exposures had less severe or less chronic maltreatment experiences than youth with single maltreatment-type experiences.

In a study seeking to parse variance in outcomes based on different indicators of chronicity of maltreatment, the number of maltreatment incidents and a chronicity measure defined by calendar duration of maltreatment experiences accounted for the most variance across a range of outcomes (English, Graham et al., 2005). Therefore, in construction of indicators of chronicity of abuse, both the frequency of event occurrences and the number of different instances or forms of maltreatment within a broader categorization of type of maltreatment are important to consider, along with severity and type.

Youth age at the time they report their maltreatment experiences is important to take into account when defining maltreatment exposure. Generally, children under the age of three are particularly vulnerable to child abuse and neglect as these young children have the highest rates of maltreatment compared to other age ranges (US Department of Health and Human Services, 2016). Abuse during early childhood may be particularly detrimental due to its impact during sensitive developmental windows (e.g., Dunn et al., 2013); however, the accuracy of reports during this critical time period is limited. While young children who come into foster care may have suffered extremely detrimental abuse, younger children may be limited in their abilities to provide an accurate report of these experiences for a few reasons. Young children may provide reports of past abuse that are inaccurate, either due to poorly developed recall strategies (Greenhoot, 2011) or potentially due to a lack of understanding that they had indeed been abused or maltreated. Conversely, older youth may have had more opportunity for different types of maltreatment exposures and different levels of maltreatment severity. In addition, youth may be most affected by the abuse they have experienced most recently, making their reports of psychosocial problems a reflection of current distress. Further, older youth may be better at remembering and reporting their maltreatment histories, as they could have been told information about their early childhood, and they may have better memory capabilities for more recent maltreatment. Yet, these facilitative factors for reporting by adolescents are met with issues related to the potential unreliability of their retrospective reports of abuse that occurred in early childhood (e.g., Greenhoot, 2011). A final reason why the child’s age at the time of the report is important to consider is that the definitions of what constitutes as neglect change as a child ages. For example, leaving an 8-year-old unsupervised for several hours may be inappropriate, while leaving a 16-year-old for the same amount of time may not be (Stein, Rhys, Hicks, & Gorin, 2009). Thus, age is important to consider when trying to comprehensively assess maltreatment.

Prior research has identified gender differences in rates of various maltreatment types, as well as differences in the strength of associations between maltreatment and youth outcomes. Age and gender can also be interrelated regarding differences in rates of maltreatment. Based on national data, younger males (<1–5 years) were more likely to be victimized at higher rates compared to girls in the same age range. However, older females (i.e., 6–10 years and 11–17 years) demonstrated higher rates of child maltreatment compared to males in the same age range (US Department of Health and Human Services, 2016). Regarding type, females are more likely to experience sexual abuse than males (Hines & Finkelhor, 2007), and males are more likely to experience physical abuse than females (Thompson, Kingree, & Desai, 2004). Regarding outcomes, differences in neuroendocrine profiles were found amongst male and female 7- to 10-year-olds exposed to maltreatment, with females typically exhibiting profiles indicative of hypo-neuroendocrine response following chronic maltreatment exposure, and males exhibiting profiles indicative of hyper-neuroendocrine response (Doom, Cicchetti, Rogosch, & Dackis, 2013). In a longitudinal study of internalizing and externalizing problems in children exposed to maltreatment, Godinet, Li, and Berg (2014) found that males exhibited more internalizing and externalizing problems concurrently, while females’ problems exacerbated over time. Physical abuse, while more prevalent amongst males, may be more detrimental for women over time, as women with physical abuse histories are more likely to acquire a mental health problem in adulthood than men (Thompson, Kingree, & Desai, 2004). What remains unknown is how measurement of maltreatment may vary across male and female youth reporters.

Residential facility placements are typically reserved for youth with extreme behavioral issues or youth who were unable to remain safe and behaviorally controlled within a traditional foster home placement (Barth, 2002). Researchers have speculated that foster youth in residential care may have more severe abuse experiences, which may have resulted in more extreme behavioral and emotional problems (Baker, Kurland, Curtis, Alexander, & Papa-Lentini, 2007). For example, rates of child abuse and neglect are remarkably high amongst all children in residential treatment centers, with about half of children with physical abuse histories and a third with sexual abuse histories (Connor, Doerfler, Toscano, Volungis, & Steingard, 2004). In addition, residential placement may exacerbate behavioral and emotional problems. For example, some sites employ inflexible point-and-level systems that may inadvertently increase the behaviors they seek to ameliorate (e.g., Mohr, Martin, Olson, & Pumariega, 2009). While youth in foster care can transition in and out of residential facilities to more traditional foster homes, placement type remains an important factor to consider when evaluating measurement of maltreatment.

Variability in the strength and direction of associations between maltreatment and various outcomes may be explained, in part, by the failure of many measurement approaches to account for the complexity inherent within youths’ maltreatment experiences. For example, if two studies focused on outcomes related to physical abuse, with one study accounting for factors such as severity and co-occurrences of other of other types of abuse, and the other study ignoring these factors, results may lead to grossly different conclusions about the impact of physical abuse. In response to calls for more comprehensive measurement of maltreatment across its various dimensions (Manly, 2005), the present study seeks to test whether a coherent latent construct of maltreatment can be attained in a sample of youth in foster care while also accounting for various maltreatment dimensions and other important contextual factors. Specifically, the primary goal was to evaluate a latent measurement model of youth self-report of maltreatment, inclusive of severity and chronicity across types of maltreatment in a sample of youth in foster care. A second goal of the study was to evaluate whether the maltreatment measurement model would suffice across age groups, genders, and placement types. It was hypothesized that a two-factor latent model of severity of maltreatment and chronicity of maltreatment indicated by the various types of maltreatment (i.e., neglect and physical, sexual, and psychological abuse) would provide adequate fit to the data. Additionally, the best-fitting maltreatment model was expected to be invariant across age, gender, and foster placement type.

Section snippets

Participants

Participants were 500 youth between the ages of 8 and 18 years who were currently enrolled in foster care in an urban county in the Midwest. Participants were enrolled in the SPARK (Studying Pathways to Adjustment and Resilience) project, a longitudinal examination of resilience in youth exposed to child maltreatment. The gender divide of the sample was approximately equal (47% female), with an average age of 12.99 years (SD = 2.95 years). Youth were primarily Black/African American (45%), with

Results

Within this sample, youth demonstrated a wide range of maltreatment experiences, with the majority of youth reporting multiple types of maltreatment (92%; Mean = 2.96 and SD = 0.95) as well as a number of different types of experiences within maltreatment type. The most common form of maltreatment reported within this sample was psychological abuse, followed by physical abuse and neglect (see Table 1). All observed indicator scores were examined across age, gender, and placement groups (see Table 2

Discussion

The present study is one of the first examinations of a CFA of a latent measurement model of severity and chronicity of maltreatment across physical, sexual, and psychological abuse and neglect experiences in foster youth. As such this research moves the field of study on maltreatment closer to a comprehensive approach to measurement of maltreatment experiences in youth and offers a framework for future studies to build upon when examining shared variance across maltreatment types. Based on the

Disclosure

Drs. Gabrielli, Jackson, Tunno, and Hambrick report no biomedical financial interests or potential conflicts of interest.

References (47)

  • S.J. Kaplan et al.

    Child and adolescent abuse and neglect research: A review of the past 10 years. Part I: Physical and emotional abuse and neglect

    Journal of the American Academy of Child & Adolescent Psychiatry

    (1999)
  • A.J. Litrownik et al.

    Measuring the severity of child maltreatment

    Child Abuse & Neglect

    (2005)
  • J.T. Manly

    Advances in research definitions of child maltreatment

    Child Abuse & Neglect

    (2005)
  • R. Mills et al.

    Child maltreatment and adolescent mental health problems in a large birth cohort

    Child Abuse & Neglect

    (2013)
  • P.B. Moran et al.

    Associations between types of maltreatment and substance use during adolescence

    Child Abuse & Neglect

    (2004)
  • M. O’Hara et al.

    Children neglected: Where cumulative risk theory fails

    Child Abuse & Neglect

    (2015)
  • K.C. Pears et al.

    Psychosocial and cognitive functioning of children with specific profiles of maltreatment

    Child Abuse & Neglect

    (2008)
  • C.M. Perez et al.

    Childhood victimization and long-term intellectual and academic outcomes

    Child Abuse & Neglect

    (1994)
  • H. Taussig

    Risk behaviours in maltreated youth placed in foster care. A longitudinal study of protective and vulnerability factors

    Child Abuse & Neglect

    (2002)
  • A.J. Baker et al.

    Mental health and behavioral problems of youth in the child welfare system: Residential treatment centers compared to therapeutic foster care in the Odyssey Project population

    Child Welfare

    (2007)
  • D. Barnett et al.

    Defining child maltreatment: The interface between policy and research

  • R.P. Barth

    Institutions vs. foster homes: The empirical base for the second century of debate

    (2002)
  • K.E. Bolger et al.

    Developmental pathways from child maltreatment to peer rejection

    Child Development

    (2001)
  • Cited by (25)

    • Relationship between Adverse Childhood Experiences and substance use in youth offenders in Singapore

      2021, Child Abuse and Neglect
      Citation Excerpt :

      Two, is an establishment of a standardized retrospective classification for severity of maltreatment or ACEs. This is a huge undertaking and may involve the development of scales that accurately capture retrospective self-reports or conducting latent measurement modelling (Gabrielli, Jackson, Tunno, & Hambrick, 2017). Although a novel finding from this study is the characterization of the effects of frequency of ACEs on delinquent behaviors such as substance use, the effects of chronicity and severity of ACEs remain elusive.

    • Maltreatment and youth self-representations in residential care: The moderating role of individual and placement variables

      2020, Children and Youth Services Review
      Citation Excerpt :

      That responsiveness and safety in childhood predicts later social behavior and mental health (Stronach, Toth, Rogosch, Oshri, Manly, & Cicchetti, 2011). On the other hand, maltreated children and youth who were removed from their families and continue to experience relatively little emotional care in residential care facilities, have developmental problems of various kinds throughout life (Cho & Jackson, 2016; Gabrielli, Jackson, Tunno, & Hambrick, 2017; Jackson, Gabrielli, Fleming, Tunno, Makanui, 2014). Studies consistently identify that youth (i.e., young people between 10 and 24 years old) in residential care have more mental health problems and psychopathology than the general population.

    View all citing articles on Scopus

    This research was supported by funding from the National Institutes of Mental Health, RO1 grant MH079252-03 as well as funding from the National Institute on Drug Abuse (NIDA), F31 grant DA034423 awarded to Joy Gabrielli. The writing of this manuscript was supported in part by Joy Gabrielli’s participation in the NIDA T32 fellowship training grant DA037202.

    1

    Joy Gabrielli is now a postdoctoral research fellow at the Department of Biomedical Data Science, Geisel School of Medicine at Dartmouth, 1 Medical Center Drive, Lebanon, NH 03766, USA.

    2

    Angela Tunno is now a postdoctoral fellow at Department of Psychiatry and Behavioral Sciences, Duke University Medical Center, 2608 Erwin Road, Durham, NC 27705, USA.

    View full text