Maltreatment Histories of Foster Youth Exiting Out-of-Home Care Through Emancipation

Abstract

Little is known about maltreatment among foster youth transitioning to adulthood. Multiple entries into out-of-home care and unsuccessful attempts at reunification may nevertheless reflect extended exposure to chronic maltreatment and multiple types of victimization. This study used administrative data from the Illinois Department of Children and Family Services to identify all unduplicated allegations of maltreatment in a cohort of 801 foster youth transitioning to adulthood in the state of Illinois. A latent variable modeling approach generated profiles of maltreatment based on substantiated and unsubstantiated reports of maltreatment taken from state administrative data. Four indicators of maltreatment were included in the latent class analysis: multiple types of maltreatment, predominant type of maltreatment, chronicity, and number of different perpetrators. The analysis identified four subpopulations of foster youth in relation to maltreatment. Study findings highlight the heterogeneity of maltreatment in the lives of foster youth transitioning to adulthood and draw attention to a need to raise awareness among service providers to screen for chronic maltreatment and multiple types of victimization.

Keywords

maltreatment foster youth aging out emancipation latent class analysis

Each year more than 20,000 foster youth exit out-of-home care at the age of majority through emancipation or “aging out” (U.S. Department of Health and Human Services [DHHS], 2013a). Research finds that many foster youth face poor prospects during adulthood, including under education, criminal justice system involvement, and homelessness (Barth, 1990; Courtney, Piliavin, Grogan-Kaylor, & Nesmith, 2001; Dworsky & Courtney, 2009; Fowler, Toro, & Miles, 2009; Goerge et al., 2002; McMillen et al., 2005). Such outcomes are troubling because of the associated risks for lifetime health, economic, and social disadvantages and the subsequent costs that fall upon communities (White House Council for Community Solutions, 2012). Although late adolescence and early adulthood are considered critical transition years for altering behaviors associated with the lifetime risks (Dahl, 2004; Masten, Obradović, & Burt, 2006), the science on how to restore adaptive functioning and promote development remains incomplete. This knowledge gap hinders the ability of policy makers to plan and distribute resources to support this important group of young adults.

Experiences leading up to placement in out-of-home care is one area in need of greater attention (Mersky & Janczewski, 2012). The extent to which maltreatment and other types of adversity leave foster youth vulnerable to poor outcomes is poorly understood. Part of the problem is that studies have not typically included information on maltreatment or other family risks. One key limitation among the handful of studies that include indicators of maltreatment (Cusick, Havlicek, & Courtney, 2012; Hook & Courtney, 2011; Lee, Courtney, & Hook, 2012) is the use of variables that measure individual types of maltreatment even though exposure to a single type of maltreatment is rare (Clemmons, Walsh, DiLillo, & Messman-Moore, 2007; Higgins & McCabe, 1998, 2000; Lau et al., 2005; Manly, Kim, Rogosch, & Cicchetti, 2001). Such methodological limitations make understanding the effects of specific combinations of maltreatment and their timing difficult to isolate within individuals experiencing chronic maltreatment. Multiple types of victimization may also be more predictive of worse long-term adjustment and functional outcomes than single types of maltreatment (Anda, Dube, & Giles, 2006; Barnes, Noll, Putnam, & Trickett, 2009; Clemmons et al., 2007; English et al., 2005; Finkelhor, Ormrod, & Turner, 2007a; Finkelhor, Ormrod, Turner, & Holt, 2009; Finkelhor, Turner, Hamby, & Ormrod, 2011). Such dichotomous measures may therefore mask important variation that could enhance clinical interventions.

The lack of attention given to maltreatment in research is surprising, considering that the abuse histories of foster youth are believed to be extensive (Courtney & Heuring, 2005). Multiple entries into out-of-home care along with unsuccessful attempts at reunification may reflect prolonged exposure to intractable family problems and ecological risks. Recent work that has identified a high rate of exposure to traumatic events at the onset of adulthood heightens concerns about accumulating exposure to violence and victimization during out-of-home care (Salazar, Keller, Gowen, & Courtney, 2013). One study of 141 foster youth transitioning to adulthood in Wisconsin found that 33% of the sample self-reported neglect by a substitute caregiver and 17% self-reported being sexually assaulted by someone other than a substitute care provider during out-of-home care (Courtney et al., 2001). Several studies have identified a lifetime prevalence rate of posttraumatic stress disorder that is twice as high as same aged peers in the general population (Keller, Salazar, & Courtney, 2010; McMillen et al., 2005; Merikangas et al., 2010). These preliminary studies suggest maltreatment prior to and during out-of-home care warrants further study.

Under the Child and Family Services Improvement and Innovation Act (P.L. 112-34) of 2011, states are increasingly being encouraged to screen for emotional trauma associated with maltreatment and removal from home. Currently, a number of barriers impede trauma screening, including a lack of training of state and private agency caseworkers to effectively conduct screenings and incorporate the information into case planning (Conradi & Kisiel, 2013). It is unclear what kinds of efforts are being directed toward older adolescents who are without permanence. Both research and practice could benefit from identifying the extent of maltreatment experienced by this subgroup to augment existing efforts at the federal policy level to improve their prospects in adulthood.

This study builds on the limited research on maltreatment of foster youth transitioning to adulthood and seeks to inform understanding of the heterogeneity of maltreatment prior to and during out-of-home care. A latent variable modeling approach was used to generate profiles of maltreatment in a sample of foster youth based on substantiated and unsubstantiated allegations of maltreatment in state administrative data (Gibson, 1959; Goodman, 1974; Magidson & Vermunt, 2004; McCutcheon, 1987). The relationships between identified profiles of maltreatment and indicators of out-of-home care are additionally evaluated. The primary research question is whether distinct profiles of foster youth can be identified based on observed dimensions of allegations of maltreatment. If so, what do these profiles look like? A secondary question asks how these profiles are related to demographic characteristics and experiences in out-of-home care. The overarching goal of this study is to explore the extent, continuity, and types of maltreatment experienced by a sample of foster youth exiting care through emancipation.

Method

Sample and Procedures

The data for this study came from administrative records on child maltreatment investigations and out-of-home care placements from the Illinois Department of Children and Family Services. The study population included a cohort of foster youth who turned 18 in foster care during fiscal year 2008. The sample was selected from the state-integrated database based upon the following criteria: (1) in out-of-home care at age 17; (2) in care for at least 1 year; and (3) out-of-home care entry due to child abuse and/or neglect as opposed to child behavior or parent–child problems. These criteria identified 897 youth. From the full sample, 51 cases were omitted due to exit from out-of-home care in the year prior to age 18 and no indication of a subsequent reentry. Of these 51 cases, 22 exited through reunification, 13 exited through adoption or guardianship, and 11 exited for other reasons. An additional 45 cases were omitted due to missing maltreatment records, leaving a total of 801 foster youth in the study sample.

These 801 youth were the subject of 9,041 allegations of maltreatment, including 51.8% that were substantiated and 48.6% that were unsubstantiated prior to and during any time in out-of-home care. Substantiated and unsubstantiated reports were not differentiated in the current study because evidence suggests differences between substantiated and unsubstantiated cases may not be very large with respect to the recurrence of maltreatment (Drake, Jonson-Reid, Way, & Chung, 2003) or risk to children (Dubowitz & Bennett, 2007; English, Marshall, Coghlan, Brummel, & Orme, 2002; Jonson-Reid, Drake, Kim, Porterfield, & Han, 2004; Kohl, Jonson-Reid, & Drake, 2009).

The sample comprises more males (55.3%) than females (44.7%). A majority of the sample was African American (71.2%), followed by Caucasian (23.2%) and other race or ethnicity (5.6%), which is consistent with the racial makeup of children placed in the child welfare system in Illinois overall. Nearly 61.8% of the cases originated in one urban region (n = 495), with 15.7% in the central region, 11.4% in the northern region, 8.1% in southern region, and 3.0% out of state. The mean age at first allegation of maltreatment was 5.4 years. The mean age at first entry into out-of-home care was 8.2 years (M = 8.23, standard deviation [SD] = 4.82). More than one quarter of the sample (28.5%) had two or more entries in out-of-home care.

Measures

Research on child maltreatment has increasingly drawn attention to the need to conceptualize and measure abuse and neglect as multidimensional experiences (English et al., 2005; Higgins & McCabe, 1998, 2000, 2001; Lau et al., 2005; Manly et al., 2001; Wolfe & McGee, 1994). The analysis examined four indicators of allegations of maltreatment, substantiated and unsubstantiated, prior to, or during any spell in out-of-home care.

Number of types of maltreatment

A categorical variable indicates whether there were “one to two” types of maltreatment; “three to four” types of maltreatment, or “five to eight” types of maltreatment reported in each case history. There were eight possible types of allegations in the data: sexual abuse, physical abuse, neglect by lack of supervision, neglect by failure to provide, environmental neglect, substance exposed infant, emotional abuse, and risk of harm.

Predominant type of maltreatment

A categorical variable for predominant type of maltreatment reflects whether the more frequently reported type of maltreatment involved (1) sexual abuse, (2) physical abuse, or (3) neglect. To create the measure, the three indicators for neglect (i.e., lack of supervision, failure to provide, and environmental neglect) were combined to create a variable for any neglect. Next, the remaining types of maltreatment within a case record were counted. The type of maltreatment with the highest frequency was coded as the predominant type of maltreatment. There were 101 cases (12.6%) in which the types of maltreatment overlapped and a predominant type of maltreatment could not be identified. In 58 of these cases, predominant type of maltreatment type was coded as sexual abuse, physical abuse, or any neglect if these types of maltreatment were tied with a corresponding risk of harm. In 40 cases, the frequency of sexual abuse, physical abuse, or any neglect allegations were tied. These were coded as sexual abuse if the frequency of sexual abuse was tied with physical abuse or any neglect and coded as physical abuse if the frequency of physical abuse was tied with any neglect. There were three cases in which the frequency of emotional abuse or substance-exposed infant was tied with risk of harm for neglect. These cases were coded as any neglect.

Chronicity

A categorical variable for chronicity was based on five child developmental stages ranging from birth to age 17. These include (1) birth (0 to 1 years), (2) toddlerhood (2 years); (3) early childhood (3 to 5 years); (4) middle childhood (6 to 11 years), and (5) adolescence (12 to 17 years). Chronicity was coded as situational if one or more allegations were reported in a single developmental period, limited if allegations of maltreatment occurred in any two developmental periods, and extended if one or more allegations of maltreatment were reported in any three or more developmental periods.

Number of unique perpetrators

A categorical variable represents the number of unique perpetrators identified in the administrative data. Fifteen types of alleged perpetrators existed in the administrative data: (1) biological parent (mother and/or father), (2) step-parent, (3) paramour (i.e., an unmarried significant other of biological parent), (4) nonrelated parent substitute, (5) aunt/uncle, (6) grandparent, (7) sibling, (8) school staff, (9) day-care provider, (10) health provider, (11) youth recreation staff, (12) foster or adoptive parent, (13) group care provider, (14) relative care provider, and (15) other/unknown. The range of possible perpetrators for this study was from 1 to 9 and was coded as “1” for one type of perpetrator, “2” for two types of perpetrators, and “3 or more” for three or more distinct perpetrators in a case history.

Analytic Strategy

Latent class analysis (LCA) was used to determine whether multiple dimensions of maltreatment could be used to classify profiles of maltreatment among foster youth exiting care through emancipation. LCA attempts to segregate the data into classes within which the observed variables used to measure maltreatment histories are statistically independent of one another. In the current analysis, LCA was used to identify the smallest number of classes required to account for the estimated associations between the observed variables. Latent Gold 4.5 (Vermunt & Magidson, 2008) was used to conduct all analyses. Latent Gold employs an iterative maximum likelihood procedure to optimize classification of individual cases into latent classes (Goodman, 1974; Vermunt & Magidson, 2005).

Decisions about model evaluation and selection were based on a combination of statistical criteria, parsimony, and interpretability (Magidson & Vermunt, 2004). A variety of tools can be used to assess the model fit to the data. The Latent Gold program uses the likelihood ratio χ² statistic L ² to indicate how well a latent class model fits the observed data. In situations involving sparse data where the average expected cell count is small, the bootstrap approach can be used to estimate the p value (Magidson & Vermunt, 2004; Vermunt & Magidson, 2005). The bootstrap p value is the proportion of the replication samples with a higher L ² than in the original sample (Vermunt & Magdinson, 2005).

An additional approach to assessing model fit when sparse data exist assesses the fit of the model using an information criterion weighing model fit and parsimony. The principal of parsimony states that all else being equal, models with fewer parameters are preferred to more complex models. The most widely used are the Akaike’s information criterion (AIC; Akaike, 1987) and the Bayesian information criterion (BIC; Schwarz, 1978). As penalized fit statistics, the AIC and BIC impose a penalty based on the number of parameters estimated. For all information criterions, a smaller value indicates a better fit or a more optimal balance between model fit and parsimony. However, because of various penalties associated with each criterion, they often do not identify the same model as optimal and thus are best for narrowing down a set of options (Collins & Lanza, 2010).

An informal approach for assessing model fit compares the L ² associated with the LCA models with the baseline value L ² (H₀) to ascertain the percentage reduction in L ² (Magidson & Vermunt, 2004). This approach complements the more statistically precise L ² and BIC/AIC approaches. The final model should have high classification quality and practical utility, which is assessed in terms of adequate class sizes and theoretical interpretability. LCA yields posterior probabilities that indicate the likelihood of membership of every case in each of the classes. These probabilities of latent class membership sum to 1 (within rounding error). Cases are then assigned to the class for which they had the highest probability.

In order to verify that the empirically derived latent classes are not products of the indicators used in the LCA, validation analyses include some variables similar to the indicators used for classification (Keller, Cusick, & Courtney, 2007). These validation variables are selected as indicators of several domains considered to be relevant to understanding the latent classes. If the empirically derived latent classes are different, a consistent pattern of differences should be apparent in the majority of validation variables. If statistical significance is observed in the majority of variables, the classification procedure is considered to have high validity.

Results

Model Selection and Case Assignment

The model fit statistics for six different models are presented in Table 1. The fit statistics suggested that at least four classes are necessary to provide a good fit to the data. As shown in Table 1, the p value was less than .05 for the one- (p < .0001), two- (p < .001), and three-class (p < .002) models, thereby rejecting the null hypothesis of model and data equivalence. Among models for which the p value is greater than .05 (i.e., four-, five-, and six-class models), the best fitting model is the four-cluster model (p value = .78 and number of parameters = 26). Because the total number of cells, or the unique combinations of observed variable category values, is large compared to the sample size, the quality of the model fit is explored in greater detail by deriving nonparametric bootstrap p values (Magidson & Vermunt, 2004). As is observed in Table 1, the least discrepancy between the model and the data occurs in the four-class solution with the smallest p value. The AIC and the BIC are also lowest for the four-class model. Latent Gold permits a conditional bootstrap option that tests whether a more restricted model provides a statistically significant improvement in model fit (Vermunt & Magidson, 2005). If the difference in model fit between the two models is insignificant, then the more restricted model is preferable because it is more parsimonious. The p value associated with the conditional bootstrapped estimate of the increase in classes from a four-class model to a five-class model is .25 with a standard error of 0.02. Since p > .05, this means that the five-class model does not provide a significant improvement over the four-class model. Therefore, the four-class model is preferred because it provides the greatest certainty in classification and yields classes that are substantively informative.

Table 1.

Model-Fit Statistics Comparisons.

Model	L ²	df	# Parameters	p Value	Bootstrap p value	BIC_LL	AIC_LL	% Reduction in L ²
One-class	649.13	72	8	<.001	<.001	6,564.97	6,527.48	—
Two-class	157.57	66	14	<.001	<.001	6,113.53	6,047.92	75.7
Three-class	96.84	60	20	.002	.004	6,092.91	5,999.19	85.1
Four-class	45.71	54	26	.78	.78	6,081.89	5,960.06	92.9
Five-class	35.22	58	32	.91	.91	6,111.52	5,961.58	94.6
Six-class	30.49	42	38	.91	.77	6,146.90	5,968.84	95.3

Note. BIC = Bayesian information criterion; AIC = Akaike’s information criterion.

For each individual case, LCA yields posterior probabilities that indicate the likelihood of membership in each of the classes. Each individual in the sample is assigned to a specific maltreatment class for which the latent class model-based posterior probability is highest (modal class).

Class Description and Profiles

Parameter estimates for the four-class model are shown in Table 2. The first row shows the relative proportion of the full sample in each class. The remaining rows present the probabilities of the variables used in the LCA.

Table 2.

Profiles of Latent Classes on Classification Variables.^a

	Class 1	Class 2	Class 3	Class 4
	Chronic maltreatment	Predominant abuse	Situational maltreatment	Predominant neglect
Proportion of sample	.376	.269	.195	.159
Multiple type of maltreatment (1–8)
1–2 types	.017	.244	.838	.289
3–4 types	.418	.656	.160	.632
5+ types	.565	.100	.002	.079
Predominant type of allegation
Sexual abuse	.160	.284	.170	.001
Physical abuse	.202	.271	.124	.081
Neglect	.638	.445	.706	.918
Chronicity (1–5)
One developmental period (“Situational”)	.002	.349	.826	.023
Two developmental periods (“Limited”)	.195	.610	.172	.523
Three+ developmental periods (“Extended”)	.803	.041	.002	.454
Number of perpetrators (1–9)
One person	.001	.014	.756	.419
Two people	.089	.353	.239	.532
Three+ people	.910	.633	.005	.049

^aThe values in Table 2 represent the proportion of each class.

Latent class 1

The largest class, representing 37% of the sample, is marked by extensive involvement with child protective services for multiple types of maltreatment. The foster youth in this group are characterized by high (>.5) probability of having allegations of five or more distinct types of maltreatment. The modal type of predominant maltreatment is neglect. The foster youth in this class are further differentiated from the other classes by high probability of extended chronicity (i.e., allegations of maltreatment occur in three or more developmental periods). The members in Class 1 are further characterized by high probability of having allegations committed by three or more different perpetrators. This suggests that Class 1 is best labeled as “chronically maltreated.”

Latent class 2

The next largest class, representing 26% of the sample, is distinguished by predominant type of allegation of maltreatment. The foster youth in Class 2 are characterized by a high probability of predominant sexual abuse and physical abuse among substantiated and unsubstantiated allegations. Although the probability of neglect in Class 2 is nontrivial, it is lower than any of the other three classes (<.5). Members in Class 2 are also characterized by high probability of having allegations representing three to four distinct types of maltreatment. Members in Class 2 (M = 1.63, SD = 0.48) are lower than the sample average (M = 2.33, SD = 1.12) with respect to chronicity. Although the mode is three or more distinct perpetrators in allegations, the remainder of the distribution is skewed toward two perpetrators. Highest probability of predominant sexual or physical abuse suggests that Class 2 is best labeled as “predominant abuse.”

Latent class 3

The third class, representing 19% of the sample, is distinguished from the other classes by having the lowest probabilities of maltreatment in three of the four observed dimensions. For instance, the members in Class 3 are characterized by the highest probability of experiencing one to two types of maltreatment as compared to the other classes. The average number of maltreatment types observed in Class 3 (M = 1.67, SD = 0.71) is half that of the full sample (M = 3.46, SD = 1.49). The modal type of predominant abuse is neglect, with the remainder skewed toward sexual abuse. In fact, this profile has the second highest probability of predominant neglect, behind Class 4 (see subsequently). This profile (M = 1.19, SD = 0.71) is also less than the sample average (M = 2.89, SD = 1.62) with respect to having an allegation of maltreatment perpetrated by different types of people. Members in Class 3 also have the highest probability across each of the classes in experiencing an allegation in one developmental period. This suggests that Class 3 is best labeled “situational maltreatment.”

Latent class 4

The smallest class, representing 15% of the sample, is distinguished by having the highest probability of predominant neglect. The probabilities of predominant sexual and physical abuse are lowest in Class 4 when compared to the other classes. Although the mode is three to four types of maltreatment, the remainder is mostly skewed toward one or two types of maltreatment. This profile (M = 2.52, SD = 0.72) is close to the sample average (M = 2.33, SD = 1.62) with respect to chronicity. The average number of perpetrators is lower for the members in this profile (M = 1.66, SD = 0.47) than the full sample average (M = 2.89, SD = 1.62). For these reasons, Class 4 is best labeled as “predominant neglect.”

Validation of Classes

Table 3 presents all variables considered for validation purposes and details how the four classes compare. The table also presents F values for tests of independence and results of Tukey’s (honest significant difference) post hoc test procedures to assess statistically significant pairwise comparisons. As shown in Table 3, lower proportions of African American youth are found in the chronically maltreated and predominant abuse classes than the situational maltreatment and the predominant neglect classes. Although youth from all regions of the state are represented in all of the classes, higher proportions of youth from an urban region (Cook County) are found in the predominant neglect class than in the chronically maltreated class.

Table 3.

Comparisons Across the Latent Classes on Variables From Relevant Domains.

	Class 1	Class 2	Class 3	Class 4
	Chronically maltreated (n = 290)	Predominant abuse (n = 210)	Situational maltreatment (n = 155)	Predominant neglect (n = 146)	Total sample (N = 801)	F or χ²statistic	Significant comparisons
Demographic characteristics
Gender (male)	53.8%	54.8%	51.0%	63.7%	55.3%	1.88
Race						30.18***
African American	64.1%	67.1%	78.1%	83.6%	71.2%		3,4 > 1; 4 > 2
White	30.3%	23.3%	17.4%	15.1%	23.2%		1 > 3,4
Other	5.5%	9.5%	4.5%	1.4%	5.6%		2 > 4
Region (YY)	55.5%	62.4%	65.2%	69.9%	61.8%		4 > 1
Maltreatment history
Age at first report (years)	2.5 (2.5)	7.3 (3.9)	9.5 (4.8)	3.8 (2.9)	5.4 (4.5)	163.19***	3 > 1,2,4; 2 > 1,4; 4 > 1
Total # allegations (1–48)	18.8 (10.6)	8.6 (5.5)	3.1 (2.3)	8.8 (5.4)	11.3 (9.6)	177.92***	1 > 2,3,4; 3 > 2,4
# Sexual abuse	1.8 (2.8)	1.1 (1.8)	0.2 (0.6)	0.1 (0.3)	1.0 (2.0)	37.94***	1 > 2,3,4; 2 > 3,4;
# Physical abuse	2.4 (3.1)	1.2 (1.8)	0.3 (0.7)	0.4 (0.9)	1.3 (2.3)	47.32***	1 > 2,3,4; 2 > 3,4
# Neglect	8.7 (6.8)	3.1 (3.5)	1.8 (2.1)	6.0 (4.6)	5.4 (5.7)	84.48***	1 > 2,3,4; 4 > 2,3
# Lack of supervision	3.8 (3.0)	1.6 (2.2)	1.2 (1.6)	2.9 (2.2)	2.6 (2.6)	50.96***	1 > 2,3,4; 4 > 2,3
# Environmental neglect	3.8 (4.7)	1.1 (1.9)	0.4 (0.9)	2.7 (3.5)	2.2 (3.6)	43.27***	1 > 2,3,4; 4 > 2,3
# Other neglect	1.1 (2.1)	0.3 (0.9)	0.1 (0.5)	0.4 (0.8)	0.6 (1.5)	18.99***	1>2,3,4
# Substance-exposed infant	0.3 (0.6)	0.1 (0.5)	0.0 (0.2)	0.2 (0.4)	0.2 (0.5)	7.36***	1 > 2,3; 4 > 3
# Emotional	0.3 (0.9)	0.0 (0.2)	0.0 (0.2)	0.1 (0.6)	0.1 (0.6)	10.49***	1 > 2,3,4
# Risk of harm (any)	5.4 (4.3)	3.0 (3.0)	0.7 (1.1)	2.0 (1.9)	3.2 (3.6)	85.46***	1 > 2,3,4; 2 > 3,4; 4 > 3
Any founded allegation by substitute care provider	18.3%	14.3%	1.3%	3.4%	11.2%		1,2 > 3,4
Out-of-home care
Age at first entry (years; 0–15)	7.1 (4.9)	9.0 (4.5)	9.6 (4.8)	8.0 (4.5)	8.2 (4.8)	10.52***	1 > 2,3; 3 > 4
Any reentry	36.2%	22.9%	18.1%	32.2%	28.5%		1 > 2,3; 4 > 3
# Spells (1–5)	1.5 (0.8)	1.2 (0.5)	1.2 (0.5)	1.4 (0.7)	1.4 (0.7)	9.77***	1,4 > 2,3;
First placement type						18.50
Relative	31.7%	27.6%	39.4%	36.3%	33.0%
Foster	32.8%	29.5%	20.6%	23.3%	27.8%		1 > 3
Congregate	16.9%	16.2%	16.8%	15.8%	16.5%
Shelter	9.0%	15.7%	14.8%	16.4%	13.2%
Adverse	9.7%	11.0%	8.4%	8.4%	9.5%
Length at first placement (days; 1–6,436)	291 (595)	262 (604)	557 (1,119)	302 (708)	337 (753)	5.68**	3 > 1,2,4
Any siblings in care	74.5%	64.3%	54.2%	72.6%	67.5%		1 > 3

With respect to the maltreatment variables, findings suggest distinct differences across the classes in relation to the age at first allegation reported, total number of allegations, types of allegations, and relationship to the alleged perpetrator. These differences generally validate the classification scheme. Consistent with their profile for having the most extensive maltreatment experiences, members of the chronically maltreated class are the most likely to have a first allegation of maltreatment at the youngest age. The members of the chronically maltreated class also had the highest frequency of allegations of maltreatment, followed by the predominant neglect class, with approximately half as many allegations on average. A pattern is also evident in the total types of maltreatments. The chronically maltreated class was the most likely to have a substantiated allegation of maltreatment by a substitute caregiver (i.e., foster parent, residential care provider, or a licensed relative), followed by the predominant abuse class.

Several out-of-home care variables varied according to class. Members in the chronically maltreated class first entered out-of-home care at the youngest age across each of the classes, followed by the predominant neglect class. With respect to reentry into out-of-home care, the situational maltreatment and predominant neglect classes were less likely than the chronically maltreated and the predominant abuse classes to reenter out-of-home care. The chronically maltreated class was more likely to be first placed into a traditional foster home than the situational maltreatment class. It is interesting to note that the situational maltreatment class had a substantially longer duration in their first placement when compared to the other classes.

Discussion

This study is the first to examine variation in maltreatment histories among a sample of foster youth exiting out-of-home care through emancipation in one state. The importance of knowing more about differences between foster youth who suffered situational experiences of maltreatment and those that experienced more of a chronic pattern is made clearer. The chronicity of allegations in the lives of the young people in this sample nevertheless underscores the years that many of these individuals spent in challenging circumstances. Although these findings raise important issues when it comes to assessment and interventions within child welfare systems, the limits of this study’s findings should be considered carefully.

The primary reason for caution in interpreting this study’s findings has to do with limitations in official maltreatment data. These data are based on the reports to child protective services and may not represent a child’s entire maltreatment history. Youth may have had experiences of maltreatment that were not reported to protective agencies; therefore, the information in this study may not provide an accurate picture of the full extent of maltreatment. A related limitation has to do with the specific definitions used in this study to create indicators of maltreatment experiences. It is possible that different indicators of maltreatment would result in different classifications. For instance, one indicator that was not available in the administrative data was severity of maltreatment. This information would arguably strengthen understanding of the characterization of maltreatment events that have been alleged. Nationwide variation in the definition of child abuse and neglect may also limit the applicability of this study’s findings to other states. The state analyzed in this study has one of the lowest rates of emotional abuse compared to other states in the United States (Hamarman, Pope, & Czaja, 2002). Any irregularities in reporting certain types of maltreatment would be expected to influence a study’s results. With these limitations in mind, several findings warrant further examination.

That the largest class experienced a pattern of recurring allegations of maltreatment is troubling. The young people in the chronically maltreated class came to the attention of child protective services early, and allegations of multiple types of maltreatment by different perpetrators accumulated prior to and during out-of-home care. Although any report of maltreatment introduces risks to child safety and well-being, the cumulative impact of multiple types of maltreatment is expected to predict worse outcomes across a number of domains in childhood (English et al., 2005; Higgins & McCabe, 1998, 2000, 2001; Lau et al., 2005) and adulthood (Currie & Widom, 2010; Jonson-Reid, Kohl, & Drake, 2012). Of concern is a pattern of child maltreatment and exposure to family problems that may lead to increased vulnerability for a broad range of future victimizations within and outside of families (Finkelhor, Ormrod, & Turner, 2007b). This suggests that efforts to identify youth and decrease their vulnerability to other types of victimization should be an important priority of any child welfare intervention.

The data are unable to explain why members in this class experienced such a high rate of allegations of maltreatment. A few speculations are offered. The vast majority of the allegations are related to neglect. Child protective agencies may not place children reported for neglect into out-of-home care until after allegations accumulate and/or there is substantial evidence of serious risk of harm to a child. These cases could represent low-level neglect cases that required multiple reports to get enough attention from child protective agencies to warrant intervention. It could also be that there was not enough evidence to substantiate allegations of maltreatment. Because substantiation has historically been used as a gateway to service provision, it may be that these cases were investigated many times before they were opened for services. Regardless of the reasons for chronic maltreatment, the findings from this study call into question the validity of decisions that leave such young children to accumulate multiple referrals for maltreatment. Targeted prevention that attempts to better link child protective systems with family service systems when early concerns about neglect cannot be substantiated may hold promise for improving protection of at-risk children (Daro & Dodge, 2009).

The predominant abuse class reflects predominant physical and sexual abuse, although the co-occurrence of abuse with other types of maltreatment is also prevalent. Not all children who have been abused experience symptoms of trauma. However, physical and sexual victimization are particularly high-impact forms of child victimization. The high rate of victimization in out-of-home care that is experienced by the members of this class may also mean that predominant abuse heightens exposure to other types of victimization. The findings from this study suggest that clinicians providing services to foster youth with predominant histories of abuse should also assess for other types of maltreatment and for traumatic stress. Further research is needed to determine whether the patterns of co-occurrence of specific forms of victimization within predominant victimization types lead to varied diagnostic profiles, impairments, and treatment outcomes (Ford, Wasser, & Connor, 2011).

The situational maltreatment class is in many ways different from the other classes. The individuals in this class came to the attention of child protective agencies relatively late when compared to the other classes, and their first entry into out-of-home care was also later. Those in the situational maltreatment profile also had the lowest number of allegations and the lowest rate of previous entry into out-of-home care, on average. Given how different this profile is from the other profiles, it may be comprised, to a large extent, of older youth who bring significant emotional and behavioral difficulties into the child welfare system under the category of neglect. This is a group that shows up in other studies, including National Survey of Child and Adolescent Well-Being, where many teens entering out-of-home care appear to have had limited maltreatment histories (U.S. DHHS, 2008). They may also be a group that is especially challenging to reunify with their families. Although the members in this class are expected to have significant emotional and behavioral needs, research suggests that many may not receive services comprehensive enough to sufficiently restore their functioning (U.S. DHHS, 2008) or prepare them for adulthood (U.S. Government Accounting Office, 2008).

The smallest class, representing predominant neglect, suggests continuous exposure to multiple types of neglectful experiences. Like the chronically maltreated class, the members in this profile came to the attention of child protective agencies early and, on average, experienced recurring allegations of maltreatment. Although maltreated children in general perform more poorly in school, neglected children appear to have the highest risk of developmental delays and academic challenges (Cicchetti & Valentino, 2006; Kendall-Tackett & Eckenrode, 1996; Manly, Lynch, Oshri, Herzog, & Wortel, 2013). There are surprisingly few empirically supported prevention interventions specifically targeting children who have experienced extensive neglect even though neglect remains the predominant reason leading to placement into out-of-home care for the majority of children. Some evidence suggests that specific types of play therapy and therapeutic day care or preschool programs may have benefits for some neglected children (Allin, Wathen, & MacMillan, 2005). The members in this class may also benefit from comprehensive assessments of developmental needs, often and early, and a strategic approach to promoting developmental milestones during out-of-home care (Masten, 2006).

One conclusion from this study’s findings is the need to monitor state performance in relation to reducing the incidence of child abuse and neglect in out-of-home care for older adolescents without permanence. Although reducing the incidence of child abuse and/or neglect in out-of-home care currently reflects a national outcome measure, the findings from this study suggest that older adolescents in out-of-home care may be at heightened risk for recurring victimization during child welfare system involvement. That 1 of the every 10 foster youth in this sample had a substantiated allegation of maltreatment by a substitute care provider suggests a need to redouble efforts to support, train, and monitor out-of-home caregivers. Although the rate of abuse in out-of-home care is not as high as the rate reported by Courtney, Piliavin, Grogan-Kaylor, and Nesmith (2001) or Pecora et al. (2005), it is substantially higher than the national incidence rate reported by the U.S. DHHS (2013b). Specific training for foster parents that helps them provide nurturing care to children who exhibit challenging behaviors may be helpful for developing and sustaining supportive relationships (Dozier et al., 2009).

The findings of this study confirm that multiple types of maltreatment are common among this sample of foster youth exiting care through emancipation. Less than 10% of the sample had a single type of maltreatment. Because experiencing multiple types of maltreatment predicts long-term adjustment (Higgins & McCabe, 2000) and functional outcomes (Smith & Thornberry, 1995), it is likely that independent living programming for foster youth transitioning to adulthood would benefit from incorporating more of a maltreatment or trauma-informed perspective. Such a perspective would draw from empirically informed and empirically validated approaches to psychological intervention for psychological trauma, such as trauma-focused cognitive behavioral therapy (Cohen, Mannarino, & Deblinger, 2006), trauma affect regulation guide for education and therapy (Ford & Russo, 2006), and integrative treatment of complex trauma (Briere & Lanktree, 2011), to provide education to foster youth about the effects of maltreatment and/or traumatic experiences; to normalize a need for help seeking; and to prevent victimizations in adulthood. However, the extent to which independent living programs integrate such approaches into services is unclear. Increasing knowledge of how independent living service providers address needs that may fall outside of the traditional services that are offered (i.e., vocational training) is needed. Orienting independent living programs to maltreatment and trauma has the potential to promote coping strategies and help-seeking behaviors that encourage youth to take advantage of support that may be expanding during the transition to adulthood.

Footnotes

Acknowledgment

The author would like to thank Mark E. Courtney and Theodore P. Cross for their helpful comments on multiple aspects of this study, including interpretation of the study findings and comments on previous drafts of the article.

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

Akaike

(1987). Factor analysis and AIC. Psychometrika, 52, 317–332.

Allin

Wathen

C. N.

MacMillan

(2005). Treatment of child neglect: A systematic review. Canadian Journal of Psychiatry, 50, 497–504.

Anda

R. F.

Dube

S. R.

Giles

(2006). The enduring effects of abuse and related adverse experiences in childhood. European Archives of Psychiatry and Clinical Neuroscience, 256, 174–186.

Barnes

J. E.

Noll

J. G.

Putnam

F. W.

Trickett

P. K.

(2009). Sexual and physical re-victimization among victims of severe childhood sexual abuse. Child Abuse & Neglect, 33, 412–420.

Barth

R. P.

(1990). On their own: The experiences of youth after foster care. Child and Adolescent Social Work Journal, 7, 419–440.

Briere

J. N.

Lanktree

C. B.

(2011). Treating complex trauma in children and youth adults. Thousand Oaks, CA: Sage.

Cicchetti

Valentino

(2006). An ecological-transactional perspective on child maltreatment: Failure of the average expectable environment and its influence on child development. Developmental Psychopathology, 3, 129–201.

Clemmons

J. C.

Walsh

DiLillo

Messman-Moore

T. L.

(2007). Unique and combined contributions of multiple child abuse types and abuse severity to adult trauma symptomatology. Child Maltreatment, 12, 172–181.

Cohen

J. A.

Mannarino

A. P.

Deblinger

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

10.

Collins

L. M.

Lanza

S. T.

(2010). Latent class and latent transition analysis with applications in the social, behavioral, and health sciences. Hoboken, NJ: John Wiley.

11.

Conradi

Kisiel

(2013). Trauma screening within the child welfare system. In T. LaLiberti & Tracey Crudo (Eds.), CW360: A Comprehensive Look at a Prevalent Child Welfare Issues, Winter (p.21). St. Paul: University of Minnesota, School of Social Work.

12.

Courtney

M. E.

Heuring

D. H.

(2005). The transition to adulthood for youth “aging out” of the foster care system. In Wayne Osgood

Michael Foster

Flanagan

Ruth

(Eds.), On your own without a net: The transition to adulthood for vulnerable populations (pp. 27–67). Chicago, IL: The University of Chicago Press.

13.

Courtney

M. E.

Piliavin

Grogan-Kaylor

Nesmith

(2001). Foster youth transitions to adulthood: A longitudinal view of youth leaving care. Child Welfare, 80, 685–718.

14.

Currie

Widom

C. S.

(2010). Long-term consequences of child abuse and neglect on adult economic well-being. Child Maltreatment, 15, 111–120.

15.

Cusick

G. R.

Havlicek

J. R.

Courtney

M. E.

(2012). Risk for arrest: The role of social bonds in protecting foster youth making the transition to adulthood. American Journal of Orthopsychiatry, 82, 19–31.

16.

Dahl

R. E.

(2004). Adolescent brain development: A period of vulnerabilities and opportunities. Keynote address. Annals of the New York Academy of Sciences, 1021, 1–22.

17.

Daro

Dodge

K. E.

(2009). Creating community responsibility for child protection: Possibilities and challenges. Future of Children, 19, 67–93.

18.

Dozier

Lindhiem

Lewis

Bick

Bernard

Peloso

(2009). Effects of a foster parent training program on young children’s attachment behaviors: Preliminary evidence from a randomized clinical trial. Child and Adolescent Social Work Journal, 26, 321–332.

19.

Drake

Jonson-Reid

Way

Chung

(2003). Substantiation and recidivism. Child Maltreatment, 8, 248–260.

20.

Dubowitz

Bennett

(2007). Physical abuse and neglect of children. The Lancet, 369, 1891–1899.

21.

Dworsky

Courtney

M. E.

(2009). Homelessness and the transition from foster care to adulthood. Child Welfare, 88, 23–56.

22.

English

D. J.

Graham

J. C.

Litrownik

A. J.

Everson

Bangdiwala

S. I.

(2005). Defining maltreatment chronicity: Are there differences in child outcomes? Child Abuse & Neglect, 29, 575–595.

23.

English

D. J.

Marshall

D. B.

Coghlan

Brummel

Orme

(2002). Causes and consequences of the substantiation decision in Washington state child protective services. Children and Youth Services Review, 24, 817–851.

24.

English

D. J.

Upadhyaya

M. P.

Litrownik

A. J.

Marshall

J. M.

Runyan

D. K.

Graham

J. C.

Dubowitz

(2005). Maltreatment's wake: The relationship of maltreatment dimensions to child outcomes. Child Abuse & Neglect, 29, 597–619.

25.

Finkelhor

Ormrod

R. K.

Turner

H. A.

(2007a). Poly-victimization: A neglected component in child victimization. Child Abuse & Neglect, 31, 7–26.

26.

Finkelhor

Ormrod

R. K.

Turner

H. A.

(2007b). Re-victimization patterns in a national longitudinal sample of children and youth. Child Abuse & Neglect, 31, 479–502.

27.

Finkelhor

Ormrod

Turner

Holt

(2009). Pathways to poly-victimization. Child Maltreatment, 14, 316–329.

28.

Finkelhor

Turner

H. A.

Hamby

S. L.

Ormrod

(2011). Polyvictimization: Children's exposure to multiple types of violence, crime, and abuse. washington, D.C.: U.S. Department of Justice, Office of Justice Programs, Office of Juvenile Justice and Delinquency Prevention.

29.

Ford

J. D.

Russo

(2006). Trauma-focused, present-centered, emotional self-regulation approach to integrated treatment for posttraumatic stress and addiction: Trauma Adaptive Recovery Group Education and Therapy (TARGET). American Journal of Psychotherapy, 60, 335–355.

30.

Ford

J. D.

Wasser

Connor

D. F.

(2011). Identifying and determining the symptom severity associated with polyvictimization among psychiatrically impaired children in the outpatient setting. Child Maltreatment, 16, 216–226.

31.

Fowler

P.J.

Toro

P.A.

Miles

B. W.

(2009). Pathways to and from homelessness and associated psychosocial outcomes among adolescents leaving the foster care system. American Journal of Public Health, 99, 1453–1458.

32.

Gibson

W. A.

(1959). Three multivariate models: Factor analysis, latent structure analysis, and latent profile analysis. Psychometrika, 24, 229–252.

33.

Goerge

R. M.

Bilaver

Lee

B. J.

Needell

Brookhart

Jackman

(2002). Employment outcomes for youth aging out of foster care. Chicago, IL: Chapin Hall Center for Children, University of Chicago.

34.

Goodman

L. A.

(1974). Exploratory latent structure analysis using both identifiable and unidentifiable models. Biometrika, 61, 215–231.

35.

Hamarman

Pope

K. H.

Czaja

S. J.

(2002). Emotional abuse in children: Variations in legal definitions and rates across the United States. Child Maltreatment, 7, 303–311.

36.

Higgins

D. J.

McCabe

M. P.

(1998). Parent perceptions of maltreatment and adjustment in children. Journal of Family Studies, 4, 53–76.

37.

Higgins

D. J.

McCabe

M. P.

(2000). Multi-type maltreatment and the long-term adjustment of adults. Child Abuse Review, 9, 6–18.

38.

Higgins

D. J.

McCabe

M. P.

(2001). Multiple forms of child abuse and neglect: Adult retrospective reports. Aggression and Violent Behavior, 6, 547–578.

39.

Hook

J. L.

Courtney

M. E.

(2011). Employment outcomes of former foster youth as young adults: The importance of human, personal, and social capital. Children and Youth Services Review, 33, 1855–1865.

40.

Jonson-Reid

Drake

Kim

Porterfield

Han

(2004). A prospective analysis of the relationship between reported child maltreatment and special education eligibility among poor children. Child Maltreatment, 9, 382–394.

41.

Jonson-Reid

Kohl

P. L.

Drake

(2012). Child and adult outcomes of chronic child maltreatment. Pediatrics, 129, 839–845.

42.

Keller

T. E.

Cusick

G. R.

Courtney

M. E.

(2007). Approaching the transition to adulthood: Distinctive profiles of adolescents aging out of the child welfare system. Social Service Review, 81, 453–484.

43.

Keller

T. E.

Salazar

A. M.

Courtney

M. E.

(2010). Prevalence and timing of diagnosable mental health, alcohol, and substance use problems among older adolescents in the child welfare system. Children and Youth Services Review, 32, 626–634.

44.

Kendall-Tackett

K. A.

Eckenrode

(1996). The effects of neglect on academic achievement and disciplinary problems: A developmental perspective. Child Abuse & Neglect, 20, 161–169.

45.

Kohl

P. L.

Jonson-Reid

Drake

(2009). Time to leave substantiation behind findings from a national probability study. Child Maltreatment, 14, 17–26.

46.

Lau

A. S.

Leeb

R. T.

English

Graham

J. C.

Briggs

E. C.

Brody

K. E.

Marshall

J. M.

(2005). What's in a name? A comparison of methods for classifying predominant type of maltreatment. Child Abuse & Neglect, 29, 533–551.

47.

Lee

J. S.

Courtney

M. E.

Hook

J. L.

(2012). Formal bonds during the transition to adulthood: Extended foster care support and criminal/legal involvement. Journal of Public Child Welfare, 6, 255–279.

48.

Magidson

Vermunt

J. K.

(2004). Latent class models. In Kaplan

(Ed.), The sage handbook of quantitative methodology for the social sciences (pp. 173–190). Thousand Oaks, CA: Sage.

49.

Manly

J. T.

Kim

J. E.

Rogosch

F. A.

Cicchetti

(2001). Dimensions of child maltreatment and children's adjustment: Contributions of developmental timing and subtype. Development and Psychopathology, 13, 759–782.

50.

Manly

J. T.

Lynch

Oshri

Herzog

Wortel

S. N.

(2013). The impact of neglect on initial adaptation to school. Child Maltreatment, 18, 155–170.

51.

Masten

A. S.

(2006). Promoting resilience in development: A general framework for systems of care. In Flynn

R. J.

Dudding

P. M.

Barber

J. G.

(Eds.), Promoting resilience in child welfare (pp. 3–17). Ottawa, ON: University of Ottawa Press.

52.

Masten

A. S.

Obradović

Burt

K. B.

(2006). Resilience in emerging adulthood: Developmental perspectives on continuity and transformation. In Arnett

J. J.

Tanner

J. L.

(Eds.), Emerging adults in America: Coming of age in the 21st century (pp. 173–190). Washington, DC: American Psychological Association Press.

53.

McCutcheon

A. L.

(1987). Latent class analysis. Thousand Oaks, CA: Sage.

54.

McMillen

J. C.

Zima

B. T.

Scott

L. D.

Jr Auslander

W. F.

Munson

M. R.

Ollie

M. T.

Spitznagel

E. L.

(2005). Prevalence of psychiatric disorders among older youths in the foster care system. Journal of the American Academy of Child & Adolescent Psychiatry, 44, 88–95.

55.

Merikangas

K. R.

Burstein

Swanson

S. A.

Avenevoli

Cui

L.,

… Swendsen

(2010). Lifetime prevalence of mental disorders in US adolescents: Results from the national comorbidity survey replication-adolescent supplement (NCS-A). Journal of the American Academy of Child & Adolescent Psychiatry, 49, 980–989.

56.

Mersky

J. P.

Janczewski

(2012). Adult well-being of foster care alumni: Comparisons to other child welfare recipients and a non-child welfare sample in a high-risk, urban setting. Children and Youth Services Review, 35, 367–376.

57.

Pecora

P. J.

Kessler

R. C.

Williams

O'Brien

Downs

A. C.

English

D.,

… Holmes

(2005). Improving family foster care: Findings from the northwest foster care alumni study. Seattle, WA: Casey Family Programs.

58.

Salazar

A. M.

Keller

T. E.

Gowen

L. K.

Courtney

M. E.

(2013). Trauma exposure and PTSD among older adolescents in foster care. Social Psychiatry and Psychiatric Epidemiology, 48, 545–551.

59.

Schwarz

(1978). Estimating the dimension of a model. The Annals of Statistics, 6, 461–464.

60.

Smith

Thornberry

T. P.

(1995). The relationship between childhood maltreatment and adolescent involvement in delinquency. Criminology, 33, 451–481.

61.

U.S. Department of Health and Human Services. (2008). Adolescents involved with child welfare: A transition to adulthood. Washington, DC: Author.

62.

U.S. Department of Health and Human Services. (2013a). The AFCARS report: Preliminary FY 2012 estimates as of July 2013. (No. 20). Washington, DC: Author.

63.

U.S. Department of Health and Human Services. (2013b). Child welfare outcomes 2008 to 2011: Report to Congress. Washington, DC: Author.

64.

U.S. Government Accounting Office. (2008). Disconnected youth: Federal action could address some of the challenges faced by local programs that reconnect youth to education and employment. (No. GAO-08-313). Washington, DC: Author.

65.

Vermunt

Magidson

(2005). Latent GOLD 4.0 user's guide. Belmont, MA: Statistical Innovations.

66.

Vermunt

Magidson

(2008). Latent GOLD 4.5. Belmont, MA: Statistical Innovations.

67.

White House Council for Community Solutions. (2012). Final report: Community solutions for opportunity youth. Washington, DC: Corporation for National Community Service.

68.

Wolfe

D. A.

McGee

(1994). Dimensions of child maltreatment and their relationship to adolescent adjustment. Development and Psychopathology, 6, 165–165.