Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Oct 12.
Published in final edited form as: Child Adolesc Social Work J. 2022 Mar 22;39(5):573–581. doi: 10.1007/s10560-022-00834-8

Parent and Caregiver Relationships and Mental Health Symptom Profiles of Youth in Foster Care

Lenore M McWey 1, Ming Cui 1, Armeda Stevenson Wojciak 2
PMCID: PMC10569441  NIHMSID: NIHMS1885293  PMID: 37829099

Abstract

Adolescents in foster care can be quite resilient, yet they also are at risk for developing internalizing and externalizing mental health concerns. Positive family relationships are central to adolescent mental health, and these relationships can be complex for youth in foster care placements. Accordingly, there can be significant heterogeneity in the mental health symptoms of youth in foster care. The aims of this study were to identify distinct subgroups of youth patterns of internalizing and externalizing symptoms and determine the extent to which positive biological and foster parent relationships predicted profiles of low youth mental health symptoms. Using data from the National Survey of Child and Adolescent Well-being II (N = 343) and a person-centered analytic approach, results revealed four distinct profiles of youths’ mental health symptoms, those with: high internalizing and high externalizing symptoms, high internalizing and moderate externalizing symptoms, moderate internalizing and moderate externalizing symptoms, and low internalizing and low externalizing symptoms. Additionally, youth with better relationships with biological parents were less likely to be in the high symptoms group compared to the low symptoms group. These findings can be used to inform targeted intervention efforts aimed to lessen the mental health symptoms of youth in foster care.

Keywords: Adolescents, Child welfare, Foster care, Latent profile analysis, Mental health, Person-centered

Adolescents in foster care are often quite resilient (Davidson-Arad & Navaro-Bitton, 2015), yet they are also vulnerable to a number of risks including the development of internalizing and externalizing mental health symptoms (e.g., Bronsard et al., 2016; Clausen et al., 1998; Heflinger et al., 2000; Munson et al., 2020). Positive parental relationships have been recognized as central to adolescent mental health for many years, regardless of whether or not youth are in foster care (Steinberg, 2001). Family relationships can be particularly complex, however, for youth in foster care placements because their families include both biological and foster parents. Although the mental health concerns of youth in foster care are widely documented, significantly less research examines the potential protective role of positive parental relationships (Brook et al., 2015). Moreover, there is significant heterogeneity in youth mental health, and more person-centered analyses are needed to understand differences, particularly among youth in foster care placements (Rebbe et al., 2017). The aims of this study, therefore, were to (a) contribute to the literature by revealing distinct subgroups of youth patterns of internalizing and externalizing mental health symptoms using a person-centered analytic approach; and (b) determine the extent to which positive biological and foster parent relationships were associated with profiles of low internalizing and externalizing symptoms of adolescents in foster care.

Mental Health Concerns of Youth in Foster Care and Person-Centered Analyses

Internalizing and externalizing symptoms are common mental health concerns among youth in foster care (e.g., Bronsard et al., 2016; Clausen et al., 1998; Heflinger et al., 2000; Munson et al., 2020). Internalizing symptoms involve signs of depression, anxiety, and emotional withdrawal. Externalizing symptoms include aggressive and delinquent behaviors. The potential long-term consequences of these mental health symptoms can be serious, and therefore, often the focus of research (Hawk et al., 2020). Most research, particularly studies specific to the mental health of youth in foster care, involves variable-centered statistical approaches (e.g., regression, correlational analyses). These variable-centered approaches are valuable in helping understand sample trends as a whole. Yet, there can be significant heterogeneity in the expression of mental health symptoms from person to person (Chan et al., 2008). As such, person-centered approaches, or analyses that test how constructs group across individuals, can complement variable-centered findings to distinguish different groups of youth from one another based on their levels of mental health symptoms (Logan-Greene et al., 2012).

Person-centered approaches have been applied to the child welfare context, albeit infrequently. For instance, with data from the National Survey of Child and Adolescent Well-being (NSCAW)—the data used in the present study—scholars found that different profiles of maternal risk factors (e.g., hardship, substance abuse, domestic violence) predicted future maltreatment reports (Kim et al., 2020). Specific to youth outcomes, with a community sample of adolescents in England, researchers conducted latent profile analysis (LPA) to determine profiles of maltreatment risk, then applied multivariate regression analyses to test associations between levels of maltreatment exposure and youth mental health outcomes (Cecil et al., 2014). Results demonstrated distinct youth profiles of maltreatment risk, with those in the high maltreatment exposure profile demonstrating higher internalizing and externalizing symptoms; however, only profiles of maltreatment were determined—not profiles of mental health symptoms. Arguably in research most closely related to the present study, scholars examined adverse childhood experiences (ACEs) and mental health and life course outcomes using the Midwest study of former foster youth (Rebbe et al., 2017). Results supported 3 distinct classes of former foster youth in terms of their past ACEs risks. Mapping outcomes on to those profiles, youth in the high adversity group demonstrated higher depressive symptoms and engagement in property crime compared to youth in the low adversity profile. Of note, profiles of mental symptoms were not tested in that study either. It is important to identify subgroups of youth with varying degrees of mental health symptoms using person-centered approaches in order to inform future research and treatment (Chan et al., 2008; Petersen et al., 2019); however, no known study has used a person-centered approach to identify profiles of internalizing and externalizing symptoms of youth in foster care. This study sought to fill that gap.

Family Relationships

Positive family relationships can be a protective factor for youth in foster care (Brook et al., 2015). Indeed, former foster youth who had a positive relationship with either a biological or foster parent reported better well-being than youth who reported feeling isolated or disconnected from family (Collins et al., 2010). With regard to biological parent relationships specifically, 9–13-year-old youths in foster care reported high levels of relationship quality with their biological mothers and there was a significant association between biological mother–child relationship quality and the youths’ internalizing symptoms while in foster care (Milan & Pinderhughes, 2000). A systematic review of research on biological parent–child relationships of youth aging out of foster care found that a “substantial proportion” had ongoing relationships with their biological parents while they were in foster care, and they maintained those relationships upon aging out (Havlicek, 2021, p. 10). As such, positive biological parent relationships can be important for the well-being of youth in foster care.

Current kinship or foster caregiver relationships also are important for youth in foster care. For instance, one study involving a sample of adolescents in foster care found a link between their caregivers’ involvement and youths’ mental health, such that youth who reported caregivers to be highly involved in their lives also reported lower externalizing symptoms (Rayburn et al., 2018). Moreover, other scholars have demonstrated an association between youths’ reports of their current caregiver relationship quality and caregivers’ reports of the externalizing symptoms of youth in foster care (Cooley et al., 2015). Therefore, current caregiver relationships also may be protective.

Risk and Resilience Framework and Study Purpose

The risk and resilience framework (Daniel & Wassell, 2002) can aid in our understanding of different youth mental health profiles and potential protective factors. According to the framework, youth may psychologically respond differently to the same challenging situation and differences, at least in part, can be due to interpersonal family dynamics (Daniel & Wassell, 2002; Jenson & Fraser, 2006). Indeed, positive family relationships may be key protective factors that can help support youth’s mental health. Applying the risk and resilience framework to the understanding of internalizing and externalizing symptoms of youth in foster care placements specifically, youths’ psychological responses to their family contexts will likely differ. These differences result in significant heterogeneity in youths’ expressions of mental health symptoms. Some youth might have high internalizing and externalizing symptoms, others might demonstrate high externalizing and low internalizing symptoms, still others might have few symptoms overall. In addition, although some youth might demonstrate concerning symptoms because of adverse family experiences, family factors can also be protective (Daniel & Wassell, 2002); however, the potential buffering effects of positive caregiver relationships have not been adequately tested in foster care research (Cooley et al., 2015). As such, it is important to identify patterns of mental health symptoms of youth in foster care using person-centered analytic approaches like latent profile analysis (LPA) and test relationship predictors of different symptoms profiles to inform targeted prevention and intervention efforts. Based on theory and extant research (e.g., Cecil et al., 2014; Collins et al., 2010; Cooley et al., 2015), we hypothesized that among a sample of adolescents in foster care (a) there would be different profiles of mental health symptoms; and (b) relationships with biological parents and current caregivers would be beneficial, in that youth with higher quality relationships would have lower mental health symptom profiles.

Method

Sample and Procedures

Data came from the National Survey on Child and Adolescent Well-Being II (NSCAW II)—a study of families involved with the child welfare system. The NSCAW II team recruited a U.S. nationally representative sample using a stratified sampling design whereby they identified distinct U.S. sampling sections, established sampling units representing distinct regions, then randomly selecting families from each region (Dowd et al., 2010). For the purposes of this study, we used Wave 1 weighted data.

Because adolescents completed the measures of interest for this study, the analytic sample included youth ages 11 to 17 years old (M = 14.09) residing in foster care (N = 343) at Wave 1. Please see Table 1. Participants were racially and ethnically diverse with 42% (n = 139) Black, 41% (n = 135) White, and 17% (n = 69) Asian, American Indian, Alaskan Native, multiracial, or another categorization. Regarding ethnicity, 26% (n = 88) were Hispanic. The sample included a near equal split of males and females (48% and 52% respectively). In terms of their types of foster care placements, 46% were in traditional foster homes, 32% in kinship settings, and the remainder of the youth in other placements (e.g., group homes, residential care).

Table 1.

Descriptive information on the sample and study variables

Variables (n) M or % (%) SD Min Max
Youth mental health symptoms
 Anxiety/depression (339) 5.58 5.35 0 27
 Withdrawn (339) 3.72 2.71 0 14
 Somatic complaints (339) 2.46 2.96 0 14
 Delinquent behavior (339) 4.14 3.43 0 18
 Aggressive behavior (339) 9.42 6.67 0 32
Relationship with biological parents (219) 11.24 2.57 3 15
Relationship with foster parents (318) 40.29 6.78 19 48
Demographic characteristics
 Age (343) 14.09 1.86 11 17
 Sex (Female) (343) 52.2%
Open in a new tab

Measures

Youth Internalizing and Externalizing Symptoms

The Youth Self Report (YSR; Achenbach & Rescorla, 2001) was used to assess internalizing and externalizing symptoms. The YSR is widely used to assess youth mental health symptoms. The measure, consisting of 112 items, has two broadband scales—internalizing and externalizing symptoms—and each broadband scale has empirically-based syndrome scales. The internalizing symptoms subscales include anxious/depressed, withdrawn, and somatic complaints. The externalizing symptoms subscales include aggressive behaviors, and delinquent behaviors. Youth reported the extent to which they experienced specific symptoms. Response options range from 0 (not true) to 2 (very often or often true) and higher scores indicate higher symptoms. For the current study, the Cronbach alphas were α = 0.95 for anxious/depressed, α = 0.96 withdrawn, α = 0.98 somatic complaints, α = 0.98 aggressive behaviors, and α = 0.98 delinquent behaviors.

Relationship with Biological Parents

Youth responded to items asking about their relationships with their biological parents. For the purposes of this study, we included their responses to 6 items such as how often they do fun things with their mother and how often they talk with her about important things. The same questions were also asked about their fathers. Response options ranged from 0 (none of the time) to 5 (all of the time). Reponses were summed and the mean of the scores for mothers and fathers was used in analyses. Because the mean level was used, the score could reflect an average of both parents, mother only (with score for father missing), or father only (with score for mother missing). The Cronbach alpha was acceptable at α = 0.94.

Relationship with Caregivers

The NSCAW team adapted the Rochester Assessment Package for Schools-Student (RAPS-S) to a 12-item measure assessing the youth’s perceptions of their relationship with their current caregiver. The measure assessed relationship domains of emotional security, involvement, autonomy, and structure. Example items include “My [caregiver] enjoys spending time with me” and “My [caregiver] does a lot to help me,” with potential answers ranging from 1 (not true at all) to 4 (very true). Some items are reverse coded and higher scores signal better a relationship with their caregiver. The alpha coefficient was α = 0.89.

Covariates

The sex of the youth was coded as 1 = male and 2 = female. The age of youth was assessed in years.

Analytic Strategy

The Vermunt’s 3-step approach (Bakk et al., 2013; Vermunt, 2010) was used to identify latent classes/profiles of youth internalizing and externalizing symptoms. The Vermunt’s 3-step approach is an improved version of the naive 3-step approach (e.g., failure to account for classification errors) and has some advantages over the 1-step approach (e.g., overly complex or unstable models and not robust to model specification) (Bolck et al., 2004; Clark & Muthen, 2009; Croon, 2002). Specifically, the Vermunt’s 3-step approach treats assigned classes (for dichotomous variables) or profiles (for continuous variables) as a nominal latent class indicator, thus predictors (e.g., relationships with biological parents and foster parents) can predict the latent class membership using the assigned class as the sole indicator variable. The main advantage of this approach is that Step 3 parameters are estimated conditionally on Step 1 parameters rather than simultaneously, along with the advantage of having a more intuitive sense of class classification.

The five subscales of youth internalizing and externalizing symptoms, relationships with biological parents and foster parents, along with covariates (sex and age) were entered in the Vermunt’s 3-step model. Listwise deletion was used. As with other types of latent class/profile analysis, the decision of number of classes (k) was made based on the following criteria: smaller values of Akaike Information Criterion (AIC; Akaike, 1974), Bayesian Information Criterion (BIC; Schwarz, 1978), and the sample size adjusted Bayesian Information Criterion (ssBIC; Sclove, 1987); bigger values of entropy (ranges from 0 to 1, Ramaswamy et al., 1993); and significant Adjusted Lo–Mendell–Rubin Likelihood Ratio Tests (A-LMR LRT; Lo et al., 2001, p < 0.05 suggests that the k-class model fits significantly better than the k-1 class solution).

The association between relationships with biological parents and foster parents and youth internalizing and externalizing symptoms were examined in Step 3. Odds ratio tests of categorical latent variable multinomial logistic regressions were utilized to regress likelihood of class membership on relationships with biological and foster parents (along with covariates sex and age).

Results

Table 1 provides the means and standard deviations for the variables of interests. Table 2 provides the selection criteria for the latent class/profile analyses. Based on the class selection criteria (i.e., smaller ICs, bigger entropy, and significant A-LMR), the 4-class solution was selected as the best fitting model. Table 2 provides the comparisons of the criteria from k = 2 to k = 5 groups. Even though the 5-class solution provided smaller ICs, the A-LMR tests suggested non-significant likelihood ratio test (meaning 5-class does not fit better than the 4-class solution). Also, the 5-class solution would produce two small classes (a class of 10 and a class of 17). Upon further examination, the 4-class solution also provided meaningful classes with sufficient sample sizes in each class.

Table 2.

Selection indices for latent class/profile selection

Model AIC BIC ssBIC Entropy A-LMR
2-Class 8935.04 8996.25 8945.50 .92 p = .00
3-Class 8758.69 8842.87 8773.08 .83 p = .03
4-Class 8682.58 8789.71 8700.89 .87 p = .04
5-Class 8613.48 8743.56 8635.71 .85 p = .30
Open in a new tab

AIC Akaike information criterion, BIC Bayesian information criterion, ssBIC sample size adjusted BIC, and A-LMR Adjusted Lo–Mendell–Rubin likelihood ratio test (LRT)

The best solution is noted in bold

Figure 1 depicts the estimated means for each subscale for the 4-class solution. Based on this comparative information from each class, individuals in Class 1 (N = 130, 38%) were moderate on all five subscales of youth internalizing and externalizing symptoms. We labeled this class “Class 1—moderate internalizing and moderate externalizing symptoms.” Individuals in Class 2 (N = 28, 8%) had relatively high levels of internalizing symptoms and moderate levels of externalizing symptoms. We labeled this class “Class 2—high internalizing and moderate externalizing symptoms.” Individuals in Class 3 (N = 165, 48%) had comparatively lower levels of both internalizing and externalizing symptoms. We labeled this group “Class 3—low internalizing and low externalizing symptoms.” Finally, individuals in Class 4 (N = 20, 6%) were high in both internalizing and externalizing symptoms. We labeled this group “Class 4—high internalizing and high externalizing symptoms.”

Fig. 1.

Fig. 1

Open in a new tab

Classes of youth internalizing and externalizing symptoms

To test the links between relationships with biological and foster parents and class membership of youth internalizing and externalizing symptoms, Table 3 provides the estimates and odds ratios from tests of categorical latent variable multinomial logistics regression using the 3-step Vermunt’s approach. Class 3 (low internalizing and low externalizing symptoms) was the reference group. The results suggested that youth with better relationship with biological parents were less likely (OR 0.79, 21% less likely, p < 0.05) to be in class 4 (high internalizing and high externalizing symptoms) than in class 3 (low internalizing and low externalizing symptoms). Further, being a female, as compared to being a male, was associated with a higher likelihood of being in class 4 (high internalizing and high externalizing symptoms) than in class 3 (low internalizing and low externalizing symptoms) (OR 6.70, p < 0.01).

Table 3.

Categorical latent variable multinomial logistic regression

Class 1 Class 2 Class 4
Relationship w/Bio.Parents b (S.E.) − 0.08 (0.07) − 0.18 (0.12) − 0.24 (0.11)*
OR 0.92 0.83 0.79*
Relationship w/Foster Parents b (S.E.) 0.01 (0.01) − 0.00 (0.01) − 0.01 (− 0.01)
OR 1.01 1.00 0.99
Sex b (S.E.) 0.54 (0.36) 0.68 (0.64) 1.90 (0.81)**
OR 1,72 1.98 6.70**
Age b (S.E.) 0.19 (0.10) 0.17 (0.18) 0.08 (0.18)
OR 1.21 1.19 1.09
Open in a new tab

Class 3 is the reference group. Class 1 = moderate internalizing and moderate externalizing symptoms. Class 2 = high internalizing and moderate externalizing symptoms. Class 3 = low internalizing and low externalizing symptoms. Class 4 = high internalizing and high externalizing symptoms. Sex: 1 = male, 2 = female

*

p < .05

**

p < .01

Findings with the covariates also indicated that youth with a higher quality relationship with their biological parent were less likely to be in Class 4—high internalizing and high externalizing symptoms as compared to be in Class 3—low internalizing and low externalizing symptoms, indicating the potential protective role of positive relationships with biological parents in youths’ mental health. The youth’s sex had a significant role in predicting group membership with females being more likely to be in Class 4 than in Class 3.

Discussion

The purposes of this study were to: (a) to use a person-centered approach to identify latent classes of internalizing and externalizing mental health symptoms of youth in foster care, and (b) to explore whether higher relationship quality with biological parents or current caregivers was predictive of youth mental health profiles. The LPA resulted in 4 distinct subgroups of mental health symptoms: Class 1—moderate internalizing and moderate externalizing symptoms, Class 2—high internalizing and moderate externalizing symptoms, Class 3—low internalizing and low externalizing symptoms, and Class 4—high internalizing and high externalizing symptoms. Findings from this study support past research demonstrating heterogeneity of mental health symptoms (e.g., Chan et al., 2008), and extend this work to youth in foster care. Moreover, findings point to the importance of positive relationships with biological parents in buffering against mental health symptoms.

A meta-analysis examining the prevalence of distinct mental health concerns of youth in foster care indicated that roughly half of the youth represented in the reviewed studies met the criteria for at least one mental health disorder (Bronsard et al., 2016). Indeed, there is a wealth of research documenting the mental health concerns of youth in foster care (e.g., Bronsard et al., 2016; Clausen et al., 1998; Heflinger et al., 2000; Munson et al., 2020). This study adds to the literature by examining multiple mental health symptoms simultaneously and demonstrating that differing mental health symptoms distinguish groups of youth. Results indicated that 48% of youth had comparatively low internalizing and externalizing symptoms and 38% had moderate internalizing and externalizing symptoms. Together that means that 86% of the youth in this sample demonstrated low to moderate mental health symptoms across both internalizing and externalizing domains. Despite the risks that many youths experienced that led to their placement in foster care, it appears many also were resilient.

A novel contribution of this study was the application of latent profile analyses to youths’ mental health symptoms. The use of person-centered approaches for research specific to the foster care context is relatively infrequent but growing (Cecil et al., 2014; Kim et al., 2020; Rebbe et al., 2017). In past studies, scholars have identified profiles of maltreatment risk (Cecil et al., 2014) and ACE exposure and mental health (Rebbe et al., 2017). Findings from the present study provide a nuanced view of the mental health symptom profiles of youth in foster care. Youth in the high internalizing and high externalizing symptoms profile demonstrated highs in anxious and depressed symptoms and delinquent and/or aggressive behaviors compared to youth in the low internalizing and low externalizing symptoms profile, whose anxious and depressed symptoms, withdrawal, somatic complaints, and delinquent and/or aggressive behaviors were all relatively low. There are practical takeaways from the profile findings. For instance, it is noteworthy that youth in Class 4 had both high internalizing and high externalizing symptoms (i.e., aggression and anxiety and depression). This is important because a youth’s externalizing symptoms might gain more immediate attention by caregivers, caseworkers, and providers; however, this group of youth were also struggling with symptoms of depression. Relatedly, a notable difference distinguishing Class 1 (moderate internalizing and moderate externalizing symptoms) from Class 2 (high internalizing and moderate externalizing symptoms) was youths’ somatic complaints. As such, even if the troublesome symptoms presented by youth in foster care seem obvious, it is important to conduct thorough assessments to capture a full spectrum of mental health symptoms. Doing so can inform treatment aimed to improve mental health symptoms.

Additionally, and similar to prior research (e.g., Milan & Pinderhughes, 2000), findings from our study indicated that the quality of relationship between the youth and their biological parents was protective. Specifically, quality relationships with biological parents reduced youth’s odds of being in the high internalizing and high externalizing symptoms group versus the low internalizing and low externalizing symptoms group. This suggests that promoting relationship quality with a youth’s biological parents, in cases where it is deemed safe to do so, might be an important buffering factor associated with adolescent wellbeing, and showcases the interplay between risk and protective factors. Promoting protective factors, like relationship quality, linked to improved mental health has other potential consequences as well. For example, foster care research indicates significant associations between youths' mental health problems and their placement instability, and placement instability is associated with poorer longer-term youth outcomes (Konijn et al., 2019). As such, supporting positive relationships with biological parents may provide multiple long-term benefits for youth in foster care.

Interestingly, a youth’s relationship with their current caregiver was not a significant predictor of group membership. This differs from past research indicating that relationships with caregivers also can be a protective factor for youth in foster care (Cooley et al., 2015; Rayburn et al., 2018); however this study considered both biological and current caregiver relationships and little past research has done so. As such, these findings could suggest the importance of having at least one positive caregiver relationship. Future research should replicate this study with a larger sample size to see if the pattern of findings remains the same and if there is perhaps an interactive effect with the biological parent and current caregiver relationships within these mental health profiles.

Further, being female as compared to male was associated with a greater likelihood of being in Class 4-high internalizing and high externalizing symptoms group. This finding might be surprising at first glance, particularly with regard to females’ externalizing symptoms, however, females are more likely to be exposed to trauma such as sexual abuse or interpersonal traumas perpetrated by someone close to them (Christiansen & Hansen, 2015; Goldberg & Freyd, 2006; Tolin & Foa, 2008), and therefore may demonstrate both high internalizing and externalizing mental health symptoms as a result.

Limitations and Future Research

There are important limitations to consider while interpreting the results of this study. The first consideration is the complex nature of the foster care system. This study identified four distinct subgroups of mental health symptoms; however, the analytic sample size did not afford the power to account for all the potential variability each youth may experience leading up to their foster care placement or their experiences while in care. Scholars interested in continuing to identify the nuances of youth mental health profiles within the complex foster care context could also consider the type of maltreatment youth experienced, how long they have been in foster care, mental health symptoms when entering care, and the frequency of contact they have with their biological parents. Similarly, a third of the youth in the sample were in kinship placements. Kinship placements are valuable because they can help maintain family connections (Kiraly & Humphreys, 2016). However, in this study, we do not know how placement type affected youths’ access to their biological parents and to what extent preexisting relationships between current caregivers and biological parents influenced youths’ mental health profiles. Future research examining the access to and frequency of contact with biological parents is warranted.

Also, there were missing data in the measurement of the relationship with biological parents. There are possible explanations for the missingness. For example, some youth may not have known their biological parents or had biological parents who were deceased. Though there were minimal missing data in the latent profiles of youth mental health symptoms, we used listwise deletion when examining the association between relationships with parents and profile membership because of the potential systematic missingness. Therefore, the findings of this study were limited to youth who knew and could report on their relationships with their biological parents. Caution should be used in interpreting these findings to avoid overgeneralization of the implications. Relatedly, due to the moderate sample size, this study only included sex and age as key covariates because their effects were demonstrated by previous studies of youth mental health (e.g., McWey et al., 2018). Future studies need to consider other important covariates, such as race, ethnicity and placement type, to ensure that the associations being tested are not spurious.

Moreover, there are a host of potential protective factors that may be associated with youth mental health and this study focused only and specifically on biological and current caregiver relationships. We did so because scholars suggest that this is an under studied potential protective factor in foster care research (Cooley et al., 2015), however, we acknowledge that there are other intra-, inter-, and environmental considerations that also may help explain why some youth may cope better than others to the foster care context.

Lastly, we opted to use of the Youth Self-Report (YSR) of mental health symptoms. Some scholars have advocated for including youth report when assessing their mental health, particularly given the discrepancies between youth and caregiver reports (De Los Reyes & Kazdin, 2005; Lyons et al., 2002). However, the adults’ report of child mental health also is available. Scholars interested in comparing the concordance of mental health profiles could examine both youth and adult reports.

Conclusions

This study used a person-centered analytic approach to identify four unique subgroups of mental health symptoms among a national sample of youth in foster care. This analytic approach compliments findings from variable centered analyses (e.g., regressions and correlations) and revealed that mental health symptoms distinguish different groups of youth—some youth with comparatively high internalizing and externalizing symptoms, some with low, and others with moderate levels. Moreover, this study also found that positive relationships with biological parents were associated with significantly lower odds of being in the high internalizing and externalizing symptoms group, providing support for the protective nature of family relationships for youth in foster care. These findings highlight the importance of conducting thorough mental health assessments for youth in foster care to capture the full spectrum of symptoms and promoting biological parent relationships whenever safely possible to do so.

Acknowledgements

This document includes data from the National Survey on Child and Adolescent Well-Being, which was developed under contract with the Administration on Children, Youth, and Families, U.S. Department of Health and Human Services (ACYF/DHHS). The data have been provided by the National Data Archive on Child Abuse and Neglect. The information and opinions expressed herein reflect solely the position of the authors. Nothing herein should be construed to indicate the support or endorsement of its content by ACYF/DHHS.

Funding

Research reported in this publication was supported by the Eunice Kennedy Shriver National Institute of Child Health & Human Development of the National Institutes of Health under Award Number R03HD099424. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Footnotes

Conflict of interest The authors declare that they have no conflict of interest.

Ethical Approval The study was approved by the Florida State University institutional review board ethics committee and the authors certify that the study was performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards.

Data Availability

Data are from the National Survey on Child and Adolescent Well-Being and are provided by the National Data Archive on Child Abuse and Neglect to approved users.

References

  1. Achenbach TM, & Rescorla LA (1991, 2001). Manual for the ASEBA school-age forms & profiles. University of Vermont, Research Center for Children, Youth, & Families. [Google Scholar]
  2. Akaike H (1974). A new look at the statistical model identification. IEEE Transactions on Automatic Control, 19, 716–723. 10.1109/TAC.1974.1100705 [DOI] [Google Scholar]
  3. Bakk Z, Tekle FB, & Vermunt JK (2013). Estimating the association between latent class membership and external variables using bias adjusted three-step approaches. In Liao TF (Ed.), Sociological methodology (pp. 272–311). Sage. [Google Scholar]
  4. Bolck A, Croon MA, & Hagenaars JA (2004). Estimating latent structure models with categorical variables: One-step versus three-step estimators. Political Analysis, 12, 3–27. [Google Scholar]
  5. Bronsard G, Alessandrini M, Fond G, Loundou A, Auquier P, Tordjman S, & Boyer L (2016). The prevalence of mental disorders among children and adolescents in the child welfare system: A systematic review and meta-analysis. Medicine (baltimore), 95(7), e2622–e2622. 10.1097/MD.0000000000002622 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Brook J, Rifenbark GG, Boulton A, Little TD, & McDonald TP (2015). Risk and protective factors for drug use among youth living in foster care. Child & Adolescent Social Work Journal, 32, 155–165. 10.1007/s10560-014-0345-5 [DOI] [Google Scholar]
  7. Cecil CAM, Viding E, Barker ED, Guiney J, & McCrory EJ (2014). Double disadvantage: The influence of childhood maltreatment and community violence exposure on adolescent mental health. Journal of Child Psychology and Psychiatry, 55, 839–848. 10.1111/jcpp.12213 [DOI] [PubMed] [Google Scholar]
  8. Chan YF, Dennis ML, & Funk RR (2008). Prevalence and comorbidity of major internalizing and externalizing problems among adolescents and adults presenting to substance abuse treatment. Journal of Substance Abuse Treatment, 34, 14–24. 10.1016/j.jsat.2006.12.031 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Christiansen DM, & Hansen M (2015). Accounting for sex differences in PTSD: A multi-variable mediation model. European Journal of Psychotraumatology, 6(1), 26068. 10.3402/ejpt.v6.26068 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Clark S, & Muthén B (2009). Relating latent class analysis results to variables not included in the analysis. Unpublished manuscript retrieved from https://www.statmodel.com/download/relatinglca.pdf [Google Scholar]
  11. Clausen JM, Landsverk J, Ganger W, Chadwick D, & Litrownik A (1998). Mental health problems of children in foster care. Journal of Child and Family Studies, 7(3), 283–296. 10.1023/A:1022989411119 [DOI] [Google Scholar]
  12. Collins ME, Spencer R, & Ward R (2010). Supporting youth in the transition from foster care: Formal and informal connections. Child Welfare, 89, 125–143. [PubMed] [Google Scholar]
  13. Cooley M, Wojciak A, Farineau H, & Mullis A (2015). The association between perception of relationship with caregivers and behaviours of youth in foster care: A child and caregiver perspective. Journal of Social Work Practice, 29, 205–221. 10.1080/02650533.2014.933405 [DOI] [Google Scholar]
  14. Croon M (2002). Using predicted latent scores in general latent structure models. In Marcoulides GA & Moustaki I (Eds.), Latent variable and latent structure models (pp. 195–223). Erlbaum Associates. [Google Scholar]
  15. Daniel B, & Wassell S (2002). Early years assessing and promoting resilience in vulnerable children (Vol. 1). J Kingsley Publishers. [Google Scholar]
  16. Davidson-Arad B, & Navaro-Bitton I (2015). Resilience among adolescents in foster care. Children and Youth Services Review, 59, 63–70. 10.1016/j.childyouth.2015.09.023 [DOI] [Google Scholar]
  17. Dowd K, Dolan M, Wallin J, Miller KA, Biemer P, Aragon-Logan E, … & Smith K (2010). National survey of child and adolescent well-being II: Data file user’s manual restricted release version. National Data Archive on Child Abuse and Neglect. [Google Scholar]
  18. Goldberg LR, & Freyd JJ (2006). Self-reports of potentially traumatic experiences in an adult community sample: Gender differences and test-retest stabilities of the items in a brief betrayal-trauma survey. Journal of Trauma and Dissociation, 7, 39–63. 10.1300/J229v07n03_04 [DOI] [PubMed] [Google Scholar]
  19. Havlicek J (2021). Systematic review of birth parent-foster youth relationships before and after aging out of foster care. Children and Youth Services Review. 10.1016/j.childyouth.2020.105643 [DOI] [Google Scholar]
  20. Hawk BN, Timmer SG, Armendariz LAF, Boys DK, & Urquiza AJ (2020). Improving behaviors and placement stability for young foster children: An open trial of Parent-Child Care (PC-CARE) in the child welfare system. Children and Youth Services Review, 119, 9. 10.1016/j.childyouth.2020.105614 [DOI] [Google Scholar]
  21. Heflinger CA, Simpkins CG, & Combs-Orme T (2000). Using the CBCL to determine the clinical status of children in state custody. Children and Youth Services Review, 22, 55–73. [Google Scholar]
  22. Jenson JM, & Fraser MW (2006). A risk and resilience perspective. Social policy for children and families Sage. [Google Scholar]
  23. Kim H, Jonson-Reid M, Kohl P, Chiang C, Drake B, Brown D, McBride T, & Guo S (2020). Latent class analysis risk profiles: An effective method to predict a first re-report of maltreatment? Evaluation and Program Planning, 80, 101792. 10.1016/j.evalprogplan.2020.101792 [DOI] [PubMed] [Google Scholar]
  24. Kiraly M, & Humphreys C (2016). “It”s about the whole family’: Family contact for children in kinship care. Child and Family Social Work, 21(2), 228–239. 10.1111/cfs.12140 [DOI] [Google Scholar]
  25. Konijn C, Admiraal S, Baart J, van Rooij F, Stams G-J, Colonnesi C, Lindauer R, & Assink M (2019). Foster care placement instability: A meta-analytic review. Children and Youth Services Review, 96, 483–499. 10.1016/j.childyouth.2018.12.002 [DOI] [Google Scholar]
  26. Lo Y, Mendell N, & Rubin D (2001). Testing the number of components in a normal mixture. Biometrika, 88, 767–778. 10.1037/a0020077 [DOI] [Google Scholar]
  27. Logan-Greene P, Nurius PS, & Thompson EA (2012). Distinct stress and resource profiles among at-risk adolescents: Implications for violence and other problem behaviors. Child & Adolescent Social Work Journal, 29, 373–390. 10.1007/s10560-012-0269-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. De Los Reyes A, & Kazdin AE (2005). Informant discrepancies in the assessment of childhood psychopathology: A critical review, theoretical framework, and recommendations for further study. Psychological Bulletin, 131, 483–509. [DOI] [PubMed] [Google Scholar]
  29. Lyons JS, Rawal P, Yeh I, & Traacy P (2002). Use of measurement audits in outcomes management. Journal of Behavioral Health Services & Research, 29, 75–80. [DOI] [PubMed] [Google Scholar]
  30. McWey L, Cui M, Cooper A, & Ledermann T (2018). Caregiver-adolescent disagreement on the mental health of youth in foster care: The moderating role of the caregiving relationship. Child Maltreatment, 23, 294–302. 10.1177/1077559518769375 [DOI] [PubMed] [Google Scholar]
  31. Milan SE, & Pinderhughes EE (2000). Factors influencing maltreated children’s early adjustment in foster care. Development and Psychopathology, 12, 63–81. 10.1017/S0954579400001048 [DOI] [PubMed] [Google Scholar]
  32. Munson MR, Katz CC, Okpych NJ, & Courtney ME (2020). Mental health management among older youth in foster care: Service utilization and preparedness. Journal of Adolescent Health, 67(2), 225–231. 10.1016/j.jadohealth.2020.02.017 [DOI] [PubMed] [Google Scholar]
  33. Petersen KJ, Qualter P, & Humphrey N (2019). The application of latent class analysis for investigating population child mental health: A systematic review. Frontiers in Psychology, 10, 1214. 10.3389/fpsyg.2019.01214 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Ramaswamy V, DeSarbo WS, Reibstein DJ, & Robinson WT (1993). An empirical pooling approach for estimating marketing mix elasticities with PIMS data. Marketing Science, 12, 103–124. 10.1287/mksc.12.1.103 [DOI] [Google Scholar]
  35. Rayburn AD, Withers MC, & McWey LM (2018). The importance of the caregiver and adolescent relationship for mental health outcomes among youth in foster care. Journal of Family Violence, 33, 43–52. 10.1007/s10896-017-9933-4 [DOI] [Google Scholar]
  36. Rebbe R, Nurius PS, Ahrens KR, & Courtney ME (2017). Adverse childhood experiences among youth aging out of foster care: A latent class analysis. Children and Youth Services Review, 74, 108–116. 10.1016/j.childyouth.2017.02.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Schwarz G (1978). Estimating the dimension of a model. Annals of Statistics, 6, 461–464. 10.2307/2958889 [DOI] [Google Scholar]
  38. Sclove SL (1987). Application of model-selection criteria to some problems in multivariate analysis. Psychometrika, 52(333), 343. 10.1007/BF02294360 [DOI] [Google Scholar]
  39. Steinberg L (2001). We know some things: Parent-adolescent relationships in retrospect and prospect. Journal of Research on Adolescence, 11, 1–19. [Google Scholar]
  40. Tolin DF, & Foa EB (2008). Sex differences in trauma and post-traumatic stress disorder: A quantitative review of 25 years of research. Psychological Trauma: Theory, Research, Practice and Policy. 10.1037/1942-9681.S.1.37 [DOI] [PubMed] [Google Scholar]
  41. Vermunt JK (2010). Latent class modeling with covariates: Two improved three-step approaches. Political Analysis, 18, 450–469. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

Data are from the National Survey on Child and Adolescent Well-Being and are provided by the National Data Archive on Child Abuse and Neglect to approved users.

RESOURCES