Matern Child Health J DOI 10.1007/s10995-014-1525-9
The Influence of Caregiver Depression on Children in Non-relative Foster Care Versus Kinship Care Placements Antonio Garcia • Amanda O’Reilly • Meredith Matone • Minseop Kim • Jin Long David M. Rubin
•
Ó Springer Science+Business Media New York 2014
Abstract Little is known about how the challenges faced by caregivers influence the variation in social, emotional, and behavioral (SEB) outcomes of youth placed in kinship versus non-relative foster care. This study examined SEB symptoms among youth in kinship and non-relative foster care settings, hypothesizing that changes in caregiver depression would modify children’s change in behavior over time. Child Behavior Checklist (CBCL) assessments of 199 children placed with kinship and non-relative foster care providers in a Mid-Atlantic city were conducted at time of placement and 6–12 months post-placement. Linear regression estimated CBCL change scores for youth across placement type and caregiver depression trajectories. Kinship caregivers were more likely to become depressed or remained depressed than non-relative foster caregivers. Youth in kinship care always exhibited better change in SEB outcomes than youth in non-relative foster care, but these positive outcomes were principally observed among families where caregivers demonstrated a reduction
A. Garcia (&) M. Kim School of Social Policy and Practice, University of Pennsylvania, 3701 Locust Walk, Philadelphia, PA 19104-6214, USA e-mail: [email protected] A. O’Reilly Division of Children and Families, State of New Jersey, Trenton, NJ, USA M. Matone J. Long D. M. Rubin PolicyLab, The Children’s Hospital of Philadelphia, Philadelphia, PA, USA D. M. Rubin Department of Pediatrics at the Perelman School of Medicine, The University of Pennsylvania, Philadelphia, PA, USA
in depression over time or were never depressed. Adjusted change scores for non-relative foster care youth were always negative, with the most negative scores among youth whose caregivers became depressed over time. Caregiver well-being may modify the influence of placement setting on SEB outcomes for youth placed into outof-home care. Findings lend to policy relevance for child welfare systems that seek kinship settings as a panacea to the challenges faced by youth, without allocating resources to address caregiver needs. Keywords Maltreatment School age children Maternal depression Social-emotional problems Prospective study
Introduction Over 400,000 children in the U.S. were placed in nonrelative foster care or kinship care in 2011 due to maltreatment [1]. Twenty-seven percent of these children were placed in kinship care. Kinship care as an alternative to non-relative foster care placements have grown considerably over the last two decades; by 2010, 2.7 million children in this country were living in kinship arrangement, a 70 % increase from just 20 years prior [2]. As the numbers of children in kinship care have grown, prior studies have demonstrated a mixed picture of social, emotional, and behavioral (SEB) outcomes among these youth. On the one hand, several studies show that children in kinship care experience less SEB problems compared to children placed in non-relative foster care [3–6]. For example, one study shows that children in kinship care experience less externalizing behaviors after 18 months of placement [7]. Another study reports that regardless of prior placement history, children in kinship care are 50 % less likely to
123
Matern Child Health J
exhibit behavioral problems 3 years after placement compared to children in non-relative foster care [6]. However, rates of SEB problems among children in kinship care exceed rates of behavioral problems relative to children living in poverty with birth parents [6, 8]. Thus, it is imperative to examine why children placed in kinship care do not fully benefit from being placed in a home that preserves family connectedness [9]. Prior research illuminates why children in kinship care may be at an increased risk to experience SEBs relative to children in the general population. First and foremost are their trauma histories. Due to severe negligent treatment and/or abuse, young children who enter out-of-home care are more likely than those without exposure to multiple forms and types of maltreatment to be diagnosed with elevated behavioral and mental health problems and developmental delays [10, 11]. Aside from maltreatment history, socio-demographic risk factors are often pervasive within families, and thus children placed with kin retain exposure to risk factors for poor well-being, perhaps at higher rates than peers placed in non-relative foster care. For example, kinship caregivers in the U.S. are older [9, 12], earn less income than foster care providers, and are more likely to live in impoverished conditions [9, 13]. These disparities are likely associated with lower educational attainment [14] and higher rates of single parenting [14–16] among these caregivers. Even with comparatively fewer socio-demographic resources, kinship caregivers are likely to provide care for the same number of children as non-relative foster care providers do [13, 17]. Caregiver responsibilities are further challenged by a host of risk factors for adverse child SEB outcomes originating prior to out-of-home placement, including parental mental health problems, poor parenting behavior, substance abuse, parental criminal involvement, and chronic maltreatment [18–23]. Collectively, it is not known the extent to which these stressors may take a negative toll on the caregivers’ own mental health. Several studies show that grandparents raising grandchildren experience increased depression compared to grandparents not caring for their grandchildren [24–26]. Little is known about how these stressors, often exacerbated by mental health symptomatology faced by caregivers, influence variation in behavioral outcomes of youth placed in kinship versus non-relative foster care. The degree that caregiver health might influence the outcomes of youth placed in kinship care would have immediate policy relevance, suggesting that placement with kin might need to be accompanied by greater attention to caregiver health to augment benefits to youth. The current study examined behavioral assessments of youth placed in both settings at baseline (time of out-of-home placement) and the second interview conducted between 6 and
123
12 months after placement, hypothesizing that changes in caregiver depression would modify children’s change in behavior over time and across kinship and non-relative foster settings, controlling for maltreatment history and child and family characteristics.
Methods Study Design and Participants The sample was drawn from a prospective, longitudinal cohort of 405 children between ages 3 and 8 years who were recruited from a large, Mid-Atlantic city’s child welfare system upon a new placement into out-of-home care from 2006 to 2008. Children were followed for 24 months from their entry into out-of-home care; with caregiver interviews conducted each time a placement change occurred or every 6 months within stable placements. The sample was restricted to those children (N = 199) placed with foster and kinship care providers who remained with their initial caregiver between baseline and the second interview. Measures The primary outcome measures were child behavioral change scores from baseline to second interview assessed with the Child Behavior Checklist (CBCL). The CBCL is a widely used, valid, and reliable measure of clinically pervasive symptoms of externalizing and/or internalizing behaviors evaluated by caregivers [27]. Each item operates on a 3-point Likert scale (0 = not true, 1 = somewhat or sometimes true, and 2 = very true or often true) to rate presence of symptoms [27]. Non-relative foster and kinship caregivers were administered the CBCL at baseline and at each subsequent interview. For each child, internalizing, externalizing, and total percentile change scores were calculated by subtracting time 1 (baseline) percentiles from time 2 percentiles. The primary exposures for the study were the placement setting and the assessment of caregiver depression between time 1 and 2. Placement setting was operationalized as non-relative foster care versus kinship care setting. Caregiver depression was assessed using the single item of the validated SF-12 Health Survey of adult health [28, 29] that asked about feeling blue or sad. The SF-12 survey is a widely used, validated 12-item scale of mental and physical health domains derived from the SF-36 Health Survey. The SF-12 is scored along individual items as well as summary scores for physical and mental health domains [28–31]. For the purposes of the current study, depression was operationalized as a single survey item regarding the
Matern Child Health J Table 1 Descriptive characteristics of youth entering new placements in non-relative foster care and kinship care settings Total (n = 199) n (%)
Foster care (n = 60) n (%)
Kinship care (n = 139) n (%)
p
Caregiver race Caucasian African American Latino and ‘‘other’’ Sexual abuse (yes)
27 (13.7)
6 (10)
21 (15.3)
.32
146 (74.1) 24 (12.2)
42 (70) 12 (20)
104 (75.9) 12 (8.8)
.38 .03
38 (19)
16 (27)
22 (16)
Caregiver income \$ 19,999
13 (23.6)
43 (32.6)
$ 20,000–$ 39,999
68 (36.4)
26 (47.3)
42 (31.8)
C$40,000
63 (33.7)
16 (29.1)
47 (35.6)
56 (30)
25 (42)
40 (29)
Children under 9 Three or more children \9 years
.07 .13
.06 65 (33)
Risk factors§ Three or less risk factors
114 (58)
31 (52)
83 (61)
83 (42)
29 (48)
54 (39)
101 (51)
38 (63)
63 (45)
Became depressed
25 (13)
4 (7)
21 (15)
Improved Remained depressed
25 (13) 48 (24)
8 (13) 10 (17)
17 (12) 38 (27)
4? risk factors Caregiver depression (between T1 and T2) Never depressed
CBCL
.24 .07
Mean (SD)
Mean (SD)
Mean (SD)
Internalizing % change score
-3.53 (30.93)
4.63 (29.50)
-7.05 (30.97)
.01
Externalizing % change score
0.66 (28.98)
6.15 (27.70)
-1.71 (29.29)
.08
Total % change score
-2.35 (29.44)
4.85 (28.68)
-5.45 (29.32)
.02
§
Risk factors include parental substance abuse, mental health, and incarceration; and domestic violence, homelessness, and prior foster care placement
extent caregivers’ felt ‘‘downhearted and blue’’ on a scale from 1 (none of the time) to 5 (most of the time). Caregivers who reported feeling downhearted and blue ‘‘a little’’, ‘‘some’’, ‘‘most’’, or ‘‘all of the time’’ were denoted as depressed and others who reported ‘‘none of the time’’ were denoted as not depressed. Because the response to the depressive symptoms item may vary with time, and given the concern that caregiver depression might also influence the likelihood of reporting child behavioral problems, we created strata of caregiver depression status over time to use as a priori effect modifier in later models. The four strata included ‘‘never depressed,’’ ‘‘became depressed’’, ‘‘improved,’’ and ‘‘remained depressed’’ to indicate the change between time 1 and time 2 depression status. At each interview, kinship and non-relative foster caregivers provided data on their income and the number of children in out-of-home placement in the household. Case record data included prior history of abuse, placement setting (kinship or non-relative foster care), and biological parental risk factors that increased the likelihood of out-of-
home placement (parental substance abuse, mental health, incarceration; as well as domestic violence and homelessness). We opted to only control for sexual abuse (versus controlling for all types of maltreatment), given its unique nature and the finding that neglect and physical abuse were highly prevalent in the sample (data not shown). Analysis Plan Linear regression estimated behavioral percentile change scores for youth across placement type and caregiver depression categories between baseline and follow-up, controlling for sexual abuse history, duration of time in days between interviews, and the following baseline sociodemographic characteristics: (1) caregiver income (\$20,000; $20,000–40,000; [$40,000) (2) providing care for a large number (i.e., C3) of young children under 9 (33 % of the caregivers met this criteria), and (3) number of parental risk factors, dichotomized into a binary variable based on mean distributions of the sample (42 %
123
Matern Child Health J
experienced four or more risk factors). Models were elaborated for the change in total score, externalizing subscore, and internalizing sub-score percentages of the CBCL between time 1 and 2. Based on a priori hypothesis, an interaction term for placement setting and caregiver depression trajectory was included in all models. This interaction term was retained only for the change in total score model where the interaction term was significant at the p \ .05 level. Post-estimation commands (margins in STATA) were then used to estimate predicted change scores on the CBCL. Although the binary representation of a single item was easy to operationalize, we also acknowledged the need for a sensitivity analysis to validate its use against the full SF12 mental health summary scores at time 1 and 2. We derived our analogous four different categories of depression trajectory over time by relying upon scoring procedures set forth by Vilagut et al. [32]. Specifically, their sensitivity and specificity analyses found that the optimal cutoff score for 30-day depressive disorders in the general population was 45.6. Thus, kinship and non-relative caregivers who scored below 45.6 were denoted as having depressive symptoms and those above 45.6 were denoted as not having depressive symptoms. Caregiver status as ‘‘never depressed,’’ ‘‘became depressed,’’ ‘‘improved,’’ or ‘‘remained depressed’’ was determined by their score as above or below 45.6 at times 1 and 2. All results were consistent with this summary score specification (estimates not reported), indicating our findings to be robust to the single item measure within the SF-12. In the end, because of the simplicity of the single-item with respect to categorizing change over time, we report our model results with that variable. Statistical analyses were performed using Stata 10.0 software [33]. The study was approved by the institutional review board of The Children’s Hospital of Philadelphia.
Results Descriptive characteristics for the sample of youth and care providers in the study are reported in Table 1. Among youth who remained stable in placement between baseline and second interview, 70 % (n = 139) were in kinship care, and 30 % (n = 60) were in non-relative foster care placement. Most of the caregivers identified as African American (74 %). Foster providers (42 %) were more likely than non-relative foster care (29 %) providers to care for three or more foster children. Children placed in nonrelative foster care settings (27 %) were more likely to have been sexually abused than those placed in kinship care (16 %). While a higher percentage of kinship families lived in poverty across all income categories than non-
123
relative foster care families, the differences were not statistically significant. Moreover, a higher percentage of kinship caregivers became depressed between time 1 and 2 (15 vs. 7 %) and remained depressed over the period from time 1 and time 2 (27 vs. 17 %) than non-relative foster caregivers. Table 1 also shows that children in kinship care were more likely to experience positive CBCL behavioral change scores than those placed in non-relative foster care. Children placed in kinship care were more likely to benefit from decreased internalizing symptoms (change score -7.05, SD = 30.97) than children placed in non-relative foster care (change score ?4.63, SD = 29.50). Children in kinship settings were more likely to benefit from a reduction in total CBCL scores (change score -5.45, SD = 29.32) relative to children in non-relative foster care (change score ?4.85, SD = 28.68). Children in kinship households were more likely to reveal fewer externalizing behaviors (change score -1.71, SD = 29.29) compared to children in non-relative foster care (change score ?6.15, SD = 27.70). Apart from trajectories of maternal depression varying across placement settings, there were also consistent differences across those trajectories in the average CBCL change scores (Table 2). In general, across all caregiver depression categories (never depressed, became depressed, improved, remained depressed), the change in scores for youth in kinship care always exhibited better internalizing, externalizing, and total behavior change scores than youth in non-relative foster care settings. Noteworthy is that children placed with kinship providers (change score -9.24, SD = 29.4) who were never depressed were more likely to experience a significant decrease in internalizing behaviors between baseline and follow-up than children placed with non-relative foster providers (change score ?4.03, SD = 32.84). Moreover, youth in kinship care (change score -14.65, SD = 24.12) whose caregivers’ depressive symptomatology got better over time were also more likely to benefit from a decrease in externalizing behaviors than those placed in non-relative foster care (change score ?19.42, SD = 10.88); a similar pattern approached significance for the youth’s total CBCL scores. Maternal depression was not the only characteristic that varied across households of youth in non-relative foster versus kinship care (Table 2). For example, a higher proportion of children placed with non-relative foster providers (vs. kinship providers) who were always depressed were sexually abused prior to out-of-home placement. In addition, non-relative foster providers who cared for more children were more likely to become depressed. There were no significant differences in caregiver income or number of risk factors by placement status. Multivariate regression revealed that placement setting and trajectory of caregiver depression influenced the
-9.24 (29.40)
4.03 (32.84)
4.95 (30.10)
4.32 (30.36)
Internalizing % change score
Externalizing % change score
Total % change score
.131
.243
.038
.688
.766
.955
.331
p
24.50 (12.58)
12.50 (17.94)
24.25 (22.68)
Mean (SD)
75.00
25.00
75.00
25.00
0.00 75.00
0.00
Foster (n = 4) %
3.95 (30.73)
9.90 (35.02)
-0.81 (33.87)
Mean (SD)
38.10
61.90
19.05
33.33
38.10 28.57
4.76
Kinship (n = 21) %
Became depressed (n = 25)
.207
.888
.172
.173
.022
.165
.656
p
7.00 (28.75)
10.88 (19.42)
-4.25 (25.41)
Mean (SD)
62.50
37.50
37.50
16.67
33.33 50.00
37.50
Foster (n = 8) %
-20.76 (35.24)
-14.65 (24.12)
-15.47 (41.19)
Mean (SD)
25.00
75.00
11.76
35.29
41.18 23.53
41.18
Kinship (n = 17) %
Improved (n = 25)
.065
.016
.488
.074
.134
.450
.861
p
-2.70 (25.88)
4.40 (29.30)
6.20 (18.79)
Mean (SD)
60.00
40.00
44.44
40.00
0.00 60.00
40.00
Foster (n = 10) %
-4.34 (22.05)
-1.05 (22.10)
-3.11 (26.36)
Mean (SD)
52.63
47.37
28.95
45.95
24.32 29.73
10.53
Kinship (n = 38) %
Remained depressed (n = 48)
.841
.520
.302
.677
.370
.106
.026
p
Caregivers were asked if they felt blue a little or most of the time at Time 1 (baseline) and Time 2 (6 months later). The four strata include never depressed, became depressed, improved, and remained depressed between these two time periods. Characteristics of youth across non-relative foster care and kinship care households are reported within each of the four strata. The Child Behavioral Checklist (CBCL) externalizing, internalizing, and total change scores were calculated by subtracting Time 2 from Time 1 data
-5.13 (30.04)
-2.49 (31.31)
Mean (SD)
35.48
64.52
Mean (SD)
39.47
4? risk factors
CBCL
60.53
36.51
29.82
28.57
39.47
33.33 36.84
15.87
Kinship (n = 68) %
31.43 40.00
23.68
Three or less risk factors
Risk factors
Three or more children \9 years
Children under 9
[$40,000
\$ 19,999 $ 20,000–$ 39,999
Caregiver income
Sexual abuse (yes)
Foster (n = 38) %
Never depressed (n = 101)
Table 2 Characteristics of youth across non-relative foster care and kinship care households, stratified by trajectories of caregiver depression over the first 6 months in placement
Matern Child Health J
123
Matern Child Health J
Change in Total CBCL Scores
Expected Change Score
* 25 15
Foster Kinship
5 -5 * -15 *** -25 Never depressed
Became depressed
Improved
Remained depressed
Change in CBCL Internalizing Subscores *
Expected Change Score
15 10 5 0 -5
Foster
-10 ***
Kinship
-15
**
-20 Never depressed
Became depressed
Improved
Remained depressed
Change in CBCL Externalizing Subscores Expected Change Score
20
**
15 10 5
change in CBCL scores for youth over time. The a priori interaction terms were only retained in the total CBCL score model (F = 2.16(7,173), p \ .05), indicating that the relationship of change scores was different within categories of caregiver depression trajectory. Proportional differences in change scores, on the contrary, were not significantly different in models using externalizing and internalizing sub-scores, so the interaction terms in those models were not included. Across all models, youth in kinship care settings more likely to experience better SEB outcomes over time than youth in non-relative foster care settings, controlling for baseline socio-demographic characteristics, sexual abuse history, and duration between interviews (Fig. 1). However, standardized change scores were only significantly improved for some groups of youth. The most significant worsened predicted change scores occurred for non-relative foster care youth whose caregivers became depressed (change score ?25.50, 95 % CI -1.49, ?52.48, p \ .10), while youth in kinship care whose caregivers improved over time experienced the greatest reduction in behavioral problems (change score -18.09, 95 % CI -31.21, -4.96, p \ .01). There were other significant standardized change scores as well. For example, internalizing behaviors decreased substantially among kinship children whose caregivers’ were never depressed (change score -10.03, 95 % CI -17.12, -2.95, p \ .01), and for those caregivers who eventually improved (change score -14.19, 95 % CI -26.84, -1.54, p \ .05). Externalizing behaviors also decreased among children who were placed with kinship caregivers’ whose depressive symptoms improved (change score -9.67, 95 % CI -21.06, ?1.73, p \ .10). On the other hand, youth in non-relative foster care settings whose caregivers became depressed were likely to experience an increase in internalizing (change score ?13.31, 95 % CI -0.93, ?27.55, p \ .10) and externalizing symptoms (change score ?15.99, 95 % CI ?3.15, ?28.82, p \ .05).
0 -5
Foster
Discussion
Kinship
-10 * -15 Never depressed
Became depressed
Improved
Remained depressed
Fig. 1 Total CBCL, externalizing, and internalizing percentile change scores for youth entering new placements of non-relative foster care and kinship care in a Large Urban Mid-Atlantic City. Note CBCL = Child Behavior Checklist. Caregivers were asked if they felt downhearted and blue at time 1 (baseline) and time 2 (6 months later). The four strata include never depressed, became depressed, improved, and remained depressed between time 1 and 2. The expected change scores adjusted for baseline score, caregiver income, the number of children in household under 9, 4? risk factors versus 0–3 risk factors, sexual abuse, and duration between interviews. *p \ .10, **p \ .05, ***p \ .01
123
The purpose of this study was to understand whether potential advantages for youth who are placed with relatives in kinship settings versus non-relatives in foster care are conferred to all youth, or just to subsets of youth whose households were better prepared to meet their needs. While prior studies reference how youth are faring in kinship versus non-relative foster care settings, we hypothesized that setting alone would not confer an advantage to youth, but instead would be modified by household characteristics, particularly the emotional health and well-being of the caregiver. By demonstrating that advantages to youth in kinship care were conferred mostly to those households in
Matern Child Health J
which the caregiver was never depressed or in which the caregiver’s depression improved over time, this study provides policy relevance for child welfare systems that seek kinship settings as a panacea to the challenges faced by their youth, without further investment in strengthening the households after placement. The high rates of depression exhibited in kinship versus non-relative foster care households likewise illustrates why previous studies suggested that kinship children experience worse SEB outcomes than children in the general population [6]. While this is the first study to our knowledge that examined whether changes in caregivers’ depressive symptomatology influences SEB outcomes between younger children placed in non-relative foster versus kinship placement settings, it adds to a literature that has illustrated the impact of maternal depression on child wellbeing. Horwitz et al. [11] found that young children (19-to36 months) investigated for maltreatment who lived with a depressed caregiver were nearly three times more likely to suffer from elevated SEB outcomes compared with children who did not live with a depressed caregiver. Another recent study found that youth between the ages of 12 and 17.5 years referred to the child welfare system are over two times more likely to report anxiety when their caregivers report depression as compared to caregivers who do not endorse depressive symptomatology [34]. Drawing from a large sample of low income families (n = 4,895) first reported to child welfare in 1993 or 1994 and followed until 2009, Kohl et al. [35] found that 67 % of children had a new maltreatment report. However, rates of reporting (80–90 %) and out-of-home placements (more than double) were much higher for children of mothers with mental illness. Irrespective of the different research questions and sampling methods, the findings support the conclusion that child welfare systems that seek to influence SEB outcomes in their youth must prioritize policies and practices that address the mental health needs of caregivers. Implications for Practice and Policy The findings related to caregiver health provide important practice and policy implications to support kinship caregivers and strengthen benefits to children. Using data from the National Survey of America’s Families, Ehrle and Geen [12] found that many kinship caregivers were not receiving public assistance and services (e.g. Medicaid and public assistance), despite the fact that they were more likely than non-kin to suffer from poor mental health. In light of these barriers and our findings, policies that require human service providers and medical professionals alike to routinely assess for need of mental health services for kinship caregivers should be prioritized. If necessary, referrals to evidence-based, culturally congruent services should
ensue. Public health and child welfare caseworkers, for example, may need to conduct periodic home visits to the caregivers’ home to assess need and provide appropriate referrals for services. The Fostering Connections to Success and Increasing Adoption Act of 2008 provides funding to states for guardianship assistance payments and authorizes grants to local child welfare agencies for kinship navigator programs designed to provide support services to kinship care families [36]. These resources may facilitate access to caregiver services once needs are identified. Our study also has direct implications for effectively working with African American families, who are disproportionately over-represented in the foster care system [37– 39]. In the current study, three-fourths of the caregivers and most of the children (particularly those in kinship care) were identified as African American. Harris and Skyles [17] argue, that kinship care is misused in child welfare practice with African American children, given that children in these placements experience lower rates of reunification than children placed in foster care. Others contend that kinship care is a response to ameliorate family separation [40]. Whatever the case, when reunification is not a viable option, caseworkers must identify viable, enduring kinship placement settings—settings in which caregivers receive efficacious interventions to promote positive mental health outcomes for their sake and the sake of the children they care for. Added supports to accomplish these goals may facilitate permanency in a timely manner for many of the African American children who are placed in out-of-home care. Limitations While the current study provides a valuable contribution to the growing body of literature on SEB outcomes among children placed in kinship settings, there are a few limitations. First, the non-relative foster care sample is relatively small compared to the number of children in the kinship group. The variability may have led to insufficient power to detect significant differences between groups and across depression trajectories. Second, we concede that our maternal depression measure is imperfect, yet our findings were robust to the methodology by which we used the SF12 mental health summary scores and single items, making us more confident in the nature of the findings. Third, the current study did not include a comparison sub-sample of children without abuse histories. Nonetheless, the purpose of the current study was to examine variation within the high risk group. Fourth, parental characteristics are one of many factors that may influence SEB outcomes among children with trauma histories. For example, the organizational culture and climate of child welfare agencies, access to health and mental health services, and
123
Matern Child Health J
neighborhood and community characteristics (e.g., poverty, crime, gang violence) where children reside may directly and/or indirectly influence caregiver and child outcomes. Finally, results should be interpreted with the acknowledgement that caregiver reporting of child behavior in the presence of depression is complex. Prior research indicates that depressed caregivers may be more likely to endorse negative SEB outcomes [41]. At the same time, children may be less likely to exhibit negative SEB outcomes in the presence of maternal depression [42, 43].
3.
Future Directions
5.
Although most research has been observational, the findings of this and other studies that illustrated the impact of caregiver health on the SEB outcomes of child welfareinvolved youth would suggest that intervention studies is a logical next step. That is, what is the potential impact of caregiver-directed interventions within kinship settings in promoting positive SEB outcomes and placement stability for youth placed in out-of-home settings? There is also a need to disentangle whether the relationship between organizational and community factors and SEB outcomes among youth in kinship settings are mediated by caregiver health and mental health outcomes. Such efforts may illuminate concrete recommendations for evidenceinformed policy and practice. Our study is the first to address the interaction between caregiver status and SEB outcomes among at-risk youth placed in kinship and non-relative foster care settings. Minimally, providers and policymakers must allocate more resources to promote positive outcomes among kinship caregivers and must resist the temptation to view the placement into kinship care as a panacea itself. Findings lend strong support that such efforts will play a significant role in improving SEB outcomes among children who experience traumatic histories and may strengthen the advantages in placement stability conferred to youth in kinship settings. Acknowledgments We thank the Philadelphia Department of Human Services and the foster care agencies and providers for their contributions to this study. Robin Hernandez-Mekonnen contributed to primary data collection, and Christina Kang-Yi assisted with data preparation and analysis. This project was supported by a grant from the William Penn Foundation and a fellowship to Dr. Rubin from the Stoneleigh Foundation.
2.
4.
6.
7.
8.
9. 10.
11.
12.
13.
14. 15.
16.
17.
18.
References 19. 1. U.S. Department of Health and Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau. The AFCARS report: preliminary estimates for FY 2011 as of July 2012. [Cited 2013 May
123
20.
23]. http://www.acf.hhs.gov/sites/default/files/cb/afcarsreport19. pdf. Annie E. Casey Foundation. (2012). Kids Count Data Book. [Cited 2013 Oct 24]. http://www.aecf.org/*/media/Pubs/Initia tives/KIDS%20COUNT/123/2012KIDSCOUNTDataBook/KIDS COUNT2012DataBookFullReport.pdf. Chipungu, S. S., Everett, J. E., Verduik, M. J., & Jones, H. (1998). Children placed in foster care with relatives: A Multi-site Study. Washington, DC: U.S. Department of Health and Human Services. Frank, D., Augustyn, M., Knight, W., Pell, T., & Zuckerman, B. (2001). Growth, development and behavior in early childhood following prenatal cocaine exposure: A systematic review. JAMA, 285(12), 1613–1625. Jones, J. K. (1998). Functioning and adjustment of children in kinship care versus nonrelative foster family care placements. Buffalo: University of New York at Buffalo. (dissertation). Rubin, D. M., Downes, K. J., O’Reilly, A. R., Mekonnen, R., Luan, X., & Localio, R. (2008). Impact of kinship care on behavioral well-being for children in out-of-home care. Archives of Pediatrics & Adolescent Medicine, 162(6), 550–556. Barth, R. P., Shenyang, G., & McCrae, J. S. (2007). Propensity score matching strategies for evaluating the success of child and family service programs. Research on Social Work Practice, 18(3), 212–222. Kemp, C. Children living with relatives struggle with physical, mental health issues. American Academy of Pediatrics News, 2010 May 2. [Cited 2013 March 21]. http://aapnews.aappublica tions.org/cgi/content/full/aapnews.20100502-3v1?rss=1. Berrick, J. D. (1998). When children cannot remain home: Foster family care and kinship care. Future of Children, 8(1), 72–87. Stahmer, A. C., Leslie, L. K., Hurlburt, M., et al. (2005). Developmental and behavioral needs and service use for young children in child welfare. Pediatrics, 116(4), 891–900. Horwitz, S. M., Hurlburt, M. S., Heneghan, A., et al. (2012). Mental health problems in young children investigated by U.S. child welfare agencies. Journal of the American Academy of Child & Adolescent Psychiatry, 51(6), 572–581. Ehrle, J., & Geen, R. (2002). Kin and non-kin foster care-findings from a national survey. Children and Youth Services Review, 24(1–2), 15–35. Berrick, J., Barth, R., & Needell, B. (1994). A comparison of kinship foster homes and foster family homes: Implications for kinship foster care as family preservation. Children and Youth Services Review, 16(1–2), 33–63. Barth, R., Courtney, M., Berrick, J., & Albert, V. (1994). From child abuse to permanency planning. New York, NY: Aldine de Gruyter. Harden, B. J., Clyman, R. B., Kriebel, D. K., & Lyons, M. E. (2004). Kith and kin care: Parental attitudes and resources of foster and relative caregivers. Children and Youth Services Review, 26(7), 657–671. Scannapieco, M., Hegar, R. L., & McAlpine, C. (1997). Kinship care and foster care: A comparison of characteristics and outcomes. Families in Society, 78(5), 480–488. Harris, M. S., & Skyles, A. (2008). Kinship care for African American children: Disproportionate and disadvantageous. Journal of Family Issues, 29(8), 1013–1030. Cheng, T. C., & Lo, C. C. (2011). A longitudinal analysis of some risk and protective factors in marijuana use by adolescents receiving child welfare services. Children and Youth Services Review, 33(9), 1667–1672. Grogan-Kaylor, A., Ruffolo, M. C., Ortega, R. M., & Clarke, J. (2008). Behaviors of youth involved in the child welfare system. Child Abuse & Neglect, 32(1), 35–49. Harden, B. J., & Whittaker, J. V. (2011). The early home environment and developmental outcomes for young children in the
Matern Child Health J
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
child welfare system. Children and Youth Services Review, 33(8), 1392–1403. Horwitz, S. M., Hurlburt, M. S., Cohen, S. D., Zhang, J., & Landsverk, J. (2011). Predictors of placement for children who initially remained in their homes after an investigation for abuse or neglect. Child Abuse & Neglect, 35(3), 188–198. Sprang, G., Clark, J. J., & Staton-Tindall, M. (2010). Caregiver substance use and trauma exposure in young children. Families in Society, 91(4), 401–407. Hines, A. M., Lemon, K., Wyatt, P., & Merdinger, J. (2004). Factors related to the disproportionate involvement of children of color in the child welfare system: A review and emerging themes. Children and Youth Services Review, 26(6), 507–527. Fuller-Thomson, E., & Minkler, M. (2000). African American grandparents raising grandchildren: A national profile of demographic and health characteristics. Health and Social Work, 25(2), 109–118. Fuller-Thomson, E., Minkler, M., & Driver, D. (1997). A profile of grandparents raising grandchildren in the United States. Gerontologist, 37(3), 406–411. Whitley, D. M., Kelley, S. J., & Sipe, T. A. (2001). Grandmothers raising grandchildren: Are they at increased risk for health problems? Health and Social Work, 26(2), 105–114. Achenbach, T. M. (1991). Manual for child behavior checklist 4-18. Burlington, VT: University of Vermont, Department of Psychiatry. Gandek, B., Ware, J. E., Aaronson, N. K., et al. (1998). Crossvalidation of item selection and scoring for the SF-12 health survey in nine countries: results from the IQOLA Project. Journal of Clinical Epidemiology, 51(11), 1171–1178. Ware, J. E, Jr, Kosinski, M., & Keller, S. D. (1996). A 12-Item short-form health survey: Construction of scales and preliminary tests of reliability and validity. Medical Care, 34(3), 220–233. Brazier, J. E., & Robert, J. (2004). The estimation of a preference-based measure of health from the SF-12. Medical Care, 42(9), 851–859. Jenkinson, C., Layte, R., Jenkinson, D., et al. (1997). A shorter form health survey: Can the SF-12 replicate results from the SF36 in longitudinal studies. Journal of Public Health, 19(2), 179–186. Vilagut, G., Forero, C. G., Pinto-Meza, A., et al. (2013). The mental component of the Short-Form 12 Health Survey (SF-12)
33. 34.
35.
36.
37.
38.
39. 40.
41.
42.
43.
as a measure of depressive disorders in the general population: Results with three alternative scoring methods. Value in Health, 16(4), 564–573. StataCorp. (2007). Stata Statistical Software: Release 10. College Station, TX: StataCorp LP. Heneghan, A., Stein, R. E. K., Hurlburt, M. S., et al. (2013). Mental health problems in teens investigated by U.S. child welfare agencies. Journal of Adolescent Health, 52(5), 634–640. Kohl, P. L., Jonson-Reid, M., & Drake, B. (2011). Maternal mental illness and the safety and stability of maltreated children. Child Abuse & Neglect, 35(5), 309–318. Fostering Connections to Success and Increasing Adoption Act of 2008 P.L. 110-351. Child Welfare Information Gateway Web site. [Published 2008; cited 2013 March 21]. http://www.child welfare.gov/systemwide/laws_policies/federal/index.cfm?event= federalLegislation.viewLegis&id=121. Hill, R. (2006). Synthesis of research on disproportionality in child welfare: An update. [Cited 2013 March 21]. http://www.cssp.org/ reform/child-welfare/other-resources/synthesis-of-research-on-dis proportionality-robert-hill.pdf. Knott, T., & Donovan, K. (2010). Disproportionate representation of African-American children in foster care: Secondary analysis of the National Child Abuse and Neglect Data System, 2005. Children and Youth Services Review, 32(5), 679–684. Roberts, D. (2002). Shattered bonds: The color of child welfare. New York, NY: Civitas Books. Scannapieco, M., & Jackson, S. (1996). Kinship care: The African American response to family preservation. Social Work, 41(2), 190–196. Boyle, M., & Pickles, A. (1997). Influence of maternal depressive symptoms on ratings of childhood behavior. Journal of Abnormal Child Psychology, 25, 399–412. Hops, H., Biglan, A., Sherman, L., Arthur, J., Friedman, L., & Osteen, V. (1987). Home observations of family interactions of depressed women. Journal of Consulting and Clinical Psychology, 55(3), 341–346. Webster-Stratton, C., & Hammond, M. (1988). Maternal depression and its relationship to life stress, perceptions of child-behavior problems, parenting behaviors, and child conduct problems. Journal of Abnormal Child Psychology, 16(3), 299–315.
123
The Influence of Caregiver Depression on Children in Non-relative Foster Care Versus Kinship Care Placements Antonio Garcia • Amanda O’Reilly • Meredith Matone • Minseop Kim • Jin Long David M. Rubin
•
Ó Springer Science+Business Media New York 2014
Abstract Little is known about how the challenges faced by caregivers influence the variation in social, emotional, and behavioral (SEB) outcomes of youth placed in kinship versus non-relative foster care. This study examined SEB symptoms among youth in kinship and non-relative foster care settings, hypothesizing that changes in caregiver depression would modify children’s change in behavior over time. Child Behavior Checklist (CBCL) assessments of 199 children placed with kinship and non-relative foster care providers in a Mid-Atlantic city were conducted at time of placement and 6–12 months post-placement. Linear regression estimated CBCL change scores for youth across placement type and caregiver depression trajectories. Kinship caregivers were more likely to become depressed or remained depressed than non-relative foster caregivers. Youth in kinship care always exhibited better change in SEB outcomes than youth in non-relative foster care, but these positive outcomes were principally observed among families where caregivers demonstrated a reduction
A. Garcia (&) M. Kim School of Social Policy and Practice, University of Pennsylvania, 3701 Locust Walk, Philadelphia, PA 19104-6214, USA e-mail: [email protected] A. O’Reilly Division of Children and Families, State of New Jersey, Trenton, NJ, USA M. Matone J. Long D. M. Rubin PolicyLab, The Children’s Hospital of Philadelphia, Philadelphia, PA, USA D. M. Rubin Department of Pediatrics at the Perelman School of Medicine, The University of Pennsylvania, Philadelphia, PA, USA
in depression over time or were never depressed. Adjusted change scores for non-relative foster care youth were always negative, with the most negative scores among youth whose caregivers became depressed over time. Caregiver well-being may modify the influence of placement setting on SEB outcomes for youth placed into outof-home care. Findings lend to policy relevance for child welfare systems that seek kinship settings as a panacea to the challenges faced by youth, without allocating resources to address caregiver needs. Keywords Maltreatment School age children Maternal depression Social-emotional problems Prospective study
Introduction Over 400,000 children in the U.S. were placed in nonrelative foster care or kinship care in 2011 due to maltreatment [1]. Twenty-seven percent of these children were placed in kinship care. Kinship care as an alternative to non-relative foster care placements have grown considerably over the last two decades; by 2010, 2.7 million children in this country were living in kinship arrangement, a 70 % increase from just 20 years prior [2]. As the numbers of children in kinship care have grown, prior studies have demonstrated a mixed picture of social, emotional, and behavioral (SEB) outcomes among these youth. On the one hand, several studies show that children in kinship care experience less SEB problems compared to children placed in non-relative foster care [3–6]. For example, one study shows that children in kinship care experience less externalizing behaviors after 18 months of placement [7]. Another study reports that regardless of prior placement history, children in kinship care are 50 % less likely to
123
Matern Child Health J
exhibit behavioral problems 3 years after placement compared to children in non-relative foster care [6]. However, rates of SEB problems among children in kinship care exceed rates of behavioral problems relative to children living in poverty with birth parents [6, 8]. Thus, it is imperative to examine why children placed in kinship care do not fully benefit from being placed in a home that preserves family connectedness [9]. Prior research illuminates why children in kinship care may be at an increased risk to experience SEBs relative to children in the general population. First and foremost are their trauma histories. Due to severe negligent treatment and/or abuse, young children who enter out-of-home care are more likely than those without exposure to multiple forms and types of maltreatment to be diagnosed with elevated behavioral and mental health problems and developmental delays [10, 11]. Aside from maltreatment history, socio-demographic risk factors are often pervasive within families, and thus children placed with kin retain exposure to risk factors for poor well-being, perhaps at higher rates than peers placed in non-relative foster care. For example, kinship caregivers in the U.S. are older [9, 12], earn less income than foster care providers, and are more likely to live in impoverished conditions [9, 13]. These disparities are likely associated with lower educational attainment [14] and higher rates of single parenting [14–16] among these caregivers. Even with comparatively fewer socio-demographic resources, kinship caregivers are likely to provide care for the same number of children as non-relative foster care providers do [13, 17]. Caregiver responsibilities are further challenged by a host of risk factors for adverse child SEB outcomes originating prior to out-of-home placement, including parental mental health problems, poor parenting behavior, substance abuse, parental criminal involvement, and chronic maltreatment [18–23]. Collectively, it is not known the extent to which these stressors may take a negative toll on the caregivers’ own mental health. Several studies show that grandparents raising grandchildren experience increased depression compared to grandparents not caring for their grandchildren [24–26]. Little is known about how these stressors, often exacerbated by mental health symptomatology faced by caregivers, influence variation in behavioral outcomes of youth placed in kinship versus non-relative foster care. The degree that caregiver health might influence the outcomes of youth placed in kinship care would have immediate policy relevance, suggesting that placement with kin might need to be accompanied by greater attention to caregiver health to augment benefits to youth. The current study examined behavioral assessments of youth placed in both settings at baseline (time of out-of-home placement) and the second interview conducted between 6 and
123
12 months after placement, hypothesizing that changes in caregiver depression would modify children’s change in behavior over time and across kinship and non-relative foster settings, controlling for maltreatment history and child and family characteristics.
Methods Study Design and Participants The sample was drawn from a prospective, longitudinal cohort of 405 children between ages 3 and 8 years who were recruited from a large, Mid-Atlantic city’s child welfare system upon a new placement into out-of-home care from 2006 to 2008. Children were followed for 24 months from their entry into out-of-home care; with caregiver interviews conducted each time a placement change occurred or every 6 months within stable placements. The sample was restricted to those children (N = 199) placed with foster and kinship care providers who remained with their initial caregiver between baseline and the second interview. Measures The primary outcome measures were child behavioral change scores from baseline to second interview assessed with the Child Behavior Checklist (CBCL). The CBCL is a widely used, valid, and reliable measure of clinically pervasive symptoms of externalizing and/or internalizing behaviors evaluated by caregivers [27]. Each item operates on a 3-point Likert scale (0 = not true, 1 = somewhat or sometimes true, and 2 = very true or often true) to rate presence of symptoms [27]. Non-relative foster and kinship caregivers were administered the CBCL at baseline and at each subsequent interview. For each child, internalizing, externalizing, and total percentile change scores were calculated by subtracting time 1 (baseline) percentiles from time 2 percentiles. The primary exposures for the study were the placement setting and the assessment of caregiver depression between time 1 and 2. Placement setting was operationalized as non-relative foster care versus kinship care setting. Caregiver depression was assessed using the single item of the validated SF-12 Health Survey of adult health [28, 29] that asked about feeling blue or sad. The SF-12 survey is a widely used, validated 12-item scale of mental and physical health domains derived from the SF-36 Health Survey. The SF-12 is scored along individual items as well as summary scores for physical and mental health domains [28–31]. For the purposes of the current study, depression was operationalized as a single survey item regarding the
Matern Child Health J Table 1 Descriptive characteristics of youth entering new placements in non-relative foster care and kinship care settings Total (n = 199) n (%)
Foster care (n = 60) n (%)
Kinship care (n = 139) n (%)
p
Caregiver race Caucasian African American Latino and ‘‘other’’ Sexual abuse (yes)
27 (13.7)
6 (10)
21 (15.3)
.32
146 (74.1) 24 (12.2)
42 (70) 12 (20)
104 (75.9) 12 (8.8)
.38 .03
38 (19)
16 (27)
22 (16)
Caregiver income \$ 19,999
13 (23.6)
43 (32.6)
$ 20,000–$ 39,999
68 (36.4)
26 (47.3)
42 (31.8)
C$40,000
63 (33.7)
16 (29.1)
47 (35.6)
56 (30)
25 (42)
40 (29)
Children under 9 Three or more children \9 years
.07 .13
.06 65 (33)
Risk factors§ Three or less risk factors
114 (58)
31 (52)
83 (61)
83 (42)
29 (48)
54 (39)
101 (51)
38 (63)
63 (45)
Became depressed
25 (13)
4 (7)
21 (15)
Improved Remained depressed
25 (13) 48 (24)
8 (13) 10 (17)
17 (12) 38 (27)
4? risk factors Caregiver depression (between T1 and T2) Never depressed
CBCL
.24 .07
Mean (SD)
Mean (SD)
Mean (SD)
Internalizing % change score
-3.53 (30.93)
4.63 (29.50)
-7.05 (30.97)
.01
Externalizing % change score
0.66 (28.98)
6.15 (27.70)
-1.71 (29.29)
.08
Total % change score
-2.35 (29.44)
4.85 (28.68)
-5.45 (29.32)
.02
§
Risk factors include parental substance abuse, mental health, and incarceration; and domestic violence, homelessness, and prior foster care placement
extent caregivers’ felt ‘‘downhearted and blue’’ on a scale from 1 (none of the time) to 5 (most of the time). Caregivers who reported feeling downhearted and blue ‘‘a little’’, ‘‘some’’, ‘‘most’’, or ‘‘all of the time’’ were denoted as depressed and others who reported ‘‘none of the time’’ were denoted as not depressed. Because the response to the depressive symptoms item may vary with time, and given the concern that caregiver depression might also influence the likelihood of reporting child behavioral problems, we created strata of caregiver depression status over time to use as a priori effect modifier in later models. The four strata included ‘‘never depressed,’’ ‘‘became depressed’’, ‘‘improved,’’ and ‘‘remained depressed’’ to indicate the change between time 1 and time 2 depression status. At each interview, kinship and non-relative foster caregivers provided data on their income and the number of children in out-of-home placement in the household. Case record data included prior history of abuse, placement setting (kinship or non-relative foster care), and biological parental risk factors that increased the likelihood of out-of-
home placement (parental substance abuse, mental health, incarceration; as well as domestic violence and homelessness). We opted to only control for sexual abuse (versus controlling for all types of maltreatment), given its unique nature and the finding that neglect and physical abuse were highly prevalent in the sample (data not shown). Analysis Plan Linear regression estimated behavioral percentile change scores for youth across placement type and caregiver depression categories between baseline and follow-up, controlling for sexual abuse history, duration of time in days between interviews, and the following baseline sociodemographic characteristics: (1) caregiver income (\$20,000; $20,000–40,000; [$40,000) (2) providing care for a large number (i.e., C3) of young children under 9 (33 % of the caregivers met this criteria), and (3) number of parental risk factors, dichotomized into a binary variable based on mean distributions of the sample (42 %
123
Matern Child Health J
experienced four or more risk factors). Models were elaborated for the change in total score, externalizing subscore, and internalizing sub-score percentages of the CBCL between time 1 and 2. Based on a priori hypothesis, an interaction term for placement setting and caregiver depression trajectory was included in all models. This interaction term was retained only for the change in total score model where the interaction term was significant at the p \ .05 level. Post-estimation commands (margins in STATA) were then used to estimate predicted change scores on the CBCL. Although the binary representation of a single item was easy to operationalize, we also acknowledged the need for a sensitivity analysis to validate its use against the full SF12 mental health summary scores at time 1 and 2. We derived our analogous four different categories of depression trajectory over time by relying upon scoring procedures set forth by Vilagut et al. [32]. Specifically, their sensitivity and specificity analyses found that the optimal cutoff score for 30-day depressive disorders in the general population was 45.6. Thus, kinship and non-relative caregivers who scored below 45.6 were denoted as having depressive symptoms and those above 45.6 were denoted as not having depressive symptoms. Caregiver status as ‘‘never depressed,’’ ‘‘became depressed,’’ ‘‘improved,’’ or ‘‘remained depressed’’ was determined by their score as above or below 45.6 at times 1 and 2. All results were consistent with this summary score specification (estimates not reported), indicating our findings to be robust to the single item measure within the SF-12. In the end, because of the simplicity of the single-item with respect to categorizing change over time, we report our model results with that variable. Statistical analyses were performed using Stata 10.0 software [33]. The study was approved by the institutional review board of The Children’s Hospital of Philadelphia.
Results Descriptive characteristics for the sample of youth and care providers in the study are reported in Table 1. Among youth who remained stable in placement between baseline and second interview, 70 % (n = 139) were in kinship care, and 30 % (n = 60) were in non-relative foster care placement. Most of the caregivers identified as African American (74 %). Foster providers (42 %) were more likely than non-relative foster care (29 %) providers to care for three or more foster children. Children placed in nonrelative foster care settings (27 %) were more likely to have been sexually abused than those placed in kinship care (16 %). While a higher percentage of kinship families lived in poverty across all income categories than non-
123
relative foster care families, the differences were not statistically significant. Moreover, a higher percentage of kinship caregivers became depressed between time 1 and 2 (15 vs. 7 %) and remained depressed over the period from time 1 and time 2 (27 vs. 17 %) than non-relative foster caregivers. Table 1 also shows that children in kinship care were more likely to experience positive CBCL behavioral change scores than those placed in non-relative foster care. Children placed in kinship care were more likely to benefit from decreased internalizing symptoms (change score -7.05, SD = 30.97) than children placed in non-relative foster care (change score ?4.63, SD = 29.50). Children in kinship settings were more likely to benefit from a reduction in total CBCL scores (change score -5.45, SD = 29.32) relative to children in non-relative foster care (change score ?4.85, SD = 28.68). Children in kinship households were more likely to reveal fewer externalizing behaviors (change score -1.71, SD = 29.29) compared to children in non-relative foster care (change score ?6.15, SD = 27.70). Apart from trajectories of maternal depression varying across placement settings, there were also consistent differences across those trajectories in the average CBCL change scores (Table 2). In general, across all caregiver depression categories (never depressed, became depressed, improved, remained depressed), the change in scores for youth in kinship care always exhibited better internalizing, externalizing, and total behavior change scores than youth in non-relative foster care settings. Noteworthy is that children placed with kinship providers (change score -9.24, SD = 29.4) who were never depressed were more likely to experience a significant decrease in internalizing behaviors between baseline and follow-up than children placed with non-relative foster providers (change score ?4.03, SD = 32.84). Moreover, youth in kinship care (change score -14.65, SD = 24.12) whose caregivers’ depressive symptomatology got better over time were also more likely to benefit from a decrease in externalizing behaviors than those placed in non-relative foster care (change score ?19.42, SD = 10.88); a similar pattern approached significance for the youth’s total CBCL scores. Maternal depression was not the only characteristic that varied across households of youth in non-relative foster versus kinship care (Table 2). For example, a higher proportion of children placed with non-relative foster providers (vs. kinship providers) who were always depressed were sexually abused prior to out-of-home placement. In addition, non-relative foster providers who cared for more children were more likely to become depressed. There were no significant differences in caregiver income or number of risk factors by placement status. Multivariate regression revealed that placement setting and trajectory of caregiver depression influenced the
-9.24 (29.40)
4.03 (32.84)
4.95 (30.10)
4.32 (30.36)
Internalizing % change score
Externalizing % change score
Total % change score
.131
.243
.038
.688
.766
.955
.331
p
24.50 (12.58)
12.50 (17.94)
24.25 (22.68)
Mean (SD)
75.00
25.00
75.00
25.00
0.00 75.00
0.00
Foster (n = 4) %
3.95 (30.73)
9.90 (35.02)
-0.81 (33.87)
Mean (SD)
38.10
61.90
19.05
33.33
38.10 28.57
4.76
Kinship (n = 21) %
Became depressed (n = 25)
.207
.888
.172
.173
.022
.165
.656
p
7.00 (28.75)
10.88 (19.42)
-4.25 (25.41)
Mean (SD)
62.50
37.50
37.50
16.67
33.33 50.00
37.50
Foster (n = 8) %
-20.76 (35.24)
-14.65 (24.12)
-15.47 (41.19)
Mean (SD)
25.00
75.00
11.76
35.29
41.18 23.53
41.18
Kinship (n = 17) %
Improved (n = 25)
.065
.016
.488
.074
.134
.450
.861
p
-2.70 (25.88)
4.40 (29.30)
6.20 (18.79)
Mean (SD)
60.00
40.00
44.44
40.00
0.00 60.00
40.00
Foster (n = 10) %
-4.34 (22.05)
-1.05 (22.10)
-3.11 (26.36)
Mean (SD)
52.63
47.37
28.95
45.95
24.32 29.73
10.53
Kinship (n = 38) %
Remained depressed (n = 48)
.841
.520
.302
.677
.370
.106
.026
p
Caregivers were asked if they felt blue a little or most of the time at Time 1 (baseline) and Time 2 (6 months later). The four strata include never depressed, became depressed, improved, and remained depressed between these two time periods. Characteristics of youth across non-relative foster care and kinship care households are reported within each of the four strata. The Child Behavioral Checklist (CBCL) externalizing, internalizing, and total change scores were calculated by subtracting Time 2 from Time 1 data
-5.13 (30.04)
-2.49 (31.31)
Mean (SD)
35.48
64.52
Mean (SD)
39.47
4? risk factors
CBCL
60.53
36.51
29.82
28.57
39.47
33.33 36.84
15.87
Kinship (n = 68) %
31.43 40.00
23.68
Three or less risk factors
Risk factors
Three or more children \9 years
Children under 9
[$40,000
\$ 19,999 $ 20,000–$ 39,999
Caregiver income
Sexual abuse (yes)
Foster (n = 38) %
Never depressed (n = 101)
Table 2 Characteristics of youth across non-relative foster care and kinship care households, stratified by trajectories of caregiver depression over the first 6 months in placement
Matern Child Health J
123
Matern Child Health J
Change in Total CBCL Scores
Expected Change Score
* 25 15
Foster Kinship
5 -5 * -15 *** -25 Never depressed
Became depressed
Improved
Remained depressed
Change in CBCL Internalizing Subscores *
Expected Change Score
15 10 5 0 -5
Foster
-10 ***
Kinship
-15
**
-20 Never depressed
Became depressed
Improved
Remained depressed
Change in CBCL Externalizing Subscores Expected Change Score
20
**
15 10 5
change in CBCL scores for youth over time. The a priori interaction terms were only retained in the total CBCL score model (F = 2.16(7,173), p \ .05), indicating that the relationship of change scores was different within categories of caregiver depression trajectory. Proportional differences in change scores, on the contrary, were not significantly different in models using externalizing and internalizing sub-scores, so the interaction terms in those models were not included. Across all models, youth in kinship care settings more likely to experience better SEB outcomes over time than youth in non-relative foster care settings, controlling for baseline socio-demographic characteristics, sexual abuse history, and duration between interviews (Fig. 1). However, standardized change scores were only significantly improved for some groups of youth. The most significant worsened predicted change scores occurred for non-relative foster care youth whose caregivers became depressed (change score ?25.50, 95 % CI -1.49, ?52.48, p \ .10), while youth in kinship care whose caregivers improved over time experienced the greatest reduction in behavioral problems (change score -18.09, 95 % CI -31.21, -4.96, p \ .01). There were other significant standardized change scores as well. For example, internalizing behaviors decreased substantially among kinship children whose caregivers’ were never depressed (change score -10.03, 95 % CI -17.12, -2.95, p \ .01), and for those caregivers who eventually improved (change score -14.19, 95 % CI -26.84, -1.54, p \ .05). Externalizing behaviors also decreased among children who were placed with kinship caregivers’ whose depressive symptoms improved (change score -9.67, 95 % CI -21.06, ?1.73, p \ .10). On the other hand, youth in non-relative foster care settings whose caregivers became depressed were likely to experience an increase in internalizing (change score ?13.31, 95 % CI -0.93, ?27.55, p \ .10) and externalizing symptoms (change score ?15.99, 95 % CI ?3.15, ?28.82, p \ .05).
0 -5
Foster
Discussion
Kinship
-10 * -15 Never depressed
Became depressed
Improved
Remained depressed
Fig. 1 Total CBCL, externalizing, and internalizing percentile change scores for youth entering new placements of non-relative foster care and kinship care in a Large Urban Mid-Atlantic City. Note CBCL = Child Behavior Checklist. Caregivers were asked if they felt downhearted and blue at time 1 (baseline) and time 2 (6 months later). The four strata include never depressed, became depressed, improved, and remained depressed between time 1 and 2. The expected change scores adjusted for baseline score, caregiver income, the number of children in household under 9, 4? risk factors versus 0–3 risk factors, sexual abuse, and duration between interviews. *p \ .10, **p \ .05, ***p \ .01
123
The purpose of this study was to understand whether potential advantages for youth who are placed with relatives in kinship settings versus non-relatives in foster care are conferred to all youth, or just to subsets of youth whose households were better prepared to meet their needs. While prior studies reference how youth are faring in kinship versus non-relative foster care settings, we hypothesized that setting alone would not confer an advantage to youth, but instead would be modified by household characteristics, particularly the emotional health and well-being of the caregiver. By demonstrating that advantages to youth in kinship care were conferred mostly to those households in
Matern Child Health J
which the caregiver was never depressed or in which the caregiver’s depression improved over time, this study provides policy relevance for child welfare systems that seek kinship settings as a panacea to the challenges faced by their youth, without further investment in strengthening the households after placement. The high rates of depression exhibited in kinship versus non-relative foster care households likewise illustrates why previous studies suggested that kinship children experience worse SEB outcomes than children in the general population [6]. While this is the first study to our knowledge that examined whether changes in caregivers’ depressive symptomatology influences SEB outcomes between younger children placed in non-relative foster versus kinship placement settings, it adds to a literature that has illustrated the impact of maternal depression on child wellbeing. Horwitz et al. [11] found that young children (19-to36 months) investigated for maltreatment who lived with a depressed caregiver were nearly three times more likely to suffer from elevated SEB outcomes compared with children who did not live with a depressed caregiver. Another recent study found that youth between the ages of 12 and 17.5 years referred to the child welfare system are over two times more likely to report anxiety when their caregivers report depression as compared to caregivers who do not endorse depressive symptomatology [34]. Drawing from a large sample of low income families (n = 4,895) first reported to child welfare in 1993 or 1994 and followed until 2009, Kohl et al. [35] found that 67 % of children had a new maltreatment report. However, rates of reporting (80–90 %) and out-of-home placements (more than double) were much higher for children of mothers with mental illness. Irrespective of the different research questions and sampling methods, the findings support the conclusion that child welfare systems that seek to influence SEB outcomes in their youth must prioritize policies and practices that address the mental health needs of caregivers. Implications for Practice and Policy The findings related to caregiver health provide important practice and policy implications to support kinship caregivers and strengthen benefits to children. Using data from the National Survey of America’s Families, Ehrle and Geen [12] found that many kinship caregivers were not receiving public assistance and services (e.g. Medicaid and public assistance), despite the fact that they were more likely than non-kin to suffer from poor mental health. In light of these barriers and our findings, policies that require human service providers and medical professionals alike to routinely assess for need of mental health services for kinship caregivers should be prioritized. If necessary, referrals to evidence-based, culturally congruent services should
ensue. Public health and child welfare caseworkers, for example, may need to conduct periodic home visits to the caregivers’ home to assess need and provide appropriate referrals for services. The Fostering Connections to Success and Increasing Adoption Act of 2008 provides funding to states for guardianship assistance payments and authorizes grants to local child welfare agencies for kinship navigator programs designed to provide support services to kinship care families [36]. These resources may facilitate access to caregiver services once needs are identified. Our study also has direct implications for effectively working with African American families, who are disproportionately over-represented in the foster care system [37– 39]. In the current study, three-fourths of the caregivers and most of the children (particularly those in kinship care) were identified as African American. Harris and Skyles [17] argue, that kinship care is misused in child welfare practice with African American children, given that children in these placements experience lower rates of reunification than children placed in foster care. Others contend that kinship care is a response to ameliorate family separation [40]. Whatever the case, when reunification is not a viable option, caseworkers must identify viable, enduring kinship placement settings—settings in which caregivers receive efficacious interventions to promote positive mental health outcomes for their sake and the sake of the children they care for. Added supports to accomplish these goals may facilitate permanency in a timely manner for many of the African American children who are placed in out-of-home care. Limitations While the current study provides a valuable contribution to the growing body of literature on SEB outcomes among children placed in kinship settings, there are a few limitations. First, the non-relative foster care sample is relatively small compared to the number of children in the kinship group. The variability may have led to insufficient power to detect significant differences between groups and across depression trajectories. Second, we concede that our maternal depression measure is imperfect, yet our findings were robust to the methodology by which we used the SF12 mental health summary scores and single items, making us more confident in the nature of the findings. Third, the current study did not include a comparison sub-sample of children without abuse histories. Nonetheless, the purpose of the current study was to examine variation within the high risk group. Fourth, parental characteristics are one of many factors that may influence SEB outcomes among children with trauma histories. For example, the organizational culture and climate of child welfare agencies, access to health and mental health services, and
123
Matern Child Health J
neighborhood and community characteristics (e.g., poverty, crime, gang violence) where children reside may directly and/or indirectly influence caregiver and child outcomes. Finally, results should be interpreted with the acknowledgement that caregiver reporting of child behavior in the presence of depression is complex. Prior research indicates that depressed caregivers may be more likely to endorse negative SEB outcomes [41]. At the same time, children may be less likely to exhibit negative SEB outcomes in the presence of maternal depression [42, 43].
3.
Future Directions
5.
Although most research has been observational, the findings of this and other studies that illustrated the impact of caregiver health on the SEB outcomes of child welfareinvolved youth would suggest that intervention studies is a logical next step. That is, what is the potential impact of caregiver-directed interventions within kinship settings in promoting positive SEB outcomes and placement stability for youth placed in out-of-home settings? There is also a need to disentangle whether the relationship between organizational and community factors and SEB outcomes among youth in kinship settings are mediated by caregiver health and mental health outcomes. Such efforts may illuminate concrete recommendations for evidenceinformed policy and practice. Our study is the first to address the interaction between caregiver status and SEB outcomes among at-risk youth placed in kinship and non-relative foster care settings. Minimally, providers and policymakers must allocate more resources to promote positive outcomes among kinship caregivers and must resist the temptation to view the placement into kinship care as a panacea itself. Findings lend strong support that such efforts will play a significant role in improving SEB outcomes among children who experience traumatic histories and may strengthen the advantages in placement stability conferred to youth in kinship settings. Acknowledgments We thank the Philadelphia Department of Human Services and the foster care agencies and providers for their contributions to this study. Robin Hernandez-Mekonnen contributed to primary data collection, and Christina Kang-Yi assisted with data preparation and analysis. This project was supported by a grant from the William Penn Foundation and a fellowship to Dr. Rubin from the Stoneleigh Foundation.
2.
4.
6.
7.
8.
9. 10.
11.
12.
13.
14. 15.
16.
17.
18.
References 19. 1. U.S. Department of Health and Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau. The AFCARS report: preliminary estimates for FY 2011 as of July 2012. [Cited 2013 May
123
20.
23]. http://www.acf.hhs.gov/sites/default/files/cb/afcarsreport19. pdf. Annie E. Casey Foundation. (2012). Kids Count Data Book. [Cited 2013 Oct 24]. http://www.aecf.org/*/media/Pubs/Initia tives/KIDS%20COUNT/123/2012KIDSCOUNTDataBook/KIDS COUNT2012DataBookFullReport.pdf. Chipungu, S. S., Everett, J. E., Verduik, M. J., & Jones, H. (1998). Children placed in foster care with relatives: A Multi-site Study. Washington, DC: U.S. Department of Health and Human Services. Frank, D., Augustyn, M., Knight, W., Pell, T., & Zuckerman, B. (2001). Growth, development and behavior in early childhood following prenatal cocaine exposure: A systematic review. JAMA, 285(12), 1613–1625. Jones, J. K. (1998). Functioning and adjustment of children in kinship care versus nonrelative foster family care placements. Buffalo: University of New York at Buffalo. (dissertation). Rubin, D. M., Downes, K. J., O’Reilly, A. R., Mekonnen, R., Luan, X., & Localio, R. (2008). Impact of kinship care on behavioral well-being for children in out-of-home care. Archives of Pediatrics & Adolescent Medicine, 162(6), 550–556. Barth, R. P., Shenyang, G., & McCrae, J. S. (2007). Propensity score matching strategies for evaluating the success of child and family service programs. Research on Social Work Practice, 18(3), 212–222. Kemp, C. Children living with relatives struggle with physical, mental health issues. American Academy of Pediatrics News, 2010 May 2. [Cited 2013 March 21]. http://aapnews.aappublica tions.org/cgi/content/full/aapnews.20100502-3v1?rss=1. Berrick, J. D. (1998). When children cannot remain home: Foster family care and kinship care. Future of Children, 8(1), 72–87. Stahmer, A. C., Leslie, L. K., Hurlburt, M., et al. (2005). Developmental and behavioral needs and service use for young children in child welfare. Pediatrics, 116(4), 891–900. Horwitz, S. M., Hurlburt, M. S., Heneghan, A., et al. (2012). Mental health problems in young children investigated by U.S. child welfare agencies. Journal of the American Academy of Child & Adolescent Psychiatry, 51(6), 572–581. Ehrle, J., & Geen, R. (2002). Kin and non-kin foster care-findings from a national survey. Children and Youth Services Review, 24(1–2), 15–35. Berrick, J., Barth, R., & Needell, B. (1994). A comparison of kinship foster homes and foster family homes: Implications for kinship foster care as family preservation. Children and Youth Services Review, 16(1–2), 33–63. Barth, R., Courtney, M., Berrick, J., & Albert, V. (1994). From child abuse to permanency planning. New York, NY: Aldine de Gruyter. Harden, B. J., Clyman, R. B., Kriebel, D. K., & Lyons, M. E. (2004). Kith and kin care: Parental attitudes and resources of foster and relative caregivers. Children and Youth Services Review, 26(7), 657–671. Scannapieco, M., Hegar, R. L., & McAlpine, C. (1997). Kinship care and foster care: A comparison of characteristics and outcomes. Families in Society, 78(5), 480–488. Harris, M. S., & Skyles, A. (2008). Kinship care for African American children: Disproportionate and disadvantageous. Journal of Family Issues, 29(8), 1013–1030. Cheng, T. C., & Lo, C. C. (2011). A longitudinal analysis of some risk and protective factors in marijuana use by adolescents receiving child welfare services. Children and Youth Services Review, 33(9), 1667–1672. Grogan-Kaylor, A., Ruffolo, M. C., Ortega, R. M., & Clarke, J. (2008). Behaviors of youth involved in the child welfare system. Child Abuse & Neglect, 32(1), 35–49. Harden, B. J., & Whittaker, J. V. (2011). The early home environment and developmental outcomes for young children in the
Matern Child Health J
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
child welfare system. Children and Youth Services Review, 33(8), 1392–1403. Horwitz, S. M., Hurlburt, M. S., Cohen, S. D., Zhang, J., & Landsverk, J. (2011). Predictors of placement for children who initially remained in their homes after an investigation for abuse or neglect. Child Abuse & Neglect, 35(3), 188–198. Sprang, G., Clark, J. J., & Staton-Tindall, M. (2010). Caregiver substance use and trauma exposure in young children. Families in Society, 91(4), 401–407. Hines, A. M., Lemon, K., Wyatt, P., & Merdinger, J. (2004). Factors related to the disproportionate involvement of children of color in the child welfare system: A review and emerging themes. Children and Youth Services Review, 26(6), 507–527. Fuller-Thomson, E., & Minkler, M. (2000). African American grandparents raising grandchildren: A national profile of demographic and health characteristics. Health and Social Work, 25(2), 109–118. Fuller-Thomson, E., Minkler, M., & Driver, D. (1997). A profile of grandparents raising grandchildren in the United States. Gerontologist, 37(3), 406–411. Whitley, D. M., Kelley, S. J., & Sipe, T. A. (2001). Grandmothers raising grandchildren: Are they at increased risk for health problems? Health and Social Work, 26(2), 105–114. Achenbach, T. M. (1991). Manual for child behavior checklist 4-18. Burlington, VT: University of Vermont, Department of Psychiatry. Gandek, B., Ware, J. E., Aaronson, N. K., et al. (1998). Crossvalidation of item selection and scoring for the SF-12 health survey in nine countries: results from the IQOLA Project. Journal of Clinical Epidemiology, 51(11), 1171–1178. Ware, J. E, Jr, Kosinski, M., & Keller, S. D. (1996). A 12-Item short-form health survey: Construction of scales and preliminary tests of reliability and validity. Medical Care, 34(3), 220–233. Brazier, J. E., & Robert, J. (2004). The estimation of a preference-based measure of health from the SF-12. Medical Care, 42(9), 851–859. Jenkinson, C., Layte, R., Jenkinson, D., et al. (1997). A shorter form health survey: Can the SF-12 replicate results from the SF36 in longitudinal studies. Journal of Public Health, 19(2), 179–186. Vilagut, G., Forero, C. G., Pinto-Meza, A., et al. (2013). The mental component of the Short-Form 12 Health Survey (SF-12)
33. 34.
35.
36.
37.
38.
39. 40.
41.
42.
43.
as a measure of depressive disorders in the general population: Results with three alternative scoring methods. Value in Health, 16(4), 564–573. StataCorp. (2007). Stata Statistical Software: Release 10. College Station, TX: StataCorp LP. Heneghan, A., Stein, R. E. K., Hurlburt, M. S., et al. (2013). Mental health problems in teens investigated by U.S. child welfare agencies. Journal of Adolescent Health, 52(5), 634–640. Kohl, P. L., Jonson-Reid, M., & Drake, B. (2011). Maternal mental illness and the safety and stability of maltreated children. Child Abuse & Neglect, 35(5), 309–318. Fostering Connections to Success and Increasing Adoption Act of 2008 P.L. 110-351. Child Welfare Information Gateway Web site. [Published 2008; cited 2013 March 21]. http://www.child welfare.gov/systemwide/laws_policies/federal/index.cfm?event= federalLegislation.viewLegis&id=121. Hill, R. (2006). Synthesis of research on disproportionality in child welfare: An update. [Cited 2013 March 21]. http://www.cssp.org/ reform/child-welfare/other-resources/synthesis-of-research-on-dis proportionality-robert-hill.pdf. Knott, T., & Donovan, K. (2010). Disproportionate representation of African-American children in foster care: Secondary analysis of the National Child Abuse and Neglect Data System, 2005. Children and Youth Services Review, 32(5), 679–684. Roberts, D. (2002). Shattered bonds: The color of child welfare. New York, NY: Civitas Books. Scannapieco, M., & Jackson, S. (1996). Kinship care: The African American response to family preservation. Social Work, 41(2), 190–196. Boyle, M., & Pickles, A. (1997). Influence of maternal depressive symptoms on ratings of childhood behavior. Journal of Abnormal Child Psychology, 25, 399–412. Hops, H., Biglan, A., Sherman, L., Arthur, J., Friedman, L., & Osteen, V. (1987). Home observations of family interactions of depressed women. Journal of Consulting and Clinical Psychology, 55(3), 341–346. Webster-Stratton, C., & Hammond, M. (1988). Maternal depression and its relationship to life stress, perceptions of child-behavior problems, parenting behaviors, and child conduct problems. Journal of Abnormal Child Psychology, 16(3), 299–315.
123
