Lacement type and contextual factors on internalizing and externalizing behaviors after controlling for child demographics, type of abuse that led to placement out of the home, and a change in the child’s living situation between baseline and 18 month follow-up. Table 2 presented parameter estimates in the final model by outcome with model fit information. The model provided adequate fit to the data; the value of the chi-square test was not zero, and the RMSEA and SRMR fell below .05. The model explained half of the variance in both internalizing and externalizing symptoms at 18 months. A significant and moderate correlation existed between internalizing and externalizing behaviors at 18 months. Baseline externalizing problems significantly predicted both internalizing and externalizing at 18 months, whereas internalizing symptoms at baseline had a specific effect on later internalizing problems in this sample. Internalizing symptoms among African American foster youth were not predicted by placement type after controlling for other child and placement characteristics; youth placed in kincare had similar rates of caregiver-reported emotional problems as youth in other outof-home settings. Internalizing symptoms at 18 SC144 web months were predicted by baseline levels of both internalizing and externalizing problems, suggesting cross symptom domain influences on mental health. In addition, youth placed in more problematic neighborhoods exhibited significantly greater internalizing symptoms at 18 months, regardless of placement type. No interactions were significant after accounting for these main effects. The model also tested effects on caregiver-rated externalizing problems at 18 months. Symptom levels were predicted by youth baseline externalizing problems, suggestingJ Soc Serv Res. Author manuscript; available in PMC 2016 February 25.Rufa and FowlerPagestability in behavioral problems among youth. Unlike internalizing problems, no cross symptom domain influence occurred indicating specificity in behavior problems over time. Significant main effects existed for both a change of living situation in the 18 months between baseline and Metformin (hydrochloride) web follow-up interview and neighborhood problems (p < .05), such that youth who moved between interviews as well as those residing in worse neighborhoods exhibited higher externalizing symptoms. Additionally, a significant two-way and three-way interaction existed between placement characteristics predicting changes in externalizing problems at 18 months. Caregiver age and physical health interacted to predict behavior problems, and this effect occurred differently by out-of-home placement type. As displayed in Figure 2, youth placed in kinship care displayed a decrease in behavior problems when placed with older caregivers who experienced fewer health problems; the reverse was true for those in nonkinship placement settings who experienced decreased behavior problems when placed with older caregivers in poorer health.Author Manuscript Author Manuscript Author Manuscript Author ManuscriptDiscussionThis study tests the effects of kinship foster care on African American youth mental health over time in context of placement and community characteristics. Hypotheses predicted that youth placed in kinship foster care would show decreases in internalizing and externalizing problems over time, especially when placed in homes with more family and community supports. Findings demonstrate complex relationships involved in out-of-.Lacement type and contextual factors on internalizing and externalizing behaviors after controlling for child demographics, type of abuse that led to placement out of the home, and a change in the child's living situation between baseline and 18 month follow-up. Table 2 presented parameter estimates in the final model by outcome with model fit information. The model provided adequate fit to the data; the value of the chi-square test was not zero, and the RMSEA and SRMR fell below .05. The model explained half of the variance in both internalizing and externalizing symptoms at 18 months. A significant and moderate correlation existed between internalizing and externalizing behaviors at 18 months. Baseline externalizing problems significantly predicted both internalizing and externalizing at 18 months, whereas internalizing symptoms at baseline had a specific effect on later internalizing problems in this sample. Internalizing symptoms among African American foster youth were not predicted by placement type after controlling for other child and placement characteristics; youth placed in kincare had similar rates of caregiver-reported emotional problems as youth in other outof-home settings. Internalizing symptoms at 18 months were predicted by baseline levels of both internalizing and externalizing problems, suggesting cross symptom domain influences on mental health. In addition, youth placed in more problematic neighborhoods exhibited significantly greater internalizing symptoms at 18 months, regardless of placement type. No interactions were significant after accounting for these main effects. The model also tested effects on caregiver-rated externalizing problems at 18 months. Symptom levels were predicted by youth baseline externalizing problems, suggestingJ Soc Serv Res. Author manuscript; available in PMC 2016 February 25.Rufa and FowlerPagestability in behavioral problems among youth. Unlike internalizing problems, no cross symptom domain influence occurred indicating specificity in behavior problems over time. Significant main effects existed for both a change of living situation in the 18 months between baseline and follow-up interview and neighborhood problems (p < .05), such that youth who moved between interviews as well as those residing in worse neighborhoods exhibited higher externalizing symptoms. Additionally, a significant two-way and three-way interaction existed between placement characteristics predicting changes in externalizing problems at 18 months. Caregiver age and physical health interacted to predict behavior problems, and this effect occurred differently by out-of-home placement type. As displayed in Figure 2, youth placed in kinship care displayed a decrease in behavior problems when placed with older caregivers who experienced fewer health problems; the reverse was true for those in nonkinship placement settings who experienced decreased behavior problems when placed with older caregivers in poorer health.Author Manuscript Author Manuscript Author Manuscript Author ManuscriptDiscussionThis study tests the effects of kinship foster care on African American youth mental health over time in context of placement and community characteristics. Hypotheses predicted that youth placed in kinship foster care would show decreases in internalizing and externalizing problems over time, especially when placed in homes with more family and community supports. Findings demonstrate complex relationships involved in out-of-.