Kinship Care and Behavioral Outcomes for Children in the Child Welfare System

1. Kinship Care and Behavioral Outcomes for Children in the Child Welfare System David Rubin, MD MSCE Director or Research & Policy Safe Place: The Center for Child Protection & Health

2. In many states in the US, child welfare workers are mandated to prioritize the placement of the child with kin rather than with non-relatives. This occurs despite scant and conflicting evidence to show how placement of a child in KC, when compared with placement in GFC, influences the child’s well-being long-term.In many states in the US, child welfare workers are mandated to prioritize the placement of the child with kin rather than with non-relatives. This occurs despite scant and conflicting evidence to show how placement of a child in KC, when compared with placement in GFC, influences the child’s well-being long-term.

3. Background: Competing theories on kinship care: “Blood is thicker than water” Stronger attachment between caregiver and child Less disruptive “The apple doesn’t fall far from the tree” Kin may share the same risk factors as the birth parents Unrestricted and unregulated contact with birth parents Currently, there are two competing theories regarding the effects of kinship care. On the one hand, blood may be thicker than water: children in kinship care can form stronger attachments with a known relative. And, literature has shown that placement into KC is less disruptive for children and that children in kinship care change homes less frequently. However, others worry that the apple may not fall far from the tree – kin may share the same risk factors as the child’s birth parents and so kinship homes maybe less safe with frequent and often unsupervised access to birth parents.Currently, there are two competing theories regarding the effects of kinship care. On the one hand, blood may be thicker than water: children in kinship care can form stronger attachments with a known relative. And, literature has shown that placement into KC is less disruptive for children and that children in kinship care change homes less frequently. However, others worry that the apple may not fall far from the tree – kin may share the same risk factors as the child’s birth parents and so kinship homes maybe less safe with frequent and often unsupervised access to birth parents.

4. Methods Data Source: National Survey of Child & Adolescent Well-Being (NSCAW) Study Population: Children who entered kinship care or non-relative foster care following report To answer this question, we used data from the Nation Survey of Child & Adolescent Well-Being, a longitudinal study using a nationally representative sample of about 5500 children reported for abuse or maltreatment. Our study population included only those children initially placed into either kinship care or general foster care following a report.To answer this question, we used data from the Nation Survey of Child & Adolescent Well-Being, a longitudinal study using a nationally representative sample of about 5500 children reported for abuse or maltreatment. Our study population included only those children initially placed into either kinship care or general foster care following a report.

5. Gender: 54% female 46% male Age: 28% < 2 years 50% 2-10 years 22% > 10 years Race & Ethnicity: 50% White 38% African-American 13% Hispanic Abuse type: 57% neglect/abandonment 19% physical abuse 9% sexual abuse 1,309 children were included in our study, a slight majority of which were females. About 40% of children were reported to be African-American and 13% reported being Hispanic. Half of children were between 2 and 10 years of age, while a majority were reported due to neglect or abandonment. 1,309 children were included in our study, a slight majority of which were females. About 40% of children were reported to be African-American and 13% reported being Hispanic. Half of children were between 2 and 10 years of age, while a majority were reported due to neglect or abandonment.

6. Initial Placement Setting 50% kinship care 50% non-relative foster care 36-month CBCL Scores 38% had scores indicating behavioral problems Half of our population started in kinship care and half started in general foster care. Among all children, slightly more then two fifths were early stable, while just under 25% were unstable over the 36 months. At 36 months, almost 2 out of 5 children had a CBCL score in the borderline or clinical range. Half of our population started in kinship care and half started in general foster care. Among all children, slightly more then two fifths were early stable, while just under 25% were unstable over the 36 months. At 36 months, almost 2 out of 5 children had a CBCL score in the borderline or clinical range.

7. Were children who were placed into kinship care different from those who were placed into non-relative foster care? Having described our total population briefly, we first asked, “Were children who were placed into KC different from those who enter GFC?”Having described our total population briefly, we first asked, “Were children who were placed into KC different from those who enter GFC?”

8. Although the children initially placed in GFC were similar to children in KC for a majority of baseline variables, there were some important differences. On this graph, children in GFC are in blue and children initially placed in KC are in orange. What we can see is that children who entered kinship care were less likely to be using mental health services at baseline and less likely to have had a birth parent with a serious behavioral or mental health problem. As well, children who entered kinship care were less likely to have had an abnormal baseline behavioral well-being score.Although the children initially placed in GFC were similar to children in KC for a majority of baseline variables, there were some important differences. On this graph, children in GFC are in blue and children initially placed in KC are in orange. What we can see is that children who entered kinship care were less likely to be using mental health services at baseline and less likely to have had a birth parent with a serious behavioral or mental health problem. As well, children who entered kinship care were less likely to have had an abnormal baseline behavioral well-being score.

9. Now we’re looking at the baseline risk for placement instability, from our propensity scores, among our two groups. The red column represents the percentage of children at low risk for placement instability, the gray column represents those at medium risk, and the blue column is the high risk children. Among those children that started in GFC, we can see that there were about equal numbers at low, medium, and high risk for instability. However, among the children initially placed in kinship care, we can see that a majority were at low risk for placement instability at baseline.Now we’re looking at the baseline risk for placement instability, from our propensity scores, among our two groups. The red column represents the percentage of children at low risk for placement instability, the gray column represents those at medium risk, and the blue column is the high risk children. Among those children that started in GFC, we can see that there were about equal numbers at low, medium, and high risk for instability. However, among the children initially placed in kinship care, we can see that a majority were at low risk for placement instability at baseline.

10. Did children in kinship care have more stability than children in non-relative foster care? Next, we asked, “Were children in KC have more stability than children in foster care?”Next, we asked, “Were children in KC have more stability than children in foster care?”

11. This graph shows the children’s actual placement stability over the 36 months. The blue column represents those children who achieved early stability, the gray column represents those that achieved late stability, and the orange column represents those that were unstable. What we can see from the data is that the GFC population was generally evenly divided among the placement stability categories. However, the kinship care population was significantly more likely to achieve early stability, and much less likely to have been unstable over the 36 months.This graph shows the children’s actual placement stability over the 36 months. The blue column represents those children who achieved early stability, the gray column represents those that achieved late stability, and the orange column represents those that were unstable. What we can see from the data is that the GFC population was generally evenly divided among the placement stability categories. However, the kinship care population was significantly more likely to achieve early stability, and much less likely to have been unstable over the 36 months.

12. Controlling for the lower baseline risk and increased placement stability among children in kinship care, were behavioral outcomes different between children in kinship and non-relative foster care? So, taking into account children in KC’s lower baseline risk and increased likelihood for stability, did children in KC have different behavioral outcomes than children in GFC?So, taking into account children in KC’s lower baseline risk and increased likelihood for stability, did children in KC have different behavioral outcomes than children in GFC?

13. To look at these results another way, we used marginal standardization to derive predicted probabilities for behavior problems at 36 months, stratifying by placement setting and stability and adjusting for baseline risk and a child’s reunification status. On this graph, again, early stable is blue, late stable is gray, and unstable is orange. What we can see is that had all children been early stable and in GFC, the probability of behavior problems at 36 months would have been about 39%. If all children had been in KC and early stable the probability of behavior problems in this group would have been only 28%. This represents a 46% increase among early stable GFC children compared to those in KC who are early stable. We see similar trends for the other stability categories as well: Late stable children in GFC would be 43% more likely than children in KC to have behavior problems at 36 months. And, unstable children in GFC would be 36% more likely to have behavior problems.To look at these results another way, we used marginal standardization to derive predicted probabilities for behavior problems at 36 months, stratifying by placement setting and stability and adjusting for baseline risk and a child’s reunification status. On this graph, again, early stable is blue, late stable is gray, and unstable is orange. What we can see is that had all children been early stable and in GFC, the probability of behavior problems at 36 months would have been about 39%. If all children had been in KC and early stable the probability of behavior problems in this group would have been only 28%. This represents a 46% increase among early stable GFC children compared to those in KC who are early stable. We see similar trends for the other stability categories as well: Late stable children in GFC would be 43% more likely than children in KC to have behavior problems at 36 months. And, unstable children in GFC would be 36% more likely to have behavior problems.

14. Some words of caution Although children in kinship had fewer behavioral problems than children in non-relative foster care, their rates of behavioral problems were greater than other children in general. Kinship care is not a realistic option for all children who enter out-of-home care. Reporting bias among the kinship care population might have explained some of the difference. There were several limitations to our study. On the one hand, there was the potential for unobserved confounding which could have explained the differences we found between the KC and GFC groups. However, we controlled for a child’s baseline attributes to limit these effects. It is also possible that kinship caregivers may be less likely to report problems in the children they care for. However, since we controlled for a child’s baseline behavior score, we attempted to minimize the possible effects of this type of bias. There were several limitations to our study. On the one hand, there was the potential for unobserved confounding which could have explained the differences we found between the KC and GFC groups. However, we controlled for a child’s baseline attributes to limit these effects. It is also possible that kinship caregivers may be less likely to report problems in the children they care for. However, since we controlled for a child’s baseline behavior score, we attempted to minimize the possible effects of this type of bias.

15. When a child is reported for abuse or maltreatment, it is sometimes necessary to remove the child from their home in order to ensure their safety. When this occurs, the vast majority of children are placed into either general foster care, with non-relatives, or into kinship care. When a child is reported for abuse or maltreatment, it is sometimes necessary to remove the child from their home in order to ensure their safety. When this occurs, the vast majority of children are placed into either general foster care, with non-relatives, or into kinship care.

16. When a child is reported for abuse or maltreatment, it is sometimes necessary to remove the child from their home in order to ensure their safety. When this occurs, the vast majority of children are placed into either general foster care, with non-relatives, or into kinship care. When a child is reported for abuse or maltreatment, it is sometimes necessary to remove the child from their home in order to ensure their safety. When this occurs, the vast majority of children are placed into either general foster care, with non-relatives, or into kinship care.

17. When a child is reported for abuse or maltreatment, it is sometimes necessary to remove the child from their home in order to ensure their safety. When this occurs, the vast majority of children are placed into either general foster care, with non-relatives, or into kinship care. When a child is reported for abuse or maltreatment, it is sometimes necessary to remove the child from their home in order to ensure their safety. When this occurs, the vast majority of children are placed into either general foster care, with non-relatives, or into kinship care.

18. When a child is reported for abuse or maltreatment, it is sometimes necessary to remove the child from their home in order to ensure their safety. When this occurs, the vast majority of children are placed into either general foster care, with non-relatives, or into kinship care. When a child is reported for abuse or maltreatment, it is sometimes necessary to remove the child from their home in order to ensure their safety. When this occurs, the vast majority of children are placed into either general foster care, with non-relatives, or into kinship care.

19. When a child is reported for abuse or maltreatment, it is sometimes necessary to remove the child from their home in order to ensure their safety. When this occurs, the vast majority of children are placed into either general foster care, with non-relatives, or into kinship care. When a child is reported for abuse or maltreatment, it is sometimes necessary to remove the child from their home in order to ensure their safety. When this occurs, the vast majority of children are placed into either general foster care, with non-relatives, or into kinship care.

20. Conclusions Children raised in kinship settings had more stability than children in non-relative foster care Kinship care conferred benefits to children beyond the increased stability that was achieved Kinship care conferred benefits to children beyond simply the increased stability that was achieved Placement into kinship care was protective of the development of behavior problems, even if the moved occurred later From a system perspective, these findings provide some of the most compelling evidence to date to support efforts to place children with kin From a pediatric perspective, physicians might consider taking a more active role following maltreatment reports to help identify available and willing kin, and pay greater attention to the needs of children and families involved in kinship care Kinship care conferred benefits to children beyond simply the increased stability that was achieved Placement into kinship care was protective of the development of behavior problems, even if the moved occurred later From a system perspective, these findings provide some of the most compelling evidence to date to support efforts to place children with kin From a pediatric perspective, physicians might consider taking a more active role following maltreatment reports to help identify available and willing kin, and pay greater attention to the needs of children and families involved in kinship care

21. Implications: Provides empirical data to support efforts to improve the early placement of children with kin when appropriate Also reinforces the need to provide better services (via accessible navigator programs) among kinship families, which can only support efforts to achieve stability and maintain permanency Kinship care conferred benefits to children beyond simply the increased stability that was achieved Placement into kinship care was protective of the development of behavior problems, even if the moved occurred later From a system perspective, these findings provide some of the most compelling evidence to date to support efforts to place children with kin From a pediatric perspective, physicians might consider taking a more active role following maltreatment reports to help identify available and willing kin, and pay greater attention to the needs of children and families involved in kinship care Kinship care conferred benefits to children beyond simply the increased stability that was achieved Placement into kinship care was protective of the development of behavior problems, even if the moved occurred later From a system perspective, these findings provide some of the most compelling evidence to date to support efforts to place children with kin From a pediatric perspective, physicians might consider taking a more active role following maltreatment reports to help identify available and willing kin, and pay greater attention to the needs of children and families involved in kinship care

22. Acknowledgements: Funding: Dr. Rubin’s work was supported through a K23 mentored career development award from NICHD (1 K23 HD045748-01A1) and a supplemental grant from the Office of Research & Planning of the Adminstration of Children & Families Special thanks to: Kevin Downes Amanda O’Reilly MPH Xianqun Luan MS A. Russell Localio JD PhD Robin Mekonnen MSW I would like to acknowledge our funders: NICHD and the Doris Duke Charitable Foundation. And, I would to thank all those individuals that have also contributed to this project.I would like to acknowledge our funders: NICHD and the Doris Duke Charitable Foundation. And, I would to thank all those individuals that have also contributed to this project.

23. Thank you.