|Home | About | Journals | Submit | Contact Us | Français|
Increasingly, children in need of out-of-home care are being placed in kinship care, yet few studies have followed their placement histories longitudinally to determine if these children constitute a homogeneous group or heterogeneous subgroups. This study of 484 children in kinship care in San Diego County, California, indicates that children in kinship care have markedly different sociodemographic and maltreatment histories, as well as heterogeneous placement experiences.
Over the last decade, the number of children in out-of-home care in the United States has increased substantially. In 1995, over 482,000 children and youths were in the out-of-home care system [Barbell 1997]. These children are usually placed in one of three types of settings: (1) with a nonkinship foster family (family foster care); (2) in group care; and (3) with extended family members (kinship care).
Children in need of out-of-home care have increasingly been placed in kinship care in the last several years, and it is now recognized as the fastest growing sector of out-of-home placement by child welfare agencies themselves [Gleeson & Craig 1994]. Although national statistics regarding the number of children in kinship care are not available, estimates are that over one-fourth of the approximately 500,000 American children in care, many of them young children, are living with kin [Barbell 1997; Klee et al. 1997]. Data from several states document this increasing use of kinship care; over 45% of foster children in both Illinois and California in the early 1990s were placed with kin [Barth et al. 1994; Wulczyn & Goerge 1992; Dubowitz et al. 1993].
This burgeoning interest in kinship placements by the child welfare system results from two major trends in the delivery of services to children in need of care. First, the use of kinship care reflects the growing crisis between supply and demand in child welfare; the number of children entering out-of-home care is increasing concurrent with a decrease in the number of nonkinship foster parents accepting children [Child Welfare League of America 1997; Berrick & Barth 1994]. Second, the increased use of kinship placement reflects legislative and policy initiatives over the last two decades that have focused on family preservation and family ties. This emphasis was reiterated in the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PL. 104-193), which requires states to consider giving preference to an adult relative over a nonrelated caregiver when determining placement [Children’s Defense Fund 1996].
Research suggests that kinship care placement may provide potential advantages for children in need of care. Children placed with kin are more often placed with their siblings [Duerr-Berrick et al. 1994] and have a closer connection with members of their immediate family [Duerr-Berrick et al. 1994; Le Prohn 1994]. Kinship caregivers appear to be more willing to focus on the separation and loss the child experiences when removed from his or her home than are nonrelative caregivers [Le Prohn 1994]. Children in kinship care also move less frequently and enjoy more stability in their out-of-home placement [Duerr-Berrick et al. 1994; Iglehart 1994]. They also may experience less maltreatment while in out-of-home care than do children in nonrelative family foster care [Zuravin et al. 1997].
A less optimistic view of kinship care also exists. Children in kinship care typically are not reunified with their families as often or as early as children in nonrelative family foster care [Duerr-Berrick et al. 1994; Wulczyn & Goerge 1992]. Delay in reunification for children placed in kinship care may be due to multiple factors, including bureaucratic child welfare systems, financial obstacles, and the lack of significant involvement of kinship caregivers in service planning (Barth et al. 1994; Gleeson et al. 1997]. In addition, cohort studies indicate that many children in kinship care do not obtain needed medical and mental health services [Dubowitz et al. 1993; Dubowitz et al. 1994].
Despite the widespread use of kinship care, research on kinship care overall has not kept pace with its use as a placement option by the child welfare system [Duerr-Berrick & Barth 1994; Dubowitz 1994]. Little information has been published on how children fare in kinship care. Studies in the literature are few and have often lacked representative samples and adequate comparison groups [Duerr-Berrick & Barth 1994]. In addition, most of the published studies in the literature on kinship care have tended to be cross-sectional in design, identifying children in kinship care based on their placement status at a single point in time [Dubowitz et al. 1994; Iglehart 1994; Benedict et al. 1996; Feigelman et al. 1995; Dubowitz et al. 1993; Link 1996]. Yet, the experience of many children while in out-of-home care often involves more than one placement setting during their entire length of stay in out-of-home placement. Reliance on point in time, cross-sectional data limits our ability to understand and track the dynamic nature of out-of-home care. Several researchers have stressed the importance of longitudinal studies that identify patterns of placement experiences for children in care and have developed databases and methodologies for investigating children’s placement pathways through the child welfare system. [Barth et al. 1994; Goerge et al. 1994; Pavkov et al. 1995; Courtney & Needell 1997]. The present study extends the work of these researchers and investigates longitudinal out-of-home care placement histories of children to determine if children in kinship care constitute a homogeneous group or several heterogeneous subgroups of children with distinct sociodemographic and maltreatment characteristics and different placement experiences. A clear understanding of what constitutes kinship care is a necessary prerequisite for addressing pressing policy questions regarding kinship care and its impact on child well-being.
Child welfare researchers have stressed that there are two types of kinship care: formal kinship care, where custody for a child has been legally transferred by the court to the caregiver or the state child welfare agency, and informal kinship care, in which an informal arrangement without legal involvement has been facilitated by the child welfare system or family members themselves [Dubowitz 1994; Harden et al. 1997]. This study focuses solely on children in formal kinship care who were placed in out-of-home care by court order. It describes the longitudinal placement experiences of a cohort of children in San Diego County and propose a conceptual framework for identifying subgroups of children in kinship care placements.
The sample of children for this study was taken from a larger cohort of 5,331 children ages birth to 16 years of age, residing in San Diego County, who were removed from their homes between May 1990 and October 1991. Of these, 1,078 children (1) were made dependents of the court and dispositioned to out-of home care, (2) were placed in San Diego County, (3) remained in out-of-home care five months after entry into the out-of-home care system, (4) had Department of Social Services (DSS) data available, and (5) were entered into a longitudinal study of children in out-of-home care funded by the National Institute of Mental Health and the National Center on Child Abuse and Neglect. Each child was followed for approximately 18 months from his or her entry into out-of-home care, regardless of whether reunification with the biological parents occurred during the 18-month study period.
Of the 1,078 children with placement data, 484 (45%) had spent some portion of the 18-month study period in kinship care and were included in the current analysis of kinship care. The mean age of the sample of children with kinship placements was 4.8 years of age (SD = 4.3, median = 4.0); 41% of the sample were two years old or younger. Forty-nine percent of the sample were female. The ethnic distribution was 33% Caucasian, 41% African American, 22% Latino, and 4% Asian and other ethnic groups. Fifteen percent had been placed by the court in out-of-home care because of physical abuse, 7% because of sexual abuse, 47% because of caregiver absence or neglect, 8% because of emotional abuse, 7% because of multiple types of abuse, and 16% for other reasons.
Data for this study were obtained from the Social Services Reporting System (SSRS) database, made available to the study by the Children’s Services Bureau, Department of Social Services, San Diego County. The data were coded by trained research staff and entered into a project database. Data collected included sociodemographic characteristics, type of maltreatment, and placement information.
Sociodemographic data collected included information on each child’s age, gender, and ethnicity. Type of maltreatment information was abstracted from the DSS case files. All types of maltreatment substantiated and discussed in the case files were coded by trained research assistants. Thus, participants could have received codes for more than one type of maltreatment. The data used in the present study, however, referred only to the initiating episode leading to out-of-home placement and did not necessarily reflect the child’s lifetime maltreatment history. Seven maltreatment categories were created from these data (see table 1). Children who had experienced either sexual, physical, or emotional abuse with or without caregiver absence and/or neglect were distinguished from children who had experienced multiple types of abuse (sexual, physical, and emotional) or children who were removed from the home solely for caregiver absence and/or neglect. Data on children removed for positive toxicology screens at delivery or for protective custody issues were also kept distinct.
With respect to placement information, each placement setting a child experienced was recorded as either (1) placement with a relative, (2) placement with a foster parent, (3) placement in a group home, (4) placement in a medical facility, (5) placement in a psychiatric facility, or (6) other (including placement in the San Diego receiving facility, adoption, reunification, juvenile hall, jail, AWOL, or death). Two unique codes were then created for each child to capture his/her placement experiences. First, a six-digit placement combination code was created for each child that assigned a dichotomous variable indicating whether or not a child had ever been placed in each of the six possible placement settings listed above during the study follow-up period. For example, a child who had only experienced kinship care during the study follow-up period would receive a code of 1-0-0-0-0-0. Placement codes were then clustered to determine a typology of placement combinations. Second, these data were also used to create a longitudinal history of each child’s placement experiences for the 18-month period following the date of out-of-home placement that captured both the longitudinal sequencing of each child’s placement history as well as the duration of time spent in each placement setting. Both the placement code combinations and the longitudinal history of each child’s placement experiences for the 18-month follow-up period were used to (1) determine subgroups of kinship care based on placement histories, (2) characterize each subgroup’s unique sociodemographic and maltreatment characteristics, and (3) describe each subgroup’s placement experiences.
The placement code combinations generated indicated that the critical placement experiences for determining subgroups of children in kinship care were placement with kin, placement in family foster care, and placement in restrictive settings. Based on these placement code combinations, three distinct subgroups of children in kinship care were identified:
Of the 484 children placed in kinship care during the 18-month study follow-up period, 92 children (19%) were placed only with kin. Three hundred forty-eight children (72%) spent time both in kinship and in family foster care. The remaining 44 children (9%) spent some of the 18-month study period in a more restrictive setting, in either a group home or psychiatric facility or in both. Of these 44 children with placement experiences in a more restrictive setting, nine children (20%) spent the remaining time in kinship care; the other 35 children (80%) were placed both in kinship and in nonrelative family foster care.
As can be seen in table 1, the three subgroups of children differ substantially from each other in terms of sociodemographic and maltreatment characteristics. Statistically significant differences were found between groups for age (p < .001), ethnicity (p = 0.003), and type of maltreatment (p = 0.009) but not for gender. Group 1 children, whose placement experiences were solely in kinship care, were on average 5.5 years of age (SD = 4.54, median = 5.0 years) and more commonly female. Children in Group 2, with placement experiences in both kinship and family foster care, averaged a slightly younger age (4.1 years, SD = 3.93, median = 3.0 years) than children in Group 1. In comparison. Group 3 children, whose placement histories included more restrictive settings, tended to be older than children in Groups 1 and 2 (average age = 9.1, SD = 4.27, median = 10.0 years) and were more commonly male. Ethnicity differed substantially between the three groups. Almost 50% of the children in kinship care were African American, while 50% of the children placed in restrictive settings were Caucasian. In a separate analysis from that shown in table 1, the proportion of children in each group from different racial/ethnic backgrounds was examined. Of note, over 80% of the children of Latino origin in the study sample were Group 2 children, with both kin and family foster care placement experiences. The three groups also differed in terms of maltreatment histories. Over 50% of Group 1 children had experienced caregiver neglect or absence, while over 50% of the Group 3 children had histories of sexual, physical, emotional or multiple types of abuse.
Table 2 presents information related to the placement histories of children in each of the subgroups, addressing which of the six types of placements the children experienced, the children’s length of stay in different placement settings, and the children’s placement stability. The first variable addressed in table 2 is the first placement type children experienced in out-of-home care. Since children in the out-of-home care system in San Diego enter the system through a single receiving facility, placement in the receiving facility was not included in the evaluation. Likewise, placement in a medical facility was not included because almost all of the hospitalizations were following birth and were not true dispositional “placements” in the out-of-home care system. In addition, because Group 1 children were only placed in kinship care settings, first placement type was analyzed only for those children in Groups 2 and 3. Interestingly, over 92% of the Group 2 children (those with both kin and family foster care placements), were initially placed in family foster care settings. Similarly, only a small proportion of the Group 3 children began in kinship care; nearly 39% were initially dispositioned to family foster care and over 43% were initially placed in group homes or psychiatric in-patient facilities. These differences in first placement type between Groups 2 and 3 were statistically significant (p < 0.001).
Length of stay in first placement also differed significantly for the three groups (p < 0.001) (see table 2). Children in Group 1 experienced much more stability than those in Groups 2 or 3, spending an average of 375.5 days (SD = 196.1, median = 475 days) in their first placement setting over the 18-month (550 days) follow-up period of the study. Children in Group 2 had a much shorter period of stay in their initial placement setting, predominantly family foster care, with an average length of stay of 39.6 days (SD = 70.2, median = 20 days). For the Group 3 children, length of stay in first placement setting averaged 53.4 days (SD = 82.3, median = 22.5 days) but was not found to be statistically different from the length of stay for Group 2 children in post-hoc comparisons.
Because the shortened first placement stay for Groups 2 and 3 may reflect use of emergency shelter care following discharge from the county receiving facility, second placement experiences were also examined. Second placement experiences for children in each of the three groups were different (p < 0.001). Group 1 was the only group where there were children who did not have a second placement (n = 68). Of the other 24 children in Group 1, 22 (92%) had a second placement in kinship care, one was placed with a legal guardian, and the other was returned to the county receiving facility. Although 117 (34%) of Group 2 children had a second placement in family foster care, 218 (63%) moved into kinship care. Twelve children in Group 2 (3%) returned to the county receiving facility following their first out-of-home placement and one child entered a medical facility. Children in Group 3 entered a variety of placement settings, including family foster care (39%), kinship care (34%), group home (11%), psychiatric hospital (5%), and the county receiving facility (9%). One child in Group 3 was a runaway. Length of stay in second placement setting also was statistically significant, with children in Group 3 averaging fewer days than children in Group 2 (p < 0.001) despite truncation of length of stay at 18 months (550 days) because of the study timeline.
In addition, length of stay in first kinship placement differed for the three groups. Children in Group 1 stayed an average of 375.5 days (SD = 196.1, median = 475 days) in their first kinship placement in comparison to 295.5 days (SD = 183.3, median = 286 days) for children in Group 2 and 147.4 days (SD = 141, median = 99.5 days) for children in Group 3 (p < 0.001). While this partially reflects the fact that measures of length of stay were truncated at 18 months (550 days) given the length of study follow-up, the differences found between Groups 2 and 3 suggest other factors influence length of stay in first kinship placement as well.
Placement stability was addressed by determining the total number of kinship placements and total number of overall placements each child experienced during his or her length of stay in out-of-home care during the 18-month study follow-up period. Children in Group 3 had more kinship placements than children in Groups 1 and 2, with over 35% of children having two or more kinship placements (p = 0.001). Interestingly, children in Group 2 showed the most stability in kinship placements once a kinship placement was determined. In terms of total number of placement locations children experienced (including placements in the receiving facility and medical facilities), the sample of 484 children averaged 4.2 total placement changes (median = 4.0) during the study period. The three groups differed significantly regarding total number of placements, with children in kinship care only averaging 2.4 placements as compared to 4.3 placements in Group 2 and 6.8 placements in Group 3 (p < 0.001).
In addition, percent of time in kinship care placement for each child was determined to be statistically different among the three groups (p < 0.001), as graphically presented in figure 1. While one-quarter of the children in Group 2 spent less than 50% of their time in kinship care, over half the group spent 75% or more time in kinship care. Percent of time in kinship care placement was much more variable for Group 3 children; nearly 60% spent less than 50% of time in kinship care over the 18-month study follow-up period. Almost 75% of the Group 3 children spent less than 30% of the study period in restrictive settings indicating that many of Group 3 children move often between kinship, family foster care, and restrictive settings.
Placement outcomes for this study included whether or not children were reunified with their biological parents, and the total length of stay in out-of-home placement for those children who were reunified. Of note, no children in the sample were adopted during the study period. Of the 484 children with some kinship care experience, 165 (34%) were reunified during the study period. A total of 137 children (28%) were reunified from a relative placement setting. Interestingly, eight of the reunified children reentered foster care during the study period. Although not statistically significant, seven of these children were from Group 2.
Table 3 shows relevant variables regarding placement for the three subgroups of children in kinship care. While 39.1% of children in Group 1 and 35.9% of children in Group 2 were reunified with their parents within the 18-month study period, the percent of children was much lower (9.1%) in Group 3 (p = 0.001). For all three groups, length of stay in out-of-home care placement for those children who were reunified averaged 10 months or greater, although the range was much broader for children in Group 1.
In separate multivariate analyses not shown in the tables, the relationship between group membership and the outcome variables remained significant when controlling for age, ethnicity, and of maltreatment.
This study identified several distinct groupings of children in kinship care and suggested a methodology for examination of placement experiences. From the data described above, it is clear that children in San Diego County placed in kinship care have remarkably different sociodemographic and maltreatment histories as well as heterogeneous placement experiences. Children who were placed only with kin were predominantly young and had been placed in care due to caregiver neglect and/or absence. A second group of children entered kinship care or family foster care following placement in family foster care with a professional foster parent. These children tended to be younger than children placed solely with kin and to have experienced a wide range of maltreatment. Children who had placement histories including restrictive settings were older and had experienced sexual, physical, emotional, and multiple types of abuse more often than other children in care.
The concept of “kinship care” is complex and not well-represented by a dichotomous variable. The complexity of placement experiences for children in kinship care is more accurately characterized by a longitudinal perspective that relates types of placements with placement pathways. Children in kinship care, when examined longitudinally, thus naturally cluster into three subgroups based on the most restrictive placement they experience. These three subgroups include (1) children solely in kinship care without experience in family foster care or restrictive settings, (2) children in kinship and family foster care, and (3) children in kinship care with histories of placements in residential centers or psychiatric hospitals.
Of particular interest is the role that ethnicity appears to play in determining different placement pathway patterns. Over 80% of children of Latino origin have both kinship and nonrelative family foster care placement experiences. These children are minimally represented in subgroups of children in kinship care containing both children placed solely with kin or children with restrictive placement experiences. This may reflect difficulties identifying kin residing in the area or, given San Diego County’s proximity to the border, difficulties with the legal status of kin of Latino children. The heavy use of kin in African American families may result from cultural patterns of extended family involvement with children prior to entering out-of-home care and possible prior use of informal private arrangements with kin prior to out-of-home placement. Parenting by kin has historically been a survival strategy used by many African American families [Stark 1974]. In comparison, both Asian and Caucasian children are over-represented among the children with restrictive placement experiences, compared with African American or Latino children. Ethnicity also appears to play a role in length of stay in different placement settings. Clearly, the relationship between ethnicity and out-of-home placement setting deserves further investigation.
The policy implications of acknowledging the complexity and heterogeneity of kinship care experiences are multiple. Conclusions drawn about children in kinship care with respect to involvement of the child welfare system with families, permanency planning, and placement outcomes need to control for the variable kinship care experiences of children in foster care. How placement patterns interact with physical, emotional, and developmental problems in determining service use by children in care also needs to be further investigated. Lastly, the critical impact of the total number of placement changes these children are experiencing over time, a perspective often lost in cross-sectional studies, should not be neglected.
This article presents basic descriptive work regarding kinship care experiences for children in San Diego County and thus may be limited in its generalizability to other areas of the country. California, however, does have the largest population in out-of-home care of any state, coordinating out-of-home care for approximately 18% of the nation’s children in care [Courtney & Needell 1997]. Further work needs to be done to determine if similar patterns of heterogeneity are found in other areas of the country.