- Email: [email protected]
Crisis nursery services and foster care prevention: An exploratory study
Children and Youth Services Review 61 (2016) 311–316
Contents lists available at ScienceDirect
Children and Youth Services Review journal homepage: www.elsevier.com/locate/childyouth
Crisis nursery services and foster care prevention: An exploratory study David Crampton a,⁎, Susan Yoon b a b
Jack, Joseph and Morton Mandel School of Applied Social Sciences, Case Western Reserve University, USA The College of Social Work, The Ohio State University, USA
a r t i c l e
i n f o
Article history: Received 25 March 2015 Received in revised form 1 January 2016 Accepted 1 January 2016 Available online 6 January 2016 Keywords: Crisis nursery services Case management Parenting education Foster care Race/ethnicity
a b s t r a c t Objective: To examine whether receiving crisis nursery services reduces the chances of subsequent placement of children into foster care. Method: The study includes a sample of 322 children from the administrative data from a crisis nursery in Ohio. Generalized estimating equations analysis was conducted to assess the relationship between crisis nursery services (case management and parenting education) and subsequent foster care placement. Results: The generalized estimating equations analyses indicated that case management and parenting education provided from the crisis nursery were associated with reduced likelihood of subsequent foster care for children who were placed in the crisis nursery. Foster care placement during the previous 12 months and Caucasian race were also associated with higher odds of subsequent foster care placement. Conclusions: Crisis nursery services delivered with case management and parenting education may be an effecive intervention to reduce children's subsequent foster care placement. © 2016 Elsevier Ltd. All rights reserved.
1. Introduction Abuse and neglect in early childhood pose an immediate threat to children's health and safety as well as to their long-term development. Early childhood maltreatment prevention programs and interventions provide an efficacious opportunity to intervene while young children's brains are most responsive to improved surroundings (Klein & Jones Harden, 2011). Although there is a growing research base for understanding the ecological experiences of young children in the child welfare system, we know less about how to intervene effectively in these cases. Crisis nurseries, which provide temporary child care and various support services to families of children at risk of abuse and neglect, may play a critical role in improving child welfare service outcomes, including number of out-of-home placements. However, only few studies have been conducted to investigate how and to what extent crisis nursery services may be associated with these improved outcomes. Therefore, this exploratory study addresses the need for additional research by examining how crisis nursery services, parenting education and case management in particular, may reduce the risk of out-of-home placements. Crisis intervention theory offers a useful framework for understanding how instant and targeted services may assist families at risk and prevent further negative consequences. Crisis intervention theory posits that families are most amenable to change when faced with an ⁎ Corresponding author at: Jack, Joseph and Morton Mandel School of Applied Social Sciences, Case Western Reserve University, 11235 Bellflower Road, Cleveland, OH 44106, USA. E-mail address: [email protected] (D. Crampton).
http://dx.doi.org/10.1016/j.childyouth.2016.01.001 0190-7409/© 2016 Elsevier Ltd. All rights reserved.
overwhelming life crisis and emphasizes the importance of providing immediate and focused services over a short period (Parad & Caplan, 1960). In line with crisis intervention theory, the crisis nursery model assumes that some families experience a short-term crisis that puts their children at risk of child abuse and neglect, and by receiving respite, parents can focus on their crisis and reduce the risk (ARCH, 1994; Cole, Wehrmann, Dewar, & Swinford, 2005). “Crisis nursery” refers to a type of respite care that is designed to provide temporary child care for young children at risk of abuse and neglect and to offer an array of support services to the families and caregivers of these children (ARCH, 1994). Previous studies that evaluated the effects of crisis nursery services and respite care on child welfare service outcomes have yielded mixed findings. Some studies have suggested that crisis nursery services may be associated with improved child welfare service outcomes, including fewer out-of-home placements (e.g. Bruns & Bujrchard, 2000; Cole & Hernandez, 2011; Cowen, 1998). For example, in a study of 240 families who received crisis child care in rural Iowa counties, there was a significant decrease in the reported incidence of child maltreatment in the rural counties with crisis child care programs compared with counties that did not offer crisis care programs (Cowen, 1998). In a controlled, longitudinal study of the effectiveness of respite care services, families who received respite care had significantly fewer incidents of out-of-home placement than the control group (Bruns & Bujrchard, 2000). Similarly, Home and Darveau-Fournier (1995) found that respite care (planned and crisis) for families in a high-risk neighborhood resulted in reduced incidents of foster care placement, keeping the children in the local community. In an evaluation study of five crisis nurseries in Illinois, children whose families received crisis nursery
312
D. Crampton, S. Yoon / Children and Youth Services Review 61 (2016) 311–316
services prior to foster care placement had significantly higher odds of being reunited with their biological families compared with children whose families received only foster care services (Cole & Hernandez, 2011). On the contrary, there were some studies that found no significant link between crisis nursery services/respite care and better child welfare service outcomes. For example, Aniol, Mullins, Page, Boyd, and Chaney (2004) found no significant relationship between respite care and child abuse potential in parents of children with developmental disabilities. Similar to this study, Cole and Hernandez (2008) also reported that crisis nursery services did not have a significant effect on potential for abuse and neglect. These inconsistent findings warrant further investigation as to whether crisis nursery services are associated with improved child welfare service outcomes, including reduced foster care placement. According to the ARCH National Respite Network and Resource Center (ARCH, 2014), previous research on respite services is limited in terms of its scope, depth, and methodological rigor. Cole and Hernandez (2011) have identified several challenges to evaluating whether crisis nurseries can prevent and reduce child maltreatment and foster care. For example, the respite services are often extremely short in duration (in some cases, just a few hours) and therefore are unlikely to have sustained benefits. In addition, the crisis nature of the need for services makes random selection of families into respite unethical. Even many quasiexperimental designs (e.g., waiting list designs) are not feasible in a crisis situation. Cole and colleagues have acknowledged the need for more objective measures of improvement (Cole & Hernandez, 2008, 2011) and responded to this need by conducting research that links crisis nursery services to public child welfare services. For example, they have used this linked data method to show how the receipt of crisis nursery services may have positive effects on the children's ultimate placement outcome after foster care placement (Cole & Hernandez, 2011). The present study takes this approach of linking the administrative data obtained from a crisis nursery to public child welfare records. Although crisis nursery services, including parenting education and case management, may play a pivotal role in preventing and reducing child abuse and neglect, more research is needed concerning how the services help children at risk. This investigation builds on previous crisis nursery research to explore how crisis nursery services delivered in a different state may relate to fewer foster care placements. More specifically, through a study of a crisis nursery, we explore an engagement model suggested by the social workers in the crisis nursery that a short-term crisis presents an opportunity for social workers to engage with caregivers and to use that engagement to encourage the caregivers to participate in parent education and case management services that may have long-term benefits that extend beyond the resolution of the crisis that lead the caregivers to seek emergency help. Although services offered by crisis nurseries may vary, ranging from parent education to mentoring to case management and aftercare, this study focused on two core services provided by the crisis nursery: parenting education and case management. To our knowledge, no studies to date have examined how other supportive services beyond temporary respite child care, such as parenting education and case management, may be associated with subsequent foster care placement. We hypothesized that families who are more engaged with the recommended parenting education and case management services would have a lower risk of their children entering foster care. 2. Methods 2.1. Subjects A secondary data analysis was conducted using administrative data provided by Providence House, a crisis nursery in Cleveland, Ohio. The study sample included all families (186 families and 322 children)
who received respite services from the crisis nursery between 2006 and 2009. 2.2. Procedures An administrative data file, which includes demographics, referral sources, reason for referral, type of crisis nursery services received, and the dates of inquiry, placement and discharge, was obtained from the crisis nursery. It was found that many families had numerous contacts with the agency and many children were placed in the agency multiple times. Thus, by matching the mother's first and last name, the mother's date of birth, the child's first and last name, and the child's date of birth, we removed duplicates and built a new file that had a single record per child. To assess the relationship between families receiving crisis nursery services and their involvement with the local public child welfare agency (DCFS, the Department of Children and Family Services), including foster care placement, the crisis nursery data on children and their mothers were matched to the Childhood Integrated Longitudinal Data (CHILD) System. The CHILD is maintained by the Center on Urban Poverty and Community Development at Case Western Reserve University as an evaluation tool for Cuyahoga County's early childhood initiative, which is called the Invest in Children (IIC) initiative. The CHILD system integrates birth records with records from many health and human services agencies to create a longitudinal record for all children who reside in Cuyahoga County from birth through age 6 and who appear in one or more of the agencies' data systems. At the time of this study, the CHILD system included all children with any DCFS involvement from July 1999 through 2010. We used probabilistic matching to match the crisis nursery children to the CHILD. The variables used for matching included child's date of birth, child's first name, child's last name, mother's date of birth, mother's first name, mother's last name, street name, street number, city, zip code, sex, race, and social security number. Once the matching was completed, we used the CHILD system to identify which children in the crisis nursery records had subsequent foster care placement. 2.3. Measures 2.3.1. Dependent variable The dependent variable of interest was children's foster care placement at any time after their first contact with the crisis nursery. It was assessed using the public child welfare records in the CHILD system and coded 0 = no and 1 = yes. 2.3.2. Independent variables Focal independent variables in the present study included two types of services (case management, parenting education) provided by the crisis nursery. For case management and parenting education variables, the following categories were used to code the receipt of recommended services: 1 = participation, 2 = nonparticipation, and 3 = exception/ not recommended. The participation group indicates families who received the recommended crisis nursery services (i.e., case management, parenting education), whereas the nonparticipation group indicates families who refused the recommended crisis nursery services. The exception/not recommended group indicates families who were not recommended for case management and/or parenting education. The nonparticipation group in each variable was used as the reference group for the generalized estimating equations (GEE) analyses in this study. 2.3.3. Control variables Based on theoretical relations and previous research, several covariates were identified and tested. They included child gender, child race, child age, number of children in the household, mother's age, length of placement in the crisis nursery, reasons for referral to the crisis
D. Crampton, S. Yoon / Children and Youth Services Review 61 (2016) 311–316
nursery (i.e., primary crisis stated by family), referral source, level of DCFS involvement (active or limited open case, custody, no involvement), substantiated maltreatment case, and ongoing maltreatment case. Three covariates that were correlated with outcomes at p ≤ .20 were included in the analysis: foster care placement during the previous 12 months, child race/ethnicity, and referral source. Foster care placement during the previous 12 months refers to a child having been in foster care some time during the 12 months before the first inquiry to the crisis nursery and was assessed using the public child welfare records in the CHILD system. Child race/ethnicity was self-identified by the caregivers using the following response categories: African American, Caucasian, Hispanic, biracial, Asian, and Middle Eastern. Because of the small sample size (consisting of less than 3% of the whole sample), Hispanic, Asian, and Middle Eastern children were excluded from the analyses. Three categories (African American, Caucasian, biracial) were included in the analyses and were dummy coded as African American and Caucasian, using biracial as the reference group. The referral source variable originally consisted of 11 categories, but was aggregated into two categories (official and unofficial). An official referral source included DCFS, social services centers, medical and mental health centers, and hospitals; an unofficial referral source included self-referral and others.
2.4. Statistical analyses Univariate frequencies and descriptive statistics including mean, standard deviation, minimum, maximum, skewness, and kurtosis were computed for predictors and outcome to understand the descriptive characteristics of the sample. Bivariate correlations between each predictor and the outcome were calculated to get a sense of the nature and strength of the bivariate relationships. Bivariate correlations between predictors were also inspected to check multicollinearity. Correlation above .70 was seen as a warning sign for multicollinearity (Tabachnick & Fidell, 2013). Chi-square analyses were conducted in order to assess bivariate relationships between two crisis nursery services and subsequent foster care. The generalized estimating equations (GEE) method as implemented in SAS Proc GENMOD (SAS, v. 9.2; SAS, 2009) was utilized to assess whether crisis nursery services are significant predictors of the subsequent placement of children into foster care. The GEE method is appropriate for the study because the outcome of interest is binary: whether a child had subsequent foster care (0 = no and 1 = yes). Furthermore, GEE estimators are appropriate for the study data as the data included sibling samples which may consist of nonindependent observations. The GEE method effectively addresses the issue of nonindependent and cluster-correlated information by accounting for correlation within groups (Hardin & Hilbe, 2012).
Table 1 Child characteristics (N = 322). n (%) Child gender, male Child race African American Caucasian Biracial Hispanic Middle Eastern Substantiated incident during the past 12 months, yes Ongoing DCFS case during the past 12 months, yes Foster care placement during the past 12 months, yes Child's age at admission (months) Length of placement in the crisis nursery (days)
M
SD
172 (53.4) 219 (68.2) 35 (10.9) 56 (17.4) 11 (3.4) 1 (0.3) 36 (11.2) 162 (50.3) 29 (9.0) 31.05 25.27
21.08 20.72
313
3. Results 3.1. Sample characteristics Characteristics of the children involved in this study are presented in Table 1. A total of 322 children were placed for crisis nursery services between 2006 and 2009. Slightly more than half of the children were male. The children were primarily African American, and the remainders were Caucasian and biracial. Children's age ranged from 0 to 7 years, with a mean age of 2. Approximately 9% of them had been in foster care during the previous 12 months before their family first contacted the crisis nursery. The length of placement in the crisis nursery ranged from a minimum of 0 to a maximum of 64 days, with a mean of 25 days. Table 2 shows the characteristics of the families of the children who were placed in the crisis nursery. A total of 186 families had at least one child placed at least once for crisis nursery services. Among these 186 families, slightly more than half received case management from the crisis nursery, and almost one third participated in parenting education recommended by the nursery staff. Primary referral sources included self-referral and DCFS. At the time of placement, the families were asked to identify their primary reasons for referral and presenting concerns. The primary reasons include inadequate living arrangements such as homelessness and unsafe living conditions, medical related reasons, mental health and substance abuse problems, and child at risk of abuse or neglect. The average number of children aged 5 and under in the families was approximately two (M = 1.65, SD = 0.81). Mothers ranged in age from 16 to 49, with a mean age of 26. 3.2. Bivariate relationship between crisis nursery services and foster care placement Chi-square tests and bivariate correlations were conducted to assess the relation between crisis nursery services and subsequent foster care. Both case management and parenting education were significantly associated with subsequent foster care placement (see Table 3). First, there was a significant difference in foster care placement by receipt of case management services (χ2 = 21.926, p b .001). Children whose
Table 2 Family characteristics (N = 186). n (%) Referral source Department of Children & Family Services Community Women's Center Domestic violence center Early start Laura's Home (a family homeless shelter) Metropolitan Medical Center Self-referral University hospital West Side Catholic Center Other Reasons for referral Child at risk of abuse or neglect Domestic violence Inadequate living arrangement Medical-related reasons Mental health/substance abuse Case management Participation Nonparticipation Exception/not recommended Parenting education Participation Nonparticipation Exception/not recommended Number of children under the age of 6 Mother's age at the time of inquiry
M
SD
1.65 26.06
0.81 6.30
36 (19.4) 10 (5.4) 6 (3.2) 1 (0.5) 1 (0.5) 6 (3.2) 67 (36.0) 5 (2.7) 9 (4.8) 45 (24.2) 17 (9.1) 15 (8.1) 81 (43.6) 43 (23.1) 29 (16.1) 97 (52.2) 31 (16.7) 58 (31.2) 62 (33.3) 32 (17.2) 92 (49.5)
314
D. Crampton, S. Yoon / Children and Youth Services Review 61 (2016) 311–316
Table 3 Crisis nursery services and subsequent foster care placement (N = 322). Subsequent foster care n (%)
χ2
p
Yes
No
Case management Participation Nonparticipation Exception/not recommended
37 (20.3) 24 (43.6) 9 (10.6)
145 (79.7) 31 (56.4) 76 (89.4)
21.926
b.001
Parenting education Participation Nonparticipation Exception/not recommended
22 (18.0) 24 (41.4) 24 (16.9)
100 (82.0) 34 (58.6) 118 (83.1)
16.089
b.001
families received the recommended case management were less likely to experience subsequent foster care compared with the families who did not receive case management. Similarly, children with families who participated in parenting education were less likely to experience subsequent foster care than children with the families who did not participate. Length of placement in the crisis nursery was not significantly associated with subsequent foster care (r = .054, p = .338). 3.3. Relationship between crisis nursery services and subsequent foster care Table 4 presents the significant effects of crisis nursery services on predicting the odds of subsequent foster care placement among children who have been placed in the crisis nursery. Because case management and parenting education were strongly correlated (r = .74), suggesting potential multicollinearity issues, the effects of these two predictors were assessed separately. Two separate GEE models were run, one model (Model 1) with case management and other covariates and the other one (Model 2) with parenting education and other covariates. While acknowledging that one way of dealing with multicollinearity is to combine predictors that are highly correlated (Pituch, Whittaker, & Stevens, 2013), this approach was not appropriate for our data. Since both case management and parenting education variables were recorded using complex coding systems (e.g., are the services recommended to the client? and if recommended, did the client participate?), simply aggregating these two variables would result in a loss of meaningful information. The GEE results indicated that children whose parents received the recommended case management had 65% lower odds of subsequent foster care placement compared with children whose parents refused the recommended case management services (OR = .353, p = .04). Table 4 GEE results predicting the log odds of subsequent foster care placement (N = 322). Model 1 Estimate Foster care during the previous 12 months African American Caucasian Official referral source Case management participation Not recommended for case management Parenting education participation Not recommended for parenting education OR = odds ratio. ⁎ p ≤ .05. ⁎⁎ p ≤ .01. ⁎⁎⁎ p ≤ .001. ⁎⁎⁎⁎ p ≤ .10.
Model 2 S.E. OR
Estimate
S.E. OR
1.698⁎
.67
5.47
1.467⁎
.64
4.34
1.191⁎⁎⁎⁎ 2.171⁎⁎ 0.636 −1.043⁎
.62 .75 .43 .50
3.29 8.77 1.89 .353
1.172⁎⁎⁎⁎ .64 2.068⁎⁎ .74 0.705⁎⁎⁎⁎ .42
3.23 7.92 2.02
−2.126⁎⁎⁎
.62
.119 −1.046⁎
.50
.351
−1.331⁎
.55
.264
Children from the families who were not recommended for case management also had 88% lower odds of subsequent foster care placement than children in the nonparticipant group (OR = .119, p b .001). Similar results were found in Model 2 in terms of the effects of parenting education. After adjusting for confounders, children whose parents participated in the recommended parenting education had 65% lower odds of subsequent foster care placement than children whose parents were nonparticipants (OR = .351, p = .04). Also, children whose parents were exempted from parenting education had 74% lower odds of subsequent foster care placement than children in the non-participant group (OR = .264, p = .02). In terms of other significant predictors of subsequent foster care, Models 1 and 2 yielded similar results. Children who had previous foster care placement during the 12 months before receiving crisis nursery services were approximately 4–5 times more likely to experience subsequent foster care placement than children without foster care placement during the previous 12 months (Model 1: OR = 5.464, p = .01; Model 2: OR = 4.338, p = .02). The odds of subsequent foster care placement were approximately 8–9 times higher for Caucasian children than biracial children (Model 1: OR = 8.768, p = .004; Model 2: OR = 7.909, p = .0005). African American identity marginally increased the odds of subsequent foster care (approximately 3 times higher) than biracial identity (Model 1: OR = 3.291, p = .06; Model 2: OR = 3.227, p = .06). In Model 2, referral from mandated reporters and official agencies marginally increased the likelihood of subsequent foster care placement (OR = 2.023, p = .09). 4. Discussion For nearly 40 years, Providence House has provided respite services as a means of reducing child abuse and neglect. More recently, the staff have developed and offered parenting education and case management services in addition to respite services. The staff believes that the relationships they form with caregivers during respite services provide an opportunity to engage with the caregivers and persuade them to participate in long-term services, when those services are appropriate. This study suggests that when caregivers participate in recommended crisis nursery services, the risk of their children entering foster care decreases. Given the finding that case management and parenting education reduce the odds of subsequent foster care, we asked the crisis nursery staff what were some of the reasons for families' nonparticipation in case management and parenting education. Possible reasons included employment, transportation, feeling overwhelmed, other children in the home, lack of trust in the effectiveness of the service, and being overprogrammed by other service agencies. When asked about what they do to encourage and improve family participation and engagement, the staff replied that they provide weekly bus passes to help the families participate in services at Providence House and address their needs in their community (get to work/school, other community appointments). In addition, the available hours of Providence House service provision is extensive to accommodate the family's busy schedules: Monday through Friday 9:00 am–7:00 pm and Saturdays 9:00 am–5:00 pm, three Saturdays a month. Previous research has identified that providers being nonjudgmental and showing empathy to the family, explaining potential benefits of services, discussing family expectations for service process and outcomes, and collaborating with the family to develop a plan to address practical and psychological engagement challenges or concerns may increase family engagement and participation in interventions and service programs (Ingoldsby, 2010). The finding that foster care placement during the previous 12 months is a risk factor for subsequent foster care placement is unfortunate but not surprising. A review of the research on risk factors for reentering foster care suggests that there are not yet any interventions that have demonstrated an ability to reduce re-entry (Kimberlin, Anthony, & Austin, 2009). At least in this study, crisis nursery services
D. Crampton, S. Yoon / Children and Youth Services Review 61 (2016) 311–316
also struggled to prevent repeat foster care placement. However, there were some examples of families who had children in foster care prior to receiving crisis nursery services who did not have subsequent foster care placement. A review of these cases might suggest some strategies that could be enhanced to reduce re-entry into foster care. For example, if the families with more subsequent foster care are more likely to have a risk factor such as housing instability, the crisis nursery could try to provide more housing services as a way to reduce foster care. The crisis nursery staff emphasized that, in general, families using their services have complex needs and are often overwhelmed, so it may be difficult to add other services that could reduce subsequent foster care. The findings related to racial differences were perhaps the most surprising. Previous studies have yielded mixed results regarding racial and ethnic differences in out-of-home placements, with some studies suggesting overrepresentation of African American children and others showing no racial differences. For example, some studies suggest that African American children investigated for child maltreatment have a higher risk of subsequent foster care placement (Bhatti-Sinclair & Sutcliffe, 2012; Knott & Donovan, 2010). Findings from a study that examined the effect of race/ethnicity on foster care placement in California showed that African American children were more likely to be removed from their caregivers and placed in foster care compared with Caucasian or Hispanic children, even after controlling for other factors such as age, reason for maltreatment, and neighborhood poverty (Needell, Brookhart, & Lee, 2003). Conversely, findings from other studies indicate no racial/ethnic differences after controlling for other family and environmental characteristics (Putnam-Hornstein, Needell, King, & Johnson-Motoyama, 2013). Research on biracial children is much more limited but also suggests that biracial children are at higher risk than Caucasian children (Fusco & Rautkis, 2011; Fusco, Rauktis, McCrae, Cunningham, & Bradley-King, 2010). Therefore, our findings that Caucasian children have the highest risk and biracial children have the lowest risk is somewhat inconsistent with previous research and warrants further investigation. We note that families with biracial children are least likely to be referred by DCFS and most likely to participate in parenting services, while Caucasian families are more likely to be referred by DCFS and less likely to participate in parenting and case management; however, race remained significant in our GEE model, even after controlling for these factors. The authors had a discussion with the crisis nursery staff regarding these race findings. The staff indicated that the families with biracial children often come in with compound crises and have much less social support than African American and/or Caucasian families. We informed the crisis nursery staff that there is some empirical evidence in the literature that families with biracial children are more socially isolated than other families (Fusco & Rautkis, 2011) and asked their opinions about biracial families being least likely to have children placed in foster care because their low social support makes them more willing to accept help and services from Providence house. The staff indicated that it is difficult to evaluate this connection without asking the biracial families themselves but they have noticed that both maternal and paternal sides of biracial families are more supportive of the parent receiving and utilizing crisis nursery services. The staff also reported that Caucasian families have a high risk; they are more likely to have strained family support and are less likely to fully utilize the crisis nursery services. Further qualitative interviews and focus groups with families are warranted in order to obtain more insights about what other factors might account for these racial differences. 4.1. Strengths and limitations Several limitations in this study should be noted. One major limitation is that this was a cross-sectional study and thus any causal links between parenting education, case management, and foster care placement could not be identified. In addition, this secondary study was limited to those variables that existed in the administrative data
315
set and thus only limited number of covariates were included in the models to control for possible pre-existing differences between the families who did and did not participate in case management and parenting education. Some potential confounding factors, including type of maltreatment, severity of maltreatment, presence of substance abuse problems, presence of IPV, and family poverty were not controlled for in the models because of a lack of data availability. Lastly, given that our race findings are inconsistent with dominant research that found ethnic minority as risk factors for foster care placement, the findings of the study may not be easily generalized or replicated with other samples and thus need to be verified further in future research. Despite these limitations, this study has demonstrated the value of linking a crisis nursery's administrative data with data from other service providers, specifically the DCFS. Similar studies could link crisis nursery data with homeless services, health care services, and other programs that serve vulnerable families. These studies may suggest additional partnerships or service coordination that could improve crisis nursery services. 5. Conclusions and implications Crisis nursery services delivered with case management and parenting education may be an effective intervention to reduce children's subsequent foster care placement. Furthermore, crisis nursery services may be more cost effective for public agencies and less disruptive for families compared with foster care. In addition, Caucasian children and children who have foster care experience during the previous 12 months warrant attention as they are at a heightened risk of subsequent placement into foster care. For these especially challenged families, partnerships between crisis nurseries and public child welfare agencies may be more effective than foster care alone. Efforts to improve crisis nursery services should further explore how engagement with recommended services may improve the effectiveness of crisis nursery services alone. Acknowledgments This research was funded with grants from the Doris Duke Charitable Foundation (Grant number 2010135) and several Cleveland-area foundations, including the Saint Luke's Foundation (Grant number 2010070), Lubrizol Foundation, the George W. Codrington Foundation, and two anonymous funders. We are grateful for assistance from the Providence House staff, including Annette Iwamoto, Ashley Kurz, Stacy Schiemann, Emily Shurilla, and President and CEO Natalie LeekNelson. We also appreciate the contributions of our CWRU colleagues Eun Lye Lee, Nina Lalich, and Michele Murphy. The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the study funders or Providence House. References Aniol, K., Mullins, L.L., Page, M.C., Boyd, M.L., & Chaney, J.M. (2004). The relationship between respite care and child abuse potential in parents of children with developmental disabilities: A preliminary report. Journal of Developmental and Physical Disabilities, 16(3), 273–285. ARCH (1994). National respite guidelines. Chapel Hill, NC: ARCH National Respite Network and Resource Center. ARCH (2014). Annotated bibliography of respite and crisis care studies (3rd ed.). Chapel Hill, NC: ARCH National Respite Network and Resource Center. Bhatti-Sinclair, K., & Sutcliffe, C. (2012). What determines the out-of-home placement of children in the USA? Children and Youth Services Review, 34(9), 1749–1755. Bruns, E.J., & Bujrchard, J.D. (2000). Impact of respite care services for families with children experiencing emotional and behavioral problems. Children's Services: Social Policy, Research, and Practice, 3(1), 39–61. Cole, S.A., & Hernandez, P.M. (2008). Crisis nursery outcomes for caregivers served at multiple sites in Illinois. Children and Youth Services Review, 30(4), 452–465. Cole, S.A., & Hernandez, P.M. (2011). Crisis nursery effects on child placement after foster care. Children and Youth Services Review, 33(8), 1445–1453. Cole, S.A., Wehrmann, K.C., Dewar, G., & Swinford, L. (2005). Crisis nurseries: Important services in a system of care for families and children. Children and Youth Services Review, 27(9), 995–1010. Cowen, P.S. (1998). Crisis childcare: An intervention for at-risk families. Issues in Comprehensive Pediatric Nursing, 21(3), 147–158.
316
D. Crampton, S. Yoon / Children and Youth Services Review 61 (2016) 311–316
Fusco, R.A., & Rautkis, M.E. (2011). Transracial mothering and maltreatment: Are black/ white biracial children at higher risk? Child Welfare, 91(1), 55–77. Fusco, R.A., Rauktis, M.E., McCrae, J.S., Cunningham, M.A., & Bradley-King, C.K. (2010). Aren't they just black kids? Biracial children in the child welfare system. Child and Family Social Work, 15, 441–451. Hardin, J.W., & Hilbe, J.M. (2012). Generalized estimating equations (2nd ed.). Boca Raton, FL: CRC Press. Home, A., & Darveau-Fournier, L. (1995). Respite child care: A support and empowerment strategy for families in a high-risk community. Prevention in Mental Health and Human Services, 12(1), 69–88. Ingoldsby, E.M. (2010). Review of interventions to improve family engagement and retention in parent and child mental health programs. Journal of Child and Family Studies, 19(5), 629–645. Kimberlin, S.E., Anthony, E.K., & Austin, M.J. (2009). Re-entering foster care: Trends, evidence, and implications. Children and Youth Services Review, 31(4), 471–481. Klein, S., & Jones Harden, B. (2011). Building the evidence-base regarding infants/toddlers in the child welfare system. Children and Youth Services Review, 33(8), 1333–1336.
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. Needell, B., Brookhart, M.A., & Lee, S. (2003). Black children and foster care placement in California. Children and Youth Services Review, 25(5), 393–408. Parad, H.J., & Caplan, G. (1960). A framework for studying families in crisis. Social Work, 5(3), 3–15. Pituch, K.A., Whittaker, T.A., & Stevens, J.P. (2013). Intermediate statistics: A modern approach. New York, NY: Taylor & Francis Group. Putnam-Hornstein, E., Needell, B., King, B., & Johnson-Motoyama, M. (2013). Racial and ethnic disparities: A population-based examination of risk factors for involvement with child protective services. Child Abuse & Neglect, 37(1), 33–46. SAS Institute. SAS/STAT(R) 9.2 user's guide 2nd ed., Cary, NC: Author. (2009). Tabachnick, B.G., & Fidell, L.S. (2013). Using multivariate statistics (6th ed.). Boston, MA: Pearson Education.
Contents lists available at ScienceDirect
Children and Youth Services Review journal homepage: www.elsevier.com/locate/childyouth
Crisis nursery services and foster care prevention: An exploratory study David Crampton a,⁎, Susan Yoon b a b
Jack, Joseph and Morton Mandel School of Applied Social Sciences, Case Western Reserve University, USA The College of Social Work, The Ohio State University, USA
a r t i c l e
i n f o
Article history: Received 25 March 2015 Received in revised form 1 January 2016 Accepted 1 January 2016 Available online 6 January 2016 Keywords: Crisis nursery services Case management Parenting education Foster care Race/ethnicity
a b s t r a c t Objective: To examine whether receiving crisis nursery services reduces the chances of subsequent placement of children into foster care. Method: The study includes a sample of 322 children from the administrative data from a crisis nursery in Ohio. Generalized estimating equations analysis was conducted to assess the relationship between crisis nursery services (case management and parenting education) and subsequent foster care placement. Results: The generalized estimating equations analyses indicated that case management and parenting education provided from the crisis nursery were associated with reduced likelihood of subsequent foster care for children who were placed in the crisis nursery. Foster care placement during the previous 12 months and Caucasian race were also associated with higher odds of subsequent foster care placement. Conclusions: Crisis nursery services delivered with case management and parenting education may be an effecive intervention to reduce children's subsequent foster care placement. © 2016 Elsevier Ltd. All rights reserved.
1. Introduction Abuse and neglect in early childhood pose an immediate threat to children's health and safety as well as to their long-term development. Early childhood maltreatment prevention programs and interventions provide an efficacious opportunity to intervene while young children's brains are most responsive to improved surroundings (Klein & Jones Harden, 2011). Although there is a growing research base for understanding the ecological experiences of young children in the child welfare system, we know less about how to intervene effectively in these cases. Crisis nurseries, which provide temporary child care and various support services to families of children at risk of abuse and neglect, may play a critical role in improving child welfare service outcomes, including number of out-of-home placements. However, only few studies have been conducted to investigate how and to what extent crisis nursery services may be associated with these improved outcomes. Therefore, this exploratory study addresses the need for additional research by examining how crisis nursery services, parenting education and case management in particular, may reduce the risk of out-of-home placements. Crisis intervention theory offers a useful framework for understanding how instant and targeted services may assist families at risk and prevent further negative consequences. Crisis intervention theory posits that families are most amenable to change when faced with an ⁎ Corresponding author at: Jack, Joseph and Morton Mandel School of Applied Social Sciences, Case Western Reserve University, 11235 Bellflower Road, Cleveland, OH 44106, USA. E-mail address: [email protected] (D. Crampton).
http://dx.doi.org/10.1016/j.childyouth.2016.01.001 0190-7409/© 2016 Elsevier Ltd. All rights reserved.
overwhelming life crisis and emphasizes the importance of providing immediate and focused services over a short period (Parad & Caplan, 1960). In line with crisis intervention theory, the crisis nursery model assumes that some families experience a short-term crisis that puts their children at risk of child abuse and neglect, and by receiving respite, parents can focus on their crisis and reduce the risk (ARCH, 1994; Cole, Wehrmann, Dewar, & Swinford, 2005). “Crisis nursery” refers to a type of respite care that is designed to provide temporary child care for young children at risk of abuse and neglect and to offer an array of support services to the families and caregivers of these children (ARCH, 1994). Previous studies that evaluated the effects of crisis nursery services and respite care on child welfare service outcomes have yielded mixed findings. Some studies have suggested that crisis nursery services may be associated with improved child welfare service outcomes, including fewer out-of-home placements (e.g. Bruns & Bujrchard, 2000; Cole & Hernandez, 2011; Cowen, 1998). For example, in a study of 240 families who received crisis child care in rural Iowa counties, there was a significant decrease in the reported incidence of child maltreatment in the rural counties with crisis child care programs compared with counties that did not offer crisis care programs (Cowen, 1998). In a controlled, longitudinal study of the effectiveness of respite care services, families who received respite care had significantly fewer incidents of out-of-home placement than the control group (Bruns & Bujrchard, 2000). Similarly, Home and Darveau-Fournier (1995) found that respite care (planned and crisis) for families in a high-risk neighborhood resulted in reduced incidents of foster care placement, keeping the children in the local community. In an evaluation study of five crisis nurseries in Illinois, children whose families received crisis nursery
312
D. Crampton, S. Yoon / Children and Youth Services Review 61 (2016) 311–316
services prior to foster care placement had significantly higher odds of being reunited with their biological families compared with children whose families received only foster care services (Cole & Hernandez, 2011). On the contrary, there were some studies that found no significant link between crisis nursery services/respite care and better child welfare service outcomes. For example, Aniol, Mullins, Page, Boyd, and Chaney (2004) found no significant relationship between respite care and child abuse potential in parents of children with developmental disabilities. Similar to this study, Cole and Hernandez (2008) also reported that crisis nursery services did not have a significant effect on potential for abuse and neglect. These inconsistent findings warrant further investigation as to whether crisis nursery services are associated with improved child welfare service outcomes, including reduced foster care placement. According to the ARCH National Respite Network and Resource Center (ARCH, 2014), previous research on respite services is limited in terms of its scope, depth, and methodological rigor. Cole and Hernandez (2011) have identified several challenges to evaluating whether crisis nurseries can prevent and reduce child maltreatment and foster care. For example, the respite services are often extremely short in duration (in some cases, just a few hours) and therefore are unlikely to have sustained benefits. In addition, the crisis nature of the need for services makes random selection of families into respite unethical. Even many quasiexperimental designs (e.g., waiting list designs) are not feasible in a crisis situation. Cole and colleagues have acknowledged the need for more objective measures of improvement (Cole & Hernandez, 2008, 2011) and responded to this need by conducting research that links crisis nursery services to public child welfare services. For example, they have used this linked data method to show how the receipt of crisis nursery services may have positive effects on the children's ultimate placement outcome after foster care placement (Cole & Hernandez, 2011). The present study takes this approach of linking the administrative data obtained from a crisis nursery to public child welfare records. Although crisis nursery services, including parenting education and case management, may play a pivotal role in preventing and reducing child abuse and neglect, more research is needed concerning how the services help children at risk. This investigation builds on previous crisis nursery research to explore how crisis nursery services delivered in a different state may relate to fewer foster care placements. More specifically, through a study of a crisis nursery, we explore an engagement model suggested by the social workers in the crisis nursery that a short-term crisis presents an opportunity for social workers to engage with caregivers and to use that engagement to encourage the caregivers to participate in parent education and case management services that may have long-term benefits that extend beyond the resolution of the crisis that lead the caregivers to seek emergency help. Although services offered by crisis nurseries may vary, ranging from parent education to mentoring to case management and aftercare, this study focused on two core services provided by the crisis nursery: parenting education and case management. To our knowledge, no studies to date have examined how other supportive services beyond temporary respite child care, such as parenting education and case management, may be associated with subsequent foster care placement. We hypothesized that families who are more engaged with the recommended parenting education and case management services would have a lower risk of their children entering foster care. 2. Methods 2.1. Subjects A secondary data analysis was conducted using administrative data provided by Providence House, a crisis nursery in Cleveland, Ohio. The study sample included all families (186 families and 322 children)
who received respite services from the crisis nursery between 2006 and 2009. 2.2. Procedures An administrative data file, which includes demographics, referral sources, reason for referral, type of crisis nursery services received, and the dates of inquiry, placement and discharge, was obtained from the crisis nursery. It was found that many families had numerous contacts with the agency and many children were placed in the agency multiple times. Thus, by matching the mother's first and last name, the mother's date of birth, the child's first and last name, and the child's date of birth, we removed duplicates and built a new file that had a single record per child. To assess the relationship between families receiving crisis nursery services and their involvement with the local public child welfare agency (DCFS, the Department of Children and Family Services), including foster care placement, the crisis nursery data on children and their mothers were matched to the Childhood Integrated Longitudinal Data (CHILD) System. The CHILD is maintained by the Center on Urban Poverty and Community Development at Case Western Reserve University as an evaluation tool for Cuyahoga County's early childhood initiative, which is called the Invest in Children (IIC) initiative. The CHILD system integrates birth records with records from many health and human services agencies to create a longitudinal record for all children who reside in Cuyahoga County from birth through age 6 and who appear in one or more of the agencies' data systems. At the time of this study, the CHILD system included all children with any DCFS involvement from July 1999 through 2010. We used probabilistic matching to match the crisis nursery children to the CHILD. The variables used for matching included child's date of birth, child's first name, child's last name, mother's date of birth, mother's first name, mother's last name, street name, street number, city, zip code, sex, race, and social security number. Once the matching was completed, we used the CHILD system to identify which children in the crisis nursery records had subsequent foster care placement. 2.3. Measures 2.3.1. Dependent variable The dependent variable of interest was children's foster care placement at any time after their first contact with the crisis nursery. It was assessed using the public child welfare records in the CHILD system and coded 0 = no and 1 = yes. 2.3.2. Independent variables Focal independent variables in the present study included two types of services (case management, parenting education) provided by the crisis nursery. For case management and parenting education variables, the following categories were used to code the receipt of recommended services: 1 = participation, 2 = nonparticipation, and 3 = exception/ not recommended. The participation group indicates families who received the recommended crisis nursery services (i.e., case management, parenting education), whereas the nonparticipation group indicates families who refused the recommended crisis nursery services. The exception/not recommended group indicates families who were not recommended for case management and/or parenting education. The nonparticipation group in each variable was used as the reference group for the generalized estimating equations (GEE) analyses in this study. 2.3.3. Control variables Based on theoretical relations and previous research, several covariates were identified and tested. They included child gender, child race, child age, number of children in the household, mother's age, length of placement in the crisis nursery, reasons for referral to the crisis
D. Crampton, S. Yoon / Children and Youth Services Review 61 (2016) 311–316
nursery (i.e., primary crisis stated by family), referral source, level of DCFS involvement (active or limited open case, custody, no involvement), substantiated maltreatment case, and ongoing maltreatment case. Three covariates that were correlated with outcomes at p ≤ .20 were included in the analysis: foster care placement during the previous 12 months, child race/ethnicity, and referral source. Foster care placement during the previous 12 months refers to a child having been in foster care some time during the 12 months before the first inquiry to the crisis nursery and was assessed using the public child welfare records in the CHILD system. Child race/ethnicity was self-identified by the caregivers using the following response categories: African American, Caucasian, Hispanic, biracial, Asian, and Middle Eastern. Because of the small sample size (consisting of less than 3% of the whole sample), Hispanic, Asian, and Middle Eastern children were excluded from the analyses. Three categories (African American, Caucasian, biracial) were included in the analyses and were dummy coded as African American and Caucasian, using biracial as the reference group. The referral source variable originally consisted of 11 categories, but was aggregated into two categories (official and unofficial). An official referral source included DCFS, social services centers, medical and mental health centers, and hospitals; an unofficial referral source included self-referral and others.
2.4. Statistical analyses Univariate frequencies and descriptive statistics including mean, standard deviation, minimum, maximum, skewness, and kurtosis were computed for predictors and outcome to understand the descriptive characteristics of the sample. Bivariate correlations between each predictor and the outcome were calculated to get a sense of the nature and strength of the bivariate relationships. Bivariate correlations between predictors were also inspected to check multicollinearity. Correlation above .70 was seen as a warning sign for multicollinearity (Tabachnick & Fidell, 2013). Chi-square analyses were conducted in order to assess bivariate relationships between two crisis nursery services and subsequent foster care. The generalized estimating equations (GEE) method as implemented in SAS Proc GENMOD (SAS, v. 9.2; SAS, 2009) was utilized to assess whether crisis nursery services are significant predictors of the subsequent placement of children into foster care. The GEE method is appropriate for the study because the outcome of interest is binary: whether a child had subsequent foster care (0 = no and 1 = yes). Furthermore, GEE estimators are appropriate for the study data as the data included sibling samples which may consist of nonindependent observations. The GEE method effectively addresses the issue of nonindependent and cluster-correlated information by accounting for correlation within groups (Hardin & Hilbe, 2012).
Table 1 Child characteristics (N = 322). n (%) Child gender, male Child race African American Caucasian Biracial Hispanic Middle Eastern Substantiated incident during the past 12 months, yes Ongoing DCFS case during the past 12 months, yes Foster care placement during the past 12 months, yes Child's age at admission (months) Length of placement in the crisis nursery (days)
M
SD
172 (53.4) 219 (68.2) 35 (10.9) 56 (17.4) 11 (3.4) 1 (0.3) 36 (11.2) 162 (50.3) 29 (9.0) 31.05 25.27
21.08 20.72
313
3. Results 3.1. Sample characteristics Characteristics of the children involved in this study are presented in Table 1. A total of 322 children were placed for crisis nursery services between 2006 and 2009. Slightly more than half of the children were male. The children were primarily African American, and the remainders were Caucasian and biracial. Children's age ranged from 0 to 7 years, with a mean age of 2. Approximately 9% of them had been in foster care during the previous 12 months before their family first contacted the crisis nursery. The length of placement in the crisis nursery ranged from a minimum of 0 to a maximum of 64 days, with a mean of 25 days. Table 2 shows the characteristics of the families of the children who were placed in the crisis nursery. A total of 186 families had at least one child placed at least once for crisis nursery services. Among these 186 families, slightly more than half received case management from the crisis nursery, and almost one third participated in parenting education recommended by the nursery staff. Primary referral sources included self-referral and DCFS. At the time of placement, the families were asked to identify their primary reasons for referral and presenting concerns. The primary reasons include inadequate living arrangements such as homelessness and unsafe living conditions, medical related reasons, mental health and substance abuse problems, and child at risk of abuse or neglect. The average number of children aged 5 and under in the families was approximately two (M = 1.65, SD = 0.81). Mothers ranged in age from 16 to 49, with a mean age of 26. 3.2. Bivariate relationship between crisis nursery services and foster care placement Chi-square tests and bivariate correlations were conducted to assess the relation between crisis nursery services and subsequent foster care. Both case management and parenting education were significantly associated with subsequent foster care placement (see Table 3). First, there was a significant difference in foster care placement by receipt of case management services (χ2 = 21.926, p b .001). Children whose
Table 2 Family characteristics (N = 186). n (%) Referral source Department of Children & Family Services Community Women's Center Domestic violence center Early start Laura's Home (a family homeless shelter) Metropolitan Medical Center Self-referral University hospital West Side Catholic Center Other Reasons for referral Child at risk of abuse or neglect Domestic violence Inadequate living arrangement Medical-related reasons Mental health/substance abuse Case management Participation Nonparticipation Exception/not recommended Parenting education Participation Nonparticipation Exception/not recommended Number of children under the age of 6 Mother's age at the time of inquiry
M
SD
1.65 26.06
0.81 6.30
36 (19.4) 10 (5.4) 6 (3.2) 1 (0.5) 1 (0.5) 6 (3.2) 67 (36.0) 5 (2.7) 9 (4.8) 45 (24.2) 17 (9.1) 15 (8.1) 81 (43.6) 43 (23.1) 29 (16.1) 97 (52.2) 31 (16.7) 58 (31.2) 62 (33.3) 32 (17.2) 92 (49.5)
314
D. Crampton, S. Yoon / Children and Youth Services Review 61 (2016) 311–316
Table 3 Crisis nursery services and subsequent foster care placement (N = 322). Subsequent foster care n (%)
χ2
p
Yes
No
Case management Participation Nonparticipation Exception/not recommended
37 (20.3) 24 (43.6) 9 (10.6)
145 (79.7) 31 (56.4) 76 (89.4)
21.926
b.001
Parenting education Participation Nonparticipation Exception/not recommended
22 (18.0) 24 (41.4) 24 (16.9)
100 (82.0) 34 (58.6) 118 (83.1)
16.089
b.001
families received the recommended case management were less likely to experience subsequent foster care compared with the families who did not receive case management. Similarly, children with families who participated in parenting education were less likely to experience subsequent foster care than children with the families who did not participate. Length of placement in the crisis nursery was not significantly associated with subsequent foster care (r = .054, p = .338). 3.3. Relationship between crisis nursery services and subsequent foster care Table 4 presents the significant effects of crisis nursery services on predicting the odds of subsequent foster care placement among children who have been placed in the crisis nursery. Because case management and parenting education were strongly correlated (r = .74), suggesting potential multicollinearity issues, the effects of these two predictors were assessed separately. Two separate GEE models were run, one model (Model 1) with case management and other covariates and the other one (Model 2) with parenting education and other covariates. While acknowledging that one way of dealing with multicollinearity is to combine predictors that are highly correlated (Pituch, Whittaker, & Stevens, 2013), this approach was not appropriate for our data. Since both case management and parenting education variables were recorded using complex coding systems (e.g., are the services recommended to the client? and if recommended, did the client participate?), simply aggregating these two variables would result in a loss of meaningful information. The GEE results indicated that children whose parents received the recommended case management had 65% lower odds of subsequent foster care placement compared with children whose parents refused the recommended case management services (OR = .353, p = .04). Table 4 GEE results predicting the log odds of subsequent foster care placement (N = 322). Model 1 Estimate Foster care during the previous 12 months African American Caucasian Official referral source Case management participation Not recommended for case management Parenting education participation Not recommended for parenting education OR = odds ratio. ⁎ p ≤ .05. ⁎⁎ p ≤ .01. ⁎⁎⁎ p ≤ .001. ⁎⁎⁎⁎ p ≤ .10.
Model 2 S.E. OR
Estimate
S.E. OR
1.698⁎
.67
5.47
1.467⁎
.64
4.34
1.191⁎⁎⁎⁎ 2.171⁎⁎ 0.636 −1.043⁎
.62 .75 .43 .50
3.29 8.77 1.89 .353
1.172⁎⁎⁎⁎ .64 2.068⁎⁎ .74 0.705⁎⁎⁎⁎ .42
3.23 7.92 2.02
−2.126⁎⁎⁎
.62
.119 −1.046⁎
.50
.351
−1.331⁎
.55
.264
Children from the families who were not recommended for case management also had 88% lower odds of subsequent foster care placement than children in the nonparticipant group (OR = .119, p b .001). Similar results were found in Model 2 in terms of the effects of parenting education. After adjusting for confounders, children whose parents participated in the recommended parenting education had 65% lower odds of subsequent foster care placement than children whose parents were nonparticipants (OR = .351, p = .04). Also, children whose parents were exempted from parenting education had 74% lower odds of subsequent foster care placement than children in the non-participant group (OR = .264, p = .02). In terms of other significant predictors of subsequent foster care, Models 1 and 2 yielded similar results. Children who had previous foster care placement during the 12 months before receiving crisis nursery services were approximately 4–5 times more likely to experience subsequent foster care placement than children without foster care placement during the previous 12 months (Model 1: OR = 5.464, p = .01; Model 2: OR = 4.338, p = .02). The odds of subsequent foster care placement were approximately 8–9 times higher for Caucasian children than biracial children (Model 1: OR = 8.768, p = .004; Model 2: OR = 7.909, p = .0005). African American identity marginally increased the odds of subsequent foster care (approximately 3 times higher) than biracial identity (Model 1: OR = 3.291, p = .06; Model 2: OR = 3.227, p = .06). In Model 2, referral from mandated reporters and official agencies marginally increased the likelihood of subsequent foster care placement (OR = 2.023, p = .09). 4. Discussion For nearly 40 years, Providence House has provided respite services as a means of reducing child abuse and neglect. More recently, the staff have developed and offered parenting education and case management services in addition to respite services. The staff believes that the relationships they form with caregivers during respite services provide an opportunity to engage with the caregivers and persuade them to participate in long-term services, when those services are appropriate. This study suggests that when caregivers participate in recommended crisis nursery services, the risk of their children entering foster care decreases. Given the finding that case management and parenting education reduce the odds of subsequent foster care, we asked the crisis nursery staff what were some of the reasons for families' nonparticipation in case management and parenting education. Possible reasons included employment, transportation, feeling overwhelmed, other children in the home, lack of trust in the effectiveness of the service, and being overprogrammed by other service agencies. When asked about what they do to encourage and improve family participation and engagement, the staff replied that they provide weekly bus passes to help the families participate in services at Providence House and address their needs in their community (get to work/school, other community appointments). In addition, the available hours of Providence House service provision is extensive to accommodate the family's busy schedules: Monday through Friday 9:00 am–7:00 pm and Saturdays 9:00 am–5:00 pm, three Saturdays a month. Previous research has identified that providers being nonjudgmental and showing empathy to the family, explaining potential benefits of services, discussing family expectations for service process and outcomes, and collaborating with the family to develop a plan to address practical and psychological engagement challenges or concerns may increase family engagement and participation in interventions and service programs (Ingoldsby, 2010). The finding that foster care placement during the previous 12 months is a risk factor for subsequent foster care placement is unfortunate but not surprising. A review of the research on risk factors for reentering foster care suggests that there are not yet any interventions that have demonstrated an ability to reduce re-entry (Kimberlin, Anthony, & Austin, 2009). At least in this study, crisis nursery services
D. Crampton, S. Yoon / Children and Youth Services Review 61 (2016) 311–316
also struggled to prevent repeat foster care placement. However, there were some examples of families who had children in foster care prior to receiving crisis nursery services who did not have subsequent foster care placement. A review of these cases might suggest some strategies that could be enhanced to reduce re-entry into foster care. For example, if the families with more subsequent foster care are more likely to have a risk factor such as housing instability, the crisis nursery could try to provide more housing services as a way to reduce foster care. The crisis nursery staff emphasized that, in general, families using their services have complex needs and are often overwhelmed, so it may be difficult to add other services that could reduce subsequent foster care. The findings related to racial differences were perhaps the most surprising. Previous studies have yielded mixed results regarding racial and ethnic differences in out-of-home placements, with some studies suggesting overrepresentation of African American children and others showing no racial differences. For example, some studies suggest that African American children investigated for child maltreatment have a higher risk of subsequent foster care placement (Bhatti-Sinclair & Sutcliffe, 2012; Knott & Donovan, 2010). Findings from a study that examined the effect of race/ethnicity on foster care placement in California showed that African American children were more likely to be removed from their caregivers and placed in foster care compared with Caucasian or Hispanic children, even after controlling for other factors such as age, reason for maltreatment, and neighborhood poverty (Needell, Brookhart, & Lee, 2003). Conversely, findings from other studies indicate no racial/ethnic differences after controlling for other family and environmental characteristics (Putnam-Hornstein, Needell, King, & Johnson-Motoyama, 2013). Research on biracial children is much more limited but also suggests that biracial children are at higher risk than Caucasian children (Fusco & Rautkis, 2011; Fusco, Rauktis, McCrae, Cunningham, & Bradley-King, 2010). Therefore, our findings that Caucasian children have the highest risk and biracial children have the lowest risk is somewhat inconsistent with previous research and warrants further investigation. We note that families with biracial children are least likely to be referred by DCFS and most likely to participate in parenting services, while Caucasian families are more likely to be referred by DCFS and less likely to participate in parenting and case management; however, race remained significant in our GEE model, even after controlling for these factors. The authors had a discussion with the crisis nursery staff regarding these race findings. The staff indicated that the families with biracial children often come in with compound crises and have much less social support than African American and/or Caucasian families. We informed the crisis nursery staff that there is some empirical evidence in the literature that families with biracial children are more socially isolated than other families (Fusco & Rautkis, 2011) and asked their opinions about biracial families being least likely to have children placed in foster care because their low social support makes them more willing to accept help and services from Providence house. The staff indicated that it is difficult to evaluate this connection without asking the biracial families themselves but they have noticed that both maternal and paternal sides of biracial families are more supportive of the parent receiving and utilizing crisis nursery services. The staff also reported that Caucasian families have a high risk; they are more likely to have strained family support and are less likely to fully utilize the crisis nursery services. Further qualitative interviews and focus groups with families are warranted in order to obtain more insights about what other factors might account for these racial differences. 4.1. Strengths and limitations Several limitations in this study should be noted. One major limitation is that this was a cross-sectional study and thus any causal links between parenting education, case management, and foster care placement could not be identified. In addition, this secondary study was limited to those variables that existed in the administrative data
315
set and thus only limited number of covariates were included in the models to control for possible pre-existing differences between the families who did and did not participate in case management and parenting education. Some potential confounding factors, including type of maltreatment, severity of maltreatment, presence of substance abuse problems, presence of IPV, and family poverty were not controlled for in the models because of a lack of data availability. Lastly, given that our race findings are inconsistent with dominant research that found ethnic minority as risk factors for foster care placement, the findings of the study may not be easily generalized or replicated with other samples and thus need to be verified further in future research. Despite these limitations, this study has demonstrated the value of linking a crisis nursery's administrative data with data from other service providers, specifically the DCFS. Similar studies could link crisis nursery data with homeless services, health care services, and other programs that serve vulnerable families. These studies may suggest additional partnerships or service coordination that could improve crisis nursery services. 5. Conclusions and implications Crisis nursery services delivered with case management and parenting education may be an effective intervention to reduce children's subsequent foster care placement. Furthermore, crisis nursery services may be more cost effective for public agencies and less disruptive for families compared with foster care. In addition, Caucasian children and children who have foster care experience during the previous 12 months warrant attention as they are at a heightened risk of subsequent placement into foster care. For these especially challenged families, partnerships between crisis nurseries and public child welfare agencies may be more effective than foster care alone. Efforts to improve crisis nursery services should further explore how engagement with recommended services may improve the effectiveness of crisis nursery services alone. Acknowledgments This research was funded with grants from the Doris Duke Charitable Foundation (Grant number 2010135) and several Cleveland-area foundations, including the Saint Luke's Foundation (Grant number 2010070), Lubrizol Foundation, the George W. Codrington Foundation, and two anonymous funders. We are grateful for assistance from the Providence House staff, including Annette Iwamoto, Ashley Kurz, Stacy Schiemann, Emily Shurilla, and President and CEO Natalie LeekNelson. We also appreciate the contributions of our CWRU colleagues Eun Lye Lee, Nina Lalich, and Michele Murphy. The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the study funders or Providence House. References Aniol, K., Mullins, L.L., Page, M.C., Boyd, M.L., & Chaney, J.M. (2004). The relationship between respite care and child abuse potential in parents of children with developmental disabilities: A preliminary report. Journal of Developmental and Physical Disabilities, 16(3), 273–285. ARCH (1994). National respite guidelines. Chapel Hill, NC: ARCH National Respite Network and Resource Center. ARCH (2014). Annotated bibliography of respite and crisis care studies (3rd ed.). Chapel Hill, NC: ARCH National Respite Network and Resource Center. Bhatti-Sinclair, K., & Sutcliffe, C. (2012). What determines the out-of-home placement of children in the USA? Children and Youth Services Review, 34(9), 1749–1755. Bruns, E.J., & Bujrchard, J.D. (2000). Impact of respite care services for families with children experiencing emotional and behavioral problems. Children's Services: Social Policy, Research, and Practice, 3(1), 39–61. Cole, S.A., & Hernandez, P.M. (2008). Crisis nursery outcomes for caregivers served at multiple sites in Illinois. Children and Youth Services Review, 30(4), 452–465. Cole, S.A., & Hernandez, P.M. (2011). Crisis nursery effects on child placement after foster care. Children and Youth Services Review, 33(8), 1445–1453. Cole, S.A., Wehrmann, K.C., Dewar, G., & Swinford, L. (2005). Crisis nurseries: Important services in a system of care for families and children. Children and Youth Services Review, 27(9), 995–1010. Cowen, P.S. (1998). Crisis childcare: An intervention for at-risk families. Issues in Comprehensive Pediatric Nursing, 21(3), 147–158.
316
D. Crampton, S. Yoon / Children and Youth Services Review 61 (2016) 311–316
Fusco, R.A., & Rautkis, M.E. (2011). Transracial mothering and maltreatment: Are black/ white biracial children at higher risk? Child Welfare, 91(1), 55–77. Fusco, R.A., Rauktis, M.E., McCrae, J.S., Cunningham, M.A., & Bradley-King, C.K. (2010). Aren't they just black kids? Biracial children in the child welfare system. Child and Family Social Work, 15, 441–451. Hardin, J.W., & Hilbe, J.M. (2012). Generalized estimating equations (2nd ed.). Boca Raton, FL: CRC Press. Home, A., & Darveau-Fournier, L. (1995). Respite child care: A support and empowerment strategy for families in a high-risk community. Prevention in Mental Health and Human Services, 12(1), 69–88. Ingoldsby, E.M. (2010). Review of interventions to improve family engagement and retention in parent and child mental health programs. Journal of Child and Family Studies, 19(5), 629–645. Kimberlin, S.E., Anthony, E.K., & Austin, M.J. (2009). Re-entering foster care: Trends, evidence, and implications. Children and Youth Services Review, 31(4), 471–481. Klein, S., & Jones Harden, B. (2011). Building the evidence-base regarding infants/toddlers in the child welfare system. Children and Youth Services Review, 33(8), 1333–1336.
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. Needell, B., Brookhart, M.A., & Lee, S. (2003). Black children and foster care placement in California. Children and Youth Services Review, 25(5), 393–408. Parad, H.J., & Caplan, G. (1960). A framework for studying families in crisis. Social Work, 5(3), 3–15. Pituch, K.A., Whittaker, T.A., & Stevens, J.P. (2013). Intermediate statistics: A modern approach. New York, NY: Taylor & Francis Group. Putnam-Hornstein, E., Needell, B., King, B., & Johnson-Motoyama, M. (2013). Racial and ethnic disparities: A population-based examination of risk factors for involvement with child protective services. Child Abuse & Neglect, 37(1), 33–46. SAS Institute. SAS/STAT(R) 9.2 user's guide 2nd ed., Cary, NC: Author. (2009). Tabachnick, B.G., & Fidell, L.S. (2013). Using multivariate statistics (6th ed.). Boston, MA: Pearson Education.