Racial/ethnic disproportionality in reunification across U.S. child welfare systems

Abstract

Background

Racial/ethnic disparities are persistent in referrals and removals of children into child welfare systems. Yet, less is known about disparities in reunification, and how system factors may contribute to more equitable outcomes for families of color.

Objective

This study examined racial/ethnic disparities in reunification rates across U.S. child welfare systems controlling for child- and system-factors.

Participants and setting

Data for this study came from the 2017 Adoption and Foster Care Analysis and Reporting System (AFCARS). We utilized a subsample of n = 284,382 children ages 0–5.

Methods

We used a bottom-up model building-approach to examine child- and system-factors associated with reunification. A series of multilevel models were run.

Results

Less than 3% of the variance in reunification occurred between state child welfare systems. Native American children had lower odds of reunification than White children (AOR = 0.87, p < .001), while Hispanic children had higher odds of reunification (AOR = 1.08, p < .001). Random effects were present for race/ethnicity and interaction terms between race/ethnicity and parental drug use were significant.

Conclusions

Racial/ethnic disparities are present in reunification, though these may vary across child welfare systems. Thus, future research could examine state systems that have better outcomes for families of color and examine factors that might explain these relationships.

Introduction

Reunification, or return to a biological parent, is the primary permanency goal in child welfare systems (CWS) (DePanfilis & Salus, 2003; Mallon & McCartt Hess, 2014). Reunification is guided by both legal and philosophical principles that parents have a right to raise their children, as well as research that shows that children in foster care consistently express their desire to return and remain connected to their biological families (Mallon & McCartt Hess, 2005). The federal Adoption and Safe Families Act (ASFA) of 1997 requires state CWS to make reasonable efforts to provide biological families with services that could support reunification (Mallon & McCartt Hess, 2005). Nationwide, reunification rates average about half of all foster care exits each year (U.S. Department of Health & Human Services [USDHHS], 2016). However, some research has found that racial disparities, defined as unequal treatment or outcome for children of color (Boyd, 2014), may create barriers to reunification for families of color (Cheng & Li, 2012; Leathers, Falconnier, & Spielfogel, 2010). This may be particularly pronounced among infants and young children, who already have lower odds of reunification (Sieger, 2020).

Infants and young children make up the largest group of children removed from their parents because of child maltreatment (U.S. Department of Health & Human Services [USDHHS], 2016), and reunification can present unique challenges for these children and their families (Humphreys & Kiraly, 2011). Several studies have found that infants are less likely to be reunified than older children (Akin, 2011; Connell, Katz, Saunders, & Tebes, 2006; Hayward & DePanfilis, 2007; Shaw, 2010; Wulczyn, Chen, & Courtney, 2011). Families with infants or toddlers in foster care face unique challenges and hurdles in forming bonds and relationships with children who have never lived with them, establishing and continuing routines, and consistency (Humphreys & Kiraly, 2011), and overcoming issues such as substance use. In fact, recent literature has found that infants and young children removed due to substance use were less likely to reunify than older children (Sieger, 2020). Struggles with substance use are particularly relevant for families in this group, as infants and young children have overwhelmingly been impacted by the opioid crisis, with over one-third of infants removed nationwide entering foster care because of parental substance use (American Academy of Pediatrics, 2017).

This study is guided by the family systems theory (Bowen, 1966; Broderick, 1971, 1993; Leupnitz, 1988). Family systems theory posits that the individual functioning of one member of a family can impact the entire family unit (Kerr, 2013). For children in the CWS, family systems theory recognizes the importance of supporting the parents and the entire family system (Minuchin, 2002). As such, assessments should include risk and protective factors for all family members, and decisions should be made based on families’ capability and functioning in relation to a child’s needs (Cimmarusti, 1992; Walton & Smith, 1999). To work towards reunification, interventions could include family boundaries, interactions, and communications to achieve a state of homeostasis (Sutphin, McDonough, & Schrenkel, 2013). Family systems theory also considers interactions between the family system and larger systems such as communities, neighborhoods, and states, and how each of these interactions can impact family functioning (White, 2005). As such, this study examined both child/family-level factors such as child race/ethnicity, age, and removal cause (e.g., family risk factors), as well as system-wide factors such as statewide poverty rate, violent crime rate, and proportion of the state child welfare budget spent on reunification services.

Guided by a family systems perspective, prior research has examined a variety of child, family, and environmental characteristics associated with reunification. In particular, the association between a child’s or family’s racial/ethnic identity has been examined in relation to reunification. Due to the over-representation of Black and Native American children in state care (Fong, Dettlaff, & Crocker, 2014; Ma, Fallon, & Richard, 2019), racial/ethnic disparities in reunification have been recognized and studied across CWS. Though several studies have failed to find any significant difference in reunification rates between white children and children of other racial/ethnic groups (see e.g., Lu et al., 2004; Putnam-Hornstein & Shaw, 2011; Wulczyn et al., 2011), multiple studies have demonstrated that Black children reunify less frequently than White children (Cheng & Li, 2012; Connell et al., 2006; Hayward & DePanfilis, 2007: Leathers et al., 2010; Putnam-Hornstein & Shaw, 2011). Yet, few studies have examined racial/ethnic disparities and reunification specifically for children ages 0–5. Given the relationship between child age and reunification (Sieger, 2020) and the vulnerability of infants and young children during the first years of life (Humphreys & Kiraly, 2011), it is important to understand how racial/ethnic disparities may impact reunification across CWS and state agency factors that may improve the odds of reunification.

At the family-level, several parental risk factors have been linked to reunification. Family factors negatively associated with reunification include parental incarceration (D’Andrade & Valdez, 2012; Moses, 2006), living in a single-parent household (Wittenstrom, Baumann, Fluke, Graham, & James, 2015), and inability to cope with parental role/responsibilities (Farmer, 2014). Evidence also overwhelmingly suggests that parental substance use impedes reunification (Akin, Brook, & Lloyd, 2015; Blakey, 2012; Brook, McDonald, Gregoire, Press, & Hindman, 2010; Choi, Huang, & Ryan, 2012; Lloyd & Akin, 2014; Mancuso, 2007; Shaw, 2010).

While little is known about parental substance use and reunification for families of color, substance use research has evidenced disparities in substance use treatment for Black (Acevedo et al., 2015; Alegria, Carson, Goncalves, & Keefe, 2011; Mennis & Stahler, 2016; Pinedo, 2019), Hispanic (Alegria et al., 2011; Pinedo, 2019), and Native American (Acevedo et al., 2015) parents with substance use disorders. Furthermore, evidence suggests that racial/ethnic disparities exist in substance use penalization and sentencing (Mooney et al., 2018). Therefore, it is possible that parental substance use may be particularly detrimental towards reunification among families of color, given disparities in treatment and penalization.

Structural bias across public systems—such as criminal justice, mental health, and healthcare systems—could contribute to racial/ethnic disparities in reunification. This systemic bias may create barriers at each step of the decision-making process for families of color, even before the child comes into the CWS (Bartholet, 2009). Evidence suggests this bias continues after an allegation occurs and permeates each stage of child welfare decision-making, from investigations to removals to permanency planning (Harris & Hackett, 2008). Bias may create disparities in tasks crucial for reunification, such as parental engagement and differing caseworker expectations for families of color (Miller, Cahn, & Orellana, 2012), as well as disparities in court-mandated substance use treatment and mental health treatment among families of color working to reunify (Osterling, Lee, & Hines, 2012).

Though systemic bias may contribute significantly to racial/ethnic disparities in reunification, few studies have examined agency or state system correlates of these disparities. Accounting for system-wide factors is particularly useful when examining racial/ethnic disproportionality, as it can adjust for other factors that may be impacted by systemic bias, that may drive other risk factors such as poverty rates or teen birth rates (Cram, Gulliver, Ota, & Wilson, 2015). Furthermore, prior literature has found significant variation in child welfare outcomes for families of color across each U.S. state (LaBrenz, Fong, & Cubbin, 2020). Thus, it is important to consider contextual factors—at the state CWS or agency level—that may further perpetuate racial disparities in child welfare outcomes.

One factor that has impacted reunification at the child- (Cheng & Li, 2012; Connell et al., 2006) and CWS-level (Wulczyn et al., 2011) is racial/ethnic disproportionality. Racial disproportionality refers to disparate outcomes for children of color in systems, such as child welfare. Traditionally, racial/ethnic disproportionality in child welfare has focused on the overrepresentation of children of color, particularly Black, Hispanic, or Native American children (Hill, 2007). Racial/ethnic disproportionality has been observed in CWS across several time points, such as referrals, substantiations, and child removals (Osterling, D’andrade, & Austin, 2008). Some prior studies examined county-level disparities of Black children and found that counties with higher disproportionality indexes (DI) in out-of-home placements had lower (Harris, 2014) and slower (Wulczyn et al., 2011) reunification rates overall.

Higher rates of maltreatment and child removal may impact rates of reunification. Some prior literature suggests that state-level poverty (Farrell et al., 2017) and community violence (Daley et al., 2016) may increase rates of child maltreatment. Other state variables such as Medicaid expansion have been found to be associated with reductions in reported child neglect cases (Brown et al., 2019) and reductions in family poverty rates, particularly for non-white families (Korenman, Remler, & Hyson, 2019)—a factor known to be associated with child abuse or neglect (Maguire-Jack & Font, 2017).

Notably, states choosing not to expand Medicaid have disproportionately higher concentrations of Black Americans—leaving a disproportionate number of Black Americans in the Medicaid coverage gap (Snowden & Graaf, 2019). The overrepresentation of this population in the coverage gap contributes to unchanged racial/ethnic disparities in mental health and substance use access (Creedon & Le Cook, 2016). Thus, limited access to treatment for non-white families may also contribute to racial/ethnic disparities in reunification in non-Medicaid expanding states.

Given differences in some findings on racial/ethnic disproportionality in reunification rates and the relationship between disparities in substance use treatment, mental health treatment, and family engagement, it is crucial to understand reunification outcomes for families of color while accounting for agency- or state-level policies and practices that could influence disproportionality. Researchers have attributed inconsistencies in the relationship between race/ethnicity and child welfare outcomes to variation across CWS (Dettlaff, Johnson-Motoyama, & Mariscal, 2014), suggesting that more research is needed to determine the relationship between race/ethnicity and reunification within and across diverse CWS. Notably, Wittenstrom et al. (2015) noted that racial/ethnic disparities in reunification outcomes could be better understood and modeled by including random effects in multilevel models.

Though prior literature has linked agency- or community-characteristics to reunification (Meezan & McBeath, 2008; Putnam-Hornstein & Shaw, 2011), few studies have utilized nested data to examine the relationship between child/family factors and reunification while accounting for contextual differences among CWS. A limited number of studies have used multi-level modeling approaches to understand predictors of other CWS outcomes—including removals, placements, and system re-entry (Huggins-Hoyt, Briggs, Mowbray, & Allen, 2019; LaBrenz et al., 2020; Yampolskaya, Armstrong, & King-Miller, 2011), demonstrating that child, family, organizational, and system factors interact to impact child outcomes. This suggests a need to extend these analyses to examine the relationship between child/family factors (e.g., removal cause, age, and length of stay in care), state factors (e.g., Medicaid expansion and state racial/ethnic disproportionality indexes), and reunification. Furthermore, given the unique challenges that families of infants and young children face in achieving reunification, research is needed that focuses on factors that contribute to reunification for families of children zero-to-five years.

This study is guided by the following research questions: 1) What child/family- and state - factors are associated with reunification for children ages zero-to-five years?; and 2) How do race/ethnicity and substance use relate to family reunification for these children differently across state CWS? Based on prior literature and ecological family systems theory, we hypothesized that: 1) family risk factors, such as parental substance use, inability to cope, and parental incarceration, would be negatively associated with reunification; 2) at the state-level, statewide poverty, teen birth rate, disproportionality indexes, and the violent crime rate would be negatively associated with reunification (e.g., higher state-level averages of these variables would result in lower odds of reunification), while state budget for reunification and drug rehabilitation per population would be positively associated with reunification; 3) Black children, Hispanic, and Native American children would have lower odds of reunification than their white peers; and 4) race/ethnicity would moderate the relationship between parental substance use and reunification.