Foster Care: How We Can, and Should, Do More for Maltreated Children

Abstract

Foster care provides round-the-clock substitute care for nearly 700,000 U.S. children who are temporarily or permanently separated from their family of origin each year. Each state manages its own foster care system according to federal regulations. Despite numerous large-scale federal policy reforms over the past several decades, substantial concerns remain about the experiences and outcomes of children in the foster care system. The most recent effort to reform foster care, the Family First Prevention Services Act of 2018, attempts to both reduce the use of foster care and increase the quality of care. In this report, we review how policy has shaped the experiences and outcomes of children in foster care, where policy has succeeded, and where it falls short of achieving its goals. We then identify opportunities for federal and state policy to better support the safety, health, and well-being of children in foster care.

Introduction

Nearly 700,000 children are substantiated (confirmed) as victims of maltreatment in the U.S. each year, and millions more are suspected victims or at serious risk (U.S. Department of Health and Human Services, 2020). The majority of these children need services to address physical or mental health needs, and a substantial number will also need a temporary or permanent substitute caregiver in order to protect them from future harm. Foster care provides such 24/7 care by placing children with relatives (kinship care) or non-relatives (family foster homes) or, for children needing intensive levels of care or supervision, in residential group (congregate) care settings. In 2018, more than 680,000 children spent time in foster care, most commonly for reasons of neglect (62%), parent substance abuse (36%), parent inability to cope with parenting (14%), or physical abuse (13%) (U.S. Department of Health and Human Services, 2019).

Foster care is meant to be a safe respite for children being harmed by their parents. Yet, the assumption among the public and some scholars and practitioners is that placement in foster care is inherently harmful and rarely, if ever, justified (Busso et al., 2019; Franck Meyer, 2019; Raz & Sankaran, 2019). These assumptions are rooted in very real problems of foster care systems nationwide – systems that, at times, fail to provide minimum standards of safety (e.g., Fowler & Ryan, 2020), cause long-term harm to children’s development by moving them children across dozens of temporary homes (e.g., Alder, 2020), or let children reach adulthood with no legal family and nowhere to go (Annie E. Casey Foundation, 2013). These examples do not represent the majority of children’s experiences in foster care – a large majority is not abused or neglected in care, experiences two or fewer placements, and exits to reunification, adoption or permanent placement with a relative (U.S. Department of Health and Human Services, 2016, 2019). Nor, however, can these system failures be dismissed as mere aberrations. As the dozens of active class action lawsuits against state foster care systems demonstrate, too many children in foster care are deprived of the safety, stability, and connectedness of family life (Font & Gershoff, 2020).

However, the alternative to foster care may be continued abuse, neglect, and traumatization of vulnerable children. Leaving children in homes with caregivers investigated for maltreatment can have dire consequences: a study of all children born in California between 1999 and 2006 found that children who had been the subject of a maltreatment allegation, whether substantiated or not, were 6 times more likely to die from an intentional injury and 2 times more likely to die from an unintentional injury (Putnam-Hornstein, 2011). Although maltreatment-related deaths are rare, recurrent victimization is not: by age 12, 1 in 7 U.S. children has been reported to Child Protective Services more than once, and more than 1 in 4 confirmed victims of child maltreatment is revictimized (Kim & Drake, 2019).

Since child maltreatment was first recognized by the medical establishment (Kempe et al., 1962), hundreds of studies have documented that maltreatment is linked with long-term harm to children’s physical and mental health and detrimental changes in their brain structure and functioning (Jaffee, 2017; Klika & Conte, 2018; Mallett & Schall, 2019). Evidence from epigenetic studies suggests that the experience of early maltreatment “gets under the skin” through processes such as DNA methylation (Cecil et al., 2020). There is also growing evidence that maltreatment results in long-term harm to physical systems, such as the cardiovascular system (Bakema et al., 2020) and the immune system (do Prado et al., 2017). There is no doubt that maltreatment can result in severe and lasting harm to children.

Foster care is never an ideal situation for a child, but it may be the least worst alternative for children whose parents intentionally, recklessly, or negligently harm them. Those children deserve a child welfare system that endeavors to make decisions that are in their best interests and to minimize foreseeable harms (Font & Gershoff, 2020). In this report, we discuss the current state of the foster care system, efforts to reform it, and its continued shortcomings. We then propose several policy recommendations to improve the quality of foster care and the health, safety, and stability of children in care.

Overview of Foster Care in the United States

Every state in the U.S. runs its own child welfare system, through which it investigates reports of child maltreatment, determines whether children should be removed from the home, and arranges substitute care as well as physical and mental health services to mitigate the effects of maltreatment. The definition of foster care in the Code of Federal Regulations is as follows:

Foster care means 24-hour substitute care for children placed away from their parents or guardians and for whom the State agency has placement and care responsibility. This includes, but is not limited to, placements in foster family homes, foster homes of relatives, group homes, emergency shelters, residential facilities, child care institutions, and preadoptive homes (45 C.F.R § 1355.20, 2011, p. 267).

States have the legal authority to remove children from homes when they are in imminent danger of harm under their parens patriae authority, which allows states to act on behalf of children who cannot act on their own behalf (Ventrell, 2010).

There are two main ways that children enter the foster care system. The most common process is through confidential reports of suspected maltreatment to a state or county hotline, either from individuals in the community who are mandated to report suspected maltreatment (e.g., physicians, police officers, teachers) or from other concerned citizens. A child protective services caseworker then investigates whether the allegations to determine whether the child or family requires protective services to prevent future harm. If the caseworker decides the child’s health or safety is at imminent risk of harm, they will seek an emergency protection order from the court, which is typically a family court judge. In many states, caseworkers can remove the child first and ask for the court authorization afterward, within 24 to 72 hours of the removal (O’Connell, 2016). If the court authorizes the request for removal, the child becomes a temporary ward of the state or county, which is then responsible for placing them in some form of foster care.

The second, less common, way that children can enter the foster care system is through voluntary relinquishment by parents who feel they can no longer care for their children. Parents in all states can surrender newborns to emergency personnel at certain locations (e.g., fire stations, hospitals) under state safe haven laws designed to prevent infant abandonment (Child Welfare Information Gateway, 2017b). Parents may also surrender children for whom they find it difficult to provide care, such as children with severe physical disabilities or behavioral problems (Bringewatt & Gershoff, 2010).

Once in foster care, children are entitled to a permanency plan that has as its goal the placement of the child in a permanent family, either through reunification with their birth parents, adoption, guardianship, or permanent custody by a relative (Adoption and Safe Families Act [ASFA], 1997). Most children are initially given a permanency goal of reunification with their birth parents; however, only half of children will ultimately be reunited with their birth parents, largely because the parents are not able to meet the requirements mandated by the court for the return of their children, which could include parenting classes, substance use treatment, or securing safe housing. Adoption or guardianship are considered appropriate alternatives when reunification is nonviable. The amount of time children spend in foster care ranges widely: most children who are reunified spend less than one year in foster care, whereas children who are adopted or waiting for an adoptive placement spend more time in foster care.

While in foster care, federal law requires that, to the extent possible, children be placed in the least restrictive setting near their parents’ home (42 U.S.C. §675(5), 2010) and their school of origin (42 U.S.C. §675(1)(G)(i), 2010). The law further prioritizes kinship care (42 U.S.C. §671(a)(19), 2010) and placement with siblings (42 U.S.C. §671(a)(31), 2010). Together, these preferences are intended to help children maintain existing family, school, and community connections and supports and to promote normalcy, so that children in foster care can continue to have experiences and opportunities that are similar to their peers not in care.

The Family First Prevention Services Act (FFPSA)

In 2018, Congress passed and the president signed into law what has been described as the biggest overhaul of the child welfare system in 40 years—the Family First Prevention Services Act (Bipartisan Budget Act, 2018). The FFPSA includes many provisions relevant to foster care.

Funding for Prevention Services

FFPSA allows for Title IV-E funding under the Social Security Act, which previously could only be used to pay for foster care services, adoption services, and assistance to kin guardians, to be used to prevent children from entering foster care. Specifically, funding can be used for 12 months of mental health, substance abuse, and in-home skills-based parenting services to parents whose children are “candidates for foster care”, including children in voluntary kinship placements, and children of pregnant or parenting teens in foster care. FFPSA does not fund child maltreatment prevention generally; rather, it specifically seeks to prevent foster care, largely by enhancing services for families where child maltreatment has already occurred. Key provisions of prevention services funding include:

• At least half of funding must be spent on evidence-based programs. Approved programs can target any number of outcomes and need not be shown to prevent child maltreatment.

• Title IV-E funds are a partial reimbursement of state expenditures, meaning that states may not receive any additional funds under this law unless they choose to increase or reallocate their own funding.

• Title IV-E funds are the “payer of last resort”, meaning that any service covered by Medicaid (which covers many mental health, substance abuse, and other treatments) cannot be counted toward states’ matched expenditures (Kelly, 2019).

Reduction of Congregate Care

In order to make the law cost-neutral, FFPSA limits the use of Title IV-E to reimburse states for the costs of congregate care, which is substantially more expensive than family-based foster care. Children who are assessed as requiring congregate care (such as children with severe or acute mental health or behavioral problems) must be placed in “Qualified Residential Treatment Programs.” Specific provisions include:

• Reimbursement to states for children in group care settings is limited to two weeks, unless the setting is a qualified residential treatment program.

• Qualified residential treatment programs must be licensed, must have registered or licensed nursing staff, and must create a trauma-informed treatment plan for each child.

• Children placed in residential treatment must be formally assessed within 30 days to determine if they can be moved to a less restrictive setting.

• All employees of qualified residential treatment programs must pass criminal history and child maltreatment registry checks.

Foster Home Recruitment and Licensure

FFPSA included several provisions that greatly affect family foster care homes, including a major effort to standardize foster parent and foster home requirements across states:

• The law directed the Administration for Children and Families (ACF) to identify model licensing standards for foster family homes to increase uniformity across states and reduce barriers to recruiting foster and kinship caregivers. The National Model Foster Family Home Licensing Standards were issued (Administration for Children and Families, 2019b), based on recommendations developed by the American Bar Association, the Annie E. Casey Foundation, Generations United, and the National Association for Regulatory Administration (2018). States are required to adhere to these standards.

• States can complete for up to $8 million in grants to support recruiting and retaining foster parents.

• A maximum of 6 children are allowed in a foster home (exceptions apply).

• Funding is provided to improve interstate foster placements of children.

Opportunities for Foster Care Reform through Policy

Foster care is widely characterized as more harmful than not intervening (Franck Meyer, 2019; Raz & Sankaran, 2019). There is widespread variation across and within states about the population served in foster care and the experience of being in foster care, and as a result, it is challenging to draw generalized conclusions about the effects or biases of the foster care system. There are risks of harm both in decisions to remove children from their homes and decisions to leave them; neither set of risks should be taken lightly. Yet, the most rigorous studies to date—which use sophisticated methodologies to compare children who experience foster care to children who are exposed to a Child Protective Services (CPS) investigation of maltreatment but remain in their familial home—report largely null effects of foster care (meaning, no significant differences in a given outcome for those who experienced foster care and those left in their familial homes) or a mix of positive, null, and negative effects (for a more detailed review of this issue, please see Font & Gershoff, 2020). We further caution that there is a tendency among researchers and media to apply statistics about youth aging out of foster care or youth in foster care in their late teens to all youth who spent time in foster care. This is problematic for two reasons. The first is that youth who age out of care constitute only 10% of those who experience foster care and thus their experiences and characteristics should not be generalized to all children in foster care; indeed, the median age of children exiting foster care is 7.5 (U.S. Department of Health and Human Services, 2019). The second reason is that youth in foster care at older ages were potentially exposed to more severe and long-term maltreatment and thus tend to have more significant behavioral and mental health problems, the latter of which make adoption and family foster care less likely to succeed.

In addition, concerns about the use of foster care emphasize concerns that the use of foster care is driven by race and class biases (Clifford & Silver-Greenberg, 2017; Heimpel, 2019; Milner & Kelly, 2020). Concerns about race and class bias reflect both historical and current conditions. Relevant historical contexts include (1) the use of family separation during slavery and the exclusion of Black children from child protection agencies (McGowan, 2014); (2) child welfare concerns were used as pretext for separating 25–35% of all Native American children from their families and tribes in the decades preceding the 1978 Indian Child Welfare Act (Bureau of Indian Affairs, Department of the Interior, 2016), and (3) early versions of foster care, including orphanages, largely served destitute children, with little emphasis on their safety. Currently, Native American and Black children today have higher rates of involvement with the child welfare system, including foster care placement, than White, Asian, and Hispanic children (Yi et al., 2020) and a majority are socially and economically disadvantaged (Dolan et al., 2011). Though disparities remain, rates of entry and length of stay in foster care have declined precipitously among Black children, both overall and relative to White children (U.S. Department of Health and Human Services, 2006, 2019). Current race and socioeconomic disparities in foster care entry generally mirror disproportionalities in rates of child maltreatment (Sedlak et al., 2010), adverse childhood experiences (Child Trends, 2019; Sacks & Murphy, 2018), and a range of other social, economic, and health characteristics that are associated with child maltreatment and the need for substitute care.

Regardless of the source of the disparities in foster care, it is crucial that agencies engage in efforts to ensure fair treatment of all children in the system. Agencies can and should leverage strategies to eliminate the potential influence of individual level-bias, such as “color-blinded” team decision-making for removal and reunification recommendations (Pryce et al., 2019) and ethical use of predictive risk modeling (Vaithianathan et al., 2017). In addition, there may be further opportunities to reduce the need for foster care – for example, by increasing rates of uptake and effectiveness of in-home services. However, large-scale reductions in the use of foster care – and the racial and socioeconomic disparities therein – are likely safely accomplished only by preventing child maltreatment. This requires prolonged and coordinated investments in addressing overall rates of and racial disparities in the social (e.g., teen parenthood, incarceration, use of harsh discipline) and economic (e.g., poverty, employment, housing security) circumstances that precipitate child abuse and, especially, neglect (e.g., Feely et al., 2020). Such responsibility cannot, and should not, be borne by an underfunded and crisis-oriented child welfare system. Moreover, even if successful, such efforts are unlikely to replace the need for foster care: available data indicate that the U.S. uses foster care at similar or lower rates than countries with more generous social welfare regimes, such as Sweden or Denmark (EuroChild, 2010; OECD Family Database, 2010).

Although sweeping negative generalizations about foster care may be unwarranted, the mixed but largely null findings on foster care likely mask significant heterogeneity, such that some children benefit immensely from foster care and others are worse off than if they had been left in their familial homes. Differential effects of foster care likely reflect variability in (1) the degree of harm experienced before entering foster care, (2) the experiences children have in foster care, and (3) the environments children experience after exiting foster care. Not coincidentally, these three factors are the foci of a great deal of federal and state policy efforts. We provide an overview of the role of policy in children experiences before and after foster care, but we focus largely on children’s experiences within foster care.

How Policy Shapes Children’s Experience Before Entering Foster Care

The severity and duration of abuse and neglect that children experience prior to entering foster care is an important contributor to their social and behavioral health, which in turn impacts their stability in foster care (Aarons et al., 2010), opportunities for permanency (Snowden et al., 2008), and long-term functioning (Norman et al., 2012). By the time children enter foster care, most have experienced multiple incidents of abuse and neglect and many already have serious physical health problems, cognitive and academic delays, and socioemotional disturbances (Chernoff et al., 1994; McMillen et al., 2005; Steele & Buchi, 2008). Early detection and timely intervention will reduce children’s exposure to maltreatment and, for those who need it, increase the likelihood that foster care is an effective intervention.

Early Detection:

Currently, federal and state governments fund a multitude of efforts to prevent child maltreatment and improve family functioning, which likely prevents some children from requiring foster care placement. However, when voluntary prevention services are inadequate to ensure child safety, harm to children can be minimized by timely detection and intervention. Technological advances have allowed for predictive risk modeling (PRM; also referred as predictive analytics), where sophisticated algorithms predict the risk of serious maltreatment, injury, or death to inform early intervention strategies or child protection decisions. Comprehensive uses of PRM include identification of at-risk children within the general population at the time of a child’s birth (Vaithianathan et al., 2013, 2018). Less comprehensive uses of PRM, such as the Allegheny Family Screening Tool (Vaithianathan et al., 2017), have sought to identify, once a family is referred to the child welfare system, which children are at greatest risk of future harm. Both uses of PRM face criticism related to violation of privacy, lack of transparency, and potential for bias; critics also emphasize poor performance of some of the PRM tools used to date (Church & Fairchild, 2017; Dare & Gambrill, 2017). Yet, each of these criticisms also applies to traditional decision-making processes and most PRM tools are intended to augment or inform, rather than replace, human decision-making. PRM will neither eliminate all decision-making errors nor replace the need for human judgment; however, if used cautiously and with appropriate ethical oversight, it provides another tool that can improve transparency and consistency at critical decision-making junctures.

A narrower approach to early detection is a birth match system, which identifies newborn children of parents with serious histories of child abuse or neglect perpetration or who have had their parental rights involuntarily terminated to a previous child (Shaw et al., 2013), thereby allowing states to initiate an assessment of the newborn child’s safety and to offer preventive services. By only using a parent’s prior confirmed behavior, birth match systems are narrowly targeted and more transparent than the complex algorithms used in PRM. Yet, only a handful of states have birth match systems (Matthews, 2017), even though current technology makes such systems feasible in all states. Proliferation of common-sense strategies to identify the highest-risk children before they experience serious harm may, for some, allow for early intervention that prevents the need for later foster care placement, but it will also ensure that children in need of substitute care are identified before their health and development are severely compromised.

Timely Intervention:

Improving early detection of children at risk of serious harm is a necessary strategy for child protection, but it will only be effective if detection leads to an appropriate and timely response. Agencies face enormous pressure to avoid placing children in foster care, in part because of the widespread misconceptions about the average effects of foster care on children (discussed earlier in this report). However, foster care is also more expensive than leaving children in their familial homes and the decision to place children in foster care comes with substantial liability. Federal court precedents assert that children in state custody have affirmative rights to a minimum standard of care and safety (DeShaney v. Winnebago County Dept. of Social Servs., 1988; Strassburger, 2018) and legal parents have constitutional rights to “care, custody, and control” of their children (Troxel v. Granville, 2000). These protections provide legal recourse when children enter foster care; in contrast, neither the state agency nor its employees are responsible when children suffer grievous injuries at the hands of their parents, even when CPS was aware of serious threats to the child’s safety and failed to act (DeShaney v. Winnebago County Dept. of Social Servs., 1988; Kevin Bom et al. V. Superior Court of Los Angeles County, 2020). To reduce decision-making errors in intervention, federal and state governments must move away from equating decreasing foster care populations with better outcomes for children or better system performance, and instead focus on whether children’s needs for safety and adequate care are met.

How Policy Shapes Children’s Experiences within Foster Care

For children in foster care, federal policy provides a range of guidelines and supports. We highlight a few of the most important factors and opportunities for improved outcomes.

Health and Education Services:

Children in foster care are nearly universally eligible for Medicaid (Child Welfare Information Gateway, 2015), which allows them to access needed health services, including psychiatric treatment. Yet, the unwillingness of many providers to accept Medicaid patients may have a particularly pronounced impact on children in foster care, who have more health problems than their peers. For example, only about 39% of dentists (Health Policy Institute, 2016) and 36% of psychiatrists (Holgash & Heberlein, 2019) accept Medicaid. In addition, whereas state policies that require health screenings for children entering foster care (Hayek et al., 2013; Leslie et al., 2003) may facilitate early identification of needed services or treatments, placement instability is a critical barrier to effective health care. Children with unstable placements may also have mental and physical health concerns identified or effectively treated less quickly, as health care providers and foster parents lack adequate knowledge of the child’s medical and psychosocial history, including prior treatments and diagnoses (Mekonnen et al., 2009). Where possible, agencies should arrange services in a child’s new placement and transfer all necessary medical and educational records before the child is moved. Improving the array of placements within communities may also enable children to keep their service providers despite changes in placement.

Minimum Standards for Foster Homes:

Federal policy has long required that prospective foster parents undergo criminal and child maltreatment background checks (Adam Walsh Child Protection and Safety Act, 2006) and that states institute licensure procedures that are aligned with professional standards. To address criticism by advocacy groups that states were imposing middle-class standards (e.g., minimum space requirements for each child) that were unrelated to child safety and well-being on prospective foster parents (American Bar Association, 2015) and that created barriers for licensing relatives (Beltran, 2017), the Administration for Children and Families (2019b) released a set of National Model Foster Home Licensing Standards. Many of the new standards have to do with the safety of the home, such as requiring carbon monoxide detectors, safe heating, safe drinking water, no smoking in the home, and up-to-date immunizations for all household members. Some of the standards are focused on the fitness of the foster parents and include a recent physical exam and disclosure of substance use and mental illness. All use of physical or degrading punishment is prohibited, as is any illegal substance use and any excessive legal substance use.

In addition, under the ACF standards, agencies continue to be required to run criminal and child maltreatment checks on potential caregivers and cannot license individuals with felony convictions for child abuse-related offenses, a subset of violent crimes (e.g., rape, homicide), or felony physical assaults or drug offenses in the past five years. However, there is no prohibition on licensing persons convicted of misdemeanor child abuse or violent crimes, or persons who were substantiated as having perpetrated child maltreatment through a Child Protective Services case. Given that the vast majority of child neglect and physical abuse is not even investigated by law enforcement (Cross et al., 2015), and very few of the child sexual abuse cases referred for criminal prosecution are ultimately prosecuted (Block & Williams, 2019), this is a troubling omission. In addition, the final standards only require out-of-state child abuse clearances for states of residence in the past five years. A majority of kinship caregivers are grandparents (Urban Institute, 2003) and thus are unlikely to have had any dependent children in the past five years. Given that more than 4 in 10 adults have lived in more than one state (Cohn & Morin, 2008), this lax standard could permit persons with lengthy and serious histories of child maltreatment to gain custody of vulnerable children without any scrutiny of their pasts. It seems that states are free to conduct more expansive background checks and to refuse licensure to anyone with a known history of child maltreatment, but for an effort seeking to increase consistency in licensing standards, it is unclear why lax and discretionary standards for child safety are desirable.

Foster Home Capacity:

Longstanding challenges in recruitment and retention limit the ability of agencies to set high standards for licensure or to have an adequate pool of foster homes from which to select the best match for each child. Challenges with recruitment and retention are numerous, but one issue is especially noteworthy: despite proclamations of critical shortages in foster homes, many foster homes sit empty (Wulczyn et al., 2018). This seeming paradox may in part reflect insufficient coordination between licensing agencies and placement agencies, or differences in the geographic clustering of foster family applicants and children entering foster care. However, under-used foster homes also reflect a mismatch between the placement preferences of licensed families and the characteristics of children being placed, with too few families willing or trained to foster special needs children or older children (U.S. Government Accountability Office, 2018). Unfortunately, there is little systematic data collection on states’ foster home capacity, beyond statewide counts of licensed foster homes (Chronicle of Social Change, 2018). Granular data on existing foster homes (number of children they are able to provide for, in what age ranges, and with what special needs; openness to adoption) are needed to help states target their recruitment efforts more successfully.

States can also do more to support families that provide care for children with very high levels of physical, mental, or behavioral health problems. In some cases, it is not realistic for a household where all caregivers have full-time employment to provide the level of care and supervision a child needs or to attend all of the child’s medical care, mental health care, school-related, court, and visitation appointments. Despite concerns about profit motivations among foster parents (Hardesty, 2018), foster care reimbursement payments cannot compensate for loss of paid employment (Testa & Rolock, 1999) and typically do not cover the full cost of raising a child (DeVooght et al., 2013). Moreover, because foster parents have no control over when agencies decide to move a child into or out of their home, the reimbursement payments are also too unreliable to justify leaving or scaling back paid employment. Professional foster care, wherein foster parenting becomes a form of paid employment, remains controversial. Yet, as agencies seek to move children from residential and congregate care facilities (which do have paid, full-time staff) to family-based foster homes under the FFPSA, states may find that there is an insufficient number of altruistic families who can independently afford to have a parent who can stay home in order to take foster children to all of their appointments. Moreover, professionalization could allow agencies to set high standards for approving applicants and for evaluating ongoing performance. Experimental evaluations of professional foster homes for children with specialized needs are essential to inform this debate.

Placement Type:

Through a series of federal laws, and most recently with the FFPSA, the federal government has clearly established preferences for kinship care and against congregate care. These policy preferences simultaneously reflect ideas about children’s best interests and a logistical consideration of resources. There would not be an adequate supply of foster homes to meet the needs of children entering foster care without kinship foster homes, and the high costs of congregate care may redirect limited funding away from other system priorities. Thus, requirements that children be placed with relatives whenever appropriate and that congregate care be reserved for when it is absolutely necessary seem both reasonable and prudent for logistical reasons, so long as there are no adverse consequences for children’s safety and well-being.

When a prospective kinship provider meets standards of health and safety and, ideally, has a pre-existing relationship with the child in question, placing the child with that relative is an easy choice. However, the relevant policy question is not, “Should relatives be given preference over a comparable non-relative foster home?” Rather, policy must grapple with the circumstances in which kinship care may not be in the best interests of the child (Font & Gershoff, 2020). Although current policy allows relatives to be turned down for foster care placement in select circumstances, states have recognized that application of existing preferences for kin have resulted in decisions that were foreseeably harmful to children (Center for Arizona Policy, 2018; Riley, 2019), especially when screening or oversight of kinship placements is inadequate (Jonathan R. v. Governor Justice, 2019).

Conclusions of studies comparing youth in congregate care to youth with similar initial characteristics who are in family foster homes are contradictory (Barth et al., 2007; Chow et al., 2014; James et al., 2012; Lee & Thompson, 2008; McCrae et al., 2010; Ryan et al., 2008), which likely reflects the fact that the quality of both congregate and family foster care are widely variable. In addition, many youth are placed in congregate care only after unsuccessful placements in less restrictive settings, such as kinship or non-relative family foster care, and often specifically because they are engaging in violent or destructive behaviors. Hence, it is difficult to disentangle the causal association between congregate care and child psychosocial and behavioral functioning.

Nonetheless, it is an eminently reasonable goal for children to be placed in non-restrictive settings so long as they are safe, stable, and are getting their needs met. The Family First Prevention Services Act instituted requirements that states move children out of congregate care placements into traditional family foster homes, or, if strictly necessary, into Qualified Residential Treatment Programs. The premise of this requirement, based on a report from the U.S. Children’s Bureau (2015), was that a substantial portion of children in congregate care had no mental health or behavioral needs that justified a restrictive placement, to the detriment of children’s well-being and at great financial cost (No Place to Grow up: How to Safely Reduce Reliance on Foster Care Group Homes; Senate Hearing 114–273, 2015). Yet, there are ample reasons to doubt the conclusions of the Children’s Bureau report, given the strong likelihood that the federal data used in their analysis significantly undercounts mental and behavioral health problems (Font, 2020). Moreover, states will likely face serious challenges in finding homes for the children currently in congregate care. Indeed, prior to FFPSA, California committed to scaling down congregate care and invested significant state resources in realizing their goal of moving most children in congregate care into family foster homes. Yet, California found that far fewer children than anticipated could be moved into family foster homes and that the anticipated cost savings have not materialized (Loudenback, 2019). Scaling down congregate care may be a worthwhile goal irrespective of any cost savings, but few foster parents are willing, much less sufficiently trained and supported, to provide long-term homes for the subset of youth with dangerous or antisocial behavior, such as violence against caregivers, self-harm, animal abuse, fire-setting, drug abuse, theft and vandalism, sexual aggression, and running away. Some states have found that, after closing their own congregate care facilities, they have no options but to place children with serious mental or behavioral health problems in out-of-state facilities or in hotels or other short-term options (Abramo, 2018; Child Welfare Monitor, 2020; Cushman, 2019; Jones, 2019). A string of temporary placements may be more harmful to youths’ academic progress and emotional well-being than remaining in a congregate care setting. If recruitment and retention efforts are poorly targeted or executed, the alternative to congregate care for some children will be a series of short-term foster homes with caregivers poorly equipped to meet their needs.

How Policy Shapes Children’s Experiences After Foster Care

When and to whom children exit foster care is among the areas where the federal government has provided substantive guidance and funding to states. Since 1980, the federal government has required reasonable efforts to reunify children in foster care with their birth families (Adoption Assistance and Child Welfare Act, 1980). When reasonable efforts do not result in safe reunification, the federal government promotes timely completion of adoption or guardianship. Where permanency efforts are unsuccessful, states provide resources for youth who age out of care.

Timely Permanency:

Approximately half of children exiting foster care are reunified with their birth families (U.S. Department of Health and Human Services, 2019). Because children typically do not enter foster care unless there are fairly serious problems in their home environments, it is difficult to ascertain whether a higher rate of reunification is feasible. Indeed, approximately 20–40% of reunified children experience additional maltreatment post-reunification or are placed back into foster care (Kimberlin et al., 2009). Reunified children may also have worse psychosocial and academic outcomes than other children in foster care (Biehal, 2007; Font et al., 2018).

Overall, the number of adoptions from foster care increased substantially at the turn of the century, with the Adoption and Safe Families Act (ASFA) (1997) receiving much of the credit (Golden & Macomber, 2009). However, ASFA included a combination of policies that may have impacted adoption rates and was implemented in the same time period as various state policy changes, making it difficult to ascertain the impacts of specific policy provisions (Barth et al., 2004). Financial subsidies for families adopting from foster care have been expanded multiple times to include a greater number of children, and subsidies are now available for guardianships as well (Fostering Connections to Success and Increasing Adoptions Act, 2008). Adoption subsidies can increase adoption rates and decrease time to adoption for children (Buckles, 2013; Hansen, 2007). However, older children and children with specialized needs still face many barriers to adoption, and policies that financially incentivize states to increase older child adoptions appear ineffective (Brehm, 2018). Despite substantial investments in promoting adoption for children who cannot be reunified, more than 120,000 U.S. children are waiting for adoption on any given day (U.S. Department of Health and Human Services, 2018), many of whom are older, are a member of a large sibling group, or have significant psychosocial or health care needs (Macomber, 2004; Snowden et al., 2008). Notably, some children, particularly older children, prefer not be adopted (Merritt, 2008) and should have alternative stable arrangements available to them.

When adoption is the goal, the best and most common path is when children’s foster or kinship caregivers want to adopt them: these adoptions occur more quickly, save children the difficulty of attaching to yet another parental figure, and are least likely to disrupt (Semanchin Jones & LaLiberte, 2010). However, not all foster parents are interested in adoption. Foster parents who do not adopt play a valuable and necessary role in the system, given the high rate of reunification and the frequency with which families discontinue fostering after an adoption. Nonetheless, it is not optimal for individual children who do not reunify with their birth families to be placed in a foster home that is not open to adoption.

One strategy to reduce this problem would be data-driven concurrent planning – to leverage statistical modeling to identify children with a high probability of reunification and strategically place them with foster parents who are not interested in adoption, in order to increase the potential pool of foster-to-adopt families for children who are likely to become adoption-eligible. Although ASFA encourages states to plan for the possibility that reunification does not occur (i.e., concurrent planning) and some states expressly mandate it, concurrent plans may be vague and delayed until a caseworker determines reunification is unlikely, and may not be indicative of proactive efforts to identify potential adoptive parents or guardians (D’Andrade, 2009; Frame et al., 2006). There are ethical concerns with using information about the probability of reunification to inform placement decisions, the primary concern being that the information would unfairly influence the biological parents’ services, case plan, or evaluation of their progress. Because requirements for “reasonable efforts” toward reunification are subjective, limited, and difficult to enforce (Crossley, 2002) and caseworkers in many states face heavy workloads, such data could induce caseworkers to invest less in families that are statistically unlikely to reunify. To mitigate this concern, data-driven concurrent planning could be coupled with a more clearly defined and enforced “active efforts” standard for reunification (retaining exceptions for egregious cases under ASFA). This two-pronged approach could institutionalize true concurrent planning and increase stability for children, while also providing consistent opportunities to achieve safe reunification.

States commonly use a public-facing website to share (limited, typically de-identified) profiles of children who need or may need an adoptive family (there is also a national version, AdoptUSkids.org). Prospective adoptive families can view the profiles and those with approved home studies can inquire with the child’s caseworker; profiles are strengths-based and therefore typically do not provide information about children’s behavioral health needs or other factors that families may weigh. Another common strategy is to hold “matching events” where families approved for adoption meet with children’s caseworkers to learn more about available children. In both approaches, the onus is on the prospective adoptive family to find a child, rather than the agency to find a family for each child. Prospective adoptive families may be overwhelmed by the number of children with profiles and, given the limited information in the profile or made available during a matching event, uncertain how to determine which children might be a good match for their family. Some states or agencies have systems where they disseminate prospective adoptive families’ profiles to children’s caseworkers and ask caseworkers to contact them about potential matches. In both scenarios, the system can be opaque and slow due to the number of agencies and stakeholders involved, and opportunities to match children and families may be missed. Many strategies are poorly-defined or have not been subject to evaluation to determine quality of implementation or effectiveness (California Evidence-Based Clearinghouse for Child Welfare, 2019). Nongovernmental agencies are leading the way in using algorithms to help agencies narrow in on the adoptive family that is best matched with the child’s needs and preferences (e.g., family-match.org). The efficacy of these strategies is likely to depend on the quality of data provided on children’s needs (psychosocial, medical, etc.) and preferences (family type, geographic area, etc.) and on prospective adoptive families’ characteristics, skills, and preferences.

On average, time to permanency has decreased substantially over recent decades (U.S. Department of Health and Human Services, 2006, 2018). Reducing time in foster care was a core goal of the Adoption and Safe Families Act (ASFA) (1997), following general consensus that children need to know where and to whom they belong, and that prolonged existence in limbo – not knowing whether they will go home or stay or in care or be adopted – causes undue harm. Despite evidence that children thrive in stabile and nurturing environments (Britto et al., 2017), there is not clear evidence that longer durations in foster care are harmful (Biehal, 2007; Font et al., 2018), especially if children are in safe, stable placements. Nevertheless, it may be advisable to remove inadvertent barriers to permanency, such as delays in the scheduling or progression of court hearings, especially during appeals of parental rights termination (Macomber, 2004). Aside from financial incentives for adoption and encouragement of concurrent planning, there were two primary ways in which ASFA sought to fast-track permanency. First, the “15 of 22” rule instructs states to pursue termination of parental rights after a child has been in foster care for 15 of the last 22 months, unless there are “compelling reasons” not to do so. This policy reflects a longstanding consensus of researchers and physicians regarding children’s need for stability and continuity (Miller et al., 2000). However, the delineated “compelling reasons” not to file termination of parental rights are quite broad: if the child resides with a relative, if termination is not in the child’s “best interests”, or if the state failed to make reasonable efforts to reunify the child with their birth parents (Bevan, 2009; Child Welfare Information Gateway, 2013). The first criterion exempts more than a quarter of children in foster care, and the second could be construed to exempt any number of children given the lack of consensus definition of “best interests” and that many state guidelines on “best interests” directly refer to family preservation (Child Welfare Information Gateway, 2016). Moreover, states have broadly determined that it is not in a child’s best interests to be a ‘legal orphan’, such that termination of parental rights typically does not occur until after an adoptive family is identified. However, because, prior to termination, birth parents typically retain regular visitation rights, it is difficult to conduct a statewide or multi-state search for an adoptive family without termination of parental rights. (Of course, birth parent contact may be in the child’s best interests irrespective of adoption, but the rigidity of visitation schedules may unnecessarily limit children’s options for adoptive families to a small geographic area.)

Second, ASFA modified the requirement of the 1980 Adoption Assistance and Child Welfare Act that states make reasonable efforts to reunify children in foster care. ASFA retained the essence of the reasonable efforts requirement but allowed for the requirement to be bypassed in “aggravated circumstances”, as defined by each state (Child Welfare Information Gateway, 2012). The general idea was that there are some acts or omissions on the part of parents that are so egregious that reunification would be ill-advised, irrespective of the services provided. In the congressional statements leading up to ASFA’s passing, concern about a universal requirement for reasonable efforts was clear: “Reasonable efforts have become unreasonable efforts […] to reunify families that are families in name only” (Improving the Well-Being of Abused and Neglected Children, 1996). Under ASFA, however, the bypass is discretionary -- states may (and do) continue to expend federal funds to reunify families regardless of the severity of the anteceding circumstances, including in cases of sexual abuse, chronic maltreatment, or other troubling circumstances. There is no systematic tracking of states’ use of ASFA’s provisions, but available data indicate that use of the reasonable efforts bypass is rare, whereas states commonly assert compelling reasons not to file for termination after 15 months (Berrick et al., 2008; Bevan, 2009; U.S. General Accounting Office, 2002).

Support for Youth Aging Out of Care:

Approximately 10% of children who experience foster care age out (U.S. Department of Health and Human Services, 2019), though rates of aging out approach 50% for those in foster care as teens (Annie E. Casey Foundation, 2018). Youth aging out of care have received sustained interest from policymakers and both state and federal governments have made meaningful efforts to respond to their needs. Findings from multi-state longitudinal research studies have played a major role in mobilizing policy changes; such work documented the low educational attainment and high rates of homelessness, criminal justice system involvement, teen parenthood, and other social ails experienced by youth who aged out of foster care (Courtney et al., 2005; Pecora, 2005). In addition to the efforts described above to reduce the number of youth who age out (by reducing entries to foster care, and promoting adoption and guardianship), state and federal governments provide a range of supports to youth emancipating from foster care, including extended Medicaid eligibility (Child Welfare Information Gateway, 2015), tuition assistance or waivers for public universities in at least 28 states (Parker & Sarubbi, 2017), federal grants for post-secondary education (e.g., Education and Training Vouchers), and independent living services (Fernandes-Alcantara, 2019). Services for youth aging out of foster care, including guidance in how to live independently, have generally been found to increase the likelihood that foster youth will complete high school and obtain paid employment, although the evidence is mixed and youth report that the support they receive is still not sufficient (Woodgate et al., 2017).

Among the most significant changes has been the provision of federal funding to extend foster care to age 21 (Fostering Connections to Success and Increasing Adoptions Act, 2008). To date, nearly all states offer some form of extended foster care (Child Welfare Information Gateway, 2017a) but uptake rates may be low (Kelly, 2018). Low uptake of extended foster care may reflect eligibility barriers, and youth may also decline extended foster care because they had negative experiences in care, or they dislike the requirements and restrictions. Regardless, existing research, though limited, suggests that extended foster care benefits youth in terms of delayed childbearing, increased employment, and reduced justice system contact (Courtney & Okpych, 2017). However, in other areas, benefits may be short-term; for example, extended foster care is associated with increased college enrollment and reductions in short-term homelessness, but does not improve rates of college completion or longer-term homelessness (Courtney & Hook, 2017; Dworsky & Courtney, 2010). Importantly, states’ extended foster care programs may serve different populations of foster youth and provide different types of support services; thus, findings from a handful of states may not represent national average effects. More research is needed to understand the variability in youth experiences in extended foster care across states and how different models of extended foster care affect youth outcomes.

Towards a Foster Care System That Works: Reimagining Evaluation and Accountability

In an ideal system, children in need of temporary or permanent non-parental care would be identified and assisted early in life, when they have the best opportunity to rebound developmentally. For those who require foster care, an ideal system would provide every child with high-quality, stable (relative or non-relative) foster homes and evidence-based support services, and children would exit care to a permanent setting with caregivers (birth or adoptive, or guardians) who are able and willing to meet their needs and with whom children feel safe and loved. A variety of data points illustrate how far we are from that ideal system: high rates of placement instability for children in long-term foster care, low rates of adoption among older children and children with special needs, and high rates of post-reunification victimization and reentry to foster care, to name a few. What does it take to change this picture? Throughout this report, we have offered some ideas about policy reforms, many of which draw on modern data technologies to improve early detection of children who may need foster care placement, targeted foster parent recruitment, and matching children to foster and adoptive homes. However, individual policy changes may be inadequate without a more fundamental shift in how foster care systems are evaluated and held accountable for their performance.

Evaluation

As we have argued elsewhere (Font & Gershoff, 2020), rigorous evaluation provides the best means of identifying the causes of and solutions to longstanding problems. For evaluation to be useful, it must be ongoing and integrated into the quality assurance processes of state and local agencies. In addition, evaluation should be both systemic – assessing average system performance – and individual – assessing, for each child, whether the system is meeting their needs, and if not, what needs to change in their case plan or foster care environment.

Evaluation of foster care should include both impact evaluation and process evaluation. An impact evaluation determines whether foster care –as implemented by a particular agency – improves, worsens, or has no impact on the safety and well-being of children who are abused or neglected in their home environments. Process evaluation provides insight into whether agencies are poorly implementing their foster care intervention or whether their formulation of the foster care intervention is critically flawed. For example, in most states, the core components of the foster care intervention would include, among other things, placing children in the least-restrictive setting available, achieving timely permanency, and providing services to children and families. A process evaluation would assess how consistently these core components of foster care were carried out, and whether those components were meaningful associated with child safety and well-being. Most agencies evaluate whether foster care intervention components were consistently completed but do not link those components to child outcomes. Thus, agencies may be redoubling efforts on intervention components that have no effect on – or are harmful to – child safety and well-being.

Thoughtful and rigorous evaluation is especially critical because foster care reform efforts, through federal policy, regulation, and formal guidance, have been largely additive, layering more protocols, priorities, and requirements to foster care practice over time. Examples of such additions include, but are not limited to: termination of parental rights timelines, biannual permanency review hearings, and concurrent planning (Adoption and Safe Families Act, 1997), sibling placement efforts and diligent search for relatives (Fostering Connections to Success and Increasing Adoptions Act, 2008), school stability and continuity (U.S. Department of Education & U.S. Department of Health and Human Services, 2016), tracking of monthly caseworker visits (Administration for Children and Families, 2012a), and psychotropic medication monitoring (Administration for Children and Families, 2012c). These added components gain status as core system objectives without an evaluation to determine whether and how they impact the critical child outcomes of safety and well-being. Given the limited resources and chronic staffing problems of many child welfare agencies, an evaluation framework that identifies the critical components of foster care that are most important for child safety and well-being is necessary to improve practice and effectively allocate limited resources. There are at least two significant barriers to evaluation: inadequate measures of child well-being, and lack of attention to the foster home as an intervention component.

Measuring Child Well-being:

Child well-being broadly encompasses social, emotional, cognitive and behavioral functioning, and physical health (Semanchin Jones et al., 2015) and comprises the third mandate of the child welfare system. The federal government has acknowledged that achieving safety and permanency does not mean that child well-being is also achieved (Administration for Children and Families, 2012b). However, the federal government does not track the well-being of children in foster care; the Child and Family Services Reviews, which are the primary mechanism through which child welfare systems are evaluated, focus on process evaluation indicators (e.g., length of time in care) and considers measure like needs assessment and service referral to be indicators of well-being. On the whole, child well-being is poorly measured in most child welfare data systems (Font & Gershoff, 2020; Jonson-Reid & Drake, 2016).

Several states have implemented some form of standardized assessment for child well-being, such as the Child Assessment of Needs and Strengths (CANS: The John Praed Foundation, 2020). However, such tools are expensive for systems to implement because staff are required to have training and certification, which can cost thousands of dollars per person (The John Praed Foundation, 2020). Given high rates of caseworker turnover in child welfare (Casey Family Programs, 2017; Cypher, 2001), per-worker costs accumulate quickly. Moreover, such tools may be ineffective in practice, given that organizational research identifies significant subjectivity in the use of structured decision-making tools (Bosk, 2018). Because the information entered into a child welfare assessment tool is derived from caseworkers’ case documentation and interpreted by the caseworker, it is generally possible to ‘game’ assessments to produce the desired result (Prottas, 1978). If the assessment tool has implications for the caseworker or agency—i.e., it is used for performance evaluation, or for decision-making that could increase the workload burden—the potential for manipulation of the result is heightened. Hence, caseworker-completed well-being assessment tools that are costly and time-consuming may be a suboptimal approach for improving child well-being.

Technology and the proliferation of electronic case records has created new opportunities to integrate data on child well-being across multiple systems, such as health care, juvenile justice, and education. Data contained in these systems – when children are sick or injured, when they are suspended from school or make the honor roll, and what medications and services they receive – provide critical cues about child well-being that are not manipulatable and place no additional documentation burden on caseworkers. Facilitating timely and ongoing access to such records and integrating this information into child welfare data systems is a feasible – and, in the long-term, less costly—approach to centering child well-being in decision-making about child placement and services.

Youth self-reports of their well-being can augment external data indicators on child well-being. For example, the Youth Thrive™ Survey is designed for children aged 12 to 20 and taps into five key dimensions: youth resilience, social connections, knowledge of adolescent development, concrete support in time of need, and cognitive and social emotional competence (Center for the Study of Social Policy & Metis Associates, 2018). For reasons of age and ability, not all children can provide this direct feedback, and it does not replace the need for external data indicators. However, youth have critical insights into their needs and strengths that cannot be captured by other means. States now commonly include youth in permanency planning meetings and court hearings, and many areas have youth advisory boards to provide feedback on policy and practice. To date, however, youths’ perspectives are not well-integrated into data systems or performance evaluations.

Foster Homes as a Core Component of Children’s Experiences:

Foster homes are not merely holding stations for children while their birth parents receive interventions (Barth, 2001); rather, the foster home is itself a core component of intervention with the potential to mitigate or exacerbate the harms children experienced prior to placement. Yet, this is not evident from the way in which foster care is commonly assessed in research and evaluation. In research and evaluation, measures of children’s foster care placement experiences are typically reduced to placement type (e.g., kin, non-relative family, congregate) and number of placements. Evaluation of states’ foster home capacity largely focuses on counts (recruitment and retention) rather than quality or suitability to children’s needs. Yet, understanding children’s experiences within their foster homes is critical to making real progress on issues of placement instability and negative well-being outcomes. The quality of foster homes has been the subject of a limited body of research, which has found wide variability in resources, parenting behaviors, and outcomes (Berrick, 1997; Crum, 2010; Font & Potter, 2019; Simms & Horwitz, 1996). Standardized assessments such as the Home Observation for the Measurement of the Environment (HOME) Inventory (Caldwell & Bradley, 2018) or the Structured Analysis Family Evaluation (SAFE: Consortium SAFE, 2011), have been used by some child welfare agencies in home studies for prospective foster or adoptive families (Selwyn, 2011). Additional research is needed to ascertain whether these assessments reliably predict foster or adoptive home placement stability, child safety, and child wellbeing, and to identify specific exclusion criteria based on assessment results. For such research to occur, agencies must first require standardized assessments of all potential foster and adoptive families by trained staff.

Even with effective training and minimum standards for fostering, foster parents have varying lifestyles, skills, capacities, and personalities, such that they are the optimal placement for some children but not others. Thus, forward-thinking systems are using data on children and foster homes to match children to foster homes where they are most likely to thrive. Agencies collect a great deal of information about foster parents and children and some private corporations have developed algorithms geared toward improving placement stability and children’s experiences in foster care (e.g., Foster Care Technologies, 2019). These tools have great potential to reduce the placement search burden for caseworkers and improve placement experiences and outcomes for children, but too few agencies have adequate electronic records to facilitate this process. As states move toward Comprehensive Child Welfare Information Systems (CCWIS) per federal requirements (Federal Register, 2016), ensuring the new systems enable use of these promising technologies is critical.

After Evaluation: Accountability for Change

Of the many challenges faced by the child welfare system, a recurring and essential question is, who is accountable? Two aspects of accountability are relevant: case-level accountability (liability for ensuring an individual child has a safe and supportive foster care experience) and system-level accountability (liability for addressing systemic problems that constrain the effectiveness of individuals working within the system).

Case-Level Accountability:

Decisions about placement, permanency, and services for an individual child are difficult and a person acting in good faith will not achieve the right outcome every time, just as a doctor cannot cure every patient of their ailments. That does not mean there is no basis for enforcing professional standards for caseworkers, judges, and other persons with authority over a child’s case. Over several decades, federal and state policies have signaled that the child welfare system cannot be trusted to fulfill its responsibilities. Instead, accountability for individual cases is increasingly vested in court processes, with federal and state funds redirected to parents’ attorneys, guardians ad litem, and court-appointed special advocates or CASAs (Administration for Children and Families, 2019a). Guardians ad litem, CASAs, and child welfare agency caseworkers and supervisors are all instructed to act on behalf of children’s best interests (Blome & Steib, 2008); these parties are also tasked with ensuring the judge has sufficient information to make informed decisions. The performance of CASAs and guardians ad litem may vary significantly across jurisdictions based on factors such as training, caseloads, and required activities (Lowe, 2018; Minnesota Office of the Legislative Auditor, 2018; Pitchal et al., 2009). Evaluations of CASA programs and guardians ad litem find inconsistent impacts on process outputs, such as reduced time in foster care or permanency (Calkins & Millar, 1999; Leung, 1996; Osborne et al., 2019; Pilkay & Lee, 2015) and none have evaluated impacts on child safety or well-being.

Overall, the child welfare workforce has undergone a decades-long process of deprofessionalization, in which casework is largely routinized, with checklists, structured assessments, layers of oversight, and increased compliance-related documentation (Ellett & Leighninger, 2006). Moreover, the work of child welfare, and foster care in particular, is increasingly diffused across numerous agencies and employees (Blome & Steib, 2014; Ellett & Leighninger, 2006). For a single child in a foster care placement, there may simultaneously be a public agency placement caseworker, a private agency case manager, a permanency specialist, an independent living coordinator, a licensing caseworker, and various supervisors, in addition to child and family service providers (e.g., therapists). The multitude of workers performing small roles may lead to a duplication of effort in some areas, and lack of responsiveness in others, as no individual is in command of all the facts, nor tasked with looking at the big picture. Although judges, lawyers, and guardians ad litem are tasked with big picture thinking, the stream of information to the court remains controlled by the child welfare agency (DeLanoy, 2019), and it is not clear that guardians ad litem routinely conduct fact-finding of their own or consistently visit the children on whose behalf they are employed (Minnesota Office of the Legislative Auditor, 2018; Pitchal et al., 2009). Although some degree of outsourcing and specialization is helpful, overspecialization results in a system where no one sees themselves as responsible for what happens to a child in the foster care system. The foster care system needs to invest in a professional workforce, in which caseworkers have the authority, skills, and training to manage multiple facets of a child’s case, with caseworkers’ whose expertise is respected by the courts and the public.

System-Level Accountability:

There are two strategies currently in place that emphasize accountability for system performance. First, the federal Child and Family Services Reviews (CFSRs) are conducted for each state and consist of case reviews, analyses of administrative data, and interviews with key informants (Administration for Children and Families, 2014; U.S. Department of Health and Human Services, 2011). The purpose of CFSRs is to determine how states are performing on selected measures of safety, permanency, well-being. However, as noted earlier, these are largely process rather than outcome measures and many of the indicators are unrelated to, or misleading measures of, child safety and well-being. A review of the CFSR performance measures should consider and whether each measure (1) is consistently and unequivocally linked to child safety or well-being, (2) is a valid measure of the underlying construct, and (3) has the potential to create perverse incentives or produce unintended consequences.

Notably, the CFSRs typically result in a Program Improvement Plan for every state, as no state consistently meets federal performance benchmarks (Administration for Children and Families, 2017). States are expected to show progress by their next assessment, but it is not clear whether there is any consequence to failing to improve. Most federal foster care policies are enacted under funding bills and noncompliance by states could result in funding revocation (Strassburger, 2018), though it does not appear that the federal government has ever enacted financial penalties against a state for failure to meet federal standards. It is also possible that cutting federal funding to underperforming states would further compromise children’s safety and well-being.

Class action lawsuits provide another means of forcing compliance with federal policy. Courts have recognized that children in foster care have affirmative rights to minimally adequate care under the U.S. Constitution for the duration of their custody (Griffith v. Johnston, 1990; M.D. v. Abbott, Memorandum Opinion and Verdict of the Court, 2015). Dozens of lawsuits have been filed on behalf of children in foster care, focusing on threats to safety, inadequate services, and other concerns (Font & Gershoff, 2020). Many states have opted to settle such lawsuits, although the reach of Constitutional protections in this realm remains unsettled (Strassburger, 2018). Lawsuits can be effective in forcing system reforms (Center for the Study of Social Policy, 2019), but states are given many years to meet the terms of the settlement agreement, and many remain under court monitoring for decades (Font & Gershoff, 2020). If the state fails to make substantial progress, they may be held in contempt, incur additional court costs, and – in extremely rare cases – may be placed under federal receivership (A Better Childhood, 2019). In the meantime, years have passed without meaningful improvements in the care and conditions of children in foster care.

For both the CFSRs and the settlement agreements, careful attention to performance measurement is warranted. As has been demonstrated in education and criminal justice, the selection of targets or performance goals in evaluation – either system-level or individual-level evaluation – can create incentives to manipulate the measures or distort the system in ways that undermine its fundamental purpose (Campbell, 1979). For example, a competition where child protection caseworkers could receive extra money if they closed more cases was widely criticized for incentivizing shortcuts in investigations (Jackson et al., 2017). In another example, the emphasis on placing children in kinship care may be one reason that states face credible allegations that they have compromised child safety by relying on poorly-screened unlicensed kinship placements (Dwayne B. v. Granholm, 2008; Jonathan R. v. Governor Justice, 2019). Similarly, pressure to quickly find permanent families for children may incentive placement with families who are not equipped for, or even aware of, the child’s needs, and ultimately high rates of adoption disruption and dissolution for some groups of children(Schmidt et al., 1988; Testa et al., 2015). This is not to suggest that evaluation is a lost cause; rather, evaluation must be thoughtful about the meaning of measures and the potential for distortion.

Can Privatization Promote Accountability Through Competition?

In a fair market system, accountability is encouraged through competition among multiple agencies that provide the same service. In child welfare, most states contract with private agencies to recruit, license, and provide supports to foster and adoptive families, either to supplement or (less commonly) replace public agencies’ work (Chapin Hall, 2019). Thus, it may be possible to have a quasi-competitive market, through which prospective foster families select the (public or private) agency that best meets their needs. The role of agency support and responsiveness in foster care and adoption placement disruptions and in foster home retention is widely acknowledged (MacGregor et al., 2006, 2006; Schmidt et al., 1988). However, prospective foster families have limited information upon which to evaluate agencies, and once licensed, families typically cannot change agencies without beginning the licensure process anew. To improve the accountability of agencies to foster families –and reduce foster home turnover due to poor agency performance—it may be useful to allow foster parents to change agencies without repeating the licensure process and to make reports of agency performance public, thus encouraging agencies to “compete” to recruit and retain foster families by being more responsive and providing better services and supports.

Policy Recommendations

The Family First Prevention Services Act has turned necessary attention to the foster care system in this country. However, as we have demonstrated thus far, there are a number of aspects of the foster care system that were not addressed in the legislation or subsequent regulations. We also foresee several unintended consequences of the legislation that could be mitigated with state action. Following is a summary of our policy recommendations for improving foster care.

• To effectively scale down the use of congregate care and improve the quality of foster care for all children, states need a greater number and array of placement options for children with high levels of physical, mental, or behavioral health needs. To this end:

  • - States should develop the capability to track, in real time, the capacity, location, and licensure status of all public and private foster homes, as well as the number and characteristics of children for whom those homes are available (e.g., homes that will provide for teens). These data can be used to target foster home recruitment efforts and to quickly identify placement options for children in foster care.
  • - Agencies need to assess the suitability and service/training needs of potential foster parents and the quality of potential foster homes using standardized measures. These data should be available both for research to inform recruitment, retention, and service activities and for placement decision-making.
  • - States should pursue some professional foster homes for foster children who require a stay-at-home caregiver with highly specialized skills.
  • - States and agencies must evaluate when kin placements may not be preferable to other foster care placements, such as when the child has no preexisting relationship with that kin or kin placement requires other sacrifices (e.g., separation from siblings, changing schools).
  • - States should evaluate and improve the quality of services and supports for foster and adoptive families to improve retention and placement stability. Allowing licensed foster families to change agencies without re-completing the licensure process could incentivize agencies to provide high-quality for support foster families and reduce foster family turnover.
  • - ACF should revise the National Foster Home Model Licensing Standards to explicitly prohibit the placement of a foster child with an adult who has ever been substantiated as having perpetrated child maltreatment or convicted of misdemeanor child abuse or endangerment in any U.S. state at any time in their adulthood, and to require child maltreatment background checks for all states in which a person has resided as an adult.
• With federal support, states should develop and test strategies for using data to improve decision-making for children in, or at serious risk of entering, foster care.

  • - Plans for CCWIS-compliant data systems should all include creation or expansion of birth-match and predictive risk modeling to improve early detection and intervention for children likely to enter foster or who are at high risk of maltreatment.
  • - Competitive federal funding should be made available to develop and test strategies to match children likely to be freed for adoption with foster families open to adoption, and match children likely to reunify with their parents with foster families not interested or able to adopt.
• Child welfare agencies need to be held accountable for the progress of each individual child through the foster care system as well as for state-level aggregates of children’s experiences in the system. To this end:

  • - Measures used in the Federal Child and Family Service Reviews should be evaluated to determine whether and to what extent they are associated with child safety and wellbeing. Measures that are not clearly associated with safety and wellbeing should be discarded, and new measures that capture the quality of foster homes should be added.
  • - ACF should offer competitive grant funding for states to integrate health and well-being outcomes into their child welfare data systems.
  • - Child welfare agencies should have up-to-date electronic access to medical, juvenile justice, and education records for all children with an active Child Protective Services or foster care case for the purposes of assessment, case planning, and service provision, and for research and evaluation. This will reduce delays or disruptions in health care when children enter or change placements in foster care.
  • - States should require that caseworkers and guardians ad litem independently assess and submit detailed information on child functioning and wellbeing to the court for each permanency review hearing.

Conclusion

The foster care system faces a variety of challenges in meeting the needs of vulnerable children. Efforts to reform foster care have burdened agencies and caseworkers with ever-increasing compliance-related activities, monitored by multiple layers of oversight. And yet, it is not evident that compliance with current mandates achieves meaningful improvements in child well-being. Efforts to evaluate and improve system performance emphasize metrics that ignore critical components of foster care, such as the quality and appropriateness of the foster care placement, and dismiss the unique circumstances of the individual child. These metrics pursue simplicity and uniformity at the expense of child-centered decision-making and a focus on child well-being. Rigorous evaluations of process and impact—with external metrics of child safety and well-being—are critical to true accountability and innovation.