Potential challenges associated with foster care or adoption might include

Leslie D. Leve,1 Gordon T. Harold,2 Patricia Chamberlain,1 John A. Landsverk,3 Philip A. Fisher,1,4 and Panos Vostanis2

Leslie D. Leve

1Oregon Social Learning Center, Eugene, Oregon, USA

Find articles by Leslie D. Leve

Gordon T. Harold

2University of Leicester, Leicester, UK

Find articles by Gordon T. Harold

Patricia Chamberlain

1Oregon Social Learning Center, Eugene, Oregon, USA

Find articles by Patricia Chamberlain

John A. Landsverk

3Child and Adolescent Services Research Center, San Diego, CA, USA

Find articles by John A. Landsverk

Philip A. Fisher

1Oregon Social Learning Center, Eugene, Oregon, USA

4University of Oregon, Eugene, Oregon, USA

Find articles by Philip A. Fisher

Panos Vostanis

2University of Leicester, Leicester, UK

Find articles by Panos Vostanis

Disclaimer

1Oregon Social Learning Center, Eugene, Oregon, USA

2University of Leicester, Leicester, UK

3Child and Adolescent Services Research Center, San Diego, CA, USA

4University of Oregon, Eugene, Oregon, USA

Correspondence: Leslie D. Leve, Oregon Social Learning Center, 10 Shelton McMurphey Blvd., Eugene, OR 97401, USA. Yel.: +1 541-485-2711; gro.clso@leilsel

Copyright notice

The publisher's final edited version of this article is available at J Child Psychol Psychiatry

Abstract

Background

An increasing number of children are placed in foster care g(i.e., a kin or nonkin family home other than the biological parent) due to experiences of physical, sexual, emotional, or psychological abuse, and/or neglect. Children in foster care are at increased risk for a host of negative outcomes encompassing emotional, behavioral, neurobiological, and social realms.

Methods

Areas of risk and vulnerability among foster children are described, including emotional and behavioral deficits, impaired neurobiological development, and social relationship deficits. Evidence suggesting the significance of family placement changes and prenatal exposure to substances as contributing mechanisms is presented. Based on a systematic search of the PsycINFO database (to March 2012), eight efficacious evidence-based interventions for foster families are summarized.

Findings

Although the development of evidence-based interventions that improve outcomes for foster children has lagged behind the delivery of interventions in other service sectors (e.g., mental health and educational sectors), several interventions across childhood and adolescence offer promise. Service system constraints offer both challenges and opportunities for more routine implementation of evidence-based interventions.

Conclusions

Given the increased likelihood of poor outcomes for foster children, increased efforts to understand the pathways to vulnerability and to implement interventions shown to be effective in remediating risks and improving outcomes for this population are indicated. Evaluation of efficacious interventions in countries outside of the USA is also needed.

Keywords: foster care, maltreatment, intervention, parenting, psychosocial adjustment, resilience

Introduction

International statistics suggest that an increasing number of children experience life in foster care (). Children placed in foster care have typically experienced maltreatment in the form of physical, sexual, emotional, or psychological abuse, and/or general neglect. Approximately one million cases of abuse and neglect are substantiated in the USA annually (), with approximately one in two of these children (50%) referred to live in out-of-home care (). In the UK, approximately 60,000 children are in the care of local authorities at any time, excluding children in short-term respite placements, of whom 80% live with foster carers (UK ). In this review, we describe some of the common vulnerabilities seen among foster children, including emotional and behavioral deficits, brain and neurobiological impairment, and poor social relationships with parents and peers. Next, we review eight evidence-based interventions shown to promote resilience among foster children, obtained via a systematic search of the PsycINFO database. Finally, we provide a discussion of implementation advances and challenges.

Of note, the authors of the present review have been conducting research and intervention work with children in foster care for the past 30 years, primarily in the USA but more recently through international implementation efforts in Europe, Canada, and New Zealand. Four of the evidence-based interventions reviewed here were developed by one of the present authors; we have attempted to give equal coverage of all eight interventions in this review.

It is also important to acknowledge the different and evolving policy and legal definitions of children in foster or other types of public care across the world. In some countries, including the USA, foster care includes kinship caregivers (i.e., grandparents and other family members who follow the same formal approval and monitoring requirements). In the UK, kinship care can include different types of legal status (e.g., private arrangements or children looked after by local authorities, under a Residence, or by a Special Guardianship Order); at present, such kinship caregivers tend to have less access to support and training than foster caregivers, although there is a recent trend to enhance their support systems (). Children in residential treatment centers are not included in this review, although children’s homes, group residential care facilities, orphanages, and alternative units (for younger children) still operate in many countries. In addition, youth in such care arrangements were included as the control group in two of the interventions reviewed here. Similarly, we have not included studies of adopted children, despite the important parallels with foster children.

Areas of Risk and Vulnerability for Foster Children

Background

Foster children often experience marked vulnerabilities in a number of areas. Below, we describe three areas of vulnerability that research evidence has suggested to be common among this population.

Emotional and behavioral development

A primary area of vulnerability among foster children is their mental health, marked by disruptions in emotional and behavioral development. Some of the most definitive evidence of foster children’s widespread mental health problems comes from the National Survey of Child and Adolescent Well-Being (NSCAW), a nationally representative sample of more than 6,200 children and families investigated by the child welfare system in the USA. Nearly half of these children show signs of emotional and behavioral problems, with even higher rates among children who were later placed into out-of-home care (). In a separate USA study with a large representative sample of 5- to 9-year-olds, reported that the rates of childhood psychiatric disorders (e.g., major depression, conduct disorder, and attention deficit/hyperactivity disorder) were nearly three times higher in families where potential child abuse was indicated: 49% of the children in such families were diagnosed with a psychiatric disorder (vs. 17% of the full sample). Such mental health problems can increase the likelihood of the child experiencing additional adversities like placement disruptions (). Similarly, a mental health survey of children in Great Britain indicated that foster children had significantly higher rates of disorder than children living in deprived private households (ratio of 3.7:1), but lower rates than children living in residential care (). Similar trends have been established in other countries, including Norway () and Australia ().

Increased rates of mental health problems for foster children extend into adulthood. For example, in a British study, adults with public care histories were nearly twice as likely to have seen a specialist for a mental health, drug, or alcohol problem after age 16 as compared to those who were never in the public care system (). Among USA foster children who began participating in the NSCAW during adolescence, 17% had been arrested during the previous 12-months at a follow-up in young adulthood, with arrest rates more than 4 times the national rate for 18- to 24-year-olds (). Despite the widely-documented and sustained mental health need and the associated high service costs, this population is generally underserved, particularly among ethnic minority foster children (), and this has been linked to a lack of joint care pathways and fragmentation of health and welfare services ().

Brain and neurobiological development

Foster children are also at elevated risk for disruptions in key areas of brain development. In one study of foster children and comparison children reared in low-income, nonmaltreating biological families, the foster children experienced deficits in a variety of neurocognitive functions, including poorer visuospatial processing, poorer memory skills, lower scores on intelligence tests, and less developed language capacities (). These deficits were related to specific aspects of their maltreatment history, including documented neglect or emotional abuse. Such deficits likely affect the children’s performance in school and their cognitive development; more than half of the children in the welfare system have been identified as having cognitive delays () while also evidencing pre-reading skill deficits (e.g., phonological awareness, alphabetic knowledge, and oral language ability) prior to school entry that are associated with poorer literacy skills once they begin school ().

Researchers examining brain development in foster children have identified at least two brain systems affected by early maltreatment experiences. First, the neuroendocrine stress response system, specifically the functioning of the hypothalamic-pituitary-adrenal (HPA) axis, has been shown to differ between foster and non-foster children (; ). Experiences of neglect and multiple caregiver transitions are particularly salient factors related to disruptions in the HPA system (). Second, regions in the prefrontal cortex associated with executive functioning have been shown to be affected by maltreatment and placement in foster care. Executive functioning deficits include impulsive behavior and poor decision making. The results from a study examining electrophysiological activity in the brain found that children in regular foster care (i.e., without any extra support services) showed a lack of responsiveness to feedback during an inhibitory control task relative to low-income control children or foster children who had received additional intervention services (). Together, this body of evidence suggests that children with experiences of maltreatment and placement in foster care might have enduring brain and neurobiological vulnerabilities that could affect their ability to succeed in home, school, and other social contexts.

Social relationships with parents and peers

A third area of vulnerability is foster children’s capacity to develop adaptive social relationships with caregivers and peers. Elevated levels of behavioral problems among foster children have been shown to predict elevated stress among caregivers (). Without additional supports, foster caregivers’ stress levels remain high, and they show increased stress sensitivity to children’s behavior problems over time (). Further, foster children with backgrounds of neglect and/or disordered attachment have shown increased physiological reactivity during an attachment task with their foster caregivers (), indicating that the quality of relationships with current caregivers might be compromised by experiences of prior neglect that impede the children’s abilities to regulate emotions in the context of environmental stress.

Emotional dysregulation might extend to other social contexts, including peers and difficulties establishing and maintaining positive peer relationships. For example, girls in foster care have significantly poorer peer relations at school entry than non–foster care girls (). Further, the results from several studies have shown that children with institutional or foster care histories tend to be indiscriminately friendly toward others (i.e., they readily approach individuals with whom they do not know to engage in conversations or contact, showing little social reserve; ). The results from a study of preschoolers in foster care indicate that the number of foster caregiver transitions is a factor in predicting children’s indiscriminate friendliness, with a greater number of caregivers leading to poorer inhibitory control and increased rates of indiscriminate friendliness (). Similarly, children adopted from institutional settings have been shown to have poorer peer and social relationships after a longer time in institutional care prior to adoption (; ).

Factors Associated with Foster Children’s Increased Vulnerabilities

An important starting point in examining early adverse influences is to recognize that not all children evidence expected negative outcomes; some show limited or relatively minor negative effects. What explains this difference in adaptation? The essence of adaptation in the context of adversity is captured by the scientific field of resiliency research (). Resilience is recognized as a developmental feature that captures individual differences in adaptation to specific risk contexts or developmental hazards, including maltreatment and foster care placement. The topic of individual resilience is one of considerable social, scientific, clinical, and policy importance, particularly in relation to policies that focus on the early identification, prevention, and treatment of mental health disorders and developmental impairment. Resiliency research differs from traditional concepts of risk and protection in its focus on individual variation in response to comparable experiences. Accordingly, the research focus and translation to policy application is on highlighting factors that explain individual differences in adaptation to adversity and the causal processes that they reflect, rather than on resilience as a general quality (). By highlighting the root cause of why some individuals prove resilient in the face of maltreatment, intervention studies might be directly informed by way of targeting mechanisms that facilitate adaptive responses.

Similarly, identifying the factors that explain why foster children are at elevated risk for poor psychosocial outcomes can help researchers and practitioners identify intervention opportunities. Two influences are reviewed here: placement disruptions and prenatal exposure to drugs and alcohol.

Placement disruptions

Placement and reunification failures are common, with between one third and two thirds of traditional (i.e., nonkin) foster care placements disrupting within the first 1–2 years (). Data from the NSCAW indicate that, over an 18-month period, nearly 30% of foster children experience placement instability (). Placement instability often arises from a breakdown of the child–foster caregiver relationship, but it can also result from administrative needs and policies (e.g., siblings being removed from the biological home and placed into care), although research in this area often fails to distinguish the cause of placement changes. Regardless of the underlying reason for placement changes, multiple studies have shown that placement disruptions have negative consequences for children’s emotional and behavioral development, with each change in foster home involving repeated discontinuity in caregiving experiences as well as social instability (e.g., school and peer changes); these factors are recognized as promoting negative psychological outcomes ().

Children’s externalizing behaviors (e.g., aggressive, destructive, and oppositional behavior) have been associated with placement disruptions. In a USA sample, found that, for each increase in the number of behavior problems above 6 per day, there was a 17% increase in the risk of a placement disruption within the next 12 months. Further, placement disruptions have been shown to contribute to the inhibitory control difficulties noted above (). There is also evidence from UK studies that difficult children tend to induce negative reactions in their caregivers, which can lead to a placement breakdown (). The associations between placement disruptions and behavior problems are likely bidirectional; interventions that decrease child behavior problems and increase foster family attachment and feelings of belonging might reduce the effect of behavioral problems, and increased caregiver support might reduce the number of placement disruptions (e.g., ; ).

Prenatal exposure to drugs and alcohol

A second factor associated with increased vulnerability among foster children is prenatal exposure to tobacco, drugs, and/or alcohol. noted that parental substance use is a concern in over half of child welfare families in England, with 24% exhibiting alcohol abuse and 16% exhibiting heroin abuse. More than 80% of children enter the USA foster care system due to parental substance abuse (). It has been widely documented that prenatal substance use exposure is linked to a host of poor outcomes from early childhood through adulthood. In a longitudinal study of children identified at birth with prenatal exposure, strong associations were noted between the timing, severity, and type of prenatal exposure and specific poor outcomes later in life (). Specifically, binge drinking during the first trimester was associated with severe long-term deficits in attention, memory, and cognitive processing. Similarly, maternal smoking during pregnancy had been linked to low birth weight, neurobehavioral deficits, cognitive deficits in learning and memory, and conduct problems (e.g., ; ). A few researchers have found higher rates of attention deficit/hyperactivity disorder among children exposed to nicotine prenatally (e.g., ) independent of prenatal stress. The results from studies of foster children who were prenatally exposed to substances show similar deleterious effects, including alterations in salivary cortisol response following a social stressor (e.g., giving a speech and performing mental arithmetic aloud in front of unfamiliar judges; ).

Evidence-Based Interventions That Improve Outcomes for Foster Children

The studies reviewed above clearly demonstrate that interventions that improve the well-being of foster children and their families are desperately needed. In this section, we review interventions developed and tested specifically with foster care samples and have been shown to improve at least one child outcome. As noted in a recent systematic review using the Cochrane Collaboration’s criteria (), few evidence-based programs exist for foster families; their review identified only five studies. The results from a study by the Office of Victims of Crimes suggest that only 1 of 24 mental health interventions for children who had been abused is effective (). Interventions for foster care families are unique in several regards, predominantly because the children in care have been exposed to neglectful and/or abusive parenting from a former caregiver but not from the current foster caregiver who would be involved in the intervention and is currently parenting the child. In addition, because of their histories of maltreatment, foster children are more likely to exhibit constellations of behavioral, neurobiological, and relationship vulnerabilities that pose unique challenges to caregivers: Thus, standard parenting intervention programs might not be sufficient or appropriate for foster families.

Methodology

We conducted a PsycINFO literature search to identify intervention programs that have been tested with foster care families and have been shown to be effective in improving children’s outcomes. We conducted the search in March 2012 using the terms foster care and intervention as the two required keywords in any field for all years available. All abstracts from journal articles on human populations (n = 559) were reviewed, and all articles that involved an evaluation of the efficacy of an intervention for foster children were acquired. The following inclusion criteria were applied to the results: (a) the study was a randomized controlled trial with foster children; (b) randomization occurred at the individual child level; (c) the study had a sample size of at least 15/group, making it sufficiently powered to detect replicable effects; and (d) the intervention produced at least one positive outcome for the intervention children relative to the control children. If an intervention was identified as meeting criteria (d), then all published studies of that intervention (whether showing positive or negative results) have been considered in the review. Eight interventions (most of which had multiple outcome publications) met all four inclusion criteria.

Although each identified intervention applies a different model, they share a common set of characteristics: (a) a focus on reducing known risk factors and enhancing individual strengths, (b) sensitivity to child age and developmental level, and (c) built on evidence for the mediating role of parenting in linking early adversity with child outcomes. The eight identified interventions are reviewed below, with additional information provided in Table 1. Table 1 also includes summary information about the range of effect sizes for each study and outcome, using Cohen’s effect size recommendations (r effects: small ≥ .10, medium ≥ .30, large ≥ .50; d effects: small ≥ .20, medium ≥ .50, large ≥ .80; ). We provide these general small, medium, and large effect size indicators for each outcome but caution readers that the interpretation of effect sizes is always context specific, depending on the specific outcome and other study design issues such as the reliability of the measures (); therefore, we encourage readers to refer to the original studies for more information. We did not initiate a meta-analysis given the heterogeneity of the study specifics included across our review. Rather, we summarize outcomes for each intervention separately and present general effect size information as included in the original publication or (in the absence of such information) using the data provided in the publication with an online effect size calculator (). The lack of a pre-specified analysis plan in most of the studies reviewed and our method of selective reporting of positive effects from reported analyses means that there is likely to be a reporting bias towards positive effects in this review (i.e., less attention is given to what interventions did not achieve).

Table 1

Intervention Programs Tested in Foster Care Settings with Efficacious Child Outcomes: Program Effects

Program name and primary referencesNSample and intervention characteristicsLongest intervention outcome follow-up publishedEffect Size1Mental healthBrain and neurobiological developmentRelationships: parents and peersPlacement disruptionsEarly childhoodAttachment and Biobehavioral Catch-up (ABC)
 
 46 intervention; 47 control; 48 non–foster care controlsUSA sample of children aged 15–24 months old at outcome assessment. Control condition included developmental education intended to improve cognitive skills in foster children. Also included a non–foster care control group.Post-intervention (unclear time period)S60 intervention; 60 controlUSA sample of children aged 2–21 months old at enrollment. Control condition included developmental education intended to improve cognitive skills in foster children.1 month post-intervention (or later for younger children)MMultidimensional Treatment Foster Care for Preschoolers (MTFC-P)
 
 
 
 57 intervention; 60 control; 60 non–foster care controlsUSA sample of children aged 3–6 years. Control condition consisted of foster care services-as-usual. Also included a non–foster care control group.2 years post-baselineMSSBucharest Early Intervention Project (BEIP)
 
 
 
 68 intervention; 68 control; 72 non-institutionalized controlsIntervention and control families recruited through an institutional setting in Romania when children were 5–31 months old. Control children remained in the institution at start of study (some were later placed). Also included a non-institutionalized control group.Child age 8 yearsS
LSLMiddle childhoodModified Incredible Years (IY) 40 intervention; 24 controlUSA sample of primarily neglected children aged 3–10 years. Primarily Latino and African American sample. Control condition consisted of foster care services as usual.3 months post-interventionS
MKeeping Foster Parents Trained and Supported (KEEP)
 
 359 intervention; 341 controlUSA sample of children aged 5–12 years. Large Latino and African American subsamples. Control condition consisted of foster care services as usual.11 months post-baselineSSSMiddle School Success (MSS)
 
 48 intervention; 52 controlUSA sample of girls transitioning to middle school (age 11). Control condition consisted of foster care services as usual.3 years post-baselineS
MMFostering Individualized Assistance Program (FIAP)
 
 54 intervention; 78 controlUSA sample of children aged 7–15 years with or at risk for an emotional or behavioral disorder. Control condition consisted of foster care services as usual.18 months post-baselineSSAdolescenceMultidimensional Treatment Foster Care for Adolescents (MTFC-A)
 
 
 
 37 intervention; 42 controlUSA sample of boys aged 12–17 years in juvenile justice who had been referred for out-of-home care. Control condition consisted of out-of-home care services as usual (typically, group care).2 years post-baselineM
LMS
M81 intervention; 85 control2 USA samples of girls aged 13–17 years in juvenile justice who had been referred for out-of-home care. Control condition consisted of out-of-home care services as usual (typically, group care).2 years post-baselineS
MM20 intervention; 15 controlSwedish sample of youths aged 12–18 years diagnosed with conduct disorder who were at immediate risk of out-of-home placement. Control condition consisted of foster care services as usual.2 years post-baselineS
M

Open in a separate window

1Effect sizes denote small (S), medium, (M), and large (L) effects, using Cohen’s effect size recommendations (r effects: small ≥ .10, medium ≥ .30, large ≥ .50; d effects: small ≥ .20, medium ≥ .50, large ≥ .80; ).

Early childhood

A primary developmental task during early childhood is the formation of a positive and secure attachment relationship with a supportive caregiver. This process can be disrupted when children experience maltreatment from their caregiver and multiple foster placements. Three independent interventions for young foster children demonstrate that, when foster caregivers are given appropriate support and training, children can develop healthy emotion and behavior regulation and positive, secure social relationships. One 10-session intervention, Attachment and Biobehavioral Catchup (ABC), was designed to help caregivers facilitate healthy regulation of their child’s behavior and stress-responses by teaching caregivers to be highly responsive to the child’s emotions and increasing caregivers’ provision of nurturing care and promotion of attachment security (). This intervention has been successful in normalizing stress responses (i.e., cortisol reactivity) among foster children whose caregivers were randomly assigned to the ABC intervention relative to children in a foster care control intervention condition. The ABC children were also more often secure and less often disorganized in their attachments to caregivers than were the control children, with 32% of the ABC children (vs. 57% of the control children) having a disorganized attachment to their caregiver and 52% of the ABC children (vs. 33% of the control children) having a secure attachment approximately 1 month after the intervention (). Publications demonstrating positive effects of the ABC intervention on mental health outcomes are not yet available in the literature.

A second intervention, Multidimensional Treatment Foster Care for Preschoolers (MTFC-P), uses a behavior-management approach and intensively trains, supervises, and supports foster caregivers to provide positive adult support and consistent limit setting. MTFC-P includes coordinated interventions with the child’s biological parents. Although this strength-based intervention is not focused specifically on attachment security, it has produced positive attachment outcomes. The MTFC-P children showed increased secure behavior and decreased avoidant behavior relative to the children in a regular foster care control condition with small effect sizes; the MTFC-P children had a 10% increase (vs. 6% in the control children) in rates of secure behavior over a 12-month period (). In addition, MTFC-P outcomes include significant influences on stress response systems: the intervention effectively prevented the MTFC-P children from having blunted diurnal HPA axis function, with medium effect sizes () and reduced caregiver stress (). Further, the intervention improved placement stability outcomes across a 2-year period and mitigated the risk of multiple prior foster care placements on children’s subsequent placement failures (). Compared to the MTFC-P children, the control children were 3.6 times more likely to have a permanent placement failure. Further, children with three of more prior placement failures were at even heightened risk of a permanent placement failure: the probability of an additional failed permanent placement was approximately three times larger for the control children. Similar to the ABC intervention, mental health outcomes resulting from the MTFC-P intervention are not yet available in the published literature.

A third early childhood intervention is the Bucharest Early Intervention Project (BEIP). In the BEIP, children who were institutionalized since birth were randomly assigned to continue living in an institutional setting or to be placed in foster care. The foster caregivers received ongoing support from social workers in managing challenging behavior, encouraging child-centered parenting, and organizing a support group. The intervention was focused on developing attachment relationships, facilitating language development, and providing foster parents with techniques for managing difficult child behavior (). This is an unusual intervention because intervention supports were provided until children were 54 months of age (a lengthy intervention), and the sample was institutionalized. Both of these design features may limit the generalizability of the findings from the BEIP to the general population. Nonetheless, a wide range of outcomes has been examined in the BEIP, up to 8-years later, with significant effects in multiple domains (; ; ). First, the intervention children were more likely to have secure caregiver attachments: at the 42-month follow-up, 49% of the intervention children versus 18% of the control children were securely attached. Second, the intervention children exhibited improved cognitive outcomes (mean IQ score 5.3 points higher than the control children at age 8), higher levels of attention (small effect size), and greater positive affect (large effect size) at 30–42 months. The IQ effects were only marginal for full-scale IQ scores, and the age 8 IQ outcomes were not as strong as age 42- and 54-month outcomes (). Third, the intervention children exhibited fewer internalizing disorders at 54 months: 22% of intervention children versus 44% of control children met diagnostic criteria. However, the intervention did not result in a reduction of total psychiatric symptoms for boys ().

Middle childhood

Four interventions for foster families have been shown to be effective during middle childhood. First, the Incredible Years (IY) intervention, which has been shown to be effective in populations of young at-risk children and children with conduct problems in the USA and UK (e.g., ), was modified and implemented in a middle childhood foster care sample. The modified IY intervention incorporated a coparenting component between foster and biological caregivers to expand their knowledge of each other and their child, practice open communication, and negotiate interparental conflict regarding topics such as family visitation, family routines, and discipline (). Families assigned to the intervention condition exhibited improvements in positive discipline (small effects at the end of the intervention that became large effects at a 3-month follow-up) and coparenting skills (small effects at the end of the intervention only) relative to the control families, indicating the potential of IY programs that include coparenting components to ultimately reduce child mental health problems for foster children. However, the intervention did not yield significant effects on children’s externalizing problems, which was the targeted distal outcome of this intervention.

A second intervention for middle childhood is Keeping Foster Parents Trained and Supported (KEEP). In an RCT evaluation of KEEP, foster caregivers who were receiving a new placement were randomly assigned to foster care services as usual or to the KEEP group intervention for 16 weeks. This included training, supervision, and support to foster parents in applying behavior management strategies. The results suggested that KEEP was effective in reducing child behavior problems compared to the services-as-usual control condition. In addition, improvements in child behavior problems were associated with intervention-driven improvements in parenting (). The intervention improved placement stability in two ways: by increasing the likelihood of reunification with biological, relative, or adoptive families (9% of the control children vs. 17% of the KEEP children experienced a positive placement change); and by mitigating the risk-enhancing effects of previous multiple placements (). Specifically, each additional placement that the control children experienced corresponded with a 15% increase in subsequent placement disruptions, whereas there was no association between the number of prior placements and new placement disruptions for the KEEP children. Intervention effects were not found for the likelihood of negative exits (e.g., child runaways, placement in a different foster home).

A third intervention for middle childhood, Middle School Success (MSS), is a derivative of KEEP that specifically targeted youth exiting primary school. MSS included foster caregiver and youth components, with 6 sessions over the summer prior to middle school entry and ongoing weekly sessions over the 1st year of middle school. The foster caregiver sessions were group based and behavior management oriented; the youth sessions were group based (summer) and individually based (school year) and were oriented toward skill building. Relative to a foster care services-as-usual control group, MSS youths exhibited decreased externalizing and internalizing problems at a 6-month follow-up () and at 12- to 24-month follow-ups (). For example, MSS girls displayed an average of 1.1 internalizing problems and 2.4 externalizing problems per day (vs. 1.5 and 2.9, respectively, in the control group) at the 6-month follow-up. Examination of intervention effects on prosocial behavior at 6 months were also examined, but were nonsignificant. However, the MSS resulted in increased prosocial behavior and fewer placement changes at a 12-month follow-up (.76 placement changes for the control girls vs. .33 for the MSS girls) and reduced substance use at a 36-month follow-up, specifically reduced tobacco and marijuana use ().

A fourth intervention for middle childhood is the Fostering Individualized Assistance Program (FIAP; ). The FIAP was focused on wrapping services around the child based on their individual and family needs with the goal of improving placement stability and reducing behavior and emotional problems. The intervention had four components, including a strength-based assessment, life-domain planning, clinical case management, and follow-along supports and services. Compared to the services-as-usual group, the FIAP children exhibited fewer attention, withdrawal, and total problems at an 18-month follow-up, assessed via caregiver report (). For example, attention problem scores were 7.2 for the FIAP children and were 7.9 for the control children. In addition, the FIAP children had fewer runaways and spent less time incarcerated than the control children. Increased placement stability has been found for older FIAP children only (). Although this evaluation noted numerous positive effects, many of these were specific to a subpopulation (e.g., boys or older children) or were not found when subscales or youth reports were examined.

Adolescence

One intervention has been shown to produce positive outcomes for foster adolescents: Multidimensional Treatment Foster Care for Adolescents (MTFC-A), a multicomponent program that involves individual placement with a specialized foster family (). In MTFC-A, youths are placed in community homes where foster caregivers are intensively trained, supervised, and supported to provide positive adult support and mentoring, close supervision, and consistent limit setting. MTFC-A placements typically last 6–9 months and involve coordinated interventions in the home, with peers, in educational settings, and with the adolescent’s long-term placement resource. The results from MTFC-A trials in the USA have indicated its effectiveness in reducing arrest rates and deviant peer affiliations for boys and girls, placement disruption and parenting for boys, and pregnancy rates and school engagement for girls (; ; ). For example, between baseline and a 12-month follow-up, the MTFC-A youths had spent 53 (boys) and 22 (girls) days in lockup (e.g., a detention facility) versus 129 and 56 days, respectively, for the control youths. However, intervention effects were not found for girls’ self-reported delinquency. In the 24-months following baseline, 26.9% of MTFC girls had a new pregnancy versus 46.9% of the control girls. International replication trials of MTFC-A have also shown positive results. The results from a trial in Sweden indicated significant reductions in youth-reported externalizing and internalizing behavior relative to a services-as-usual control group (). Further, depression scores were twice as high for the control youths compared to the MTFC-A youths at the 2-year follow-up.

Other Noteworthy Foster Care Interventions

In addition to the eight interventions presented in Table 1, several interventions met most (but not all) of our inclusionary criteria; therefore, we note them here as promising programs that merit additional research. First, implemented a randomized waitlist-control version of the IY program. Although their intervention-versus-control effects were not significant, their results indicated significant pre-post reductions in problem behaviors for the intervention foster children (but not for the control children). Second, adapted elements of the MTFC-A model to supplement a statewide study of treatment foster care in the USA. They randomized at the agency level and augmented existing services with supervision/support of caregivers by the supervisory staff; proactive, teaching-oriented approaches to problem behaviors; preparation of the adolescent for adulthood; and treatment of previous trauma. Their results indicated that, compared to the control youths, the intervention youths showed significant improvements in symptoms, behaviors, and strengths at a 6 month follow-up (improvements sustained at 12-months for one of the outcomes scales but not the other two). Finally, an independent quasi-experimental replication of MTFC-A in England found that young offenders assigned to MTFC-A rather than to custody had significantly lower recidivism rates and were more likely to live with their families 1 year after entering the program, but long-term effects were not sustained (). Several other foster care interventions show promise internationally but have not yet been evaluated using randomized designs (e.g., ; ).

In addition, the results from at least two randomized controlled trials with foster children have indicated nonsignificant differences between treatment and control conditions across all of the key child outcomes examined (e.g., ; ), and other failures-to-replicate may exist in the unpublished literature. An analysis of the common intervention components that result in positive intervention outcomes across multiple studies could help further refine knowledge about the core intervention components that help to improve outcomes for foster children. In addition, all but two of the studies noted in Table 1 included children and families from the USA. Evaluations in new locales with teams independent of the intervention developers are needed to examine whether these interventions remain effective when implemented outside of the original setting and across countries with varied foster care practices. This has begun for MTFC-A and MTFC-P, but more wide-spread efforts are needed on the international level.

Limitations of Prior Research

Although the eight interventions identified in Table 1 offer promise for improving a range of outcomes for foster children, there are significant limitations in the research to date, and the results from the studies described above should be interpreted with caution, particularly given that the criteria used for selecting interventions to include in this review favored the reporting of positive intervention effects (rather than negative effects). Further, a common limitation is the lack of long-term follow-up data. Three of the inventions do not have published follow-up data at 12 months or beyond (ABC, modified IY, and KEEP), making it difficult to discern whether the observed effects would sustain beyond the intervention period. In one of these interventions (modified IY), the effects dissipated quite quickly: 3 months after the intervention ended, only the positive discipline effect remained; co-parenting effects were nonsignificant, and child behavior problems did not show a significant group difference at either assessment. Further undermining confidence in the sustainability of effects, the length of time between intervention termination and the follow-up assessment was often unclear from the published work. For example, the ABC samples had a wide child age range; therefore, the outcome assessment was delayed if the child was younger at the start of the study. In addition, some children in all of the studies experienced one or more placement changes prior to the outcome assessment. Together, these limitations make it difficult to ascertain whether the identified intervention effects sustain for most children well after the intervention has ended and the child is in a permanent placement setting with new caregivers. Long-term follow-up studies of the interventions included in this review are needed to better evaluate whether initial effects are maintained over time; only two of the interventions described above have published effects beyond 24-months (BEIP and MSS).

Another issue confounding the results from some of the studies noted above is that the data collection process was not fully blinded to study condition. For example, in the BEIP study, which relied in part on observational data, it would be readily apparent whether the child was assigned to remain in the orphanage or to be placed in foster care. Another limitation common to the results from several of these studies is the lack of baseline (pre-intervention) data on one or more of the outcome measures. For example, the two primary ABC intervention outcomes were attachment and cortisol reactivity, neither of which was assessed pre-intervention. Although random assignment should remove pre-intervention differences, chance differences can be present and might confound conclusions about postintervention outcomes. Also, the effects found across studies did not consistently generalize to other measures. For example, the FIAP intervention effects on child mental health were present for caregiver reports but not for youth reports, and delinquency outcomes in MTFC-A were present for girls using days in locked settings data but not using self-reported delinquency data. The effects were also sometimes specific to subpopulations (e.g., one gender or older children; FIAP; MTFC-A) or to one set of hypothesized outcome constructs but not to another set (e.g., MSS, BEIP, and FIAP). These inconsistent findings, combined with the generally small effect sizes (and the failure to report effect sizes in many published reports), suggest that findings might not be as robust as hoped and/or might be specific to only certain outcomes.

A final limitation worth noting regarding the interventions and results noted above is that there is significant variation in the number of published outcome studies derived from each intervention. Some interventions resulted in only a single outcome paper in which multiple outcomes were presented (modified IY), whereas other interventions resulted in three or more published outcome papers (BEIP; MTFC-P; MTFC-A). A related point is that evaluations of children’s behavioral and mental health outcomes have not been published in several of the early childhood interventions (ABC and MTFC-P), and brain and neurobiological outcomes have not been published in any of the middle-childhood or adolescent interventions. It is unclear whether such effects are not present (null findings) or whether they have not been examined.

In summary, we have presented the results from eight intervention programs that have been tested using randomized trial designs at the individual level and have been shown to be effective in improving one or more outcomes for foster children. The results from most of these studies have small to moderate effect sizes that typically decrease over time (MTFC-A and BEIP are two exceptions to this pattern, with more sustained effects and some evidence of large effect sizes). Overall, effective programs are attachment focused or have evolved from parenting interventions based on social-learning frameworks. The impact of these interventions might be mediated by the foster caregiver’s skill base, past experience, training, and supports (; ). These interventions offer great opportunity for more widespread implementation of effective services for foster children, ultimately improving their well-being and outcomes and reducing the intergenerational transmission of foster-care involvement; however, additional research with pre-specified outcome analyses could provide stronger evidence of generalizability to other populations and countries.

Implementation Challenges and Opportunities

Despite the availability of evidence-based interventions for foster families, the implementation of such interventions into routine services has been met with challenges. Because child welfare is a universal catchment system for all youth with documented maltreatment, the services population is quite diverse and covers a wide age range (birth through adolescence). Many foster families have multiple children in their care, and the original biological caregiving environment might continue to be less than optimal. These system conditions can pose inherent challenges to providing effective services. Child welfare systems are subject to pressures related to high caseloads, high staff turnover, and a continual influx of new cases, creating stressed, difficult work climates that interfere with the sustained use of effective interventions (). Nonetheless, system challenges also present opportunities to improve children’s well-being.

Issues related to racial and ethnic diversity

Ethnic minorities comprise a disproportionate number of foster children in some countries. Data from the NSCAW study indicate that the foster children in the USA are 45% African American, 31% White non-Hispanic, and 17% Hispanic (). Further, disparities persist in terms of services; African American children are less likely to receive mental health services than White children and are more likely to wait longer in foster care for permanent adoption (; ). In the UK, the Adoption Research Initiative suggests that some ethnic minority children wait longer than White children for permanence or do not achieve permanence at all. The UK government has urged adoption agencies to practice placement equality among all children regardless of whether or not the child shares the same racial or cultural background as their foster family. Government Ministers have stressed that, if an adoptive family cannot be identified that closely matches the child’s ethnicity and cultural heritage, then every effort should be made to find an alternative family. Similarly, in the USA, the Multiethnic Placement Act of 1994 and the Interethnic Adoption Provisions of 1996 aim to decrease the time that minority youths wait to be adopted, prevent discrimination in adoptive and foster placement decisions, and increase the number of foster and adoptive parents of ethnic minorities. Such policy-guided changes, along with ongoing cultural and diversity training, could ultimately help remedy the disparities in service access and well-being for minority foster children.

Screening methods to increase opportunities for effective service implementation

Screening tools have been developed to help distinguish foster youth who would most benefit from enhanced intervention services targeting behavioral and mental health problems. Such screening tools could help practitioners allocate intervention services, particularly when resource and staffing budgets are limited. One tool, the Ages and Stages Questionnaire: Social-Emotional (ASQ:SE): A parent completed, child monitoring system for social-emotional behaviors.

What are the challenges of adoption?

What is the biggest challenge in fostering?

What are the common challenges faced by a family which could be similar to the challenges faced by the country?

What are some challenges when working with families?