Listen to this
The goal: By 2030, zero children in institutional care in Europe.
(UNICEF Policy Report, Keeping Families Together in Europe, September
2024)
Over the past decade, there have been repeated calls in the peer-reviewed research literature on residential care for the elimination of any form of group care for children and youth. Such calls are reiterated in numerous agency’s reports, often with references to the research reports. For example, in a 2014 “consensus statement” of the American Orthopsychiatric Association (Ortho Consensus, Dozier et al., 2014, p. 219), solicited by the Annie E. Casey Foundation and the Youth Law Center, the authors state that “Group settings should not be used as living arrangements, because of their inherently detrimental effects on the healthy development of children, regardless of age.” Before the committee that ultimately developed the United States’ Family First Act (Lindell et al., 2020) that severely limits the use of federal funds for residential care, Kohomban (2015) testified that the “…evidence is not on their [supporters of residential care] side” while referring to the Ortho Consensus (although mistakenly attributing it to the American Psychological Association). Later, in a report for the Brookings Institute, Haskins (2017, pp. 3-4) too referred to the consensus statement, saying that “There is now almost universal agreement that group or institutional care should be considered an option of last resort.”
In the conclusion to a recent summary of the Bucharest Early Intervention Project’s (BEIP) longitudinal research, Wade et al. (2022, p. 162) saw as their “single most important legacy” the “urgent need to end institutional rearing and promote high-quality and stable family placements.” Similarly, a Lancet Group Commission on Institutionalisation and Deinstitutionalisation of Children (Lancet Commission, van IJzendoorn et al., 2020, p. 703) concluded: “Every effort should be made to minimise children’s exposure to institutional care.” In a UNICEF report, Cappa et al. (2022, p. 110) claim “The toll institutionalized care exacts on nearly all aspects of child development is astounding…” citing five references, including three from the BEIP, a narrative review emphasizing the efficacy of high-quality staff training, and a report of a meta-analysis showing no difference between children in institutional and noninstitutional care when the group care was evidence-based.
In short, UNICEF, the Annie E. Casey Foundation, Casey Family Programs, a Lancet Group Commission, the Board of Directors of the American Orthopsychiatric Association (now the Global Alliance for Behavioral Health and Social Justice), and other groups and academics have called for the end of any form of group care. These calls are almost always supported with citations to empirical research and child development theory like those above. In this article, I will argue that there is a significant gap between what research really has shown us about residential care, and any firm basis for calls to eliminate it.
Problems with the Research used to Call for Elimination
Why has there been a call for the elimination of all residential care? The conventional wisdom says it is because research unequivocally shows residential care to be harmful. But does it? In my view, there are important problems with the research used to demand the elimination of residential care that make such a strong conclusion untenable (for more detailed treatments of the research mentioned in this and the next section, see Henderson, 2024). Here I will consider five sorts of problems.
The ill-defined independent variable
What is being studied in studies of residential care? A pervasive problem with the research on residential care is the failure of researchers to specify exactly what they are studying. The default description of “institution” may refer to large barracks-style orphanages, custodial facilities with barbed-wire fences, bucolic ranches, children’s homes with campuses of cottages and recreation centers, small group homes, or various other facilities with or without institutional features. Rarely do researchers provide assessments of the quality of the care. As a result, attempts at reviewing the research on residential care have included studies of very different kinds of care that vary by both type and quality. For example, an influential study from the Lancet Commission (van IJzendoorn et al., 2020), using meta-analyses, found that children in “institutions” showed relative deficits in development compared to those in “families.”
However, their meta-analyses combined hundreds of studies from a period of nearly 70 years in institutions of very different types, in many different countries, and of care of unknown quality. There is no way to attribute effects to any specific form of care from these analyses. In the rare cases when the care studied is specified, the studies have been conducted in extremely stark environments with infants. Research that does not specify the type or quality of care provided should not lead to an inference that all residential care should be eliminated.
A focus on infants
Perhaps the single most influential study of residential care is the BEIP (Nelson et al., 2014; Wade et al., 2022). A substantial investment in research personnel allowed detailed study of infants and young children who had been surrendered by their parents to group care in a form of ideological group care experiment (popular reports of this study often refer to “Romanian orphans,” but almost all the children had at least one parent with whom they sometimes had contact). The group care provided was characterized by large child-to-caregiver ratios, frequent changes in caregivers, and stark physical environments devoid of stimulation. The BEIP included a comparison of the development of a randomly selected subset of children who went to live with foster parents trained by the researchers with those who remained in the stark institutions. Children in the experimental foster families clearly did better on a range of physical and psychological measures. A reasonable conclusion from the BEIP is that infants should not be reared in poor social and physical environments.
Most studies of residential care included in the Lancet meta-analyses described above were conducted with infants or children who entered care as infants. Much of the analysis provided in the consensus statement of the Ortho Consensus, too, was based on theorizing and research concerning infants and very young children. One of the most comprehensive reviews of the potential harm of residential care is provided in Sherr et al. (2017). Nearly all the studies included there are of children who were in residential care as infants and young children. How is the research on residential care for infants and very young children relevant to residential care for older children and youth? It is not. Regarding therapeutic care for special problems, most children are not identified with serious problems during infancy or early childhood. In most of the world’s child welfare systems, non-residential care in the form of kinship and family foster care is available. In most countries, that has been true for a half century or longer.
Nonequivalent comparison groups
The BEIP includes a comparison of institutionalized young children with those who entered an experimental foster care intervention. Otherwise, there are no randomized clinical trials of residential care compared to other forms of care, including, of course, being raised in a family of origin. Experimental studies would be unethical. The studies included in the Lancet Commission meta-analyses are all non-experimental. Intact groups of unknown equivalence are compared. Comparisons of children in residential care to other forms of care are not uninformative but must be interpreted with caution. In a comprehensive review of comparisons involving residential care, Lee et al. (2011) found that: (a) family foster care has slightly more positive outcomes (on a variety of measures such as school achievement and behavior ratings) than community-based group homes; (b) in studies with juvenile offenders, therapeutic foster care is more effective than group homes; and (c) high-quality children’s homes are as effective or more effective than therapeutic foster care or non-placement services. Regardless of their limited research designs, these studies do not support the elimination of all residential care.
Anecdotes
Calls for the elimination of residential care are often accompanied by powerful anecdotes about negative experiences in residential care (e.g., Fathallah & Sullivan, 2021; No place to grow up, 2015). The problem with anecdotes is that they can be obtained for or against any kind of care, residential, family foster, or with families of origin. They are essentially case studies with all the well-known problems of selection bias, history, maturation, etc. A widely circulated but un-peer-reviewed report of research sponsored by the Annie E. Casey Foundation, Casey Family Programs, and “Think of Us,” (Fathallah & Sullivan, 2021) illustrates the problem. Researchers obtained their sample of 78 with what they describe as an “open call,” a form of “…purposive sampling, a non-probability technique that selects subjects based on predetermined participant characteristics and the objective of the study” (p. 126). Participants underwent carefully designed and conducted interviews about their experiences. However, the research product is essentially a series of anecdotes from an unrepresentative sample. Any inferences from the interviews are also limited by the fact that two-thirds of those interviewed had four or more placements, nearly a third, ten or more, including in foster families. Anecdotes, even in numbers, can tell us little about residential care.
Overgeneralizations
All the problems with the major research used to call for the elimination of all residential care are related to the ubiquitous problem of over-generalization. Findings are generalized to all group care from one age group (most frequently with infants or young children), from one type of setting (e.g., custodial care; large institutions), from one quality of setting (usually extreme environments), from one historical era to another, or from one kind of child outcome to another. The technical term for the issue of generalization is “external validity.” Most studies used to justify the elimination of all residential care have been overgeneralized. They lack the external validity required for such strong conclusions.
Skewed developmental theory
This is not the forum for an extensive discussion of the theoretical basis for understanding the influences of residential care. Perhaps it is enough to point out that those who argue against group care frequently do so because of a reliance on theories about early experience and early attachment. The BEIP and Lancet Commission explicitly rely on them and much of the argument in the Ortho Consensus statement does too. The Ortho Consensus also includes arguments from parenting and peer contagion theories, although the use of peer contagion theory distorts what we know about the positive influences of peers, especially during adolescence. Given that most of the residents of group care are pre-adolescents and adolescents, there is a noticeable absence of theorizing in the negative literature about issues such as adolescent cognitive development and education (Ackerman, 2023; Crone, 2017; Tervo-Clemmens, 2023), career development and interests (Low et al., 2005; Renninger & Hidi,, 2019), identity development (Côté, 2009; Ferrer-Wieder & Kroger, 2020), peer influences (Laursen & Veenstra, 2023; Oberle et al., 2024), and self-competence (Harter, 2012). A good example of the appropriate use of child development theory for practice with older children and youth in residential care can be found in research on the CARE model of residential care (Holden, 2024). The CARE approach carefully synthesizes Anglin’s pain theory (Anglin, 2002), Vygotsky’s social-cognitive theory (Vygotsky, 1978), and Bronfenbrenner’s bioecological theory (Bronfenbrenner & Evans, 2000) in ways that apply to older children and adolescents rather than infants and preschool children.
Is there Research to Support the Use of Residential Care?
Yes, there is evidence to support the use of high-quality residential care. The nature of the evidence does not allow the “proof” of the effectiveness of residential care. However, there are features of the findings that suggest a role for residential care in the child welfare systems of the world.
Surveys of residential care alumni and residents
McKenzie (1997, 2003), himself a former resident of a children’s home (McKenzie, 1996), conducted extensive surveys of former residents of children’s homes of the mid-20th century in the United States. McKenzie received over 2,000 usable responses in two surveys. McKenzie found that respondents were overwhelmingly happy and successful. Almost all viewed their experiences in “orphanages” positively. When asked if they would have preferred to have been in a foster family, few said they would.
In a more recent survey of 450 children and youth currently in group care, Boel-Studt et al. (2023) found strikingly similar results. Resident responses about relationships with staff, staff facilitation of relations with families of origin, safety, educational and skill development, and preparation for life transitions were positive across domains, leading Boel-Studt et al. to conclude that their results “…counter longstanding assumptions about negative experiences in residential care and widespread generalizations that have driven policy in Western cultures” (p. 6).
Direct comparisons of care in low resource settings
Two extensive studies of group care in Asia and Africa provide solid, if not definitive, support for residential care. The Positive Outcomes for Orphans (POFO; Huynh et al., 2019; Whetten et al., 2009; 2014) and the Orphan & Separated Children’s Assessments Related to their Health & Well-being Project (OSCAR; Embleton et al., 2014) both show that children and youth in group care were performing as well on indicators of physical, emotional, and cognitive development when compared to those with a surviving parent, in kinship care, or in foster care. In the OSCAR study, children in group care were more likely than the comparison children to be meeting the United Nations rights requirements for basic nutritional, medical, and educational needs and less likely to be involved in harmful labor.
Quality vs site of care
A consistent finding in research on out-of-home care is that the quality of the child’s experience is more important than the site of the care (Boel-Studt & Tobia, 2016; Folman, 2009; Hermenau et al., 2017; Holmes et al., 2017; Huynh et al., 2019; Knorth et al., 2008; Neagu, 2021). For example, in a review of residential care programs and practices, Lee (2020, p. 308) concludes that group care is not only inherently worse than other forms of care and that “…group care may produce equivalent or better outcomes than family care” and that when residential care is evidence-based there are no residential-family care differences, suggesting that it is the quality of programming that is crucial.
Change during time in care
Simple comparisons of how children and youth are doing when they enter residential care with how they are doing when they leave cannot conclusively show that residential care is the cause of the change. However, if group care is harmful, yet residents consistently leave care doing better than when they entered, an inference that residential care is doing harm is unsupportable. Real positive influences appear to be more likely. In a meta-analysis of studies of pre-post change in children and youth who have been in residential care Knorth et al. (2008) provide convincing evidence that they improve in their psychological functioning. Similar patterns of change have been reported by James (2011) and Nixon and Henderson (2022).
Do we Really Need Residential Care?
Those who wish to eliminate all residential care consider it harmful or ineffective and therefore unneeded. The research reviewed here suggests it is unlikely that such care is harmful or ineffective. However, is it possible that alternative forms of care such as family reunification, adoption, kinship care, therapeutic foster care, and family foster care can always be used instead of residential care? Two conceptual issues apply to the question: stability and resources.
Stability is the degree to which a child can find a stable home in a particular form of care. The importance of stability during development cannot be overestimated. Changes in family structure (e.g., death of a parent, divorce) are associated with problems in a wide variety of measures of cognitive development, educational achievement, and social-emotional competence. Many children and youth are in care because of instability in their families. When family reunification works, it is a preferred option. However, it often does not work, and many children and youth go back to their families for a short time, then rebound to care. Kinship care is also a good option. However, problems often run in families.
Those who foresee the end of all residential care see family foster care as a better form of care, presumably because it provides a family. Research shows, however, that family foster care itself can be a major source of instability. Children are often subjected to multiple changes in their homes, schools, neighborhoods, and peer groups. Many children face the end of a foster care placement prematurely. In one U.S. state, Cross et al. (2013) found that 20-50% children in foster care experienced at least three placements in the first year of care. In another state, Connell et al. (2006) reported an average period of only 3.9 months in a first placement. In some cultures, there is no tradition of foster care outside extended families. Even in resource-rich countries, finding stable foster care is difficult, especially with families of the same linguistic, religious, and ethnic backgrounds (Chowdhury, 2021). Estimates indicate U.S. foster parents average only a year in the system (Williams et al., 2023).
Given what we know about the effects of general changes in family structure, it comes as no surprise that instability in care placements is associated with many developmental issues. They include insecure attachments (Quirogal & Hamilton-Giachritsis, 2016), feelings of unwantedness or discrimination (Bombach et al., 2018; Unrau et al., 2008), sibling separation (Rock et al. 2015), failed family reunification (Wulczyn et al., 2020), hampered cognitive, physical, and social-emotional development (Asif et al., 2024; Konijn et al., 2019), educational achievement (Ellis & Johnson, 2024), and behavior problems (Rock et al., 2015). The latter is particularly important because increases in behavior problems are associated with subsequent failed placements.
Those who call for the end of all residential care have failed to see how good residential care can provide children and youth with stability. Good residential care can provide consistent developmentally appropriate relationships with adults who care about them (Li & Julian, 2012), a relatively stable, supportive peer group (Laursen & Veenstra, 2023), a respite from constant change in their schooling (Font & Palmer, 2024), and perhaps most importantly, a sense of belongingness in a stable neighborhood with its playgrounds, ballfields, shops and houses of worship. Stability for a time can make subsequent fostering or adoption more likely and more successful. Children and youth often do not need yet another family. They need a stable home.
In addition to stability, children and youth need resources. Some of those resources are for the children and youth themselves. Those with complex needs not readily available to families need intensive interventions (Whittaker et al., 2016). Some need therapy plus additional learning through living treatment for the “other 23 hours” (Trieschman et al., 1976). Residents who have experienced multiple placements are of special need of intense educational interventions (Del Quest et al., 2012; Huefner, 2018; Ringle et al., 2010). As indicated regarding stability, residential care can provide a strong peer group. It is unfortunate that many critics of residential care have demonized the peer group in a misapplication of peer contagion theory (Lee & Thompson, 2009). Under appropriate supervision, the peer groups in residential care can have an important role in positive development (Laursen & Veenstra, 2023). Residential care centers can also provide resources to residents’ families of origin and/or foster families. Resources that are often associated with residential care facilities include adoption services, career development, and emergency foster care. Residential site staff can help in family reunification, foster family recruitment, training, and support, and parent training.
Conclusion
The movement to eliminate all residential care is puzzling to anyone who has been involved in high-quality care. The psychology and sociology of children and families have never been more complex. Children and youth have never faced more challenges to their healthy development. We need more options in dealing with the complexities, not fewer (Riley, 2021). Most children and youth do not need to be in residential care for years on end. But for those who need stability and special resources, residential care can provide an oasis in their journeys. And those who enter care late need transitional resources.
The view that I have called the conventional wisdom, that residential care is harmful and that we know that because research says so, has been harmful to children, families, and communities. It has been harmful to those who provide and support good residential care. There are rich, powerful institutions behind the goal of eliminating all residential care. They may be successful. If they are, however, it will be despite the available research evidence, not because of it.
References
See PDF version for a full reference listing.
