Abstract
Understanding the impact that early psychosocial neglect has on the course of human development has implications for the millions of children around the world who are living in contexts of adversity. In the United States, approximately 76% of cases reported to child protective services involve neglect; worldwide, there are more than 150 million orphaned or abandoned children, including 10.5 million orphaned because of COVID-19. In much of the world, children without primary caregivers are reared in institutional settings. We review two decades of research based on the only randomized controlled trial of foster care as an alternative to institutional care. We report that children randomly assigned to continued care as usual (institutional care) suffer from persistent deficits in social, cognitive, and emotional development and show evidence of disruptions in brain development. By contrast, children randomly assigned to foster care show improvements in most domains of functioning, although the degree of recovery is in part a function of how old they were when placed into foster care and the stability of that placement. These findings have important implications for understanding critical periods in human development as well as elucidate the power of the psychosocial environment in shaping multiple domains of human development.
Although genes contribute to the basic blueprint of early brain development (particularly during the prenatal period), the course of human development is greatly influenced by experiences that transpire over the early years of life, when the pace of brain development is greatest. Indeed, by some estimates, the brain generates more than 1 million synapses per minute from the late prenatal period through the 2nd to 3rd year of life (see Anusuya & Katti, 2010; Bourgeois, 1999; Fox et al., 1999). Because the brain is constructed in part in response to an individual’s experience, the experiences children have are enormously impactful on the course of development. As J. McVicker Hunt noted, “Experience cuts both ways.” 1 Thus, children’s exposure to positive experiences biases development in a healthy direction, whereas children’s exposure to adverse experiences biases the system in an unhealthy direction. An analogy here might be that experience provides the foundation and speed upon which brains are built. Not surprisingly, then, psychosocial neglect can lead to disastrous consequences, for the simple reason that the brain lacks the “expected” experiences during critical periods of development to build that foundation and to wire adaptively. Given that 76% of the cases reported to child protective services involve neglect (U.S. Department of Health and Human Services, 2021), it is imperative that we understand how neglect shapes brain development and changes the course of human development.
For many decades, neuroscientists have studied the effects of experience—or the lack of experience—using animal models. Classic work in the visual and auditory systems, for example, demonstrated that the lack of visual or auditory input during critical periods of development led to impairments in vision and audition (e.g., Chen & Yuan, 2015; Hubel & Wiesel, 1962). Similarly, work with nonhuman primates demonstrated that infant monkeys deprived of maternal care suffered long-term impairments in their social-emotional functioning (Harlow & Suomi, 1971). Nevertheless, studying the effects of neglect on human development has been more difficult, as neglect is often embedded in a host of other adverse experiences, such as abuse, poverty, or parental substance use, making it difficult to disentangle neglect from other adverse events. Arguably, children raised in deprived institutional settings represents the purest example of psychosocial neglect, because they are fed, clothed, and sheltered in a setting that is physically safe, albeit impersonal and insensitive (King et al., 2022). The absence of nurturing care is the primary form of deprivation. It is in this context that the Bucharest Early Intervention Project (BEIP) was launched more than 20 years ago.
The BEIP is a randomized controlled trial (RCT) of foster care as an alternative to institutional care. Its RCT design (see Fig. 1) facilitates the ability to draw causal inferences about the effects of institutional rearing on child and brain development. Prior to this study, the bulk of the literature on the effects of institutionalization focused on children abandoned or orphaned at birth, placed in institutional settings, and then adopted (generally by middle-class families). These studies were correlational in nature and are subject to sample biases that do not exist in an RCT design.
Bucharest Early Intervention Project Study Design: Drawing from six different institutions in Bucharest, more than 180 children underwent a pediatric and neurological exam. We excluded children with frank genetic or neurological disorders. The remaining 136 children, comprised of 68 children from the care-as-usual group (CAUG) and 68 children from the foster-care group (FCG), underwent an extensive baseline assessment, as did 72 never-institutionalized children (NIG) we recruited from the community. Following this assessment, half of the institutionalized children were randomly assigned to continued institutional care and half to a high-quality foster-care program the study team created and financed. Average age at randomization was ~22 months. Children in all three groups were then followed up at 30, 42, and 54 months and again at 8, 12, and 16 years; a 22-year follow-up is in progress.
In this article, we first focus our attention on the cultural forces that led to more than 170,000 children being raised in state-run institutions in Romania. We then describe the experimental design adopted in the BEIP and conclude with a summary of findings to date.
History of Child Abandonment in Romania and the Rise of Institutional Care
Nicolae Ceausescu, the Communist leader of Romania from 1965 to 1989, believed that he could increase Romania’s stature by increasing the population to support massive industrial expansion. To accomplish rapid population growth, he implemented a series of decrees that included banning contraception and abortion and imposing a tax on families who had fewer than five children. This resulted in an increase in the number of children whose parents were unable to care for them, which in turn led to an enormous number of abandoned children. Ceausescu then constructed dozens of state-run institutions to warehouse these children; by January 1990, shortly after Ceausescu was deposed and executed, there were more than 170,000 children living in institutions, many in unspeakably squalid conditions (Kast & Rosapepe, 2014). These institutions were generally characterized as having a small number of caregivers take responsibility for many children, coupled with profound neglect, as illustrated in Figure 2.
An illustration of a typical Romanian orphanage in the 1980s (courtesy of Mike Carroll).
The Bucharest Early Intervention Project
The BEIP was an outgrowth of the MacArthur Foundation Research Network on Early Experience and Brain Development (https://www.macfound.org/networks/research-network-on-early-experience-brain-develop) and took place in Romania at the invitation of a member of the Romanian government. After establishing a laboratory that was physically located at St. Catherine’s, the largest of the six institutions for young children in Bucharest (in 1989, St. Catherine’s housed ~850 children less than 3 years old), a pediatric team conducted exams on more than 180 children under the age of 3. Of these, 136 met inclusion criteria (e.g., no frank neurological or genetic problems, no fetal alcohol syndrome, etc.). An additional 72 never-institutionalized children were recruited from public pediatric clinics as a comparison sample. After an extensive baseline assessment while children were living in the institutions, these 136 children were randomly assigned either to a high-quality foster care program that our team created, financed, and supported or to continued care as usual (continued institutional care). A few components of our foster care program included the following: (a) All foster parents were licensed by the government, and they received a monthly salary comparable to the per capita income in Romania at the time (by Romanian law foster parents are salaried employees); (b) BEIP social workers visited the families every 10 days initially and met weekly with the BEIP team in the United States (via Skype, by phone, or in person); (c) we provided all material support (e.g., toys, diapers); and (d) a pediatrician was on call 24 hours a day. Children were initially followed up at 30, 42, and 54 months, when the trial ended. Children were seen again at 8, 12, and 16 years; a 22-year follow-up is in progress. Figure 3 illustrates the domains we assessed.
Domains of assessment.
Conceptual framework for the project
Based on our knowledge of Greenough’s concept of experience-expectant and experience-dependent brain development (the former refers to elements of the environment likely to be experienced by all members of the species, whereas the latter refers to elements of the environment unique to individual members of the species; see Greenough et al., 1987), we assumed that for typical neural circuitry to form, the brain requires patterned, contingent input, and the lack of such input could lead to underspecification and miswiring of circuits. For children growing up in neglectful environments, contingent input is generally lacking; thus, we might expect errors in brain development among such children, which would be manifest in brain structure and functioning and therefore in behavior. Moreover, the timing of when the neglect occurs should prove important; if certain experiences fail to occur during critical periods, the consequences could be worse than if the neglect occurs at a later age.
Ethical considerations
We and others have written extensively about the ethics involved in conducting this research (Miller, 2009; Nelson et al., 2014; Zeanah et al., 2012). Because the population was exceptionally vulnerable, we ensured oversight from the institutional review boards at the three U.S. universities of the principal investigators and at Bucharest University. Some may object to adoption of an RCT design, but this was critical to our ability to draw causal inference about the efficacy of the foster-care intervention compared with the institutional-care intervention and to inform a policy debate in Romania about which form of care was preferable. Throughout, all decisions about children’s placement were made by Romanian authorities—no children were retained in institutions because of study participation. In fact, at the trial’s conclusion, more than half of the children randomly assigned to care as usual (institutional care) were living in families, either by adoption, by reunification with their biological families, or by placement into newly available government foster care. Finally, all foster families were licensed by local authorities and monitored carefully by project social workers (for details, see Nelson et al., 2014).
Summary of findings
Since the project was conceived and implemented, the findings through age 16 have been extensively reported in both the scientific and lay literatures. Rather than summarize the findings to date across multiple domains, we instead focus on three broad questions: First, was the intervention efficacious, and if so, in what domains? Second, in those domains where the intervention was efficacious, were the effects influenced by critical period timing—that is, the age at which children were removed from institutions and placed into foster care? Finally, did disruptions in care (lack of stability in placement) impact outcomes?
Efficacy of the Intervention
With few exceptions (summarized later), the intervention was largely successful in improving the developmental outcomes of children. For example, at the behavioral level, children randomly assigned to foster care showed improvements in IQ; indeed, through age 18 there was a consistent 7- to 9-point advantage in full-scale IQ in foster care compared with care as usual (see Humphreys et al., 2022). There were also improvements in language (Humphreys et al., 2020; Windsor et al., 2011), attachment security (Smyke et al., 2010), and peer relationships (Almas et al., 2015). In terms of physical development, foster care led to improvements in length and height and, if placed before 1 year of age, in head circumference (Johnson et al., 2010). Finally, in terms of brain activity, children randomly assigned to foster care showed increased electroencephalogram (EEG) alpha power at rest (Debnath et al., 2020; Marshall et al., 2004, 2008; Vanderwert et al., 2010) and modest increases in white-matter volume (Sheridan et al., 2012). In addition, by age 16, a number of regions of the prefrontal cortex were thinner in children in foster care compared with children remaining in care as usual, a pattern consistent with the normative typical thinning the cortex undergoes during adolescence (Sheridan et al., 2022), likely the effect of synaptic pruning. Finally, in terms of psychopathology, psychiatric disorders were consistently less common in children randomized to foster care (Humphreys et al., 2020; Zeanah et al., 2009).
There were, however, a few exceptions to these positive intervention effects. We observed little improvement in many executive functions through age 16 (Wade et al., 2019). Second, through age 16, we continued to observe a very high rate of attention deficit hyperactivity disorder in both the care as usual and foster care groups (approximately 20%; see Humphreys et al., 2020). Third, children in both the foster care and care-as-usual groups showed reduced gray-matter volume that did not increase with placement into foster care (Sheridan et al., 2012). We have puzzled over the lack of intervention effects in these domains, although this general pattern is consistent with the other major study of Romanian orphans adopted into families, the English Romanian Adoptees Study (see Rutter et al., 2007). One possible explanation is that the average age of placement—22 months—may have exceeded the critical period for recovering from early neglect. Another is that although our RCT design helped us avoid sample bias postrandomization, we have incomplete data regarding the prenatal care or prenatal drug or alcohol exposure that may affect the brain areas associated with executive functions. Be that as it may, we remain keenly focused on whether we will observe some recovery in these domains as children make the transition to adulthood (our 22-year follow-up).
Critical Period Timing
Because children placed in foster care ranged in age from 7 to 33 months, we were able to assess whether earlier placement was associated with improved outcomes. More favorable development for earlier-placed children would be consistent with the concept of critical periods for brain and behavioral development.
There were many domains that seemed affected by timing of placement. We note only a few examples. For EEG power (a coarse index of brain development), children placed prior to 24 months of age were significantly more likely to have higher alpha power at 8, 12, and 16 years; indeed, alpha power for children randomized to foster care before 24 months was indistinguishable from that for children who had never been institutionalized (Vanderwert et al., 2010). With regard to IQ, when children were 54 months of age, when the formal trial ended (and BEIP foster care was turned over to the government), those who were placed into foster care before 24 months had significantly higher IQs than those who were placed at 24 months or later (Nelson et al., 2007). Language also seemed to be influenced by age of placement (Windsor et al., 2011, 2013), as was security of attachment (Smyke et al., 2010).
Collectively, children placed earlier in their development had more favorable outcomes than children placed later in their development, although this varied somewhat by domain, given that different domains are influenced by different critical periods. But the overall main effect of foster-care timing suggests that there should be an urgency about removing children from adverse caregiving environments and placing them in more favorable environments. These findings are consistent with other studies supporting urgency about getting children into favorable caregiving environments (Zeanah et al., 2011).
Stability of Placement
It has been well established that children under the care of child protective services who experience multiple foster-care placements have worse developmental outcomes than those who experience stable placements. In a series of recent reports from BEIP, we examined the issue of stability of placement in foster care. We did this by examining children in foster care who had stable placements versus those who had one or more disruptions in placement. We consistently observed that children in stable foster placements did better than those with disrupted care. For example, foster-care children in stable care exhibited fewer symptoms of psychopathology (Humphreys et al., 2015) and had higher IQs (Humphreys et al., 2022). But the question is whether stability reduces psychopathology or psychopathology leads to placement disruptions. To address this as best we could, we examined psychopathology, IQ, and growth parameters at 54 months, when the trial was concluded. There were no differences at 54 months in those children who would and would not subsequently remain in their original placements.
Conclusions
Sensitive, consistent caregiving that is responsive to a child’s needs is one of the most powerful interventions that is available for children who are orphaned, abandoned, and maltreated, for the simple reason that such care is individualized and has the best interests of the child in mind. Over the course of 20 years, we have consistently demonstrated that even when a child’s physical needs are met, psychosocial neglect is deleterious to brain and behavioral development. Much like exposure to toxins, the higher the dose of toxin, and the more prolonged the exposure, the more problematic the effects. Similarly, the longer a low level of quality care is provided, the more harmful it is to children’s development. The sooner children can be placed in adequate or better caregiving environments, the more likely they are to recover, and the fuller their recovery is likely to be.
Recommended Reading
Humphreys, K. L., Gleason, M. M., Drury, S. S., Miron, D., Nelson, C. A., Fox, N. A., & Zeanah, C. H. (2015). (See References). A follow-up of a randomized controlled trial that found children in foster care had significantly fewer externalizing symptoms and disorders, particularly true of those who remained in stable placements.
Humphreys, K. L., King, L. S., Guyon-Harris, K. L., Sheridan, M. A., McLaughlin, K. A., Radulescu, A., Nelson, C. A., Fox, N. A., & Zeanah, C. H. (2022). (See References). A two-sample study that found children placed in high-quality foster care had significantly higher IQ scores than those randomized to care as usual, in part, by higher-quality caregiving and attachment security in childhood.
King, L. S., Guyon-Harris, K. L., Valadez, E. A., Radulescu, A., Fox, N. A., Nelson, C. A., Zeanah, C. H., & Humphreys, K. L. (2022). (See References). A comprehensive multilevel analysis discussing how children assigned to foster care had better cognitive and physical outcomes and less severe psychopathology than did those who received care as usual, an effect that was stable from childhood through adolescence.
Nelson, C. A., Fox, N. A., & Zeanah, C. H. (2014). (See References). A book discussing deprivation, brain development, and the struggle for recovery of abandoned children in Romania.
Nelson, C. A., Zeanah, C. H., Fox, N. A., Marshall, P. J., Smyke, A. T., & Guthrie, D. (2007). (See References). An article discussing a critical period for recovery of IQ following early psychosocial deprivation in young children.
Footnotes
Acknowledgements
The writing of this article was made possible by the John D. and Catherine T. MacArthur Foundation, the Binder Family Foundation, and the National Institute of Mental Health MH091363. The authors wish to thank the dedicated members of our team in Bucharest, Julie Staples-Watson for program development and administrative help, and the many students and postdocs who have worked on this project for the past two decades. We also thank Sheena Odabashian for editorial assistance.
Transparency
Action Editor: Robert L. Goldstone
Editor: Robert L. Goldstone
