A child stands next to a bullet shattered window during the 1989 anti-communist revolution in Timisoara, December 23, 1989.
A child stands next to a bullet shattered window during the 1989 anti-communist revolution in Timisoara, December 23, 1989.
Petar Kujundzic / Courtesy Reuters

The Bucharest Early Intervention Project (BEIP), the first-ever randomized trial of foster care for young children with a history of institutionalization, began in 2000. This was about ten years after the fall of Nicolae Ceaușescu, general secretary of the Romanian communist party. During his tenure, Ceaușescu had been convinced that the way to build a powerful economy was to increase the number of available workers for production. To that end, he instituted a number of pronatalist policies designed to increase the population of the country. These included banning abortion, outlawing contraception, and imposing a tax on families with fewer than five children.

As a result of these policies, there was a dramatic increase in the birth rate. But many poor families were forced to have more children than they could afford. As an alternative to raising children in families, the government encouraged families to place children they could not afford to care for in large, state-run institutions. Faced with widespread poverty and limited economic opportunities, many families abandoned their children at birth or soon thereafter. As a result, over the course of several decades, abandonment became an acceptable choice for tens of thousands of families, creating one of the largest per capita systems of child-rearing institutions in history. By December of 1989, when Ceaușescu was overthrown and executed, there were more than 170,000 children living in various state-run institutions in Romania.

For more than ten years after the Romanian Revolution, the rate of child abandonment remained as high as it was during the Ceaușescu era. As a result, large state-run institutions continued to operate until early in the twenty-first century, when pressure from the European Union, combined with the release of initial findings from the BEIP, led to two major changes. The first was legislation that forbade the institutionalization of children under two years of age (unless they were severely handicapped) and the second was systematic efforts to close many of these institutions in favor of either family reunification, government foster care, or placement in smaller, more family-like group centers.

To be sure, Romania’s system of institutional care for orphaned or abandoned children was not unique nor a new phenomenon. Historically, institutions for abandoned infants appeared in Europe in the Middle Ages, but they came into prominence in the nineteenth century in Western Europe in response to urbanization, war, and diseases. Today, institutional rearing is common in many parts of the world, including Asia, Central and South America, the Middle East, and, increasingly, in Africa. Orphanages were common in the United States through the first half of the twentieth century. And, with armed conflicts around the globe and disease epidemics in Africa (first HIV, and now Ebola), millions of children are being left without families.

And that is why a close look at Ceaușescu's legacy in Romania is warranted. Young people who spent their childhoods in institutions are now left to fend for themselves, often homeless and on drugs. The evidence that placement early in life into institutions has a negative impact on brain and behavioral development is now overwhelmingly clear. Policymakers in other countries are faced with a similar challenge: Institutional rearing is shown to be harmful, but few countries have enough resources to develop alternative family placements. Caring for orphaned, abandoned, and maltreated children remains a global challenge.


The network of childcare institutions in Romania was overseen by different arms of the government. In general, state-appointed physicians directed these facilities. Those born in the countryside were often brought directly to the institution. Infants born at maternity hospitals in cities were abandoned there. After spending up to several months in a maternity hospital, these infants were transferred to institutions for young children known as leagane (cradles), where they were cared for in groups of 12 to 15 by caregivers working rotating shifts. The caregivers had little education or experience in childcare. Given the demands of tending to large numbers of very young and needy children, many caregivers either remained detached, with limited emotional investment in the children they cared for, or alternatively, they selected one or two favorites and spent a disproportionate ammount of time with them. And the state-employed physicians who directed the institutions were not well trained in child development, and although they were pleased to have jobs, were often neither equipped nor motivated to foster healthy child development.

Infants in a typical Romanian institution in the late 1990s/early 2000s.

Around the age of three, children were evaluated through an ideological prism called “defectology.” Based on a Soviet idea of child development, the goal was to identify children early in life who, officials believed, were unlikely to become productive citizens. By identifying defects early, the thinking went, the state could cheaply warehouse defective children—on the grounds that they were unlikely to recover—and spend its resources on others who had more potential.

Thus, during the Ceaușescu era, institutions conducted neuropsychiatric examinations that classified children as either “normal,” having a “curable deficiency,” or being “irrecuperable.” If normal, the child was sent to a casi di copi (children’s home) that was roughly equivalent to a group home. There, the children had some private space and attended public school, but lacked warm and responsive caregiving. Those with curable deficiencies were sent to gradinite (special education kindergarten) for preschool. If the child had an identifiable handicap and was designated irrecuperable—which sometimes happened even with treatable conditions—the child was sent to set of institutions that included camin spitale and institutule neuropsihiatrice (both long-term facilities for the handicapped). These were dreadful places where children’s heads were often shaved, they were dressed in tattered clothing, and their activities were limited. Hygiene was sub-standard too, and in at least some institutions, there were frequent occurrences of physical and sexual abuse (by both peers and staff). Profoundly handicapped children were often left to lie in their own waste and there were instances in which children were tied to their cribs to prevent them from injuring themselves. In the 1980s, many children were administered plasma for vaguely defined “health reasons.” Some of the plasma became infected with HIV, leading to widespread infections among the children and a mortality rate approaching 50 percent in some institutional settings. Indeed, it is believed that, at one point, Romania had one of the highest rate of pediatric HIV infection in the world.

All of these horrors came to light with the collapse of the Ceaușescu regime. Reports on the tragic conditions of institutionalized children were picked up by the Western media and broadcast around the world. With the best of intentions, thousands of families from Western Europe and North America adopted many of these children, assuming that, if they were provided with good homes, they would recover to lead healthy, normal lives. However, some of these families quickly realized that these children were suffering from serious developmental delays and had problems forming relationships with others, profoundly diminished intellectual functioning, and various psychiatric disorders. Moreover, these problems did not always disappear with time in their new home. Children demonstrating lasting issues despite being removed from the institutions and placed into enhanced caregiving environments raised the possibility that there was a sensitive period during which exposure to profound neglect might have significant and long-lasting consequences for brain development. Further, it added urgency to removing children from institutions and placing them into families as soon as possible in order to limit long-term harm.


The notion of sensitive developmental periods was brought to the American public’s attention in 1997, when U.S. President Bill Clinton and First Lady Hillary Clinton hosted a conference on early brain development. Shortly after the conference, the John D. and Catherine T. MacArthur Foundation began to fund a research network focused on early experience and brain development. The mandate of this group (which included the authors of this paper) was to understand how early experience “gets under the skin” and influences the course of brain development. Eventually, the group conducted a range of studies on the effects of experiences on brain development, but one study in particular became focused specifically on children experiencing profound neglect from early life.

After an initial visit to Romania in 1998, we received an invitation from the government to design a study to test the efficacy of foster or family care as an alternative to institutional care. The work began by identifying a cohort of 136 very young children considered to be relatively healthy and not suffering from any obvious genetic or neurological disorder. These children were extensively assessed while living in institutions, with particular attention paid to their intellectual, social, emotional, and brain development. After this baseline assessment was performed, half of these children were randomly assigned to a high-quality foster care program that our group created, maintained, and financed. The other half were randomly assigned to care-as-usual—continued institutional care. We also recruited a comparison group of children who had never spent time in an institution and lived with their biological families in the greater Bucharest community. These three groups of children have been extensively studied for the past 14 years.

The ethics of randomly assigning children to foster care and care-as-usual were carefully and extensively discussed by members of the research network. The first ethical challenge was whether random assignment to the two arms of the study was justifiable. In the end, we believed that it was. We were comparing two interventions for abandoned children in a country that had historically employed one approach (institutional rearing) and which was considering policy changes to support an alternative approach (foster care). At the time, many child protection professionals in Romania were deeply skeptical about foster care because some of them believed that adults who would take children unrelated to them into their homes could not be trusted. In addition, there were economic incentives within national and local government agencies that favored maintaining the status quo. Because random assignment provides the most incontrovertible proof of intervention efficacy, because there had never been a randomized controlled trial of foster care versus institutional care anywhere, because Romania invited the study to inform the policy question, and because we believed that results of the study would be useful to other countries using institutions to care for orphaned, abandoned, and maltreated children, we concluded that our design was ethically justified.

To ensure that the vulnerable children we studied were fully protected, we agreed on several other principles. First, we did not interfere with the possible placement of children in either group—child protection authorities determined if children were adopted, returned to biological parents, or placed in government foster care (which later became available). All of these placements were available to children in both groups so that, when our support for the foster families ended once the children reached 54 months of age, more than half of the care-as-usual group was no longer living in institutions. Second, we used procedures and methods that posed no more than minimal risk and are routinely used throughout the world with infants and toddlers being raised in their own families. Third, the study was and continues to be reviewed by multiple oversight agencies, including the scientific ethics review boards of the investigators’ universities, by an ethics committee at the University of Bucharest, and by a data safety monitoring board in Romania.

We collected baseline data before the children were placed with foster parents or assigned to care-as-usual. This allowed us to compare institutionalized children to typical Romanian children living in families. The differences were large on virtually every measure we examined: growth, cognition and language, brain functioning, social relatedness, and competence. For example, whereas the children being reared in families had average IQs of 100, those reared in institutions had IQs clustered in the 60s and 70s (more than two standard deviations below the mean).

The children were randomly assigned to foster care or care-as-usual at, on average, 22 months of age (children ranged from 7 to 33 months). We performed follow-up examinations of the children at 30, 42, and 54 months, and then again at 8 and 12 years (we have just embarked on a follow up exam for when the children turn 16). We found that those placed in foster care had higher IQs than children assigned to care-as-usual; over the first 4.5 years of the project, we also observed a sensitive period for IQ—children placed in foster care before 24 months had markedly higher IQs at 54 months than those placed after 24 months. By ages 8 and 12, we continued to observe an intervention effect in IQ for children placed into foster care homes, but the timing of the experience, the sensitive period found earlier, was no longer there. This suggests that for a complex measure such as IQ, subsequent life experiences may override the importance of the early years. On the other hand, we found a sensitive period for EEG activity, a proxy for brain development, which was evident when children were 8 and 12 years of age. Children removed from institutions before 24 months of age had EEG activity that was similar to typically developing community children, while those randomized to care-as-usual had delayed patterns of brain activity.

The top and bottom halves display the effects of the age of placement on IQ and EEG, respectively. Regarding EEG, each of the four displays is a view of the head from the top. The colors represent the amount of the brain’s electrical activity. It is clear from the two images to the far left that the brain activity of children in the care-as-usual group and those placed in foster care after two years is identical; similarly, the two images on the far right indicate that the amount of brain activity is identical among children who had never been institutionalized and those placed in foster care before age two.

A similar pattern was shown for attachment (an assessment of the quality of the relationship a child has with his or her caregiver). Not only was attachment quality higher among children placed in foster care, it improved the most among children placed in foster care before 24 months of age. Furthermore, we determined that the quality of attachment was crucial in determining subsequent psychiatric difficulties. Children who were able to form healthy attachments were significantly protected against developing later adverse symptoms, particularly anxiety.

It is worth noting that there were some areas in which we did not see better outcomes among children in our foster care intervention. For example, at 54 months of age, the prevalence of attention deficit hyperactivity disorder among children who had ever been institutionalized was seven-fold higher than among those who had never been institutionalized. There were no differences, however, in the rates among foster care children and care-as-usual children. In addition, children placed in foster care differed little from those assigned to care-as-usual in the broad domain of executive functions—higher cognitive functions like cognitive flexibility and impulse control at 8 years of age, although some differences emerged at age 12.

As these children grew older, we sought to determine whether differences in brain development could explain the behavioral differences we had observed. When the children were 8–10 years of age, they underwent a magnetic resonance imaging (MRI) scan. From these data, we found that children who had ever been institutionalized had a dramatic reduction in the size of their brain and in the integrity of the nerve fibers that connect different regions of the brain. The children in foster care showed fewer differences in the integrity of these fibers compared to never-institutionalized children. Finally, some of the differences in brain anatomy explained the EEG differences we initially found, as well as some of the problem behaviors we observed, such as hyperactivity and deficits in attention. Thus, children who experienced greater thinning of the cortex (that is, they had lost more neurons and connections among neurons) were more likely to have attentional impairment and hyperactivity (symptoms of ADHD).


The findings from BEIP are important in several respects. First, they suggest that children who experience profound early neglect exhibit a wide range of developmental delays and disorders, which are reflected in structural and functional abnormalities in their brains. Second, in many of the areas we studied (EEG, attachment, language), we observed a sharp inflection point—a sensitive period—such that children placed in foster care before a certain age (in most cases, 15–24 months) had much better outcomes than those placed in foster care after that age. Third, there were a few areas in which foster care did not come with any observable benefits. We attribute at least some of this to the relatively late placement into foster care many of our children experienced; had we been able to place them in the first few months of their lives, we might have observed more substantial developmental gains.

Romania’s story has unique features, but our findings have important implications for millions of children currently living in institutions around the world. For example, children who experience a dramatic reduction in IQ and an increase in emotional and behavioral problems due to being reared in institutions from their earliest years, are less likely to complete high school and develop skilled positions in the workforce, achieve economic independence, enjoy lasting intimate relationships with others, or contribute meaningfully to society.

There are currently millions of children living in institutions around the globe. Many are there because of poverty, war, and disease. A public health problem of this magnitude requires a thoughtful and comprehensive response. Indeed, UNICEF has for many years attempted to address this issue. But there are no easy solutions. For example, in areas with widespread poverty or war-torn regions, there are few adults to foster orphans; in countries where a moratorium has been placed on international adoption (for example, Russia), orphaned or abandoned children may often be relegated to institutional care, since in these same countries, foster care is limited.

Even in the United States, where nearly half a million children are in foster care, the struggle to provide high-quality, trauma-informed, and developmentally appropriate foster care is an aspirational goal, despite increasing evidence that this can substantially reduce the harmful effects of adverse experiences.

This study—along with a number of others like it—has raised the alarm about the deleterious effects of children experiencing early adversity. If policymakers do not heed the warning, they risk looking the other way as millions of children around the world see their life-long potential erode.

You are reading a free article.

Subscribe to Foreign Affairs to get unlimited access.

  • Paywall-free reading of new articles and a century of archives
  • Unlock access to iOS/Android apps to save editions for offline reading
  • Six issues a year in print, online, and audio editions
Subscribe Now
  • CHARLES A. NELSON is Professor of Pediatrics and Neuroscience at Harvard Medical School and the Richard David Scott Professor of Pediatric Developmental Medicine Research at Boston Children’s Hospital.
  • NATHAN A. FOX is Distinguished University Professor in the Department of Human Development at the University of Maryland, College Park.
  • CHARLES H. ZEANAH is Mary Peters Sellars-Polchow Chair of Psychiatry, Professor of Pediatrics, Vice Chair for Child and Adolescent Psychiatry, Executive Director of the Institute of Infant and Early Childhood Mental Health, and Director of the Tulane Infant Team at Tulane University.
  • More By Charles A. Nelson
  • More By Nathan A. Fox
  • More By Charles H. Zeanah