Developmental Psychology · Research presentation

Can a family repair what an orphanage took away?

The Bucharest Early Intervention Project: the first randomized trial comparing foster care with institutional care for young children.

SCIENCE · 2007 · VOL. 318 · PP. 1937–1940 Cognitive recovery in socially deprived young children: The Bucharest Early Intervention Project Charles A. Nelson, Charles H. Zeanah, Nathan A. Fox, Peter J. Marshall, Anna T. Smyke & Donald Guthrie

Presented by Jamil Aghasiyev

Press → or Space to advance · N for speaker notes · T for timer · F for full screen

01 Background & aim

0children were living in Romanian state institutions when the Ceaușescu regime fell in 1989.

1966Contraception and abortion were banned to grow the population. Many poor families could not raise their children and gave them to the state.

CARERotating 8-hour shifts and many children per caregiver. Little chance to form an attachment to any one adult.

Aim of the study

To test, with a randomized controlled trial, whether moving institutionalized young children into foster families improves their cognitive development, and whether the age at placement matters.

Does institutional care harm cognitive development?
Can foster care help children recover?
Is there a sensitive period, so that earlier placement helps more?

Researchers had compared institutionalized and family-reared children for about 60 years, but no one had run a randomized trial before BEIP.

02 Methodology · data collection

Names drawn from a hat

68Care as usual
(stay in institution)

68Foster care
(move to a family)

72Never in an institution
(comparison)

187 children under 31 months, in all 6 Bucharest institutions for young children. 51 excluded after medical screening. 136 children enter the trial. Random assignment: half stay, half move to foster families. Plus 72 children living with their own families in Bucharest.

DesignRandomized controlled trial, longitudinal. Started in 2000. Analysed by original group assignment (intention-to-treat).
ParticipantsAbandoned children aged 6–31 months at baseline (average ≈ 21 months), most of them in institutions since birth.
InterventionHigh-quality foster care built by the project, because Bucharest had almost none. Foster parents were recruited, trained and supported by social workers.
Data collectionStandardized cognitive tests given by trained examiners: Baseline · Bayley30 mo · Bayley42 mo · Bayley54 mo · WPPSI-R Bayley gives a developmental quotient (DQ), WPPSI-R an IQ. Both have a population mean of 100.

Now the method. You can click "Next step" or press the arrow key. The team screened 187 children under 31 months in all six Bucharest institutions for young children. [→] 51 were excluded after medical screening, which left 136 children.

[→] Then the key moment: names were drawn from a hat. 68 children stayed in the institution, which is called "care as usual". 68 moved to foster families that the project itself recruited, trained and supported, because Bucharest had almost no foster care at that time.

[→] A third group of 72 children who had never lived in an institution was used for comparison. Data were collected with standardized cognitive tests: the Bayley Scales at baseline, 30 and 42 months, and the WPPSI IQ test at 54 months. Because assignment was random, any difference between the two institution groups can be explained by foster care itself.

03 Key findings · three groups

Foster care helped. It did not close the gap.

Never institutionalizedFoster careCare as usual

−39

Before the split, institutionalized children scored about 64, against about 103 for children living with their families.

+8

The foster care group scored about 8–9 points higher than the care-as-usual group at 42 and at 54 months. The difference was statistically significant.

18–28

Points by which foster children still trailed children raised in their own families. Recovery was real but partial.

Group means. Bayley DQ at baseline and 42 months, WPPSI-R IQ at 54 months.

04 Key findings · timing

Earlier placement, bigger recovery

Age when the child moved from the institution to a foster family

94.4DQ at 42 months

Move the slider. Children who left the institution before about two years recovered far more than children who left later.

A sensitive period: after about 24 months, foster placement brought scores only slightly above children who stayed in the institution. The authors suggest early childhood is a window when family experience matters most for cognitive development.

Subgroup sizes: 14, 16, 22 and 9 children. At 54 months the pattern was similar but less clear-cut (IQ 84.8, 86.7, 78.1, 71.5).

05 Significance · how the results are used today

From a research finding to national law

2000BEIP begins in BucharestThe project builds a foster care network where almost none existed.
2004Romania's Child Protection Law 272/2004Institutional care banned for children under 2, except those with severe disabilities. Researchers shared findings with officials as they emerged.
2007Results published in ScienceThe first randomized evidence that family care beats institutional care.
2009UN Guidelines for the Alternative Care of ChildrenCare for young children, especially under 3, should be family-based.
TodayRomania: foster care for all children under 7Family-based care is the required form of out-of-home care for young children.

Child welfare policyCore evidence for deinstitutionalization reforms: close large orphanages and place young children in families.

Social work and courtsPlacing a young child quickly in a permanent family is treated as urgent, because delays have a cost.

Psychology scienceThe textbook human example of a sensitive period, and a model for ethically designed trials with vulnerable children.

06 Limitations

What this study cannot tell us

DesignTiming was not randomizedAge at placement was not assigned by chance. It is mixed up with how long a child had already spent in the institution.

SampleSmall subgroupsThe timing result rests on 9–22 children per age group, and the effect was clearer at 42 months than at 54 months.

InterventionUnusually good foster careFoster families were well trained and supported by the project. Ordinary foster care may help less.

GeneralizationOne extreme settingRomanian institutions were severely depriving. Results may not apply to better orphanages or other countries.

EthicsA debated control groupLeaving children in institutions for comparison was criticized (Millum & Emanuel, 2007). The team replied that no foster system existed and no child got less than usual care.

Follow-upOnly to age 4½This paper stops at 54 months. Later BEIP papers followed the children to age 8, 12 and beyond.

Contradictory and complicating research

Beckett et al. (2006) · English & Romanian Adoptees StudyThe window may close at 6 months, not 24Romanian children adopted into UK families after 6 months of age still showed marked cognitive deficits at age 11. Adoption, not randomized, and even harsher early-1990s institutions.

Fox et al. (2011) · same BEIP team, age 8The IQ advantage became smallerThe full-scale IQ difference was only borderline (p = .07), though verbal comprehension still favoured the foster care group.

Every study has limitations. [→] First, the age at placement was not randomized, so it is mixed up with how long the child had already spent in the institution. [→] Second, the timing groups were small, only 9 to 22 children, and the timing effect was clearer at 42 months than at 54. [→] Third, the foster care was unusually good, so normal foster systems may help less.

[→] Fourth, Romanian institutions were extremely depriving, so the results may not apply everywhere. [→] Fifth, ethics: some scientists criticized leaving children in institutions as a control group. The researchers answered that no foster system existed and no child received less than usual care. [→] Finally, this paper only follows the children to four and a half years.

[→] Other research also complicates the findings. Beckett and colleagues studied Romanian children adopted into British families and found the critical cut-off was 6 months, not 2 years. So both studies agree that earlier is better, but disagree on when the window closes. That study was not randomized, and the children came from even harsher institutions. [→] And the BEIP team's own follow-up at age 8 found the IQ advantage of foster care had become smaller and only borderline significant, although verbal scores still favoured foster care.

07 Conclusion

01Early institutional deprivation seriously harms cognitive development.
02Family-based foster care produces significant but partial recovery.
03Timing matters: placement before about 24 months gave the largest gains, which suggests a sensitive period.

Can a family repair what an orphanage took away? Partly, and the earlier, the better.

Thank you. Questions?

Nelson, C. A., Zeanah, C. H., Fox, N. A., Marshall, P. J., Smyke, A. T., & Guthrie, D. (2007). Cognitive recovery in socially deprived young children: The Bucharest Early Intervention Project. Science, 318(5858), 1937–1940. https://doi.org/10.1126/science.1143921 Beckett, C., Maughan, B., Rutter, M., et al. (2006). Do the effects of early severe deprivation on cognition persist into early adolescence? Findings from the English and Romanian Adoptees Study. Child Development, 77(3), 696–711. Fox, N. A., Almas, A. N., Degnan, K. A., Nelson, C. A., & Zeanah, C. H. (2011). The effects of severe psychosocial deprivation and foster care intervention on cognitive development at 8 years of age. Journal of Child Psychology and Psychiatry, 52(9), 919–928. Millum, J., & Emanuel, E. J. (2007). The ethics of international research with abandoned children. Science, 318(5858). United Nations General Assembly. (2010). Guidelines for the Alternative Care of Children (A/RES/64/142).