Reader Comments

Post a new comment on this article

No Evidence for Orphanages Being ‘Not So Bad’

Posted by vanijzen on 24 Dec 2009 at 15:37 GMT

No Evidence for Orphanages Being ‘Not So Bad’

Comments to Whetten K, Ostermann J, Whetten RA, Pence BW, O’Donnell K, et al. (2009). A Comparison of the Wellbeing of Orphans and Abandoned Children Ages
6–12 in Institutional and Community-Based Care Settings in 5 Less Wealthy Nations. PLoS ONE 4(12): e8169. doi:10.1371/journal.pone.0008169

Marian J. Bakermans-Kranenburg, Marinus H. van IJzendoorn
Centre for Child and Family Studies,
Leiden University, The Netherlands

December 24, 2009

MJBK and MHvIJ were supported by awards from the Netherlands Organization for Scientific Research (MJBK: VIDI grant no. 452-04-306; MHvIJ: SPINOZA prize). The authors declare that no competing interests exist.

Correspondence: Marian J. Bakermans-Kranenburg or Marinus H. van IJzendoorn, Centre for Child and Family Studies, Leiden University, PO Box 9555, 2300 RB Leiden, The Netherlands (email: or

“Institutions are not so bad” according to Dr. Kathryn Whetten in her interview in the New York Times of December 18, 2009. In her recent PLoS ONE paper on the Positive Outcomes for Orphans (POFO) survey in 6 sites across 5 ‘low-resource’ countries including 1,357 institution-living and 1,480 community-living orphans and abandoned children (ages 6–12 years), she and her co-authors conclude that health, emotional and cognitive functioning, and physical growth are no worse for institution-living than community-living children. In fact, in most domains institution-living children seem to fare better.

We take issue with this provocative claim that is not supported by the data collected in the POFO study. The first issue concerns the comparability of the institution-living and community-living children. The quasi-experimental design of the survey does not protect against the influence of pre-existing differences between the institution-reared and community-raised children. The second issue pertains to the generalizability of the findings. We argue that any conclusion from the POFO study is limited to a specific age range, a specific set of countries and institutions, and a restricted set of measures on certain developmental domains.

Comparability of the institution-living and community-living groups. The POFO study uses a cross-sectional and quasi-experimental design. Without randomized assignment of subjects to conditions there is always the risk of groups being systematically different (pre-, peri-, and postnatally) in crucial areas of development before they entered orphanages or community-based child rearing settings. In the POFO study there is cause for concern regarding the composition of the community-based care group.
First, almost 65% of the community-living children have a biological mother who still is alive, in contrast to 43% of the institution group. Consequently, a remarkably high percentage of community-living children are raised by their biological parents (55%). It makes no sense to compare this group of home-reared children with institutionalized children to answer the central question of where abandoned and orphaned children should be placed as placement of these children with their biological parent is no option.
The comparisons should be based on three groups: (1) institution-reared children, (2) orphaned or abandoned children not living with their biological parent or grandparent, and (3) a comparison group of children living with (one) biological parent or grandparent. The ‘sensitivity analyses’ conducted by the authors excluding several groups (e.g., non-orphaned children) is an insufficient surrogate for these analyses.
Looking at the data on children residing in institutions for less than one year (Appendix 1B), we have reason to believe that the institution- and community-living children did differ before entrance into their respective settings. Children residing in an orphanage for only a brief period of time (less than a year) show better health, less problem behaviors, and better cognitive development than their community-living peers. These differences might easily have been present before entrance into institutional care, and thus would give the institution-reared children a head start. The large variation in child outcomes within settings may be ascribed to variations in these settings, as suggested by the authors, but might also be due to genetic or perinatal differences among the children – without random assignment we simply do not know.
Of course, a randomized control trial to compare the effectiveness of institutional versus family-type care arrangements for orphans and abandoned children is difficult to imagine. Nevertheless, it did happen. In the unique randomized control Bucharest Early Intervention Project [1] young children living in institutions were randomly assigned to continued institutional care or placement in foster care, and their development was tracked through 54 months of age. Three main findings emerged from the project. First, children reared in institutions showed greatly diminished intellectual performance (borderline mental retardation) relative to children reared in their families of origin. Second, children randomly assigned to foster care experienced significant gains in cognitive function. Lastly, the younger a child was when placed in foster care, the better the cognitive outcome. Indeed, there was a continuing "cost" to children who remained in the institution for longer periods of time. Similar findings have been reported on physical growth parameters and socio-emotional development [2,3].

Generalizability of the findings. The age range of the current study is 6-12 years. It appears from longitudinal studies of institutionalized children that most significant delays are found in the first years of life, and that institutionalized children tend to catch up after their fourth birthday (see e.g. [4], for growth parameters, and the Bucharest Early Intervention Project [1] for cognitive development). A meta-analysis of cognitive development of institutionalized children (including 75 study outcomes) showed a difference of almost 20 IQ-points between institutionalized and family-reared children [5]. But children who were younger at assessment and those who were younger at placement in the orphanage showed more delays than their older peers. The focus on the age range 6-12 years may obscure substantial differences in the years before.
Second, differences may be less visible in the specific outcomes that were included in the current study. The authors are to be commended for the inclusion of a large sample in their survey, but this necessarily goes at the cost of more time-consuming and at the same time more valid observational assessments. As has been documented in a number of studies, the formation of selective and stable attachment relationships is seriously compromised in all institutionalized care settings studied thus far [2, 6-9], and the Strengths and Difficulties Questionnaire is a meager operationalization of the children’s emotional development. Furthermore, long-term effects of institutionalization may have gone “under the skin” and be visible in, e.g., deviant stress regulation, as documented by Gunnar and her team [10] in a study on HPA axis functioning of children adopted from Romanian orphanages; for children residing in Ukrainian orphanages see [11].
Third, the selected regions and countries constitute a small and specific selection of countries where institutionalized care is present to a large extent, and this selection favors the ‘no difference’ outcome. To call countries like Cambodia, Ethiopia, Tanzania, and Kenya ‘less wealthy nations’ is a crude understatement. The aforementioned meta-analysis on cognitive development showed that in countries with extremely low Human Development Index scores there was no difference between family-reared and institutionalized children [5]. The extremely poor living conditions for institutionalized as well as community-care children in these countries may result in a floor effect without any differential effects of rearing condition. To call the well-being of the children living under these conditions ‘not so bad’ is an ethnocentric euphemism.
Fourth, and related, from various studies on institutional care [12] damage appears to be largest when caregiver arrangements are characterized by instability of caregivers and frequent shifts of arrangements during a care schedule of 24 hours, 7 days per week. Precisely that aspect of institutions appears to be less prevalent in the institutions included in the POFO study, where institutional caregivers worked on average more than 100 hours per week, and 37% of the caregivers were present continuously (168 hours per week). This is –alas– not the case in the majority of institutional care settings [1-2, 6-7], and in fact may come closer to a family-type environment but with heavier financial burdens and larger costs for the private life of the caregivers than in more common institutionalized care settings.

The Positive Outcomes for Orphans study (POFO) –nomen est omen- still has some way to go to fulfill its mission. In the meantime, we should devote our resources to the extension foster and adoptive care, and to the improvement of institutional care, which will unfortunately remain a last resort for millions of children for years to come [13].


1. Nelson CA, Zeanah CH, Fox, NA, Marshall PJ, Smyke AT et al. (2007) Cognitive recovery in socially deprived young children: The Bucharest Early Intervention Project. Science 318: 1937-1940.
2. Zeanah CH, Smyke AT, Koga S, Carlson E, The BEIP Core Group (2005) Attachment in institutionalized and community children in Romania. Child Development 76: 1015-1028.
3. Smyke AT, Zeanah CH, Fox NA, Nelson CA, Guthrie D Placement in foster care enhances quality of attachment among young institutionalized children. Child Development. In press.
4. Van IJzendoorn MH Bakermans-Kraneburg MJ Juffer F (2007) Plasticity of growth in height, weight and head circumference: Meta-analytic evidence of massive catch-up of children’s physical growth after adoption. J Dev Behav Pediatr 28: 334-343.
5. Van IJzendoorn MH, Luijk M, Juffer F (2008) IQ of children growing up in children’s homes: A meta-analysis on IQ delays in orphanages. Merrill-Palmer Quarterly 54: 341-366.
6. The St. Petersburg – USA Orphanage Research Team (2008) The effects of early socialemotional and relationship experience on the development of young orphanage children. Monographs of the Society for Research in Child Development 73: 1–262.
7. Vorria P, Papaligoura Z, Dunn J, Van IJzendoorn MH, Steele H, et al. (2003). Early experiences and attachment relationships of Greek infants raised in residential group care. Journal of Child Psychology and Psychiatry 44: 1208-1220.
8. Zeanah C H, Smyke AT, Dumitrescu A (2002) Attachment disturbances in young children. II: Indiscriminate behavior and institutional care. Journal of the American Academy of Child and Adolescent Psychiatry 41: 983-989.
9. Rutter M, Kreppner J, Sonuga-Barke E (2009) Emanuel Miller lecture: Attachment insecurity, disinhibited attachment, and attachment disorders: Where do research findings leave the concepts? Journal of Child Psychology and Psychiatry 50: 529-543.
10. Gunnar MR, Morison SJ, Chisholm K, Schuder M (2001) Salivary cortisol levels in children adopted from Romanian orphanages. Development and Psychopathology 13: 611-628.
11. Dobrova-Krol N, Van IJzendoorn MH, Bakermans-Kranenburg MJ, Cyr C, Juffer F (2008) Physical growth delays and stress dysregulation in stunted and nonstunted Ukrainian institution-reared children. Infant Behavior and Development 31: 539-553.
12. Dobrova-Krol NA, Van IJzendoorn MH, Bakermans-Kranenburg MJ, Juffer F. Effects of perinatal HIV infection and early institutional rearing on social cognitive development of children in Ukraine. Child Development. In press.
13. Bakermans-Kranenburg, M.J., Van IJzendoorn, M.H., & Juffer, F. (2008). Earlier is
better: A meta-analysis of 70 years of intervention improving cognitive
development in institutionalized children. Monographs of the Society for
Research of Child Development, 73, 279-293.

No competing interests declared.

RE: No Evidence for Orphanages Being ‘Not So Bad’

kwhetten replied to vanijzen on 04 Jan 2010 at 03:29 GMT

We thank Bakermans-Kranenburg and van IJzendoorn (MJBK and MHvIJ) for their detailed comments and would like to use this forum to reply. The quasi-experimental design, indeed, does not allow us to draw conclusions about the influence of pre-existing conditions or selection effects on our estimates. We agree with MJBK and MHvIJ that significant selection processes may influence whether or not children enter institutions, and the timing of such entry. Such processes cannot be identified in our study. Clearly, a randomized controlled trial would be best positioned to determine the effect of placement; however, children would have to be randomized into a wide variety of institutional and family care settings in order to account for the range of care characteristics observed in our study. Ethical and logistical considerations would make this an unreasonable endeavor. Quasi-experimental designs, such as the one employed in POFO, remain the only way to draw inferences about this broad range of both institutional and family care options. The Bucharest Early Intervention Project (BEIP) demonstrates that improving care early in life remediates most of the negative effects of deprivation and that these changes can be seen even in brain images. The POFO research adds to the BEIP in demonstrating that for older children being in an institution does not necessarily equal poor wellbeing. We show great variation in outcomes among children in institutions and similarly great variation among children in family care settings. The average differences between the two groups, by comparison, are small, and only longitudinal data can ascertain whether differences between the groups narrow or widen as children spend more time in institutions.

The results presented in “A Comparison of the Wellbeing of Orphans and Abandoned Children Ages 6–12 in Institutional and Community-Based Care Settings in 5 Less Wealthy Nations,” are derived from children aged 6 to 12 who have been living in their current care settings for variable amounts of time. This study does not include or apply to younger children. As noted in the paper, for children in the new OAC epidemic across southern and eastern Africa the risk of having a parent die increases with each year of age. This presents a different situation from countries where orphaning and abandonment primarily occur in infancy. Just as study results for children over the age of 6 are not generalizable to those under the age of 6, the same may be true in reverse. In countries facing the greatest burden of the OAC epidemic, identifying solutions for older children is as important as finding solutions for younger children.

Like MJBK and MHvIJ, we were concerned that single orphans living with a biological parent in a family setting have different experiences than double orphaned or abandoned children. Single orphans in family settings were included in our analyses because they are also at risk of being placed in an institution; this fact is demonstrated by the large number of institution-based children who have a surviving parent. Many would argue that the inclusion of single orphans would result, on average, in more positive outcomes for family dwelling children, making the results of our comparisons conservative. To account for the concern that including single orphans in family dwellings may influence the results, two sets of results are presented in our paper: 1) comparisons between institution-based OAC and all OAC living in family dwellings (this group includes single orphans); and 2) comparisons between institution-based OAC and only OAC who live in family dwellings without a biological parent (this group excludes single orphans). The results do not change when the family dwelling group is changed. We are presently analyzing the extent to which the variation in outcomes is affected by the presence of other family members in the household (MJBK and MHvIJ, for example, referred to grandparents).

Bakermans-Kranenburg and van IJzendoorn raise another important point that the POFO Research Team hopes others will understand: this manuscript reflects not just institutional settings, but also the community settings. The community settings in which the family dwelling children are being raised are very poor. National and international policies and interventions are desperately needed to help alleviate extreme poverty and assist with education, nutrition, shelter and health care for adults and children alike. The countries selected for inclusion in this study do not represent the poorest of the poor; they were selected because they are historically, politically, religiously and culturally diverse so that the study findings could be more generalizable.

It is important to remember that the 83 participating POFO institutions represent a random sample of eligible institutions in the affected areas. Extensive effort was put into compiling the list from which institutions were randomly selected at each site; at no site did a single source of information have a list of all institutions in the area. Without such a census of institutions it is not clear what characteristics adequately describe “a majority” of institutional care settings referred to by MJBK and MHvIJ. Researchers in the past have made the assumption that institution-based children experience a certain kind of care that includes, among other negative attributes, significant neglect and inconsistent caregiving. What this study might start to demonstrate is that less wealthy nations, in response to the orphan crisis, may be developing a great variety of group care options within the affected communities.

Institutional caregiving arrangements, especially caregiver stability and working hours, indeed appear to differ in participating POFO institutions, especially compared to the Eastern European and American institutions that have generated much of the knowledge and opinions about institutional care to-date. This is care that is not being provided, with 3 exceptions, out of the homes of families and the size of the care setting varies tremendously with aggregate outcomes not changing based on institution size. Regardless of whether or not institutional care settings in fact come closer to family type environments, resources need to be focused on a better understanding of what aspects of care – in both settings – are associated with positive outcomes and resiliency in the face of adverse circumstances so as to provide optimal care options for children facing different situations. The resources needed to improve both family based and institutional care settings will undoubtedly be enormous.

No competing interests declared.