Orphanages, CHIFF, and UNICEF Recommendations: Is There Science Behind Anti-Orphanage Positions?

When care for children without functional families has been discussed recently, a single perspective has been paramount. This is the belief that group care in institutions (orphanages) is “toxic” to good development, and that individualized foster care and adoption are the only acceptable choices. The proposed Children In Families First (CHIFF) legislation in the U.S., and the recommendations made by UNICEF about shifting from orphanages to foster homes, are both based on the claim that scientific evidence strongly supports the idea that institutional care is in and of itself harmful to children’s development.

However, the existing evidence is not nearly as strong as some proponents argue. Today, I want to present some sources of relevant information, including both older and more recent studies, and to look at their conclusions and the evidence that supports them.

Historically, it has been fairly common for children, even infants, to be cared for in groups. Valerie Fildes, in her fascinating book Wet Nursing, looked at the practice in France of sending infants to be breastfed in the countryside for the first two years of their lives. The mothers, who were usually skilled textile workers, could continue to contribute substantially to the family’s earnings, while the wet nurse did well for herself and her family by caring for multiple infants. The mothers might or might not visit their babies, and the children were badly distressed when suddenly transferred back to their family’s care and separated from the only caregiver they knew. This does not seem to us post-Bowlby readers as a very good idea, but in fact (as James Robertson was to point out much later), when the children were then cared for sympathetically, most of them did very well. A somewhat similar situation existed in World War II Britain, when quite young children were evacuated from London to unfamiliar places and multiple caregivers   ( see Churchill’s Children, by John Welshman), or were cared for in group homes like Anna Freud’s Hampstead Nurseries. Again, good developmental outcomes were the rule, rather than  a “toxic” effect lasting through the children’s lifetimes.

No one would claim that wet nursing or evacuation were interventions that we would expect to be beneficial to development, but they seem to have done very little harm, and along with ethnographic data suggest that human beings do not require a single, restricted set of experiences in order to thrive in early life.

Like other wars, World War II created many orphans, and institutions for their care were established. Even when the war orphans were close to adulthood, institutions continued to exist to help unmarried mothers to deal with pregnancy and to care for children after birth. As these institutions were in existence contemporaneously with John Bowlby’s formulation of attachment theory, and because their arrangements were in conflict with Bowlby’s (now discarded) tenet of monotropy (the need for an infant to form an attachment to a single caregiver), many questions were asked about the effects of institutionalization on early development.

One institution that received much early attention was the Metera, a Greek foundation for unmarried mothers and their babies. According to a 1960 article in a popular magazine (possibly not very accurate), an infant born in the Metera  was assigned, along with two other young infants, to two caregivers who lived in the institution, and who moved along with their three infants to a different ward if a transfer was made. Family members could visit the children, but the emphasis was on the relationships with the caregivers; adoptions were to be arranged within a few months if possible. However, by 1979, practices in the Metera seem to have changed—or perhaps they were never as positive as previously indicated. Berry Brazelton and two colleagues examined neonatal behavior of Metera infants, as compared to other groups (Pediatrics, 63(2)), and found worse performance on the part of the Metera children at birth. They attributed these difficulties to the extreme disapproval of unmarried pregnancy in Greece, and the attempts of the mothers to starve themselves before they arrived at the Metera--  but also mentioned that the infants were in unstimulating white cubicles and fed on a strict 4-hour schedule. More recently, Vorria et al (“Early experiences and attachment relationships of Greek infants raised in residential group care”, Journal of Child Psychology and Psychiatry, 2003, 43 ,pp. 1208-1220) looked at the development of children who remained in the Metera for many months, and found that an unusually high number of them showed disorganized attachment (that is, atypical behavior when a caregiver returns after a separation in a strange place, for example, freezing in place or backing toward the adult). However, some showed secure attachment. The children, who had incidentally had much lower birth weights than a control group, were less advanced cognitively than home-reared controls. Vorria et al noted that although there were claims that the Metera babies had plenty of interaction with a small number of caregivers, in fact they had little interaction in the early months, and were later moved to a pavilion where the ratio of babies to caregivers was 4-6:1, a situation where the best-trained and best-motivated caregiver would have difficulty in being sensitive and responsive to all infant signals.

The interesting point about these studies of the Metera is that although there was little question that the institutional babies fared less well than home-reared babies did, the authors did not attribute the problems to any single factor, particularly not to institutionalization in and of itself. On the contrary, they looked at characteristics of the infants’ pre-birth experiences and at specific characteristics of the institution as possible causes of poor development.

The recent research on which CHIFF and the UNICEF recommendations depend is the Bucharest Early Intervention Project (BEIP), conducted by the eminent child psychiatrist Charles Zeanah and many collaborators, and reported in a growing number of publications. The BEIP research is unique among investigations of  institutional effects on children in that it involved a randomized controlled trial (RTC), the highest level of research design, and one which does much to assure that outcomes are caused by the treatments the children experience, rather than by other unknown factors. As described by C.H. Zeanah, N.A. Fox, and C.A. Nelson in 2012 (“The Bucharest Early Intervention Project…”, Journal of Nervous and Mental Disease, 200, pp. 243-247), this project worked with 136 children 6-31 months old who were being cared for in Romanian orphanages. Children were assigned at random to remain in the institution where they already were or to go to a foster home.

BEIP authors have given clear descriptions of the resources poured into the foster homes. The work included establishment of a foster care network, as well as training of social workers who would oversee and encourage the development of relationships between the foster children and their caregivers. The social workers also received weekly consultations with expert psychologists in the U.S. It is notable, by the way, that the training and resources involved here were probably a good deal greater than those available to foster parents in the U.S. or the U.K.

Publications on the BEIP have given much less detail about the experiences of the children who were randomized to institutional care. It would be of much interest to know details of these experiences such as the ratio of infants to caregivers or the sizes of groups (these factors generally being considered to have strong effects on the outcomes of day care). The 2012 article by Zeanah et al references the well-publicized appalling conditions in Romanian institutions for children following the fall of the Ceasescu regime in 1989, but does not provide much information about ways in which orphanages might have changed (for better or worse) in the ensuing period. This lack of information about the experiences of the children randomized to the institutional treatment arm makes it difficult to know what factors actually differed between the two groups, and to what extent they were different.

An additional difficulty of design has been pointed out by Douglas Wassenaar, writing in Infant Mental Health Journal in 2006. Wassenaar noted a problem of scientific validity in the BEIP study: the fact that evaluators should have been, but were not, “blinded” to (unaware of) the treatment being received by each child, for “both the ‘soft’ psychosocial evaluations, which are notoriously subject to rater bias, and some of the more ‘objective’ physical evaluations”. Wassenaar also pointed out that this matter should have been discussed by Zeanah and other BEIP authors, as an important issue with respect to confidence in the conclusions, “particularly in view of their expressly stated bias favoring deinstitutionalization”.  

In spite of these difficulties and criticisms, BEIP participants (and many others) have continued to state general conclusions that the project has shown the advantages of foster care over institutional care for young children. Fox et al, in the Journal of Child Psychology and Psychiatry in 2011, stated that children raised in institutions exhibit lower IQ scores than those raised in family settings.

However, even setting Wassenaar’s concern about blinding aside, it’s necessary to question what the BEIP data actually show. Is it not that children who were in high-quality, resource-rich foster care did better than children who were in institutions whose quality was not clearly described but may have been abysmal? Is the conclusion not that a particular group of children, in one set of conditions, did better than another group, in a specific other set of conditions? Those conclusions are a far cry from saying that institutions are “toxic”, and that all possible institutional variants are harmful in comparison to all possible variants of adoption or foster care (including, perhaps, “mega-families”). The latter statements smell strongly of the “expressly stated bias favoring deinstitutionalization” mentioned by Wassenaar and evident in CHIFF and other proposals.

What would have happened if the training and resources lavished on the foster homes were also provided for the  Romanian institutions? That’s the question that must be answered as a step toward understanding whether institutions are “toxic” (and I would like to see that sensational “toxicity” metaphor abandoned, as it distracts from rational discussion). Until we have further information, it will not be time to say that science supports one view or another of care for children without parental care. When such statements are made, they should be recognized as the ideological positions they actually are.

Do Adopted Children Want to Breastfeed? Karleen Gribble's Report

After I commented on some recent discussions of breastfeeding for adopted children, Karleen Gribble, of the University of Western Sydney , responded and sent me a copy of her paper “Post institutionalized adopted children who seek breastfeeding from their new mothers” (Journal of Pre- and Perinatal Psychology and Health, 19(3), 217-235). I’m going to comment on Gribble’s paper today.

Let me say first that I’m thrilled to see more observational work on care events in infancy and childhood. I consider this to be a much-neglected topic, and exactly what happens during breastfeeding--  including the many variations--  is rarely given systematic attention. I attempted to do some data collection of my own some years ago, but unfortunately La Leche League (it was at one of their functions that I was observing) was not willing to give permission for this. I also vividly recall being roundly scolded by the journal editor Marc Bornstein for submitting a comment that proposed more such work. I confess that I have not been entirely evidence-based or free from speculation about breastfeeding interaction (www.psychologytoday.com/blog/child-myths/200906/breastfeeding-speculating-wildly), but this is an area where knowing what to observe depends partly on thinking through one’s own experiences and other people’s stories. Nonetheless, it’s the observation rather than the speculation that we need to attain.

In considering Gribble’s paper, it’s important to look first at the a priori assumptions it contains. The author refers frequently to breastfeeding as a causal factor in attachment and attunement, but at the same time concludes the paper by noting that attachment may need to precede breastfeeding. (It is not clear whether this implies that newborn infants who breastfeed must already be attached to their mothers.) In addition to Gribble’s statements in the paper, we need to look at the reference section and see what authors she considered to provide appropriate background for her work; these include Deborah Gray, Mary Hopkins-Best, Terry Levy and Michael Orlans, and Nancy Thomas--- all people committed  to an alternative theory of attachment rather than to Bowlby’s conventional psychosocial approach. In line with this background, Gribble references the so-called “attachment cycle” as a series of experiences that cause emotional attachment (see http://thestudyofnonsense.blogspot.com/2012/08/parsing-attachment-cycle-fox-terrier-of.html for a discussion of this issue). Like other authors who publish in JPPPH, Gribble also references ideas about children’s memories of birth and early life. All of the assumptions displayed here suggest that Gribble is very ready to accept the belief that unconscious, biologically-driven motives and behavior patterns play powerful roles in children’s development even after the first few months of life have passed. She quotes a mother as saying that her child’s need to suck was ”primal”, exemplifying the belief I have just described, and suggesting strongly that some human maternal and child behaviors are best seen as instinctive.

The assumptions I have just described are in contradiction to established conventional views of early development (and of course such conventional views can turn out to be quite wrong, but it has yet to be shown that they are wrong). Conventionally, feeding modes are not considered to be important to development of parent-child relationships, although a parent’s sensitivity and responsiveness to infant feeding cues are important, just as they are in every other area of parent-child communication. Neither is skin-to-skin experience thought to play a strong role in establishing emotional relationships. Infants are not considered to have an attachment to anyone at the time of birth, nor are they thought to have memories of birth or of early life. Biologically-determined infant social reactions are thought to be paramount in the early months, but after that learning from social interactions begins to take over. While initially biological, modes of communication become a matter of learning and therefore are strongly culturally influenced rather than instinctive.  

Understanding the assumptions of Gribble’s paper, and the ways they differ from the foundations of conventional approaches to early development, let’s go on to look at the information reported about adopted children wanting to breastfeed. Gribble reports information about 32 adopted children, of both sexes and a range of ages (ages at placement= 8 months to 12 years) and separation histories, whose adoptive mothers stated that the children asked for breastfeeding or signaled that they wanted to breastfeed. It is not clear how many of the mothers were actually interviewed by Gribble, as some of the cases were said to have been reported by social workers or drawn from published material. In addition, it is unclear what sex ratio was involved; of the five interviews reported by Gribble, four of the children were girls.

Gribble’s paper provides an interesting beginning for discussion of children’s motivation for adoptive nursing, but a much better context is needed before we can interpret this report. For example, the Wikipedia article on adoption in Australia states an average of 330 intercountry adoptions per year (the children in Gribble’s study were adopted from other countries). However, Gribble does not say over how many years her information was collected, so it is impossible to know whether the 32 children discussed were a very large or a very small proportion of similar adopted children.

In addition, Gribble does not state or even speculate upon the number of nonadopted children who, having been bottle-fed from the beginning, or having been weaned from the breast, later communicate to their mothers that they want to breastfeed. This is an important issue because of Gribble’s argument that adoptive breastfeeding facilitates attachment in children who have experienced separations. Nonadopted children have presumably had ideal opportunities to develop attachment and have not experienced serious separations, so if they signal their wish for later nursing in the same proportions as adopted children, it is hard to see what the emotional motivation for this behavior would be. A full understanding of the phenomena reported by Gribble  awaits information that would permit this comparison.

Another important unanswered question is the role of the adoptive mother’s beliefs, expectations, and caregiving behavior in creating the child’s interest in breastfeeding. Gribble has pointed out elsewhere that mothers are not likely to provide information about the atypical behavior of adoptive breastfeeding unless they trust their confidant; this suggests that the mothers have a belief system that is not entirely shared by most other people. Gribble states a belief that skin-to-skin contact is important for attachment, and describes a mother who “used skin-to-skin contact via co-bathing and a cuddle time in the evening as a way of promoting attachment”. Mothers who share this belief provide opportunities for breast contact that would not occur in the Western world in most other situations or be presented by mothers who did not share the belief. Adoptive mothers who believe in the “skin-to-skin” and “attachment cycle” system may also accept the idea that breastfeeding is important for the mother-child relationship and therefore be exceptionally ready to read child behavior as a wish to nurse. (Gribble refers to sucking on clothing as such a signal, although mouthing and sucking objects is common childhood behavior, as often seen in school-age boys who like to chew on the necks of their t-shirts.) Interestingly, Gribble also describes children who sought to breastfeed as part of “birth games” played with their adoptive mothers, suggesting that she is focused on a group of adoptive parents who are committed to an alternative theory of early development--  not surprising in light of the journal in which the study was published.

The information Gribble presents is of great interest, and certainly should be kept in mind in cases where breast-touching by adopted children is regarded as “sexualized” behavior indicative of previous sexual abuse. (If Gribble’s reports are accurate, such behavior may not be an indication of sexual experience at all.) However, interpretation of the reported cases must await contextualization by information about other adopted children and about nonadopted children as well.  Gribble’s extensive discussion of why adopted children seek breastfeeding is premature, because we have no idea whether they actually do so more than nonadopted children do, or whether their adoptive parents’ belief systems lead them to read child messages differently than they might otherwise do.

As a final comment, I want to turn to two sentences that Gribble places at the end of the article abstract and that she does not actually discuss in the body of the paper. The first sentence states that the “frequency of adopted children seeking breastfeeding is unknown, however adoption professionals should advise adoption applicants of the possibility”. Such advice, if it is being given, is certainly likely to increase the number of cases where mothers interpret ambiguous behavior as bids for breastfeeding.

In a second sentence at the end of the abstract, Gribble makes the following claim: “It may also be appropriate for adoptive mothers to pursue breastfeeding in the event that the child does not.” Nothing in the paper provides grounds for this claim, and it is most concerning to think that a group of parents who have already (as Gribble notes) experienced various disappointments and losses should be offered an additional challenge when there is no evidence that it is necessary.