I appreciate the thoughtful responses to my essay on the ethics of the Bucharest Early Intervention Project (BEIP), from its investigators, Drs. Fox, Zeanah and Nelson and from Dr. Millum, one of the bioethicists who had defended their study design.
Let me begin by considering how Zeanah, Koga, Simion, Stanescu, Tabacaru and Nelson for the BEIP Core Group responded , in 2006, to an ethical critique that year by Douglas R. Wassenaar from the Infant Mental Health Journal. After raising some important methodological issues about blinding in the randomized clinical trial design, Wassenaar asks why provisions were not included in the original BEIP protocol for transferring children out of the institutions if foster care was “proved” superior in the context of the study. He asked what would be done with the other children left behind in the institutions:
The authors (the BEIP investigators) describe the financial constraints impacting on the study and the institutional context, particularly where the authors mention that local and national authorities could not guarantee that the results of the study would be nationally implemented if the findings favored foster care, as they did. This is clearly beyond the resources of the investigators. However, it would have been valuable to know if the investigators tried to negotiate foster care placements for the institutional control group if the intervention were shown to be effective. Although considerable, the financial implications of this seem within the realm of persuasive negotiation and advocacy, and ethically very important. No evidence is presented of such efforts, and we are left with the impression that the approximately 67 control subjects, subject to the batteries of physical and psychosocial evaluations over 4 years, languish in these institutions to this day. Most ethical guidelines require that in comparative interventions, the intervention if beneficial should have been offered to the control group(s) after the study. Failure to discuss this issue is a major omission of the paper, and a major ethical concern.
Although the investigators respond that of the children, “only 12 remain as of this writing,” Wassenaar’s critique is important because it points out that the authors did not prospectively make provisions for transferring children out of the orphanages should their intervention prove beneficial because of fiscal constraints and the lack of governmental guarantees.
This is a revealing omission. It suggests that the investigators were operating on grounds circumscribed by authorities and not traditional scientific and humanitarian norms: If one determines a benefit, a data safety monitoring board would close an arm of a study down and move subjects into the therapeutic side of the randomization. The failure to describe a mechanism for transfer out of the orphanages suggests an inability to adhere to established norms given the political or economic circumstances under which the study was conducted. It suggests, from study inception, the co-optation of, as Wassenaar put it, “ethical guidelines,” despite the misgivings of the investigators.
The authors’ reply to Wassenaar was far more forthcoming than their response to my piece in Bioethics Forum. In an almost confessional tone, they wonder whether they should have done more for the children in the institutional group. Their worry is palpable:
For us, the major unresolved second guessing about the BEIP from an ethical standpoint concerns whether the study should have attempted an intervention on the institution side. Comparing foster care to enhanced institutional care was not the question posed to us by the Romanian government officials originally, however, nor was it the question most relevant from a Romanian policy perspective. Nevertheless, comparing our model of foster care to some form of enhanced institutional care would have likely reduced the ethical concerns about studying “business as usual” institutional care, an intervention for young children in which the investigators had little confidence.
Let’s dissect this passage. First, the authors admit that they did not have autonomy as scientists to design the study as they saw fit. They themselves have “unresolved second guessing” about the orphanage arm of the study suggesting that they were not in a state of equipoise about at least one arm of the study, notwithstanding their appeal to ethics experts. Ignoring their intuitions about “an intervention for young children in which” they “had little confidence” led to a perpetuation of harm and blunted a clear duty to rescue children from the squalor of the orphanages and at least bring that intervention up to standards of basic decency.
What they call this “enhanced institutional care” is a euphemism for conditions brought up to civilized standards. Only by using those improved institutional conditions in their intervention would the researchers have overcome what might be viewed as “deliberate indifference” by the children’s institutional keepers, borrowing the phrase from the U.S. Supreme Court’s decision in Estelle v. Gamble, which established the rights of prisoners to medical care. The Court held, “Deliberate indifference by prison personnel to a prisoner’s serious illness or injury constitutes cruel and unusual punishment contravening the Eighth Amendment.”
It is not a stretch to draw an analogy between Romanian children in orphanages, captive to a government’s ideology and subjected to de minimis conditions, to prisoners incarcerated in a correctional facility. Both children and prisoners have requirements that must be met for the maintenance of health and both are wholly dependent upon authorities for these services. Indeed, the needs of children are greater because they require adults and trained personnel to identify medical conditions that may require care. Given this, the continued institutionalization of some children in the “control arm” perpetuates what, in the United States, would be an unconstitutional state of deliberate indifference, a potentially “cruel and unusual punishment” that American investigators, governing institutions, and philanthropies should have found inconsistent with our law and mores. This, too, argues for a duty to rescue the children, augmenting my previous appeal to Richardson’s work on ancillary care obligations.
But that wasn’t the case because the governing standards had gone off-shore. Romanian authorities dictated study design and the questions that could be asked. As the investigators noted, “Comparing foster care to enhanced institutional care was not the question posed to us by the Romanian government officials.” That was also not the relevant question “from a Romanian policy perspective.” This admission confirms that the BIEP investigators had lost control of their study and its moral compass, all with grave consequences.
Regrettably, they conducted the study at the behest of a post-totalitarian regime and did so in a manner acceptable to that power structure. And so directed, their search for truth became a search for truths acceptable to their hosts. This cooptation and subservience to government officials distorted, in my view, study design and even what was scientifically in doubt. In the end, it turned questions of science into an exercise of extreme cultural relativism: what was true was not a question of science or developmental biology. There was no doubt that institutionalization and sensory deprivation were harmful to developing children. That truth was well-established and an issue beyond equipoise. But their hosts wanted something other than the truth, and this quest for information led to the perpetuation of avoidable harms associated with ongoing institutionalization.
Sadly, the BEIP became a quest to collect information, no matter the human cost, that would be persuasive to Romanian officials. This allegiance to governmental authorities distorted study design and led to a fixation on a randomized clinical trial design, that could never be blinded and which left some children’s neglect in plain view.
Equally troubling was the neglect of at least two alternative methodologies. The first would have been to upgrade the orphanages to a minimal standard and then compare the “enhanced” orphanage experience and foster care. This would have mitigated the toxicity of that environment and rightly focused on the effects of two “positive” interventions, that of an orphanage arm brought up to par and of foster care. The second, and in my view better alternative, would have been to place the entire cohort into foster care and gauge its effects on development based on the time the children had already been exposed to the orphanages and the timing of their removal based on their age.
In a response to my original essay, Joseph Millum, one of the consulting ethicists, even concedes that there might have been alternative methodologies that did not involve randomization, noting, “The correct answer to this question may be disputed.” But then he points to what I view as the study’s critical misstep, namely that the BEIP investigators choice of design was predicated upon the needs of the Romanian authorities: “They wanted clear data that would be compelling to Romanian policy makers and designed their study on that basis” (italics added). An alternative methodology was not on the table because it was unacceptable to the investigators’ government minders.
While the investigators’ desire to rectify the tragedy of the Romanian orphans discovered in post-Ceausescu era was laudable, it remains regrettable that they chose unethical methods to satisfy the policy requirements of the government. This deference to governmental preference, breaching what was already known by many experts about the dangers of continued institutionalization and the superiority of foster care, led some children to be randomized to an ongoing state of toxic exposure, mistakenly and tragically viewed as the “standard of care.”
The BEIP investigators counter that changes in Romanian policy justify their actions. They point to the 2004 Romanian prohibition on the institutionalization of children less than two years of age.
BEIP was uniquely positioned to provide data relevant to the question of whether foster care offered advantages over institutional care for children who were abandoned and placed in institutions in the Romanian context. With such data, we believed, Romanian policy-makers could make more informed decisions. In fact, this is exactly what happened. Romania banned the institutionalization of children under the age of two in 2004.
This passage is problematic, suggesting a clear causality between the BEIP study, with all its ethical concerns, and an outcome that all would welcome and some might even view as exculpatory for their study design. But it is not that straightforward, for a couple of reasons. First, there may have been ways other than randomization to demonstrate the heightened developmental vulnerability of children under. Second, because there was a change in policy does not mean that the BEIP was responsible for it. Social policy is a complex phenomenon and there were other forces at play at the time, such as the attention of international media and the efforts of human rights workers. Each could have contributed to the ban.
This argument may seem far-fetched and speculative, but a review of the publication history of the BEIP group suggests otherwise. The BEIP investigators assert that Romanian officials banned institutionalization of children under two because of the data from the randomized clinical trial. However, the policy changed three years before the BIEP results were published in Science in 2007. This is an important question because the investigators’ justification for the randomization – and, therefore, for the ongoing harms of institutionalization to which some children were exposed – hinge on the need for data from a randomized trial.
The gap between the ban and the study’s publication suggests that the investigators are offering a post hoc justification for the trial’s design, impossibly claiming credit for a data-based reform before the study was even completed, much less the data fully collected and the work subjected to external peer review. This suggests that results from the study were not necessary to effect changes in Romanian law. This realization makes the suffering of the children randomized to the orphanage arm all the more worrisome.
It is a worry that I hope others will consider carefully, so that the memory and methods of the BEIP are not lost to history but become an object of its scrutiny. That effort would be a fitting legacy to those children who should have been rescued from the rigors of randomization and the Bucharest Early Intervention Project.
Joseph J. Fins, M.D., M.A.C.P., is The E. William Davis, Jr., M.D. Professor of Medical Ethics; Chief, Division of Medical Ethics; Professor of Medicine; Professor of Public Health; and Professor of Medicine in Psychiatry at Weill Medical College of Cornell University. He is a member of the board of directors of The Hastings Center, a Hastings Center Fellow, and chair of its Fellows Council.
Acknowledgements: The author thanks the Sackler Institute’s Infant Psychiatry Seminar at the Payne Whitney Clinic and the Clinical Seminar Series at Rockefeller University for invitations to present aspects of this work and for their insights. He also notes the partial support of a CTSC grant to Weill Cornell Medical College #UL1RR024996.
1. C.H. Zeanah, S.F. Koga, B. Simion, A. Stanescu, C.L. Tabacaru, N.A. Fox, C.A. Nelson for the BEIP Core Group, “Response to Commentary: Ethical Dimensions of the BEIP,” Infant Mental Health Journal 27, no. 6 (2006): 581-583.
2. D. R. Wassenaar, “Commentary: Ethical Considerations in International Research Collaboration: The Bucharest Early Intervention Project,” Infant Mental Health Journal 27, no. 6 (2006):577-580.
3. Estelle, Corrections Director, et al. v. J.W. Gamble, 429 U.S. 97 (1976)
4. W.J. Rold, “Thirty Years After Estelle v. Gamble: A Legal Retrospective,” Journal of Correctional Health 14, no. 1 (2008):11-20. http://jcx.sagepub.com/content/14/1/11.short
5. H.S. Richardson, Moral Entanglements: The Ancillary-Care Obligations of Medical Researchers, (New York: Oxford University Press, 2012).
6. J.J. Fins, “A Review of Moral Entanglements: The Ancillary-Care Obligations of Medical Researchers by Henry S. Richardson, New York: Oxford University Press, 2012,” Notre Dame Philosophical Reviews, May 19, 2013. http://ndpr.nd.edu/news/40053-moral-entanglements/
7. C.A. Nelson, C.H. Zeanah, N.A. Fox, P.J. Marshall, A.T. Smyke, and D. Guthrie, “Cognitive Recovery in Socially Deprived Young Children: The Bucharest Early Intervention Project,” Science 318 (2007): 1937-1940.