The clinical research community and a number of prominent bioethicists have swiftly come to the defense of investigators conducting the SUPPORT study, in which approximately 1,300 premature infants were randomly assigned to be maintained at higher or lower levels of oxygen saturation. The study took place between 2005 and 2009, involved 22 sites and was reviewed by at least as many institutional review boards. In March, the Office of Human Research Protection (OHRP) concluded that investigators had violated the informed consent provisions of the federal regulations governing research by failing to inform parents of infants enrolled in the study about risks of retinopathy, neurological injury, and death. Results from the study revealed that infants assigned to receive the lower range of oxygen suffered higher rates of death than infants assigned to the upper range, while the latter suffered higher rates of retinopathy than the former. Defenders accuse OHRP of faulting the investigators for failing to inform parents of risks they learned about only through the study.
The central point of disagreement between defenders and critics of the study appears to be whether participants in the study were receiving medical care that was different from the care they would have received outside the study and whether participation in research therefore carried any medical risks that required risk/benefit scrutiny by IRBs or disclosure to parents of the infants enrolled. This would appear to be a factual matter about which one could obtain some clarity, but discussions of this issue have been somewhat opaque.
Part of the reason for this may be different, but unacknowledged, understandings of the concept of “standard of care,” a term used – both in the informed consent forms and the commentary about them – but rarely defined in this debate.
Defenders of the study, as well as the investigators themselves, insist that because the lower (85 percent to 89 percent) and higher (91 percent to 95 percent) ranges of oxygen saturation provided to the infants were within the standard of care, understood as a range of 85 percent to 95 percent, there were no known risks to participants in the study. David Magnus and Art Caplan write, “Given that there was variation in clinical practice [within the guideline-specified range of 85 percent to 95 percent] at the time the study was mounted, it is not clear how randomization among treatment options could have created novel risk over random physician preference.” John Lantos writes, “Every baby in the study received treatment that could have and perhaps would have been given to babies who were not enrolled in the trial.” The understanding of “standard of care” used by these commentators seems to lean toward the legal concept of that name. In an ordinary malpractice case, physicians will be liable only if they fail to provide minimally competent care, care outside the range of what reasonable physicians would provide. Appropriately, the legal standard of care is broad and often allows for many different approaches.
Critics of the study – namely OHRP and Public Citizen – use a different understanding of standard of care. They are focused on either what physicians, or in this case, institutions, typically do or what a particular physician or institution typically does, not the range of what physicians can or might do without risking legal liability or criticism from their peers. OHRP’s determination letter relating to the study indicates that, at the time the study was initiated, participating institutions allowed oxygen saturation levels to range between 85 percent and 95 percent, rather than confining levels to either the lower or upper portion of that range. This led OHRP to conclude that the level of oxygen received by infants in the study “was different from what they would have received had they not participated in the study.” Some received more oxygen than they otherwise would have, thus facing a greater, and foreseeable, risk of blindness, while others received less oxygen than they otherwise would have, thus facing a greater, and foreseeable, risk of brain injury and death. A recent analysis by Public Citizen also focuses on what the usual practice was and, based on newly released documents, reports that the SUPPORT protocol itself stated that the higher oxygen-saturation range was “more conventional.” In addition, two of the informed consent forms (the only ones that include this information) suggest that usual practices tended toward a target range of 88 percent to 92 or 94 percent.
This is the crux of the problem. Which of these two understandings is most appropriate for evaluating and disclosing risk of harm in a research context? Whether the care provided will be within the bounds of acceptable practice or whether it will be the same as what the patient would receive outside of research?
Defenders of the study have expressed outrage over OHRP’s action. Lantos criticizes OHRP (and Public Citizen) of endorsing “an anti-intellectual, unscientific approach to medical innovation” and concludes that “the real risk to babies come from reckless and ill-informed opinion about highly ethical scientific studies.” In a more recent essay, he accuses OHRP of wanting doctors to stop doing research and to rely instead on “mystical certainties.” Magnus and Caplan are more restrained. They charge that OHRP’s action “poses substantial risk” to the conduct of valuable research.
But it is not out of bounds for a research oversight system responsible for protecting subjects to ask – similar to the informed consent standards for medical treatment – “What do reasonable people who are considering enrollment in a study want to know?” If this is the question asked, most likely the answer would be that they would want to know what reasonably prudent physicians typically, or usually, or conventionally do in situations similar to that of their child’s. Given such information, they would better understand that enrolling in research means allowing physicians to do something different from the medical care they would normally receive, with potential attendant risks and benefits.
Defenders of the study suggest that because physicians didn’t know the optimal level of oxygen saturation, any particular infant’s physician, or NICU, could have reasonably chosen to hover within the lower or upper range. But did they? And if they didn’t, why not? These are questions to which the parents in the SUPPORT study had a right to answers. Because they were told that their children would receive standard of care, they couldn’t know to ask them.
Lois Shepherd, J.D., is the Peter A. Wallenborn, Jr. and Dolly F. Wallenborn Professor of Biomedical Ethics, Professor of Public Health Sciences, and Professor of Law at the University of Virginia.