Bioethics Forum Essay

Public Citizen and Misinformed Consent in Neonatal Intensive Care

Public Citizen, the so-called “citizen’s advocacy group,” continues to criticize the NIH-sponsored clinical trials of oxygen therapy for premature babies.  They followed up their April 10th letter with another, on May 8th.  The April letter was wrong on a number of counts, as I documented in Bioethics Forum on April 18th.  The latest document is worse. The authors of the previous letter, Sidney Wolfe and Michael Carome, are joined by bioethicist Ruth Macklin. Together, they manage to mangle the facts, misrepresent clinical research, and misinform the citizens for whom they claim to advocate.

Their central claim is that the “consent form failed to mention the serious, reasonably foreseeable risks related to the part of the study comparing two experimental strategies for managing oxygen in extremely premature infants. Those risks, correctly identified by OHRP, included increased risks of brain injury; an eye disease called retinopathy of prematurity, which can lead to blindness in severe cases; and death, depending on the randomized group assignment of each baby.”

The problem here is that those risks were not “correctly identified.” Neither Public Citizen nor OHRP understand the differences between, on the one hand, the risks of enrolling in the research project and, on the other, the risks associated with being an extremely premature baby. All of the risks that they talk about are risks associated with neonatal intensive care for premature babies. Not one of these risks was increased by being in the study.

The state of knowledge regarding oxygen therapy at the time this study was designed and initiated can found in many papers that were written at the time.  Ten years ago, Cole and colleagues summarized what was known about oxygen therapy for premature babies.  They wrote, “We do not understand optimal oxygenation management in extremely low gestational age neonates (<28 weeks’ gestation).  No randomized control trial has clarified the relation between retinopathy of prematurity (ROP) and blood oxygen (PaO2), transcutaneous oxygen (tcO2), or oxygen saturation (SpO2) levels.”[i]

Given that, here is a partial list of the information that the new letter insists should have been included in the consent form, but which is in fact incorrect:

  1. That enrolling babies in the study would expose them to an increased risk of brain injury compared to babies who were not in the study. In fact, there was no evidence prior to the study that oxygen therapy in the ranges being used in the study would lead to an increased risk of brain injury. And no evidence of such increased risk was found in the study.
  2. That enrolling babies in the study would expose the babies to an increased risk of retinopathy of prematurity and/or blindness. In fact, there was no evidence prior to the study that enrolling babies in the study would lead to an increased risk of retinopathy or blindness. And it did not.
  3. That enrolling babies in the study would expose babies to an increased risk of death. In fact, there was no evidence that enrolling in the study would lead to an increased risk of death. And it did not.
  4. That doctors, using clinical judgment, would have made clinical decisions that would have resulted in better outcomes than doctors following the study protocol. There was, in fact, no evidence that this was true. And, as it turned out, it was not.
  5. That the technique of adjusting oxygen saturation monitors so that they masked doctors from seeing the actual oxygen saturations within a narrow range led to more risky treatments. In fact, there was no reason to believe that this would happen if the doctors followed the protocol. And it did not happen. There was no difference in intubations between the two groups.

It would be hard to know from Public Citizen’s cascade of misinformation that babies in the SUPPORT study (or, for that matter, in similar studies carried out in many other countries)  had better outcomes than babies who were not in the study. The babies in the studies had lower rates of death, lower rates of disability, lower rates of retinopathy, and were less likely to undergo invasive procedures such as intubation than were babies who were not in the studies. The data showing this comes from comparisons of the outcomes of babies who were in the studies with any and every possible comparison group – babies who were eligible for the studies but not enrolled, babies who were treated at the same time (or a few years before or a few years after) in the same countries, and babies in other countries who had the same conditions as babies in the studies.

Take mortality rates. For babies in the study, mortality rates were 15.3 percent to16.6 percent in Canada, 16.6 percent to 19.2 percent in Australia, New Zealand, and the UK, and 16 percent to 20 percent in the US.  For babies in the U.S. who were not in the studies, mortality rates for babies born in 2003-2007 who were not in the study were 24 percent.[ii]  In Victoria, Australia, mortality rates for such babies born in 2005 were 25 percent. For all of Australia, in 2010, mortality rates were 20 percent. For all of Canada in 2010, mortality rates were 23 percent for babies at these gestational ages. That is, mortality rates for babies in the study were 20 percent to 40 percent lower than for babies who were not in the study. Babies in the studies also had lower rates of retinopathy, blindness, and cerebral palsy compared to babies born at the same gestational ages in those same countries.

Public Citizen dismisses the relevance of these data by saying, “It is not clear why the investigators think these data are important or relevant.”

So let me explain. An informed consent form is supposed to inform parents of the risks of a study. It should do so by explaining to them what is known or believed to be true. So the first criteria for judging such forms is to look at what was known at the time they were written. By those criteria, the informed consent forms used in the SUPPORT study were adequate and accurate. It is possible that, in retrospect, we come to understand new things about the risks or benefits of an innovative treatment. That can be assessed by looking at data from studies and is an additional way of evaluating whether consent forms were adequate.

Before the SUPPORT study was done, a large meta-analysis concluded that there would be no difference in death or neurocognitive impairment using restricted oxygen versus liberal oxygen. There was a suggestion that restrictive oxygen would lead to lower rates of retinopathy.   That, then, is what the consent form stated. The fact that a post-hoc analysis showed that, compared to babies who weren’t in the study, those in both arms of the study had better outcomes is interesting and important. It suggests directions for future research. It suggests that there may be benefits to being in a study, benefits that an informed consent form ought to mention.

Clearly, the studies were well-designed, they answered important clinical questions, and they did so without causing harm.

So a mystery remains. Why would a so-called “citizen advocacy group” insist that consent documents for such safe and well-designed studies ought to misinform parents of the true risks of enrolling their babies in studies? And why they would now want to shut down other studies that are also well-designed and ethically appropriate?

The explanation for that mystery was, I believe, given decades ago by Dr. William Silverman. Silverman was a pioneer of neonatology. He was one of the first doctors to warn about the toxicity of high-dose oxygen. He was a passionate believer in clinical studies to improve the safety and efficacy of neonatology. He was familiar with the types of arguments that are now being made by Public Citizen. He identified them as a belief in “mystical certainty” rather than an acceptance of “scientific uncertainty.” He noted, “Doctors are viewed suspiciously when they ask questions – a switch from their accustomed role as providers of answers.”

Neonatologists today want to investigate the safety and efficacy of commonly used but inadequately studied therapies. They develop an unprecedented international collaboration to do so. The protocols are carefully reviewed by IRBs. The consent forms are reviewed for accuracy and understandability. Parents are accurately informed of the risks and benefits of the research. The studies are done in a careful and prospective manner. Babies in the studies are protected from the risks of therapies that have not been validated and have better outcomes than babies who are not in the studies. The results of the studies could help many other babies.

But then federal agencies whose task is to insure the conduct of responsible research criticize the studies. They suggest that parents should have been warned against the studies by being given false and misleading information. They suggest that babies would be better off if doctors would ignore their scientific uncertainty, stop doing such research, and go back to the era of mystical certainties. Who would have thought we’d see the day when the Office of Human Research Protection and Public Citizen would be insisting upon misinformed consent?  With advocates like these, who needs enemies?

John B. Lantos, M.D., a Hastings Center Fellow, is director of pediatric at Children’s Mercy Hospital and the University of Missouri-Kansas City School of Medicine. He was not involved with the SUPPORT study. Children’s Mercy Hospital is now in the NICDH Neonatal Research Network but was not at the time of the SUPPORT study.



[i] Cole CH,  Wright KW, Tarnow-Mordi W, Phelps DL. Resolving our uncertainty about oxygen therapy.  Pediatrics 2003; 112:1415-9.

[ii] Stoll BJ, Hansen NI, Bell EF.  Neonatal outcomes of extremely premature infants from the NICDH neonatal research network.  Pediatrics, 2010; 126:443.

Posted by Susan Gilbert at 05/16/2013 08:48:27 AM |

Read More Like This