Premature babies and informed consent: we need to do it right

The New York Times reported April 10, 2013 that an investigation by the Department of Health and Human Services “has found that a number of prestigious universities failed to tell more than a thousand families in a government-financed study of oxygen levels for extremely premature babies that the risks could include increased chances of blindness or death.”  “Study of Babies Did Not Disclose Risks, U.S. Finds”, by Barbara Tavernise reports that the study of the use of oxygen in 1,300 infants born at 24-27 weeks of gestation during 2004-2009 and published in the  New England Journal of Medicine in 2010, “did have an effect on which infants died and which developed blindness, and that those risks were not properly communicated to the parents, depriving them of information needed to decide whether to participate,” and that these  “…conclusions were listed in great detail in a letter last month to the University of Alabama at Birmingham [UAB], the lead site in the study,” which included 23 academic institutions, including Stanford, Duke and Yale.

The science is complex. Very premature infants such as those studied are at high risk of death and other complications, including blindness. It has long been known that the use of very high concentrations of oxygen, near 100% (room air is 21% oxygen), causes them to have an increased risk of blindness. The Times notes that “Clinical treatment of premature infants has a troubled history. Attempts to treat them with higher oxygen levels that were thought to improve their odds of survival led to many cases of blindness. Premature babies need oxygen because their lungs are underdeveloped and they often need help breathing.” The study under investigation was using lower concentrations, 85-95%, in an effort to find a level that was most beneficial for survival without having an increased risk of serious outcomes, such as blindness. The response letter from the lead investigator at UAB argues that a similar (but not formal control group) of infants had a higher mortality rate than those in the study group; this is criticized by others who note that it is not valid science to retrospectively attempt to create a comparison group because they may be different in important ways. Usually, participants are randomly assigned to either receive the intervention or not, because this means that any differences that may exist, known or not, are likely to be equally present in both groups. In fact, one critic, Michael Carome of Public Citizen’s Group on Health, points to a study published in the journal Pediatrics in 2012 that noted the babies in the other group were sicker and thus more prone to die.

But what is of concern to the HHS investigators, and to me, is that the parents of the babies in the study were apparently not informed of the risk of blindness that is known to accompany high-level oxygen therapy, despite the study (and its consent form) having been approved by the Institutional Review Boards (IRBs) of all 23 participant institutions. This is shocking on the face of it, but is even more surprising given that consent forms have largely become, as the Times article quotes Arthur Caplan, head of the division of medical ethics at New York University Langone Medical Center, “...captured by worries about legal liability, so risks tend to come billowing forward like a huge fog… It’s a truth dump, so they are covered should something go wrong.” He is absolutely correct; consent forms commonly include a list of potential negative outcomes so long as to obfuscate what is important, e.g., “rash, fever, nausea, vomiting, headaches, bleeding, seizures and death”. (NOTE: This is MY rendition of what I think is typical; it is NOT the list contained on the premature infant oxygen study.) What is a patient or parent of a patient to think? How bad are these outcomes? How likely are they to occur? How can I make a decision? And, perhaps most important, how can I give a well-thought-out response, giving or withholding consent, when I am – or my baby is – terribly sick and at high risk. What usually happens, as Dr. Caplan notes, is that “…often in such emergency medical situations, parents often rely more on what doctors say in deciding whether to participate than on the fine print of a consent form.“

Much, if not most, research, even medical research, is done with people who are not at high or immediate risk; much early clinical research, as well as that in the social sciences, specifically looks at “normal volunteers” or at regular people in the community. All require informed consent by participants. There are (at least) 3 different sets of interests in any research. The researcher is interested in finding out the answer to a question, which in clinical research is usually about whether one type of treatment is better than another. The patient (or guardian), whose motivation, according to the urgency of the situation, may be to contribute to expanding medical knowledge, to earn some money (if there is an honorarium involved), to help others in the future who may confront the same problem, or, especially when the situation is very urgent and the patient very sick (as in this study) to hopefully get immediate benefit for themselves or their child. The institution in which the study is conducted (represented by their legal department) is interested in guarding, to the extent possible, against risk of lawsuits.

Interestingly, the allegation in this current study is the failure to mention the potential outcome of blindness, more a typical failing of the “bad old days” when patients were not informed at all of risks (e.g., the Tuskegee syphilis studies) than of the “bad new days” where the risk aversion of universities’ legal departments has created the type of laundry list of bad outcomes noted above. But both are bad. Consent cannot be “informed” if typical, regular, non-lawyer people cannot understand the consent form because the language is so abstruse, or if the list of potential outcomes is a jumble of serious and minor ones. This is particularly true when a patient  is critically ill, and they or their guardian is under incredible stress. This is when, as Dr. Caplan notes, they rely on what the doctors say; unfortunately this puts the doctor in the position of trying to dispassionately provide a recommendation as to whether or not to participate in a study that s/he usually feels passionately about. This is not a setting conducive to true informed consent.

The literature on informed consent is enormous, and it is a major focus of the field of medical ethics. But the key principle must be that participants be adequately informed of the risks that may exist, in language that they can easily understand. The risks must be presented with the most serious ones highlighted, along with the probability that they will occur. They must be presented openly and dispassionately, and study leaders must be available to honestly and completely answer questions. There is, unfortunately, an inverse relationship between the comprehensibility of a consent document and the degree to which it is written in legal language (as anyone who has ever tried to read anything written in legal language can believe).

Consents must be appropriate to the situation. Researchers who wish to, say, interview people outside a local supermarket to find out what they think are health problems in their community, should not have to have those people sign a 3-page consent to participate or no one would do it. Researchers who wish to study clinical interventions in high-risk patients in urgent circumstances need to provide adequate information about risks and benefits in a way that patients can understand, and not use their role as the “doctor” to “sell” a study because they believe the intervention will be beneficial even though the results are not yet in. IRBs need to take their responsibility to protect participants’ interests and be truly informed, while facilitating the conduct of research seriously. It’s not easy. Lawyers need to learn how to write in a way that folks can understand.

Clearly, leaving out a major potential complication like blindness from a consent form is wrong, but there is much more to getting true informed consent than a simple fix. But it has to happen if we are to continue to be able to conduct research in order to have new knowledge.