Questions About Informed Consent:
A Recent Case in the News

Research on human subjects is tricky business. One ethical standard, among many, is to have subjects give informed consent to the experimental procedures. Informed consent includes understanding the risks and study design (i.e., whether it is placebo-controlled).

Recently, a study on premature infants has come under attack. The study, “Target Ranges of Oxygen Saturation in Extremely Preterm Infants,” was published in the New England Journal of Medicine (2010; 362:1959-1969). The attacks have come from several quarters and are nicely summarized on the website of the Alliance for Human Research Protections ( At issue is the accuracy of the informed consent documents.

The study tested different levels of oxygen for premature infants. It was known before doing the study that high levels of blood oxygen saturation levels (around 95%) were associated with retinopathy (blindness) and two studies indicated that oxygen toxicity was associated with death as well. Of course, the authors also noted that studies from the 1950s and 1960s, as well as observational studies suggested an increased risk of mortality for infants receiving lower oxygen levels.

So what are the complaints? AHRP summarizes the informed consent issues as follows:

NONE of the 22 consent forms explained that death was a risk—especially for babies randomized to low oxygen levels;

Only two of 22 consent forms disclosed that babies randomized to high oxygen levels were at increased risk of retinopathy or blindness. 

NONE of the 22 consent forms disclosed that the pulse oximeter readings—that neonatologists rely on to determine an infants’ need for supplementary oxygen—were intentionally altered to provide the treating medical teams with either false high or false low oxygen saturation values (SOP2) (

Although the researchers noted in the original article that there was evidence of mortality for babies receiving lower oxygen saturation, this was not mentioned. It was confusing to read the researchers’ defense of the trial in a letter to the editor of the New York Times where they say,

When the study was planned, the best evidence showed that lower oxygen targets — even lower than used in the study — resulted in less eye disease without a higher death rate. The finding of a higher death rate in one study group was not anticipated (, emphasis added).

And this comment is in tension with the following comment which is culled from the original article, “Despite the increase in mortality when restrictive oxygen supplementation was used in the 1950s and 1960s and the limited data from observational studies,… (emphasis added)” It appears then, that an increased risk of death is enough of a possibility to list it in an informed consent document.

Of course, not every possibility needs to be documented in an informed consent document – if they were, this would actually compromise informed consent as it would involve “information overload,” as they say. Researchers are stuck with making prudential decisions about which risks are most important.

I reserve judgment on this specific study, but it is illustrative of several general points: (i) if you are a research subject, ask for a verbal explanation of the risks; (ii) if you are a researcher, ask yourself two questions: would you want your mother to enroll in your study, and what do you think your mother would want to know before enrolling? (iii) which risks a researcher mentions is a judgment call, but all serious risks, however improbable, should be includedCthe risks should not be translated into terms such as “slight” or “rare”. Evidence strongly suggests that subjects comprehend better risks in terms of both numerical probability (e.g., 1/10) paired with examples from everyday living (for example, “the probability of death for this procedure is similar to the probability of being attacked by a shark while swimming off the coast of Montauk Bay”). 


Stephen Napier Ph.D. CIP. (Certified IRB Professional)
Consultant, National Catholic Bioethics Center