While I was reporting on a study for The BMJ, I suddenly felt as if I’d walked through Alice’s Looking Glass.  You’ve possibly heard about the study by now: researchers found that patients treated by sleep-deprived resident doctors were no more likely to die or suffer serious complications than patients under the care of doctors working limited hours.

Of course the study seems a bit strange given what we know about doctors’ performance and sleep deprivation.  And while sleepy truckers are well-known to be deadly on the nation’s highways (as are pilots and others entrusted with the public welfare) it appears that some doctors want to believe they are immune to the biologic limits of other humans.

There are plenty of study design reasons to doubt the reliability of study outcomes, but it was the response of the editors of the New England Journal of Medicine to a question I asked about the ethics of the study that left me flummoxed.

But before I get to that, a little backstory: In November 2015, the Public Citizen’s Health Research Group and the American Medical Students Association roundly criticized the study as unethical, in part because resident physicians and their patients never gave informed consent.

That’s right; the researchers didn’t obtain informed consent from the 4,330 residents and 138,691 patients who served as guinea pigs for the study. This was apparently possible since the researchers declared the study to be a “non-human-subjects study.” The Northwestern University Institutional Review Board agreed, saying it determined prior to the start of the study that it was “non-human-subjects’ research, which of course does not require consent.”

I asked the lead author of the study, Karl Y. Bilimoria, how a study could be a “non-human-subjects study” when the outcome measured was patient deaths. He responded that the study had been approved by multiple parties, including bioethicists.  When I asked him to name relevant bioethicists, he referred me first to Dr Peter Angelos, professor of surgery and surgical ethics chief at the University of Chicago.  Angelos told me via email:

“I think that you raise a legitimate question over whether the IRB was correct in categorizing the study as not a human subjects trial, but what you absolutely cannot question is the importance of doing the study.  I would argue that even if you categorize the study as a human subjects trial, it should have been done.

OK. So, there’s a bit of strangeness here, and Dr Angelos explains that he wasn’t consulted until after the study was underway.

Of course, obtaining informed consent might have proved difficult in light of a 2010 survey that found 80% of individuals said they would request a different doctor if they learned he or she was working more than 24 hours. Or, perhaps they might also decline if they learned of certain earlier studies showing significant increases in medical error when doctors are sleep deprived.

But what left me puzzled was how the editors of the NEJM could justify publishing the article in light of their own ethics policies, which state that they will publish only studies that meet ethical standards, including informed consent.

So, I asked: Would they retract the article since test subjects never gave informed consent? The editors stood firm; a spokesperson said, “We believe the study was conducted in an ethical fashion. The investigators followed appropriate procedures, and the study protocol was reviewed by the institutional review board office at Northwestern University.”

My through-the-looking-glass moment came when I realized that although anyone can see that the study is indeed a study of human beings, as long as the researchers say it is not, the editors at the NEJM seem to be willing to go along with the pretense.

Shortly after that, I got a call back from Ezekiel Emanuel, chair of the Department of Medical Ethics and Health Policy at the University of Pennsylvania School of Medicine , who told me that although he had nothing to do with the study, he found it to be an “absolutely reasonable study.” When asked why informed consent wasn’t obtained, he too held firm, saying the study simply couldn’t have been performed if they had to obtain patient consent. “It just isn’t practical with a cluster-randomized trial,” he said.

Of course, practical or not, rules exist for a reason. And this rule, enshrined in the Nuremberg Code, which says participation of test subjects must be voluntary and informed, is a good one.

As Public Citizen’s Michael Carome warned, “the NEJM editors have signaled to the research community that the ethical lapses made by the FIRST trial researchers were acceptable, effectively encouraging similar lapses by future researchers.”

The researchers still hold that it wasn’t a human subjects research trial.

Which is sadly reminiscent of a Richard Pryor comedy when his wife catches him in bed with another woman, and Pryor’s character attempts to defuse her outrage by paraphrasing Groucho Marx, exclaiming, “Who are you gonna believe, me or your lying eyes?”

Jeanne Lenzer is a medical investigative journalist who sadly has to leave fascinating tidbits about various stories on the cutting room floor. The Backstory is an attempt to archive some of those bits and to provide a bit of insight for the public about the “behind-the-scenes” aspect of investigative journalism.

Competing interests: None declared.