Jeffrey Clark and David Harari are resident physicians at the University of Washington specializing in psychiatry.

In 1989, Sidney Zion famously wrote this about staffing practices at U.S. hospitals: “You don’t need kindergarten to know that a resident working a 36-hour shift is in no condition to make any kind of judgment call — forget about life-and-death.”

After his 18-year-old daughter, Libby, died at a New York hospital in 1984 while under the care of junior physicians stretched dangerously thin, Zion pushed to change the system. But the system has shown a stunning ability to deny the obvious. Even after a 2009 report from the prestigious Institute of Medicine confirmed Zion’s suspicion that shifts beyond 16 hours are risky, some still want to test the theory that patients are well-served by newly trained doctors who have been awake and working for 30 or more consecutive hours.

The proponents of 30-plus-hour shifts argue that new limits on residents’ duty hours have not consistently led to improvement in patient outcomes. But why would anyone expect otherwise? The standards published in 2011 by the Accreditation Council for Graduate Medical Education still allow hospitals to put residents through blistering 80-hour work weeks, while setting maximum shift lengths of only 16 hours for interns and 24 hours for more senior residents. Interns simply work shorter but more-frequent shifts. Doctors hand off patients to each other more regularly but without the training needed to manage these transitions effectively. And, by and large, hospitals have not responded to the changes with larger workforces, leaving residents no choice but to compress their daily work into shorter time periods.

Now, two randomized studies at major teaching hospitals across the country are comparing residency programs that use the 2011 duty-hour regimen with those employing a more flexible schedule. Residents in the flexible arm can work shifts of unlimited length provided they log fewer than 80 hours a week and don’t work extended shifts more often than every third night. The scientists running the studies, known as FIRST and iCompare, theorize that mortality rates could be no worse for patients, while residents might be better educated, if shift lengths of 30-plus hours were reinstated.

We believe there are significant ethical, scientific and regulatory problems with these trials. First, the trial investigators at Northwestern University (FIRST) and the University of Pennsylvania (iCompare) have exposed residents to well-documented increased risks of motor vehicle accidents and needle-stick injuries. There is also good reason to believe that overwork and sleep deprivation contribute to the epidemic of depression among resident doctors.

Second, after erroneously upholding that these studies presented only “minimal risk” to doctors and patients, the institutional review boards responsible for human research protection at the two universities waived the requirement for obtaining informed consent for all subjects. This means that patients and doctors alike did not need to be notified that they were enrolled in a study, did not have a chance to provide informed consent and could not reasonably opt out. (The two of us and our patients were not provided informed consent before being enrolled in the iCompare trial). Conducting this research without the informed consent of residents and patients violates the basic ethical principle of respect for persons.

Third, a high-quality survey sponsored by the Committee of Interns and Residents and the advocacy group Public Citizen showed that the public overwhelmingly opposes extended hours for residents. Only 1 percent of respondents approved of shifts greater than 24 hours, and four out of five said they would ask for another doctor if they found out that their resident had been awake for more than 24 hours. Treating patients with an intervention that they do not want should never be sanctioned.

Ultimately, the biggest problem is the question being studied. We already know that extended shifts are dangerous. While many people rightfully suspect that current duty-hour limits aren’t improving outcomes, these studies err in assuming that the dangers of sleep deprivation must be traded for the dangers of shared patient care. Such a zero-sum framework won’t help us improve patient care or ensure the well-being of resident physicians.

Perhaps we should start over. For more than 100 years, we have tried to train doctors to live without adequate sleep, and yet we have predictably failed to produce superhumans. Instead, we’ve created a medical culture that encourages severely sleep-deprived, impaired physicians to take care of others. Does anyone want this?

There is no reason to believe that 80-hour workweeks and shifts longer than 16 hours are associated with optimal patient or resident health. Adequate sleep is a fundamental physiological need. No amount of caffeine, prescription stimulants (as some physician leaders have advocated for) or “alertness management strategies” can adequately compensate for acute and chronic sleep deprivation.

But neither should anyone assume that more reasonable duty hours are a panacea. We also need to improve patient handoffs, increase staffing and emphasize resident education over patient volume. These things will require time and money. But as a profession, we can begin by abandoning the archaic, harmful and misguided belief that doctors are immune to normal human limitations. Until that happens, we can’t meaningfully address the parts of residency training that we can indeed change.