I'm going to design a trial to look at work hours. First, I will randomize interns without their consent or knowledge, after the match, to work 16 or 30 hour shifts. If they try to change programs, I will count this against them. 1/

Though the only substantive change will be the shift length of the intern, the only outcome I'm going to look at is mortality. Even though mortality is almost impossible to change, if there is no difference between the groups, I will declare the work hours to be equivalent. 2/

Even though medical errors have been conclusively shown to increase with longer shifts, I'm not going to measure them. That would only cramp my style. 3/

I will also not look at car crashes, needle stick injuries, or depression among interns randomized without their consent, even though others have shown these to increase as well with longer shifts. 4/

I will measure the interns' quality of life, but I will ignore their reports of negative effects on their lives, and bury these answers in all press releases. 5/

I will then use my findings to state that work hour restrictions made no difference, and we can go back to longer shifts, even though the complexity and intensity of the shifts is increasing rapidly over time. 6/

Whenever someone complains about work hours, I will point them to my study, declare work hour restrictions dead, and frame it as a patient safety issue, a thing I never measured. 7/

I will then ask everyone who trained under the old regime to quote my study at any suggestion of reform. Bonus points if they point out how much tougher they are than the new crop of residents. 8/8

Sound familiar? @ShannonOMac @GoodishIntent @ShreyaTrivediMD @laxswamy @cirseiu @arghavan_salles @drjessigold