Response concerning two papers published in the Apr 2013 JAMA on ACGME duty hour restrictions

I agree that we ought to let data guide policy on duty hours, but by my reading the data are nowhere near as clear cut as they have been made out to be by opponents of duty hour restrictions. The 2013 papers from both Desai and Sen suffer from such deep methodological flaws that they are of little use, if any, in determining the real impacts of duty hour restrictions. Both studies rely primarily on subjective, rather than objective measures of quality of care – in fact, Sen et al do not have a single objective measure – and are, by necessity, unblinded. This opens the analysis to a crippling bias that neither team addresses.

We all know that if one wishes to assess the affects of an intervention, one must either rely on objective measures of success, or be able to conceal from the participants which arm of the study they are in. It is preferable to do both, of course, but a study that does neither is nothing more than an fancy opinion poll. It is already well known that duty hour restrictions face considerable opposition within some quarters of the American medical community, and these studies merely confirm this fact. If the FDA were to approve a drug on the basis of an industry-sponsored survey that asked doctors to rate their opinion of the study drug versus a control, we would be outraged. But that is the level of evidence we are dealing with here.

It is hard to judge the impact of duty hour limits outside the context of the safety and educational reforms that were intended to accompany them. Nasca et al, writing for the ACGME in the July 2010 NEJM, remarked that duty hour limits were never meant to solve all of the problems of the American hospital system. Unfortunately, however, “[...] duty hours remained the primary focus for programs and institutions; larger changes in the learning environment that were envisioned when the duty-hour standards were instituted in 2003 never materialized.” If, during a trial comparing a new surgical procedure to medical management, it was learned that several of the operating rooms had been contaminated, we would need to repeat the study; if, in fact, the present duty hour limit regime is harming patients, we will not know whether the restrictions themselves are to blame, or whether the problem is the medical community’s failure to implement a more comprehensive reform.

But at the moment, unless there is better work out there of which I’m not aware, we can’t make any claims regarding the effect of duty hour restrictions, pro or con. On the other hand, it has been shown that interns are more likely to be involved in motor vehicle collisions when driving after working extended shifts (Barger et al, NEJM 2005). There are those who insist that a person who is too tired to drive is not too tired to care for patients, but we must have evidence to accept this counter-intuitive claim. If there’s data – not glorified opinion polls, not alarming anecdotes, not nostalgic paeans to the good old days, but real data on patient outcomes – then let’s see that data. Otherwise, I can’t agree that moving back to the old system is a viable option.

Submit comment