Work-hour restrictions in the ICU

While I haven’t read the original study, based on this small snippet, the study design appears to be simple and elegant. I can’t understand the statement by Dr. Ellen Burnham tempering results due to the fact that the time period before and after was “pretty short.” Why is a longer time period necessary? The work-hours changes were immediate including effects on continuity of care, implementation of compensatory mechanisms such as wider use of physician extenders and increased attending presence, improved resident sleep, etc. The effects on patients should be equally immediate given the acute setting the study was conducted in. This isn’t a population based screening study requiring years before a mortality benefit is apparent. ICU patients either live or die in their short time in the unit. Increasing the study time would increase the number of patients (though I assume the study had sufficient power already) but wouldn’t effect the observation period for individual patients. However, if Dr. Burnham is suggesting the observation period is too short because we haven’t allowed ample time for the compensatory measures to fail, (as evidenced by her statements in the last paragraphs) this is both speculative and irrelevant with regard to the current study. In fact, the study argues why these mechanisms shouldn’t be allowed to fail.
It also struck me how the interviewer seemed (it’s hard to be 100% certain) to miss the point entirely when he asked if the implication of the study was that the effects of reform were “nonexistent”. His expectation, it seems, was that reducing work hours would have a beneficial effect for patients (reducing errors), an argument often touted by the pro-work-hours restriction crowd. Why is that the every change in medicine has to benefit patients? Can’t we, as Dr. Prassad argues, introduce some humanity towards doctors, regardless of patient benefit, as long as it doesn’t harm anyone? And speaking of harming patients, this study suggests that the much-touted plague of medical errors killing scored of patients may also be overblown. It would be interesting to see if medical errors were reduced during the same period of time in which no mortality change was observed. It wouldn’t surprise me if errors were reduced substantially.

Er, is that patient mortality or intern mortality?