Response

To the Editor:

We appreciate the comments of Drs Sriratanaviriyakul and Albertson and Dr Rho and colleagues in response to our study of nighttime intensivist staffing.¹ We agree wholeheartedly with Drs Sriratanaviriyakul and Albertson that mortality has significant limitations as an outcome in this and all studies of ICUs, as we discussed briefly in our recent article.¹ Further, as we note in our review of critical care randomized trials, mortality is widely accepted as the primary outcome in research on critically ill patients, but it has a number of methodologic and conceptual limitations.^2,3 Indeed, our study may be the first to systematically address this limitation by directly examining patterns of limitations on life support. However, we acknowledge that this effort was exploratory in nature and does not address all the weaknesses of mortality as a critical care outcome measure. We agree that further work is needed to more fully unpack mortality as an outcome and to identify other relevant outcomes in ICU research.

We also agree with Dr Rho and colleagues about the importance of understanding the impact of nighttime intensivist staffing on education, as we have noted previously in CHEST.⁴ However, as we described in that more thorough treatment of this important issue, we do not believe the educational outcomes to be studied are yet apparent. The majority of ICUs in this country, and indeed the world, are not staffed by medical trainees; thus, the key educational issue is the ultimate preparedness of future ICU staff physicians to function independently. However, the studies of educational outcomes thus far have focused primarily on perceptions of education and autonomy, which may or may not actually correlate with competency. Therefore, as nighttime staffing patterns in ICUs continue to evolve, with implications on our future workforce, we too hope that this area of research will continue to be pursued.