iCOMPARE, what value does it add to resident duty-hour discussions?

Discussions regarding resident duty-hour restrictions have been ongoing and heated. One influential argument for restrictions has been patient safety. Two trials, FIRST and iCOMPARE, were performed to investigate this relationship with surgical and medicine training, respectively. As the authors are approaching this discussion from a medicine-based perspective, iCOMPARE will serve as the primary basis of our discussion. Results from the iCOMPARE trial comparing flexible (28-hour shifts allowed) to the original 2011 ACGME shift requirements (maximum 16 hours) were recently published in the New England Journal of Medicine. This non-inferiority trial used 30-day post-hospitalization mortality as its primary endpoint. Results met qualifications for non-inferiority, and ACGME policy was changed to allow for 28-hour shifts for medicine residents. iCOMPARE results were highly lauded and used as primary justification for extending resident duty hours. Despite this sweeping impact, few have critically evaluated what this study actually adds to the literature. Herein, we argue that serious questions regarding trial design are apparent. Most importantly, the non-inferiority margins chosen were large, and represent an ambiguous marker of resident performance. Additionally, we question the lack of both patient consenting and direct patient-reported or patient-centered outcomes within the hospital stay. As more discussion arises in the medical literature surrounding patient-reported outcomes and shared decision making, we argue that the results of iCOMPARE disregarded the patient perspective or meaningful patient outcomes in an attempt to maintain status quo. Lastly, we discuss how iCOMPARE missed the broader question of actual duty-hour restrictions, and some practical methods already in practice at some programs, which may more directly balance resident work hours with patient care and resident learning.