For more than a decade, the medical community has been concerned about continuity of care. Concern began when the Accreditation Council for Graduate Medical Education implemented national duty hour limitations in 2003. These rules became more restrictive in 2010. The current rules include an 80-hour-per-week limit, no more than 16 hours consecutively (for interns) or 28 hours (for senior residents), and at least 8 hours off between shifts. With the blessing of the Accreditation Council for Graduate Medical Education, 2 multicenter trials are beginning to study the impact of a return to less restrictive rules.
The root of the concern about duty hour limits is whether the physician fatigue that existed prior to the duty hour rules is more dangerous than the resulting discontinuity caused by shorter shifts and more hand-offs. However, an ideal shift length is not necessarily known or knowable. A 2010 systematic review examined shift length and patient outcomes, and no ideal shift length could be determined, although shorter shifts were associated with fewer errors in some studies. A more recent study showed no difference in intensive care unit patient outcomes with 12-hour, 16-hour, or 24-hour housestaff shifts.
The relationship between fatigue, shift length, hand-offs, and discontinuity is complex. It is clear that human performance suffers in the setting of too much fatigue. However, fatigue is not only a result of shift length; sleep debt, time of day, and other factors also play a role. Some of these factors are within the control of residency programs (scheduling to avoid sleep debt) and some are not (the amount of sleep individuals get when not on duty). Minimizing fatigue while minimizing discontinuity is a challenge. To understand how important discontinuity is to patient safety, we conducted the following study to examine objectively measured aspects of discontinuity and their impact on patient adverse events (AEs). We hypothesized that less continuity would increase the risk of AEs.
Methods
This prospective cohort study of physician schedules with retrospective chart review of patient outcomes was conducted at 3 sites affiliated with the internal medicine residency program: an academic Veterans Affairs Medical Center (VAMC); a community teaching hospital; and a tertiary care private teaching hospital. The community hospital also has its own transitional-year intern program (see Appendix A, available online, for program details).
We recruited housestaff assigned to general medicine services between March 16, 2009 and March 15, 2010. We did not include medical students. Physicians gave written informed consent. For a team to participate in the study, all physicians on the team had to consent.
A random sample of patients assigned to each participating team was selected each week, except for the 3 weeks surrounding the Christmas and New Year holidays. We excluded those times because we did not have staff available to collect data. We excluded patients that were 1) not on a participating team; 2) admitted directly to an intensive care unit; 3) cared for by a fourth-year medical student as an “acting intern;” 4) hospitalized for <48 hours (because many AEs require 48 hours to be labeled hospital-acquired); and 5) assigned to observation status at Froedtert Memorial Lutheran Hospital because those patients were admitted only to the hospitalist service. We obtained a waiver of informed consent for patients.
To read this article in its entirety and to view additional images please visit our website.
-Kathlyn E. Fletcher, MD, MA, Siddhartha Singh, MD, MS, Marilyn M. Schapira, MD, MPH1, Vishal Ratkalkar, MD, Alexis M. Visotcky, MS, Purushottam Laud, PhD, Christa Kallio, RN, Susan Framberg, RN, Jianing Li, PhD, Andrew Kordus, BS, Jeff Whittle, MD, MPH
This article originally appeared in the March 2016 issue of The American Journal of Medicine.
