Brief (<4 hr) sleep episodes are insufficient for restoring performance in first-year resident physicians working overnight extended-duration work shifts

Abstract

Study objectives: The Accreditation Council for Graduate Medical Education (ACGME) recently reinstated extended-duration (24-28 hr) work shifts (EDWS) for postgraduate year 1 (PGY-1) resident physicians. This study examined the relationship between overnight sleep duration during EDWS and subsequent "post-call" performance in PGY-1 resident physicians.

Methods: Thirty-four PGY-1 resident physicians (23 males; 24-32 years) were studied between 2002 and 2004 during 3-week Q3 "on-call" rotation schedules in the Medical and Cardiac Intensive Care Units at Brigham and Women's Hospital in Boston. Daily sleep logs (validated by ambulatory polysomnography) were collected and the 10 min psychomotor vigilance task (PVT) was administered every ~6 hr during each EDWS. Generalized estimating equations were used to examine the relationship between overnight sleep duration and PVT performance "post-call" (0500-1900 hr). Postcall performance during EDWS was compared with sessions matched for time-of-day and weeks-into-schedule in the same resident physician during an intervention schedule that eliminated EDWS.

Results: Resident physicians obtained an average of 1.6 ± 1.5 hr cumulative sleep overnight during EDWS (<4 hr on 92% of nights). PVT attentional failures were significantly reduced only after >4 hr sleep (p = 0.027 versus no sleep). Despite this apparent improvement, the odds of incurring >1 attentional failure were 2.72 times higher during postcall following >4 hr sleep compared with matched sessions during non-EDWS.

Conclusions: Even with >4 hr sleep overnight (8% of EDWS), performance remained significantly impaired. These findings suggest that even "strategic napping," a recommendation recently removed from ACGME guidelines, is insufficient to mitigate severe performance impairment introduced by extending duty beyond 16 hr.

Keywords: medical errors; medical training; patient care; resident physicians; sleep deprivation.

Figure 1.

Attentional failures (number of lapses, normalized by $\sqrt{n}+\sqrt{n+1}$) from all 295 sessions of the PVT included in the current analysis are plotted by sleep duration bin. Individual points within a bin have been offset to show overlapping points. The median and interquartile range (boxplots) and mean (open squares) also are plotted for each sleep duration bin.

Figure 2.

Stacked bar chart comparing the percentage of PVT sessions following 0–4 hr of spontaneous sleep and >4 hr of sleep in which the number of attentional failures fell between the designated thresholds among PGY-1 resident physicians. White bars: at or below 1 lapses; light gray bars: between 2 and 5 lapses; dark gray bars: between 6 and 10 lapses; black bars: >10 lapses.