0969 Attentional Failures Are Correlated With Serious Medical Errors In Resident Physicians

Abstract

Introduction: First year postgraduate (PGY1) resident physicians have impaired vigilance and more attentional failures working extended duration work rosters (EDWR) with 24+ hour extended‐duration shifts compared to shifts scheduled to 16 or fewer hours. We examined the impact of a rapid cycling work roster (RCWR) intervention designed to limit continuous work to 16 hours maximum in PGY2 and PGY3 resident physicians. Methods: Resident physicians participated in a multi‐center cluster‐randomized crossover clinical trial of six U.S. pediatric intensive care units. Participants worked 4‐week rotations of either EDWR or RCWR schedules. Participants (n=294, 173 females; 358 resident‐physician rotations) completed sleep and work logs daily, and the 10‐minute psychomotor vigilance task (PVT) and the Karolinska Sleepiness Scale (KSS) approximately at the start, during, and end of selected work shifts. Resident‐physician associated serious medical errors (SMEs) were identified using both direct observation and chart review and adjudicated by independent physician‐reviewer pairs (k range: 0.52‐0.67). Results: On the RCWR intervention, attentional failures (lapses in reaction time ≥500 msec), mean reaction time, and subjective sleepiness were significantly improved relative to EDWR (p<0.04). Neurobehavioral performance was better under the RCWR throughout the work shift and across the 4‐week rotation (p<0.0001). Attentional failures measured on the PVT were on average significantly correlated with resident‐physician‐related SMEs during the rotations (p=0.04), as assessed from observation and retrospective chart review. While a significantly higher rate of SMEs was observed under the RCWR, after adjusting for resident‐physician workload (i.e., ICU patients per resident physicians) in a post‐hoc exploratory analysis, RCWR had a significantly protective effect on error rates [RR 0.51 (95% C.I. 0.32 ‐0.82)] compared to EDWR, although this result depended on how workload was measured and included in our model. Conclusion: RCWR implementation reduces the substantial impairment in neurobehavioral performance and subjective sleepiness associated with the EDWR. These data, and their correlation with serious medical error rates, highlight the impairment of neurobehavioral performance that occurs when resident physicians work extended‐duration shifts, and have important implications for patient safety.