📄 medicine

Hours slept, not hours worked: predictors of next-day recovery in emergency physicians

This study of 44 emergency physicians reveals that while night shifts negatively impact next-day recovery primarily by reducing sleep duration, the amount of sleep obtained is a more significant predictor of recovery than shift length or its interaction with sleep.

Original authors: Siddharth Vipankumar Abrol Neena, Vishnunarayan Girishan Prabhu, Thomas W. Britt, Lauren A. Fowler, Kevin M. Taaffe, Emma C. Vosika, Sarah D. Meyer, Spoorthy Machineni, Emily L. Hirsh

Published 2026-10-07

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Siddharth Vipankumar Abrol Neena, Vishnunarayan Girishan Prabhu, Thomas W. Britt, Lauren A. Fowler, Kevin M. Taaffe, Emma C. Vosika, Sarah D. Meyer, Spoorthy Machineni, Emily L. Hirsh

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

In the high-stakes environment of a hospital emergency department, the work never stops. Doctors must be ready to treat patients at any hour, which means their schedules often rotate through mornings, afternoons, and the middle of the night. This constant shifting disrupts the body's natural internal clock, the biological rhythm that tells us when to be alert and when to rest. When a person stays awake during the time their body expects to sleep, or tries to sleep when their body expects to be awake, the restorative power of that rest is diminished. This is not just about feeling tired; it is about the body's ability to repair itself after the stress of a demanding shift. If a doctor cannot fully recover between shifts, their well-being suffers, and the safety of the patients they treat is at risk. Understanding exactly what helps a doctor recover—whether it is the length of the shift they just finished or the amount of sleep they managed to get afterward—is a critical question for keeping emergency care safe and effective.

A team of researchers set out to answer this question by following forty-four emergency physicians over the course of four weeks. Instead of asking doctors to guess how they felt over a long period, the team asked them to report their state every single day. Each morning, the doctors were asked to rate their recovery on a scale from zero to one hundred percent, essentially telling the researchers how rested they felt after their previous shift. The researchers also tracked how many hours each doctor slept and the specific type of shift they had worked, categorizing them as morning, afternoon, or night. By looking at the daily changes within the same individual, the study could separate the effects of a specific bad night from the general habits of a particular doctor.

The results offered a clear and somewhat surprising distinction between two common assumptions about fatigue. The study found that the most powerful predictor of how well a doctor recovered the next day was simply how much they slept. On days when a physician slept longer than their personal average, they reported significantly higher levels of recovery. In contrast, the length of the shift they had just worked did not independently predict how tired they would be the next day. While working a longer shift than usual was associated with slightly lower recovery, the difference was not statistically significant. Furthermore, the data showed no evidence that getting extra sleep could fully "buffer" or cancel out the negative effects of an unusually long shift. The two factors appeared to operate independently: more sleep helped recovery, but a longer shift did not automatically destroy it unless it also resulted in less sleep.

The type of shift played a major role, but primarily because it dictated how much sleep was possible. Night shifts were associated with the least amount of sleep, averaging just under six hours, while morning shifts allowed for the most, averaging over eight hours. Consequently, recovery scores were lowest after night shifts and highest after morning shifts. However, when the researchers accounted for the actual amount of sleep, the gap in recovery between different shift types narrowed considerably. Sleep duration explained a significant portion of why night shifts felt worse than morning ones. Yet, even after adjusting for sleep, a direct negative effect of night shifts remained. This suggests that while losing sleep is a major part of the problem, the timing of the sleep itself—occurring during the day when the body wants to be awake—also carries a unique cost that cannot be fixed by sleep duration alone.

The study concludes that protecting the opportunity for sleep is a vital lever for improving physician well-being. Schedules that allow for sufficient rest between shifts are crucial, particularly for those working at night. While extending the time between shifts might help, the research indicates that the quality and timing of the rest are just as important as the quantity. The findings suggest that simply working fewer hours is not a complete solution if the sleep that follows is still disrupted by the body's internal clock. For emergency physicians, the path to recovery lies in ensuring they can get enough sleep in a way that respects their biological rhythms, rather than just focusing on the number of hours they are on duty.

Technical Summary: Hours Slept, Not Hours Worked: Predictors of Next-Day Recovery in Emergency Physicians

Problem Statement
Emergency physicians (EPs) operate in high-demand, continuously staffed environments characterized by rotating schedules, extended shifts, and night work. These conditions often conflict with the circadian and homeostatic processes governing sleep-wake regulation, leading to circadian misalignment and accumulated sleep pressure. While existing literature establishes that insufficient sleep and shift work are linked to fatigue, burnout, and medical errors, prior evidence is predominantly cross-sectional or between-person. Such designs cannot disentangle whether a specific physician recovers less on days they sleep less, nor can they separate the effects of working a night shift from the inherent characteristics of physicians who work more night shifts. There is a critical gap in understanding the daily, within-person dynamics linking specific shift characteristics (duration and type), actual sleep duration, and next-day recovery.

Methodology
This study employed a 4-week repeated-measures field design involving 44 emergency physicians from a single hospital system in South Carolina. The analysis utilized 447 working-day observations.

Data Collection: Participants completed daily surveys assessing self-reported recovery and sleep duration (adapted from the Pittsburgh Sleep Quality Index). These self-reports were temporally aligned with administrative shift records (start/end times) to establish a sequence: Shift ( $t-1$ ) $\rightarrow$ Sleep Episode $\rightarrow$ Next-Day Recovery ( $t$ ).
Variables:
- Outcome: Self-reported recovery (0–100% scale).
- Predictors: Daily sleep duration, shift duration (hours), and shift type (Morning, Afternoon, Night).
- Covariates: Consecutive workdays.
Statistical Approach: Linear mixed-effects models with physician-specific random intercepts were used to account for the non-independence of repeated observations. Continuous predictors (sleep and shift duration) were decomposed into within-physician (day-to-day deviation from individual mean) and between-physician components to isolate daily fluctuations.
Analysis Strategy:
1. Examined the joint association of daily sleep and shift duration with recovery, including an interaction term.
2. Modeled the association between shift type and sleep duration.
3. Assessed the association between shift type and recovery, both unadjusted and adjusted for sleep duration.
4. Conducted a physician-level cluster bootstrap mediation analysis (5,000 resamples) to test if sleep duration mediated the relationship between shift type and recovery.
5. Performed a sensitivity analysis to estimate the magnitude of unmeasured confounding required to nullify the mediation effects.

Key Results

Sleep Duration and Recovery: Recovery was significantly higher on days when physicians slept longer than their personal average. Specifically, each additional hour of sleep relative to a physician's average was associated with a 4.22-percentage-point increase in recovery ( $p < .001$ ).
Shift Duration: Working a shift longer than a physician's average was associated with a slight decrease in recovery (1.82 percentage points), but this was not statistically significant ( $p = .179$ ).
Interaction: The interaction between sleep duration and shift duration was negative but not statistically significant ( $p = .093$ ), indicating insufficient evidence that the benefit of sleep varies by shift length.
Shift Type and Sleep: Sleep duration varied systematically by shift type. Relative to night shifts, afternoon shifts were associated with 1.06 additional hours of sleep, and morning shifts with 2.25 additional hours ( $p < .001$ for both).
Shift Type and Recovery: Recovery was highest after morning shifts, intermediate after afternoon shifts, and lowest after night shifts.
- Unadjusted: Morning shifts yielded 19.57 percentage points higher recovery than night shifts.
- Adjusted for Sleep: The gap narrowed but remained significant (10.74 points for Morning vs. Night; 7.78 points for Afternoon vs. Night). The Morning vs. Afternoon difference became non-significant after adjusting for sleep.
Mediation: Sleep duration partially mediated the association between shift type and recovery.
- Indirect effects were significant for all contrasts.
- Sleep duration accounted for approximately 35% of the Afternoon vs. Night difference, 46% of the Morning vs. Night difference, and 63% of the Morning vs. Afternoon difference.
- Significant direct effects remained for Morning vs. Night and Afternoon vs. Night, suggesting factors beyond sleep duration (e.g., circadian phase, social disruption) contribute to the recovery penalty of night shifts.
Sensitivity: The mediation findings were sensitive to unmeasured confounding. An unmeasured within-physician factor inducing a residual correlation of approximately $\rho = 0.35$ between sleep and recovery would be sufficient to eliminate the estimated indirect effects.

Significance and Claims
The paper claims to identify "protected sleep opportunity" and "circadian-informed scheduling" as modifiable targets for supporting EP recovery. The findings suggest that:

Sleep is a Proximal Driver: Daily sleep duration is a stronger predictor of next-day recovery than shift duration. The study supports the homeostatic model where sleep reduces accumulated pressure.
Shift Type Matters Beyond Duration: While sleep duration explains a substantial portion of the recovery deficit associated with night shifts, a significant direct effect remains. This implies that night shifts impose a "dual penalty" (reduced sleep opportunity and adverse circadian phase) that cannot be fully mitigated by sleep duration alone.
Limitations of Current Evidence: The authors explicitly state that the mediation results should be read as "consistent with a sleep pathway rather than as establishing one" due to the observational nature of the study and the sensitivity to unmeasured confounders (e.g., daily workload, stress).
Practical Implications: The study suggests that scheduling practices should prioritize inter-shift intervals long enough to accommodate a full sleep episode. However, because longer shifts did not independently show a significant negative association with recovery in this sample, and because sleep duration only partially explains the night-shift penalty, the authors caution against viewing sleep duration as a complete substitute for structural scheduling changes. They recommend complementing sleep protection with strategies addressing circadian misalignment (e.g., forward-rotating schedules, light exposure management).

The paper concludes that while it moves the field from between-person comparisons to within-person daily dynamics, the findings are associative and require further validation with objective sleep metrics and validated recovery measures to establish causal mechanisms.

Technical Summary: Hours Slept, Not Hours Worked: Predictors of Next-Day Recovery in Emergency Physicians

More like this