Sleep deprivation on shift and medical error

It is four in the morning and the resident has been awake for almost twenty hours. Another prescription arrives to be reviewed, the phone rings twice while he types, and the nurse asks about another patient in the middle of the task. He finishes, signs and moves on to the next one. Only in the morning, at handover, does someone notice that the dose of a medication was wrong. No one was harmed, but the scare remains.

This scene is common in hospitals around the world. The link between sleep and error seems obvious, and sleep deprivation does indeed impair attention, memory and mood. But research shows a more complex picture: error rarely comes from a single factor, and reducing working hours alone has not solved the problem.

This text reviews what is known about sleep deprivation during shifts, how it affects clinical reasoning, what studies show about errors and what doctors and services can do to protect patients and professionals.

Key points

• Sleep deprivation impairs attention, working memory, executive function and mood.
• Residents tend to accumulate chronic deprivation, which worsens with each night shift.
• Poor sleep, interruptions and multitasking add up in the risk of prescription error.
• Limiting working hours alone has not consistently reduced errors.
• Insomnia and irregular sleep are associated with depressive symptoms in residents.
• Recovery after a shift can take more than one night.

What lack of sleep does to the brain

Sleep deprivation can be total, when a person stays awake an entire night, or partial, when they sleep less than they need for several days. A classic review shows that both forms impair cognitive performance (Alhola and Polo-Kantola, 2007). Total deprivation affects mainly attention and working memory, but also long-term memory and decision-making. Partial deprivation affects mainly attention, especially vigilance.

The same review highlights that the response varies greatly between people and depends on factors such as age and sex (Alhola and Polo-Kantola, 2007). Some tolerate a night without sleep well and others lose a lot of performance. This variation is one of the reasons why the individual perception of "I am fine" is not a reliable guide.

A review on residents describes the typical sequence of responses to sleep loss: first mood changes, then cognition, and finally performance (Kramer, 2010). In the laboratory, the deficits are clear and reproducible. In the hospital, the situation is different, because the acute deprivation of the shift adds to the chronic deprivation the resident already carries.

Acute deprivation on top of chronic deprivation

A study in Israel assessed 33 clinical and surgical residents at two moments, at the start and after a 26-hour shift, and compared them with 18 young attending physicians who did not work night shifts (Choshen-Hillel and colleagues, 2021). The residents slept on average 3 ± 1.2 hours during the shift (Choshen-Hillel and colleagues, 2021). Even before the shift, they already showed worse executive function than the attendings, as well as more impulsivity and slower processing.

The shift with acute deprivation worsened executive function even further (Choshen-Hillel and colleagues, 2021). The residents also had lower morning cortisol and higher high-sensitivity C-reactive protein than the attendings, signs of altered stress axis and low-grade systemic inflammation. There was no difference in risk-taking tendency, assessed in another group.

In Spain, a study with residents from two public hospitals in Catalonia compared those who worked a 24-hour shift with those who did not (Morales and colleagues, 2019). The residents on the shift had significant worsening in all the physiological, performance and mood indicators analyzed. The authors stress, however, that other studies found no effect and that the appropriateness of 24-hour shifts remains under debate.

From the shift to the error: what the studies show

Linking sleep to clinical error is harder than it seems. Error depends on the task, the environment, the experience and how many barriers exist to catch it before it reaches the patient. The best studies try to measure several of these factors at the same time.

In Australia, researchers followed 36 emergency physicians for 120 hours, recording every task, interruption and episode of multitasking (Westbrook and colleagues, 2018). The physicians were interrupted 7.9 times per hour (Westbrook and colleagues, 2018). In all, 28 of them made 239 prescriptions, which contained 208 prescribing errors. During prescribing, interruptions reached 9.4 per hour (Westbrook and colleagues, 2018).

The error rate rose when the physician was interrupted (rate ratio 2.82) or was doing several tasks at once (rate ratio 1.86) while prescribing (Westbrook and colleagues, 2018). The strongest finding was about sleep: those who had slept below average in the previous 24 hours had an error rate more than 15 times higher (rate ratio 16.44, with a wide confidence interval, from 4.84 to 55.81) (Westbrook and colleagues, 2018). Working memory capacity was protective. Above-average sleep, workload and physician sex did not change the error rate.

The study shows that sleep, interruptions and multitasking do not act in isolation. A tired physician, interrupted in the middle of prescribing, combines several risk factors at once. The wide confidence interval calls for caution about the exact number, but the direction of the effect is clear.

Insomnia, hours and errors on a large scale

In Japan, a study of 5095 residents analyzed the relationship between working hours, sleep, depressive symptoms and self-reported errors (Nagasaki and colleagues, 2024). Among them, 15.5 out of every 100 slept less than 5 hours a day and 26.7 out of every 100 had insomnia (Nagasaki and colleagues, 2024). Working 90 hours or more per week was associated with fewer hours of sleep and more insomnia symptoms, compared with those working 60 to 70 hours.

The result on errors is surprising (Nagasaki and colleagues, 2024). Self-reported errors increased among residents with insomnia, but had no association with sleep duration. The authors suggest that limiting hours to avoid excessive load may improve sleep and well-being and, potentially, reduce errors associated with insomnia.

In San Francisco, a survey of internal medicine residents had responses from 125 of the 164 invited (Vidyarthi and colleagues, 2007). The most common suboptimal practices were working impaired by fatigue and forgetting to hand over information at the shift change. The residents pointed to work stress, which brings together fatigue, excessive load, distractions and lack of time, as the most frequent factor in errors. Working more than 80 hours per week was not associated with suboptimal care or with errors.

What these studies have in common is the difficulty of measuring error. Self-reported error depends on the physician noticing and admitting the failure, and someone who is tired may notice less. Observed error requires an observer present, which changes the behavior of those being observed. None of the methods is perfect, and that is why the results vary so much from one study to another.

Even so, reading them together points in one direction. Poor sleep, whether because of duration or because of insomnia, appears linked to more errors when it is measured closely. The long shift, on its own, does not always appear. This suggests that the problem lies less in the hospital clock and more in what happens to sleep and to the load within those hours.

Fatigue and surgical performance

In surgery, a systematic review analyzed 134 studies on non-muscular fatigue and surgical outcome (Reijmerink and colleagues, 2024). Among the simulator studies, 11 of 24 showed worse performance with fatigue. Among the real-life studies, 39 of 110 showed worse performance, and it appeared in only 12.5 out of every 100 of the outcome measures (Reijmerink and colleagues, 2024).

The authors explain the difference in two ways (Reijmerink and colleagues, 2024). In real life, the surgeon can compensate for fatigue with more effort and motivation. In addition, most real-life studies used the time of surgery as a stand-in for fatigue, without measuring it directly, and used outcomes with low sensitivity. The conclusion is that fatigue is a real concern, but current measures are not enough to define its impact.

Why reducing hours did not solve everything

In the United States, the death of an 18-year-old patient in a New York hospital in 1984 led to a commission that studied the case and made recommendations. They became law in the state, with hour limits and minimum supervision levels for residents, and were later adopted, in essence, by the body that accredits residency in the country (Kramer, 2010). The logic was simple: fewer hours, less fatigue, fewer errors.

The review of this history shows that the results were contradictory (Kramer, 2010). An intervention study in intensive care showed that eliminating extended shifts and reducing residents' weekly hours reduced serious medical errors. But questionnaires, observational studies and evaluations of the duty hour changes gave equivocal results. Residents generally feel better working fewer hours, but improvements in patient care often do not appear.

A Canadian narrative review proposed four explanations for the lack of a clear link between resident hours and important patient outcomes (Osborne and Parshuram, 2014). The first questions the weight of resident fatigue in generating serious errors. The second points to other determinants of fatigue beyond hours, such as workload, experience and individual characteristics. The third describes adverse effects of reducing hours that may cancel out the benefits. The fourth recalls factors that mitigate risk, such as supervision and other safety barriers.

Continuity and handoffs

Among the adverse effects of reducing hours, the most cited is the loss of continuity (Osborne and Parshuram, 2014). Shorter shifts mean less direct contact with each patient and more handoffs. Each handoff is an opportunity to lose information, and the San Francisco survey showed that forgetting to pass on information at handoff is one of the most common suboptimal practices (Vidyarthi and colleagues, 2007).

This does not mean that long hours are safe. It means that how the work is designed matters as much as the hours worked. The San Francisco survey showed that residents who felt overloaded and who spent more than half their time on non-medical tasks reported more suboptimal practices (Vidyarthi and colleagues, 2007). The authors suggest that reducing hours must come together with changes in the nature of the work.

Resident sleep, mood and mental health

Sleep deprivation does not affect only the patient. It weighs on the mood and mental health of those who work.

In Saudi Arabia, 88 junior doctors at a university hospital answered questionnaires before and after a shift (Wali and colleagues, 2013). More than 87 in 100 slept 5 hours or less during the shift. The proportion of alert doctors fell after the shift, and mood worsened among those who slept less than usual. Among those who managed to sleep the same number of hours as usual, there was no significant difference in mood.

A study of 2115 residents in the United States used wristbands to measure sleep and an app to record daily mood during the first year of residency (Fang and colleagues, 2021). Sleeping less, going to bed later and having more variation in sleep duration and wake time were associated with more depressive symptoms. The weight of irregularity was similar to that of the amount of sleep. In daily life, longer nights and more stable schedules were associated with better mood the next day.

The authors conclude that interventions aimed at sleep regularity, and not only at duration, can help residents' mental health (Fang et al., 2021). For those working night shifts, this is difficult, because the schedule changes sleep timing every week. But it shows that protecting regularity on days off also matters.

Sleep as a warning sign

A study of 69 residents followed sleep quality and well-being over one year of training (Min et al., 2015). Worsening sleep predicted a drop in well-being the following month, even accounting for each resident's overall level of sleep disturbance. The authors suggest that periodic sleep assessment can alert the program coordination and the resident to an impending drop in well-being.

In the Japanese study, fewer hours of sleep and more insomnia symptoms were also associated with more depressive symptoms (Nagasaki et al., 2024). Sleep therefore works as an early sign that something is not right. Persistent insomnia, very short sleep and marked irregularity deserve attention, as do other signs of burnout.

Recovery after the shift

How long does it take to recover from a night shift? A Swedish study measured sleep with actigraphy and a sleep diary in 15 anesthesiologists and 17 pediatricians and otolaryngologists (Malmberg et al., 2010). Average sleep was 3 hours on the shift, 7 hours on the next two nights and also on Saturday, and 6 hours after a regular workday.

Mental fatigue and the feeling of rest worsened after the shift and returned to Sunday morning levels after two nights of sleep (Malmberg et al., 2010). The authors concluded that these 16-hour night shifts allowed short-term recovery. But they warned that limited sleep duration in general, including after regular days, remains a long-term health concern.

The review on sleep deprivation notes that cognitive recovery processes are still poorly studied and seem more demanding after chronic partial restriction than after a full night without sleep (Alhola and Polo-Kantola, 2007). In other words, one night of sleep may not be enough to make up for weeks of sleeping little.

In practice, this has two consequences. The day after the shift is not a regular day, and important decisions deserve extra care. And sleeping little on days off keeps the deficit, even if each night seems "enough".

What to do on the shift and outside it

There is no single solution. The risk comes from the sum of sleep deprivation, workload, interruptions and handover failures. The measures also need to be combined.

Factor	What the studies show	What helps
Acute deprivation on the shift	Worse attention, executive function and mood	Planned nap when possible; double-checking in the early hours
Chronic deprivation	Accumulated deficit even before the shift	Protect sleep on days off
Irregular sleep	Associated with more depressive symptoms	Keep stable schedules outside the roster
Interruptions and multitasking	Increase prescribing errors	Reduce interruptions during prescribing
Insomnia	Associated with self-reported errors	Recognize and treat; do not normalize
Handover	Forgetting information is common	Structured, written handover
Recovery	May take two nights or more	Avoid important decisions the next day

For the physician, the first step is to recognize one's own limits. The feeling of "I'm fine" is unreliable after a night without sleep. In the early hours, it is worth slowing down on high-risk tasks, such as prescribing dangerous medications, and asking for a second check when possible. The emotion of the shift also weighs on reasoning, as the text on affective bias shows, and prolonged strain brings the picture closer to burnout.

For the service, the measures are about work design: reduce interruptions during prescribing, cut non-medical tasks, structure the handover and ensure supervision in the early hours. A just culture, which treats error as a system failure and not only an individual one, makes it easier to talk about fatigue without fear of punishment.

Errors that repeat

The first mistake is thinking that experience protects against sleep deprivation. The response to lack of sleep varies between people, but no one is immune, and individual perception of performance is unreliable (Alhola and Polo-Kantola, 2007).

Another common mistake is treating tiredness as a matter of character. Asking for help, swapping a risky task with a rested colleague or stopping a few minutes before a complex prescription is not weakness. It is a safety decision, of the same kind as double-checking a dose or confirming the patient's identity before a procedure.

The second mistake is looking only at the hours. Workload, interruptions, non-medical tasks and shift handovers weigh as much as the number of hours worked (Vidyarthi and colleagues, 2007; Osborne and Parshuram, 2014).

The third mistake is normalizing insomnia. It is associated with self-reported errors and with depressive symptoms and deserves attention (Nagasaki and colleagues, 2024).

The fourth mistake is treating the day after a shift as an ordinary day. Recovery can take two nights or more (Malmberg and colleagues, 2010).

The fifth mistake is blaming only the individual. When an error happens, the doctor suffers too, and the second victim needs support. Learning from an error requires looking at the system that allowed it to happen.

How NEXO helps train this reasoning

In NEXO, the case of the day is fictional, with a real clinical presentation and up to six progressive clues, and you propose the diagnosis by searching for the code in ICD-10.

The limits on using the game are described on the NEXO economy page. If you research the teaching of clinical reasoning, the page for researchers explains how to get in touch. Professionals who write or review cases find information under curation and creators, and institutions can learn about the proposal on the partnerships page. Training differential diagnosis with clues and searching for the right ICD-10 code complete this practice.

Frequently asked questions

Does sleeping little really increase medical error?

Studies point in that direction, but the relationship is not simple. In emergency physicians, below-average sleep was associated with far more prescription errors. In Japanese residents, insomnia was associated with self-reported errors, but sleep duration was not.

Does limiting working hours solve the problem?

On its own, no. Reducing hours improved residents' well-being, but the effects on errors and patient outcomes were contradictory. Workload, interruptions and shift handovers also need to change.

How long does it take to recover from a night shift?

In a Swedish study, mental fatigue and the feeling of being rested returned to normal after two nights of sleep. Chronic deprivation may require more time.

Does sleep deprivation affect the doctor's mental health?

Yes. Less sleep, irregular schedules and insomnia were associated with more depressive symptoms in residents. Worsening sleep may come before the decline in well-being.

What can I do during a shift?

Slow down on high-risk tasks, ask for a double check when possible, reduce interruptions during prescribing and hold a structured shift handover.

Take-home messages

• Sleep deprivation harms attention, memory, executive function and mood.
• Residents accumulate chronic deprivation, and the shift adds acute deprivation.
• Poor sleep, interruptions and multitasking add up in the risk of error.
• Reducing hours without changing the design of the work is not enough.
• Insomnia and irregular sleep are warning signs for mental health.
• Recovery after a shift can take more than one night.

References

1. Alhola P, Polo-Kantola P. Sleep deprivation: Impact on cognitive performance. Neuropsychiatr Dis Treat. 2007. PMID 19300585.
2. Choshen-Hillel S, Ishqer A, Mahameed F, et al. Acute and chronic sleep deprivation in residents: Cognition and stress biomarkers. Med Educ. 2021. PMID 32697336.
3. Fang Y, Forger DB, Frank E, et al. Day-to-day variability in sleep parameters and depression risk: a prospective cohort study of training physicians. NPJ Digit Med. 2021. PMID 33603132.
4. Kramer M. Sleep loss in resident physicians: the cause of medical errors? Front Neurol. 2010. PMID 21188260.
5. Malmberg B, Kecklund G, Karlson B, et al. Sleep and recovery in physicians on night call: a longitudinal field study. BMC Health Serv Res. 2010. PMID 20712854.
6. Min AA, Sbarra DA, Keim SM. Sleep disturbances predict prospective declines in resident physicians' psychological well-being. Med Educ Online. 2015. PMID 26202848.
7. Morales J, Yáñez A, Fernández-González L, et al. Stress and autonomic response to sleep deprivation in medical residents: A comparative cross-sectional study. PLoS One. 2019. PMID 30947295.
8. Nagasaki K, Kobayashi H, Nishizaki Y, et al. Association of sleep quality with duty hours, mental health, and medical errors among Japanese postgraduate residents: a cross-sectional study. Sci Rep. 2024. PMID 38233476.
9. Osborne R, Parshuram CS. Delinking resident duty hours from patient safety. BMC Med Educ. 2014. PMID 25561349.
10. Reijmerink IM, van der Laan MJ, Wietasch JKG, et al. Impact of fatigue in surgeons on performance and patient outcome: systematic review. Br J Surg. 2024. PMID 38097353.
11. Vidyarthi AR, Auerbach AD, Wachter RM, et al. The impact of duty hours on resident self reports of errors. J Gen Intern Med. 2007. PMID 17356987.
12. Wali SO, Qutah K, Abushanab L, et al. Effect of on-call-related sleep deprivation on physicians' mood and alertness. Ann Thorac Med. 2013. PMID 23439930.
13. Westbrook JI, Raban MZ, Walter SR, et al. Task errors by emergency physicians are associated with interruptions, multitasking, fatigue and working memory capacity: a prospective, direct observation study. BMJ Qual Saf. 2018. PMID 29317463.