Resting and Recharging: A Narrative Review of Strategies to Improve Sleep During Residency Training

Residency involves demanding training with long hours that may impact a resident's sleep. Sleep deprivation may occur acutely, such as during extended and nocturnal shifts, and chronically through accumulated work hours. Effects of sleep deprivation on residents and the patients under their care include decreased alertness,¹ creative thinking,² working memory capacity, and neurocognitive function.^3,4 Sleep deprivation is also associated with poor mood and increased depression,^5,6 driving risk,⁷ and increased medical errors.⁸ Sleep deprivation has been linked to resident burnout, while restful sleep can lead to a higher sense of well-being.⁹

While significant focus has been shifted to fatigue mitigation at the Accreditation Council for Graduate Medical Education (ACGME) level, including significant changes to work hours and scheduling strategies and a heightened awareness of issues related to burnout, the effectiveness of these strategies to limit fatigue is largely unknown. It is also unclear whether these interventions lead to improved patient care or resident satisfaction. The aim of this narrative review is to identify evidence-based non-pharmacological and pharmacological interventions to reduce the physiologic effects of fatigue and sleep deprivation after overnight and night shift work during residency training.

A PubMed literature search was conducted by 2 authors (J.R., L.K.D.) from inception of the database through August 30, 2021 using the terms “resident” and “sleep” in the title or abstract. This list was further narrowed by inclusion of a third search term from the following list: fatigue, sleep deprivation, intervention, duty hours, work hours, night float, night shift, wellness, napping, sleep hygiene, meditation, sedative, and stimulant. The bibliographies of the included studies and review articles were also reviewed to identify other potentially relevant studies not included in our search strategy.

The search was limited to retrospective and prospective observational studies, randomized controlled trials, systematic reviews, and meta-analyses of human subjects written and published in the English language. Studies were excluded if they were not specific to residents, interns, or fellows or if they did not investigate a treatment or intervention (eg, studied the effect of overnight work and night shift work rather than comparing intervention to control) or were reviews, surveys, case reports, case series, or editorials. Two authors (J.R., L.K.D.) assessed the articles for eligibility, and disagreements were resolved by consensus between the 2 authors. If consensus was not reached between the 2 authors, a third author (A.M.K.) provided the majority decision. The articles were organized and summarized by one author (J.R.). Due to the heterogeneity of study design and intervention, a narrative review approach was chosen with results categorized into non-pharmacological and pharmacological interventions.

Our initial search using the terms “resident” and “sleep” yielded 271 articles. This was narrowed to 28 original articles with the sequential addition of a third search term. Bibliography review of these studies yielded 4 additional articles. Given the low number of relevant articles as well as the heterogeneity of study design and intervention type, we were unable to systematically analyze the data via a meta-analysis. In order to present the most inclusive and broad range of interventions along with their limited evidence, we chose to present these results using a narrative approach with articles sorted into 2 broad categories: non-pharmacological (Table 1) and pharmacological (Table 2) interventions.

Perhaps the most widely recognized intervention to address acute and chronic sleep deprivation in resident physicians is work hour limitations. In 2003, the ACGME issued requirements that limited resident work hours to 24 consecutive hours a day and 80 total hours per week. Multiple systematic reviews have attempted to investigate the affects of the work hour limits on residents and patients. Despite the heterogeneity of existing data, systematic reviews have found generally positive effects on measures of resident well-being, including sleep, burnout and mood, and quality of life.^10-12 Mixed effects on patient safety outcomes were found which often varied based on the medical specialty evaluated.^10-12 The majority of negative findings were related to resident education, such as an inability to attend teaching conferences due to work hour requirements, decreased continuity of care, limited supervising attending availability, and reductions in case volume related to unavailability.^10-13 Whether fewer work hours translate to reduced work remains unclear, and several studies cited work compression resulting in increased stress and decreased educational time under the new restrictions.^12,14

In addition to limiting total work hours, the impact of eliminating extended duration shifts (>24 hours) to comply with ACGME restrictions has also been evaluated. Resident and patient outcomes of shortening shift length were largely similar to the overall outcomes of reducing total work hours as previously described. An observational study found increased sleep and improved performance on psychomotor vigilance testing in residents working shorter shifts when compared to extended duration shifts.¹ Multiple randomized controlled trials comparing extended duration shifts to shortened shifts (<16 hours) in intervention groups found reduced sleepiness and increased sleep duration, again at the potential cost of resident education.^15,16 There were no changes in patient mortality or medical errors after adjusting for the number of patients per resident.^15,17 Systematic reviews of eliminating extended work shifts had more favorable findings, including fewer medical errors, motor vehicle crashes, and percutaneous injuries, and improved resident quality of life.^18,19 Overall no effects were found in resident education in these studies.

Major scheduling changes in response to work hour limitations have had varying effects on resident sleep and wellness. Traditional internal medicine scheduling typically requires attendance at ambulatory clinics during inpatient rotations. Block scheduling separates these responsibilities into consecutive weeks on inpatient rotations followed by consecutive weeks in ambulatory clinics. A single center study compared traditional scheduling (attendance at ambulatory clinics duing inpatient rotations) vs block scheduling (ie, 3 weeks in an inpatient setting followed by 1 week of ambulatory clinic). Block scheduling was associated with greater total sleep time during the ambulatory week. A traditional schedule was associated with worse Epworth Sleepiness Scale (ESS) and Perceived Stress Scale (PSS) scores throughout week 4, while residents on a block schedule returned to their pre-MICU baseline ESS and PSS scores by the end of their week of ambulatory clinic.²⁰

Dedicated night coverage is another common scheduling adaptation to provide 24-hour coverage while adhering to work hour limits; however, its benefit remains unclear. A crossover study in medical interns comparing every-fourth-night call with 30-hour work limits to an intervention group of dedicated night shifts with 16-hour work limits found increased sleep in the intervention group at the cost of reduced educational opportunities, increased handoff, and both nursing and resident reduction in perceived quality of care.²¹ A prospective cohort study of residents on a general pediatric ward had conflicting results—compared to traditional call schedule with extended duration shifts, dedicated night shifts unexpectedly resulted in decreased sleep hours.²² There is limited data on the optimal schedule and duration for overnight or night shifts. However, an observational study of electroencephalographic patterns of anesthesiology residents following 6 consecutive night shifts found that time spent in restorative (deep and rapid eye movement) sleep did not return to baseline over a 3-day period, thereby suggesting insufficient time for recovery.²³

The ACGME previously recommended “strategic napping” in its 2003 work hour guidelines, which was later retracted.^24,25 Several studies aimed to evaluate if, when, and how long napping while working extended duration shifts. In one study of interns, napping ad libitum when fatigued on night shift (average 2 hours duration) was associated with improved performance on task switching attention tests but not on go/no-go attention tasks.²⁶ Another single center study of residents working 24-hour shifts found that >4 hours of napping resulted in improved performance on psychomotor vigilance testing, although improvement was not enough to fully recover to post-call performance.²⁷

Beyond ad libitum naps, several studies have evaluated scheduled and protected nap time overnight ranging from 3 to 6 hours. On extended duration shifts (>24 hours), protected nap time resulted in more total sleep and sleep efficiency, less fatigue,²⁸ improved performance on psychomotor vigilance testing,²⁹ and no changes in patient outcomes.³⁰ Protected nap time of just 20 minutes during the daytime was associated with improved performance on testing for attention failures.³¹ Despite beneficial outcomes, napping may be limited by residents' desire for continuity of care, and notably multiple studies demonstrated only a fraction of protected nap time was spent sleeping.

Paging is still the primary method of communication in many hospitals and occurs frequently throughout the night, often for non-urgent issues. The ability to rest on extended duration shifts or night shifts is often interrupted by frequent pages. At a single institution, an interventional program that filtered pages through a charge nurse resulted in a decrease in total and non-urgent pages by 50% and 75%, respectively.³² A quality improvement study developed focus groups for nighttime paging which led to a 50% decrease in total nighttime pages; however, no significant difference in total sleep time was observed.³³

The ACGME requires that residency programs educate their residents on fatigue mitigation,³⁴ but there is currently no set standard for how to teach this. A single institution study investigated the effect of the Sleep, Alertness, and Fatigue Education in Residency (SAFER) program developed by the American Academy of Sleep Medicine to improve intern sleep, but demonstrated no significant changes in sleep time after participation in the program.³⁵ Another center implemented a complete wellness program with biweekly mental health and sleep education lectures. After 1 year, residents reported decreased anxiety and sleepiness and improved quality of life.³⁶ Although a single education session is likely insufficient to change resident sleep habits, longitudinal educational programs and individual practice changes may improve resident sleep while limiting fatigue.

Caffeine is frequently used to combat fatigue during overnight and night shift work. In a survey of 12 emergency medicine residency programs, 89% reported use of caffeine during night shifts.³⁷ In a single center randomized placebo-controlled crossover study, simulated driving performance of anesthesiology residents was evaluated following 6 consecutive night shifts. The intervention group (consumed beverage containing 160 mg caffeine) demonstrated overall better performance in a high-fidelity driving simulator, struck significantly fewer obstacles, and had quicker reaction time on psychomotor vigilance testing than the control group (consumed a non-caffeinated beverage).³⁸ Improved driving performance was observed only in the later part of the driving test, and residents who consumed caffeine performed worse in the first 15 minutes of testing. Larger studies are needed to further elucidate the benefits and potential limitations of caffeine on fatigue.

Modafinil is a central nervous system stimulant that is available only by prescription in the United States and used for treatment of narcolepsy, obstructive sleep apnea, and shift work sleeping disorder. In a survey of 133 emergency medicine residents, 2.4% reported using modafinil. Of these, 64% reported improved clinical performance with modafinil, while 31% of users reported significant side effects, including insomnia, agitation or restlessness, palpitations, and nausea or anorexia.³⁹ Evaluation of the effectiveness of modafinil in adjusting to shift work in residents is limited to survey-based reports.

Nonbenzodiazepines, sometimes called “Z-drugs,” include zolpidem, eszopiclone, and zalepon. Nonbenzodiazepines differ in structure from benzodiazepines, are less likely to cause physical dependence, and are approved for the treatment of insomnia. The benefit on any sleep or performance metrics with the use of nonbenzodiazepines has not been studied specifically in resident physicians, but between 14% and 22% of emergency medicine residents report using nonbenzodiazepines to sleep following shift work.^37,39 Reported side effects include drowsiness, dizziness, headache, hallucinations, depression/mood lability, and amnesia. Although described as a safer alternative to benzodiazepines, the regular use of nonbenzodiazepines is associated with risk for physician dependence, rebound insomnia, and a single-case report of zolpidem withdrawal seizures in a resident with chronic heavy use of alcohol and zolpidem.⁴⁰

First-generation H1 antagonists are widely available over the counter primarily as allergy medications but are frequently used for their sedating side effects. No specific study of the utility of antihistamines for improving sleep has been conducted in residents, though, similar to nonbenzodiazepines, they are commonly used by residents to improve sleep. Among 226 emergency medicine residents surveyed, antihistamines were the most commonly used sleep aid (31%), followed by zolpidem (14%), melatonin (10%), and benzodiazepines (9%).³⁷ Another survey of emergency medicine residents found that 46% used a sleep aid with antihistamines (diphenhydramine and doxylamine) reported as the most commonly used.⁴¹

Melatonin is a hormone secreted by the pineal gland that regulates several physiological functions including circadian rhythms. Exogenous melatonin is widely available over the counter and is a popular sleep aid, classified in the United States by the Food and Drug Administration as a “dietary supplement.” There is no standard dosage for melatonin although typical dosages range from 1 to 10 mg.⁴² Modest benefits of melatonin were shown in night-shift doctors and nurses in a pilot double blinded, randomized, placebo-controlled crossover trial. Intervention participants took 6 mg of melatonin following their first of 3 contiguous night shifts. Compared to placebo, the intervention subjects performed better on double digit addition and reaction time testing. Psychomotor vigilance testing, sleep latency, duration, wake after sleep onset, and ESS scores were not significantly different between treatment arms.⁴³

Melatonin has been shown to have limited or equivocal effects in medical residents. In a single center randomized placebo-controlled crossover study of pediatric residents, subjects received 3 mg melatonin or placebo in the morning following night shift. Intervention residents had isolated improvement in omission errors but no differences were observed in attention, sleep, and mood compared to placebo were observed.⁴⁴ In another single center randomized placebo-controlled crossover study of emergency medicine residents, nighttime awakening and daytime drowsiness were not significantly different with melatonin compared to placebo.⁴⁵

This narrative review examined non-pharmacological and pharmacological interventions to reduce the physiologic effects of fatigue and sleep deprivation from overnight and night shift work. Non-pharmacological interventions such as strategic scheduling may improve total sleep time but at the expense of education (eg, availability to attend conferences) and continuity of patient care. Institutional changes in scheduling, faculty supervision, and innovation in technology and care may also contribute to improved resident well-being and patient safety. Although survey data suggest that pharmacological interventions are frequently used by residents to reduce fatigue and improve sleep, few have been rigorously studied in prospective randomized controlled trials. Based on limited data, the effects of melatonin on sleep are equivocal.

Other non-pharmacological interventions, such as bright light and dark exposure, mindfulness, and meditation, have been trialed to reduce fatigue in night shift workers and other medical professionals. For example, exposure to 6 hours of bright light in the workplace while adapting to night shift correlated with shifts in circadian rhythms (measured by core body temperature and salivary melatonin) by an average of 9.3 hours and 11.3 hours, respectively, compared to control group phase shifts of 4.1 hours and 5.1 hours, respectively.⁴⁶ Mindfulness and meditation are widely available and have been studied in patients with insomnia and sleep disturbances as a potential treatment modality. A meta-analysis of 6 randomized controlled trials found that mindfulness meditation improved sleep quality and decreased total wake time during sleep, but had no significant effect on sleep onset latency, total sleep time, wake after sleep onset, sleep efficiency, insomnia severity index, Pittsburgh Sleep Quality Index (PSQI) and Dysfunctional Beliefs and Attitudes about Sleep Scale.⁴⁷ When compared to standardized sleep hygiene education, one randomized controlled trial found that mindful awareness practice, a standardized course on mindfulness meditation, was more effective in improving sleep quality as measured by the PSQI (improvement by a mean of 1.1 vs 2.8, respectively).⁴⁸ A single randomized controlled trial evaluated the use of at least 10 minutes of daily app-guided (“Calm”) meditation for 8 weeks was associated with decreased daytime sleepiness in adults with sleep disturbance.⁴⁹ Although these techniques have not been studied in residents as tools for managing sleep deprivation (and thus they were excluded from our review), their success in other populations make them potentially useful adjuncts for managing shift-work related insomnia and sleep disturbances.

Our results are limited by the low number of relevant articles, as well as the heterogeneity of study design and intervention type. We were unable to systematically analyze the data via a meta-analysis. Our results are presented as a narrative review, which is inherently less rigorous and reproducible. Data on non-pharmacological interventions were limited by the heterogeneity of intervention type, sample size, length of follow-up, and reliance on self-report. Existing studies are of a pre-/post-design or are small and specialty specific, limiting their applicability to other specialties and practice settings. Larger propsective interventional studies to evaluate the impact of non-pharmacological interventions on patient-related outcomes are needed to ensure these changes are not adversely affecting patient care. Despite the apparent lack of studies supporting the use of pharmacological agents, residents may be tempted to use these medications to mitigate fatigue. Randomized controlled studies are needed to evaluate both the utility and safety of pharmacological treatments to reduce fatigue and improve sleep in resident physicians.

Non-pharmacological interventions (eg, work hour limits, block scheduling, dedicated nighttime coverage, napping, paging efficiency, and sleep hygiene education) and pharmacological interventions (melatonin, anti-histamines, non-benzodiazepines, and caffeine) have been studied and demonstrated to have varying effectiveness on improved restful sleep. Other non-pharmacological interventions, such as mindfulness meditation, have demonstrated beneficial effects for improving sleep in other study populations and should be studied in resident physicians.