Residents have two primary tasks during graduate medical training: to care for patients and to learn medicine. Residents must accomplish these overlapping tasks despite acute and chronic sleep deprivation. Extended duration on-call shifts (greater than 24 hours) impair resident performance and cause patient care errors(1). Sleep after learning enhances retention of facts, learning of motor skills, and development of insight(2). The consequences of 24-30 hour extended duration on-call shifts on memory consolidation and clinical learning in residents remain unknown; in fact, the influence of sleep upon any form of occupational learning is uncertain.  We, therefore, aim to define the relationship of sleep to learning among residents. In this study, subjects trained on clinical and nonclinical tasks, either took their typical overnight call- the call condition- or slept at home- the noncall condition- and tested 48 hours after training upon both sets of tasks.  All residents completed both a call and a noncall condition, and we compared them within-subject learning across conditions.

(should the paragraph below be saved for the discussion portion?)

In concert with their potential impairment in learning, extended shifts are a primary contributor to medical error(3), the sixth leading cause of death in America(4).    Numerous studies, summarized in a 2005 metareview, demonstrate impaired performance in clinical and nonclinical tasks associated with sleep deficit(5). The Harvard Work Hours, Health, and Safety Group (HWHHSG) gathered 17003 surveys from 2737 first-year residents in 2002-2003.  More accidents occurred while driving home from extended than from nonextended shifts (OR 2.3, CI 1.6-3.3)(6).  Fatigue-related errors increased in months requiring 5 or more calls (OR 7.5, 95% CI 7.2-7.8)(7).  Percutaneous injuries occurred more often during extended than during non-extended shifts (OR 1.61, 95% CI 1.45-1.78)(8).  The HWHHSG randomized intensive care unit first-year residents to a “traditional schedule,” where they worked typical shifts (up to 34 scheduled hours) and the atypical number of hours per week (~80) and to an “intervention schedule,” where their longest duration shifts and weekly work hours were significantly shorter (16 hours and 63 hours per week, respectively)(9). On the traditional schedule, interns experienced twice as many attentional failures (p=.02) and committed significantly more serious medical errors (136.0 vs. 100.1 errors per 1000 patient days, p<.001); teams including interns on the traditional schedule made significantly more errors than teams including interns on the intervention schedule (193.2 vs. 158.4 per 1000 patients, p<.001)(1).  

Given a growing body of evidence that sleep deficits impair the performance, the Accreditation Council for Graduate Medical Education (ACGME) imposed work hour regulations in 2003(10).  Effects of work-hour regulations upon learning remain unclear.  After the introduction of ACGME regulations, scores on the otolaryngology training examination did not change (11).  American Board of Surgery In-Training Examination scores improved.  American Board of Neurosurgery Scores declined (12). In multiple reports, faculty report dropoff in resident knowledge following ACGME regulations, while residents report little change(13;14).  2011 ACGME regulations, though stricter, permit non-first year residents to work up to 30 hours consecutively and first-year residents to work up to 80 hours per week(15).

Improved sleep-dependent memory consolidation may partially account for stable or improved test scores despite reduced hours.  Over the last twenty years, a growing body of literature has demonstrated a link between sleep and learning(16).  In typical sleep and learning experiments, subjects learn a task, either sleep or do not sleep, and then undergo testing on the learned task.  In no area of learning does sleep appear to have no benefit(17).  Sleep leads to insight into a hidden numerical rule(18).  Rapid eye movement (REM) sleep facilitates anagram solving(19).  Sleep fosters grammatical rule abstraction in infants (20)and the integration of new words into the vocabularies of adults(21).  Subjects gain insight into the “gist” of a list of words with sleep(22).  In word-pair learning, an activity of declarative memory subjects train upon sets of words and provide these words, after a period of sleep or no sleep, during a testing period.  In most cases, sleep improves word recollection(23).  The effect of sleep upon word-pair learning may increase with greater emotional salience (24), less rigorous training(25), and greater slow-wave sleep(26).  In the motor sequence task (MST), subjects repeatedly type a brief number sequence as quickly and accurately as possible and, 12 hours later- after sleeping or not sleeping- retype the same sequence over several trials.  Sequence speed and accuracy significantly (~25%) improve in the sleep condition(27).   Sleep enhances MST performance regardless of what time of day the task was learned (28), regardless of what time of day sleep occurs (28)(as long as it occurs between the training and testing session), and regardless of the level of alertness as assessed by PVT.  Length of stage II sleep correlates positively with MST learning(27). The robustness of learning may depend upon the type of learned material, the type of learning, the type of retrieval task, and the stage and length of sleep following learning(29).  

We hope to address several gaps in the literature regarding sleep and learning.  First, despite hundreds of papers assessing sleep and learning, only one addresses occupation-specific tasks of sleep and learning.(30)  Second, though daytime naps may consolidate procedural and declarative learning, the effect of brief bouts of night-time sleep is unclear.(31)  


Subjects: We presented our study during first-year resident physician (intern) orientation in the Department of Internal Medicine at Brigham and Women's Hospital in June of 2010.  Our study ran from July of 2010 to April of 2011. 65 Interns initially volunteered for the study.  We excluded ** interns whose schedules did not provide assessment windows meeting our criteria for study participation.  ** Interns began the study but did not conclude all timepoints.  Their data are excluded  ** interns, therefore, participated in our study.  There were ** women and ** men.  Mean (+/- SD) was **+/-*.  The human research committee of Partners Healthcare and Brigham and Women’s Hospital approved all procedures, and all participants provided written informed consent.

Participation schedules

Using a within-subjects design, we studied ** interns.  Interns trained on clinical and nonclinical tasks, either took their typical overnight call- the call condition- or slept at home- the noncall condition- and tested 48 hours after training.  Each intern completed both a call and a noncall condition, and we compared within-subject learning across conditions.  During the call condition, only shifts beginning in the morning and ending after greater than 24 hours qualified as “call.”  Interns at Brigham and Women’s Hospital complete many rotations that require such a call.  These include:

  •  Brigham and Women’s Hospital: Medical Intensive Care Unit, Coronary Care Unit, Oncology Service, Cardiac Service, and General Medical Service
  • Faulkner Hospital: General Medical Service
  • VA Hospital: Medical Intensive Care Unit

We required interns to have two weeks without extended shifts, overnight shifts, or evening shifts before completing each condition.  Work shifts were predicted based upon an online block schedule, generated before the internship began, and confirmed upon each intern’s beginning of each condition.  The training portion of the call condition, therefore, occurred in the day leading up to the first overnight call, following a noncall block, of a new rotation.

**Interns undergo testing 48 hours after each training session.  Immediate post-call testing could demonstrate both call-related learning deficit and call-related performance impairment.  Thus, even the appearance of poor learning associated with call could result from performance deficits known to accrue from acute sleep deprivation.  To reduce such performance deficits, we allow subjects one normal night of sleep before testing.  This night of home sleep reduces the performance effects of acute sleep restriction and isolates sleep-dependent learning.**

We scheduled each intern’s call and noncall conditions at least one month apart to minimize learning effects from one condition to the other.  When schedules’ subjects allowed, condition order was randomized.


All training and testing occurred in the same room at the Neil and Elise Wallace Simulation, Training, Research, and Technology Utilization System (STRATUS) Center at Brigham and Women’s Hospital.  Training and testing occurred between 11 am and 7 pm.  All tasks were performed on standard desktop computers with standard keyboards.  Blinds were closed, and fluorescent lights shone.

Sleep and Work Hour Measurements

For four days leading up to each condition, interns completed a previously validated(9) log of sleep and work hours.  Diary compliance leading up to training day, the number of diaries completed over **x4, was **. Between the training and testing phases of each condition, interns wore a wrist actiwatch, which provided detailed information regarding sleep-wake timing and continued to complete sleep and work logs.  Sleep diary compliance over this period was **%.  Diary-actigraphy compliance/agreement/concordance was **.


All instruments were administered during a 45 minute to 1.5-hour training session and during a 30 minute to the 1-hour testing session.  In the early portion of the study, interns completed only the motor sequence task and the visual paired associates task.  We then added (in September) the video learning task and subsequently (in November) the case learning task.  Interns who began in July and August- and who therefore completed two tasks during the first condition- completed only two tasks at their subsequent condition.  Task order was randomized between interns but remained unchanged from call training day to call testing day, noncall training day, and noncall testing day.

Motor Sequence Task (MST)

Each subject presses four numeric keys on a standard computer keyboard with the fingers of the nondominant hand, repeating the five-element sequence, 4-1-3-2-4, “as quickly and as accurately as possible” for a period of 30 s. The numeric sequence (4-1-3-2-4) is displayed at the top of the screen at all times to exclude any working memory component. Each key press produces a white dot on the screen, forming a row from left to right, rather than the number itself, to avoid feedback regarding accuracy. The computer records the key responses, and each 30 s trial is automatically scored for the number of complete sequences achieved (speed) and the number of errors made (accuracy). 30 seconds of rest follows each trial, and subjects perform 12 trials.  A complete session thus lasts 12 minutes.  The score (speed and accuracy) from the first trial of the training session is a “baseline” measure, while the averaged scores from the final three trials comprise “post-training” performance.  We define intersession improvement as a percent increase in the number of correct sequences typed from the final three training trials to the first three test trials.  Interns learn one MST (41324) sequence during one condition and another (23142) during the other condition.  Sequences are randomized by call condition.  We assessed  MST improvement, call compared to noncall conditions, within-subject.

Visual Paired Associates (VPA)

Subjects were shown 15 black-and-white face-object pairs with the name of each object displayed under the object.  Each pair was presented for 5 seconds.  Subjects then received a cued recall test in which each face was presented in random order.  Subjects were instructed to provide the correct associated word.  The correct answer was presented for 4 seconds.  If the subject entered a correct response, the pair was not presented again.  If incorrect, the subject sees the picture again.  Subjects trained until they correctly paired 80% of pictures with words.  At retest, subjects were shown the pictures again and provided answers.  Total number of words recalled correctly out of 15 at retest comprised the score.  Interns learn one face-pair set during one condition and another during the other condition.  Sets are randomized by call condition.  We assessed VPA improvement, call compared to noncall condition, within-subject.

Video learning assessment

On the training day, interns took a 20 question baseline quiz regarding lumbar puncture or paracentesis.  The paracentesis is a medical procedure requiring needle extraction of peritoneal fluid via abdominal puncture, and the lumbar puncture is a medical procedure requiring needle extraction of cerebrospinal fluid via lower back puncture.  Interns then watched a 7-10 minute New England Journal of Medicine video showing that procedure.  On the test day, interns retook the baseline quiz.  Difference between baseline quiz score and test day quiz score comprised “learning.”  Interns took a lumbar puncture quiz, watched a lumbar puncture video, and took a lumbar puncture re-quiz during one condition and took a paracentesis quiz, watched a paracentesis video, and took a paracentesis re-quiz during the other condition.  Procedures were randomized by call condition. We assessed video learning, call compared to noncall conditions, within-subject.  

Case Learning

Interns received 9 minutes to review a unique set of three cases during the testing day.  On the teaching day, interns received unlimited time to complete a free text web-based questionnaire involving 21 questions.  Interns were randomized by call condition to receive either “case set A” or “case set B.” We assessed case learning, call compared to noncall conditions, within-subject.

Dr. Mary Thorndike, Instructor in Medicine at Brigham and Women’s Hospital, and Dr. Ken Kehl, senior medicine resident at Brigham and Women’s Hospital designed these cases to approximate emailed sign-outs that interns send to one another when they transition from one-two service block to another.  Dr. Joel Katz, a residency director in the department of internal medicine, reviewed and approved these cases.

Video learning and case learning assessments are scored for accuracy on a scale of 0-21 by two blinded observers, Dr. Kehl and Dr. Thorndike.  Dr. Kehl and Dr. Thorndike discuss any inconsistencies.

Reference List

(1) Landrigan CP, Rothschild JM, Cronin JW, Kaushal R, Burdick E, Katz JT et al. Effect of reducing interns' work hours on serious medical errors in intensive care units. N Engl J Med 2004; 351(18):1838-1848.

(2) Stickgold R. How do I remember? Let me count the ways. Sleep Med Rev 2009; 13(5):305-308.

(3) Landrigan CP. Effect of lack of sleep on medical errors. In: Cappuccio FP, Miller MA, Lockley SW, editors. Sleep, Health and Society: From Aetiology to Public Health. Oxford, UK: Oxford University Press; 2010. 382-396.

(4) Institute of Medicine. To err is human: Building a safer health system. Kohn LT, Corrigan JM, Donaldson MS, editors.  1999. Washington, D.C., National Academy Press. 

Ref Type: Report

(5) Philibert I. Sleep loss and performance in residents and nonphysicians: a meta-analytic examination. Sleep 2005; 28(11):1392-1402.

(6) Barger LK, Cade BE, Ayas NT, Cronin JW, Rosner B, Speizer FE et al. Extended work shifts and the risk of motor vehicle crashes among interns. N Engl J Med 2005; 352:125-134.

(7) Barger LK, Ayas NT, Cade BE, Cronin JW, Rosner B, Speizer FE et al. Impact of extended-duration shifts on medical errors, adverse events, and attentional failures. PLoS Med 2006; 3(12):e487.

(8) Ayas NT, Barger LK, Cade BE, Hashimoto DM, Rosner B, Cronin JW et al. Extended work duration and the risk of self-reported percutaneous injuries in interns. JAMA 2006; 296(9):1055-1062.

(9) Lockley SW, Cronin JW, Evans EE, Cade BE, Lee CJ, Landrigan CP et al. Effect of reducing interns' weekly work hours on sleep and attentional failures. N Engl J Med 2004; 351(18):1829-1837.

(10) Accreditation Council for Graduate Medical Education. Report of the ACGME Work Group on Resident Duty Hours.  6-11-2002. Chicago. 

Ref Type: Report

(11) Shonka DC, Jr., Ghanem TA, Hubbard MA, Barker DA, Kesser BW. Four years of accreditation council of graduate medical education duty hour regulations: have they made a difference? Laryngoscope 2009; 119(4):635-639.

(12) Jagannathan J, Vates GE, Pouratian N, Sheehan JP, Patrie J, Grady MS et al. Impact of the Accreditation Council for Graduate Medical Education work-hour regulations on neurosurgical resident education and productivity. J Neurosurg 2009; 110(5):820-827.

(13) Dola C, Nelson L, Lauterbach J, Degefu S, Pridjian G. Eighty hour work reform: faculty and resident perceptions. Am J Obstet Gynecol 2006; 195(5):1450-1456.

(14) Vallier H, Prokuski L, Nash C, Patterson B. Effects of resident work-hour restrictions on orthopaedic education and patient care. Current Orthopaedic Practice 2009;77-86.

(15) Common Program Requirements.  1-19. 10-26-2010. 1-17-2011. 

Ref Type: Report

(16) Stickgold R, Walker MP. Sleep and memory: the ongoing debate. Sleep 2005; 28(10):1225-1227.

(17) Stickgold R. How do I remember? Let me count the ways. Sleep Med Rev 2009; 13(5):305-308.

(18) Wagner U, Gais S, Haider H, Verleger R, Born J. Sleep inspires insight. Nature 2004; 427(6972):352-355.

(19) Walker MP, Liston C, Hobson JA, Stickgold R. Cognitive flexibility across the sleep-wake cycle: REM-sleep enhancement of anagram problem. Cognitive Brain Research 2002; 14:317-324.

(20) Gomez RL, Bootzin RR, Nadel L. Naps promote abstraction in language-learning infants. Psychol Sci 2006; 17(8):670-674.

(21) Dumay N, Gaskell MG. Sleep-associated changes in the mental representation of spoken words. Psychol Sci 2007; 18(1):35-39.

(22) Payne JD, Schacter DL, Propper RE, Huang LW, Wamsley EJ, Tucker MA et al. The role of sleep in false memory formation. Neurobiol Learn Mem 2009; 92(3):327-334.

(23) Diekelmann S, Wilhelm I, Born J. The whats and whens of sleep-dependent memory consolidation. Sleep Med Rev 2009; 13(5):309-321.

(24) Wagner U, Gais S, Born J. Emotional memory formation is enhanced across sleep intervals with high amounts of rapid eye movement sleep. Learn Mem 2001; 8(2):112-119.

(25) Drosopoulos S, Schulze C, Fischer S, Born J. Sleep's function in the spontaneous recovery and consolidation of memories. J Exp Psychol Gen 2007; 136(2):169-183.

(26) Plihal W, Born J. Effects of early and late nocturnal sleep on priming and spatial memory. Psychophysiology 1999; 36(5):571-582.

(27) Walker MP, Brakefield T, Morgan A, Hobson JA, Stickgold R. Practice with sleep makes perfect: Sleep-dependent motor skill learning. Neuron 2002; 35:205-211.

(28) Walker MP, Brakefield T, Seidman J, Morgan A, Hobson JA, Stickgold R. Sleep and the time course of motor skill learning. Learn Mem 2003; 10(4):275-284.

(29) Diekelmann S, Wilhelm I, Born J. The whats and whens of sleep-dependent memory consolidation. Sleep Med Rev 2009; 13(5):309-321.

(30) Gais S, Lucas B, Born J. Sleep after learning aids memory recall. Learn Mem 2006; 13(3):259-262.

(31) Nishida M, Walker MP. Daytime naps, motor memory consolidation, and regionally specific sleep spindles. PLoS ONE 2007; 2:e341.

Figure 1: Sample sleep and work schedule of a single intern over the call and noncall conditions.

Top.  Before the call condition, the intern had two weeks of normal sleep leading up to the training condition. The intern then came into work, trained on four tasks, and subsequently worked overnight.  This intern had a brief bout of sleep during the call.  The intern completed call slept, returned to work on day 3, and underwent testing. Bottom: Before the noncall condition, the intern also had two weeks off the call.  The intern then came to work, underwent training, slept at home for the next two nights, and underwent testing on day 3.