RESEARCH Open AccessThe metabolic demands of internalmedicine residencySarah Sy1*† , Karanvir Sall1†, Erika Dempsey2, Gale Tedder1 and Kenneth Michael Madden1AbstractBackground: North American and European accreditation bodies have legislated progressively more strict workhour restrictions for residents in light of evidence that sleep deprivation leads to increased medical errors anddecreased wellbeing. The purpose of the study is to determine the physiologic demands of internal medicinetraining during residency as well as document average sleep (on- and off-call) and physical activity performed usingaccelerometers.Methods: A total of 40 internal medicine residents working on the clinical teaching unit at a single center wereenrolled in the study from November 2011 to March 2016. There were 22 subjects that completed the study andwere included in the analysis. SenseWear PRO 2 armband monitors were worn for 5 consecutive days includingone call day. The primary outcomes of the study were to quantify and compare the calories per day, steps per day,METs per hour, hours of activity, hours of sleep, and sleep efficiency for on call versus post-call and non-call days.Results: The average activity per day, calories per day, steps per day and METs per hour for the call day were 7.6 ±7.6 h, 2647.0 ± 541.1, 11,261.1 ± 2355.9, and 1.7 ± 0.2 respectively. Each of these parameters had a statisticallysignificant F statistic compared to post-call and non-call days. The subjects had a mean of 1.8 ± 2.0 h of sleep perday with a sleep efficiency of 77.3 ± 23.8% for the call day. The F statistic for sleep per day was significant with a pvalue < 0.001.Conclusion: This study shows that overnight call has a substantial impact on multiple metabolic parameters. Thesefindings have potentially important implications on future resident working hour restrictions.Keywords: Metabolic parameters of residency, Work hours, Sleep efficiency, Internal medicineBackgroundOver the past decade, there has been increasing pressureon health systems to decrease resident work hours. Theoriginal push to reform resident working hours derivedfrom the case of Libby Zion, an 18-year-old woman whopassed away under the care of exhausted residents [1].There has been increasing evidence that sleepdeprivation leads to increased medical errors andthreatens resident safety [2–5].North American and European accreditation bodieshave legislated progressively more strict work hourrestrictions. In the United States, the AccreditationCouncil for Graduate Medical Education (ACGME)implemented on July 1, 2017 changes to residentduty-hours based on recent studies that showednon-inferiority of flexible, less-restrictive duty hourscompared to standard duty hour policy [6, 7]. TheACGME has capped the work week to 80 h per weekaveraged over 4 weeks. In Europe, work hour restrictionstend to be even more conservative, with 48 h per weekbeing the latest maximum [3]. In 2011, a Quebec labourarbitrator found that 24-h call shifts violate theCanadian Charter of Rights and Freedoms and limitedshift length for all Quebec residents to 16 hours [8]. It isuncertain how this will affect work hour restrictionsacross Canada but data regarding the physical exertion,caloric expenditure and amount of sleep experienced oncall will certainly be helpful in the formulation of na-tional, provincial or institutional policies in this regard.© The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.* Correspondence: sarah.sy@alumni.ubc.ca†Sarah Sy and Karanvir Sall are co-authors.1Department of Geriatric Medicine, 2775 Laurel Street, 7th Floor, Vancouver,BC V6H 0A5, CanadaFull list of author information is available at the end of the articleSy et al. Journal of Occupational Medicine and Toxicology           (2019) 14:14 https://doi.org/10.1186/s12995-019-0234-0During their general internal medicine rotation, resi-dents at Vancouver General Hospital (VGH) work inclinical teaching units (CTU) that are comprised of asenior resident (second or third year internal medicineresidents), two first year residents (one from internalmedicine and one from another specialty), and two tothree medical students. The residents have graduatedresponsibility and participate in a 26 h on-call system. Intotal, there are 5 CTU teams, and one member fromeach team is on call each night such that there is asenior resident, 2 junior residents, and 2 to 3 medicalstudents. The purpose of this study is to determine thephysiologic demands of internal medicine trainingduring residency as well as document average sleep (on-and off-call) and physical activity performed usingaccelerometers.MethodsStudy populationA study was designed to determine the physiologicaldemands of general medicine training at a single quater-nary care centre (Vancouver General Hospital) fromNovember 2011 to March 2016. Ethics approval wasobtained through the University of British ColumbiaEthics Board. The participants were post-graduate years(PGY) one to three residents on the inpatient internalmedicine clinical teaching service, who were recruitedby email invitations and posters. The investigators weremembers of the internal medicine program but at nopoint were on service at the same time or had anyeducational authority over the residents.To be eligible for inclusion into the study, the residentmust be working 5 consecutive days on the clinicalteaching unit including one call shift, and had not previ-ously participated in this study. The residents wereexcluded if they did not work 5 consecutive days, or ifthey were taking modafinil.Study protocolThe participants met individually with an investigator toobtain written consent and receive instructions.Anthropomorphic information was self-reported and in-cluded sex, age, height, weight, and race. The subjectswere instructed to wear SenseWear PRO 2 armbandmonitors for 5 consecutive days including one call day.On non-call days, daytime responsibilities for theresidents were from 7:45 AM to 5 PM and varieddepending on their role as senior or junior resident. Thejunior residents cared for 6 to 8 patients on the teamand participated in educational teaching sessions withthe senior resident and/or attending physician. Thesenior resident was responsible for overseeing the careof the CTU team (20 patients), and teaching the juniorresidents and medical students. After 5 PM, the residentswere dismissed from clinical duties. The on call dayconsisted of being first call to the ward during thedaytime from 7:45 AM to 6 PM followed by doinginternal medicine consults from the emergency depart-ment from 6 PM to 7:45 AM. Subsequently, the post-callday started at 7:45 AM whereby the residents wouldfinish reviewing new patients and other patient issueswith an attending physician until 10 AM, at which point,they were excused from clinical duties and allowed to gohome. The residents would return to work the followingday.The SenseWear PRO 2 armband is a non-invasive,wearable monitor that employs multiple sensors includ-ing a biaxial accelerometer, heat flux, temperature, andgalvanic skin response sensors with artificial intelligencecapabilities. The armband is programed with thesubject’s sex, age, height, and weight using BodyMediaSenseWear Professional 8.1 Software (Philadelphia,USA). When worn, the armband sensors measure in 1min intervals and use BodyMedia’s algorithms to deter-mine daily energy expenditure, step count, hours ofactivity, and METs as well as information regarding sleeppatterns [9]. The SenseWear Pro 2 armband recordedphysical activity in 1-min epochs when METs were equalto or greater than 1.0. It is worn over the right tricepsmuscle and participants in previous validation studieshave not found it bothersome or to interfere withfree-living activity [10, 11]. The armband was permittedto be removed during contact with water, but notmore than 60 mins at a time. After completion of the5 days, the participants met with an investigator toreturn the device and obtain a summary of theirSenseWear data.Statistical analysisData analysis was conducted after all participants hadcompleted the study. The work days were measured in24 h periods (e.g. from 7:45 AM to 7:44 AM the nextday). The primary outcomes of the study were to quan-tify and compare the calories per day, steps per day,average METs per hour, hours of activity, hours of sleep,and sleep efficiency for on call versus post-call andnon-call days. The SenseWear device algorithm calcu-lated values for each data variable in one minuteintervals. These one-minute intervals were tallied overthe defined 24 h period (7:45 AM to 7:44 AM) to obtaina value for each day. Next for each variable, an averagewas created for non-call days, which would allow forcomparison to the call day. The data of the subjects thatcompleted the study was analyzed using R version 3.4.2(R Foundation for Statistical Computing, Vienna,Austria) ANOVA one-way repeated measures to deter-mine if there was a statistically significant differencewhen on call versus not on call.Sy et al. Journal of Occupational Medicine and Toxicology           (2019) 14:14 Page 2 of 5ResultsA total of 40 subjects were enrolled in the study fromNovember 2011 to March 2016. The prolonged enrol-ment period was in part due to the absence of a studyinvestigator to recruit new participants over a two yearperiod. The periods of active recruitment primarilyoccurred over 10 months from a pool of 150 potentialresident participants. There were 22 subjects that com-pleted the study and were included in the analysis(Fig. 1). The mean age was 31.3 years and the majority ofthe subjects were internal medicine residents. The aver-age number of hours worked in a 5 day period was 52.2h. Table 1 summarizes the subject demographics. Thereasons for not completing the 5 day research studyincluded: intolerance to the device (uncomfortable,bulky, and/or ruined clothing), and battery failure.Metabolic parameters including activity per day,calories per day, steps per day, METs per hour, hours ofsleep per day, and sleep efficiency were measured usingthe SenseWear PRO 2 armband monitors. In terms ofactivity per day, the call day had a mean of 7.6 ± 7.6 hand the ANOVA analysis showed a statistically signifi-cant p value < 0.001 compared to the post call andnon-call days (Table 2). The calories per day, stepsper day and average METs per hour for the call daywere 2647.0 ± 541.1, 11,261.1 ± 2355.9, and 1.7 ± 0.2respectively. Each of these parameters had a statisti-cally significant F statistic compared to post-call andnon-call days.With respect to sleep, on the call day, subjects had amean of 1.8 ± 2.0 h of sleep per day with a sleepefficiency of 77.3 ± 23.8%. The F statistic for sleep perday was significant with a p value < 0.001.DiscussionThis study demonstrates that overnight call has asubstantial impact on multiple metabolic parameters.The post-call and non-call days demonstrate decreasedphysical activity, METs per hour, steps per day, and calo-ries per day compared to the call day. In addition, sleepis significantly reduced on the call day to the pointwhere most residents were achieving only 1–2 h of totalsleep in 24 h. Subsequently, the residents required moresleep post call.While there is a plethora of literature that focuses onresident duty hours, there is a paucity of literature exam-ining specifically the effect of call on metabolic parame-ters in physicians and residents. Our study is incongruence with previous studies examining the effectof call on sleep. A study by Amirian et al. examiningsurgeons using actigraphy found call day to be associ-ated with reduced sleep [12]. In addition, this studyshowed that on post-call day one there was the lowestamount of daytime activity and the highest amount ofreduced wake time. Another study by Rose et al. withinternal medicine residents on a 1-in-4 call scheduledfound reduced sleep on call, but no significant increasesin sleep on post call day one [13].RecruitmentPosters and Emails2011-201640 participants enrolled22 participants completed the study18 participants failed to complete the study3 equipment failures15 participants did not wear accelerometer for the study durationFig. 1 Recruitment Flow ChartSy et al. Journal of Occupational Medicine and Toxicology           (2019) 14:14 Page 3 of 5With respect to sleep efficiency, our study revealsdecreased sleep efficiency during the call day versus thenon-call days. There are few studies that specificallyinvestigate the effect of call on sleep efficiency. A studyby Morhardt et al investigated sleep efficiency in urologyresidents during home call using Fitbit technology [14].The study found that the residents had 47% sleepefficiency and that the Stanford Sleepiness Scale did notcorrelate well with the results. In a different study by Koet al., sleep patterns of urology residents participating indifferent call schedules were investigated [15]. This studyfound that there was no difference between day shift,night shift, and 24 h call for total sleep time, lightsleep, deep sleep or number of times woken.However, sleep latency was the longest in those resi-dents participating in the day shift call. Thus far, ourstudy is the first to report decreased sleep efficiencyduring call in internal medicine residents. There is aclinical trial underway examining the impact of nightfloat on sleep patterns in anesthesiology residents(NCT03325244).Other studies focusing on parameters such as caloriesper day, METs per hour, and hours of activity could notbe found.In regard to creating policies for duty hour regulation,governing bodies have not taken into account thephysiological effects of working extended shifts orovernight call. The literature used to determine policieshave focused primarily on measures of well-being, cogni-tive function, sleep efficiency, and fatigue. Our study re-vealed that internal medicine residents experienced lesshours of sleep and decreased sleep efficiency while oncall. Previous studies have shown that lack of sleep con-tributes to increase errors and worse well-being [4, 5, 7].In our study, the residents were less physically active ontheir post-call and non-call days. They did not achievethe recommended amount of exercise per week norintensity of exercise [16]. Research has shown thatsedentary lifestyles can increase the risk of cardiovascu-lar disease, metabolic syndrome, and all cause mortality[17]. Even short bouts of sedentary activity have beenshown to have risks [18]. Therefore, the effect of work-ing hours on physiologic and metabolic parameters mayrequire more attention.Our study has several potential limitations. Slowrecruitment resulted in data collection over a widerperiod of time, which can introduce unaccountedvariables. In addition, the small sample size, single sitestudy, short study period, and limitation to primarilyinternal medicine residents make the generalizability ofresults to residents from different specialties difficult.Another limitation of the study is that the residents’baseline physical activity, the number of consultscompleted during call and the requirement for fre-quent patient reassessments based on illness severitycould not be controlled. As a result, this could im-pact the parameters measured and may be a reasonfor the large standard deviations. We were alsounable to control which day the residents decided totake call during the study period due to servicerequirements and personal preferences. Furthermore,sleep quality scores, cognitive performance measures,subjective measures of well-being, and direct observa-tion were not recorded.ConclusionsIn this single center study of internal medicineresidents, we show that overnight call has a negativeimpact on metabolic parameters including activity perTable 2 Metabolic Parameters and One-Way Repeated Measures ANOVAActivity (hours per day) Calories (per day) Steps (per day) METS (per hour) Sleep (hours per day) Sleep efficiency (%)Call Day 7.6 ± 7.6 2647.0 ± 541.1 11,261.1 ± 2355.9 1.7 ± 0.2 1.8 ± 2.0 77.3 ± 23.8Post Call Day 5.0 ± 5.3 2265.8 ± 485.3 8086.8 ± 2952.3 1.4 ± 0.1 8.7 ± 2.1 82.6 ± 6.6Non Call Day 5.9 ± 5.8 2440.7 ± 531.4 10,099.2 ± 4749.2 1.5 ± 0.2 5.9 ± 1.2 81.2 ± 7.5F-statistic 13.469 21.491 7.387 8.674 86.127 1.068P value < 0.001* < 0.001* 0.002* < 0.001* < 0.001* 0.356Table 1 Subject Demographicsn = 22Age 31.4 ± 6.3BMI 23.2 ± 4.4Female (%) 64% (n = 14)Hours worked during 5 day period 52.3 ± 6.3Ethnicity (%)Caucasian 46% (n = 10)Asian 36% (n = 8)Other 18% (n = 4)Specialty (n)Internal Medicine 20Off-service 2Residents per year (%)Junior (PGY-1) 59% (n = 13)Senior (PGY-2/3) 41% (n = 9)Smoker (%) 0%Sy et al. Journal of Occupational Medicine and Toxicology           (2019) 14:14 Page 4 of 5day, calories per day, steps per day, METs per hour,hours of sleep per day, and sleep efficiency. Furtherresearch of resident duty hours that incorporatesmetabolic parameters, patient outcomes measures,and qualitative measures would be beneficial to betterunderstand its impact on residents and implementingsustainable policy changes.AbbreviationsACGME: Accreditation Council for Graduate Medical Education; CTU: Clinicalteaching unit; PGY: Post-graduate yearAcknowledgementsNone.FundingNone.Availability of data and materialsThe datasets used and/or analyzed during the current study are availablefrom the corresponding author on reasonable request.Authors’ contributionsSS and KS recruited participants, analyzed and interpreted the data, and wrotethe manuscript. ED was part of the initial recruitment of participants andpreliminary analysis of data. GT was the clinical research coordinator for thestudy. KM conceptualized the project design, analyzed the data, and reviewedthe manuscript. All authors read and approved the final manuscript.Ethics approval and consent to participateEthics was approved by The University of British Columbia Ethics Board, ID#H11–01833. Written consent was obtained from the study participants priorto enrolment in the study.Consent for publicationNot applicable.Competing interestsThe authors declare that they have no competing interests.Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.Author details1Department of Geriatric Medicine, 2775 Laurel Street, 7th Floor, Vancouver,BC V6H 0A5, Canada. 2Department of Geriatric Medicine, Bridgeland SeniorsHealth Clinic, 1070 McDougall Road NE, Calgary, Alberta T2E 8B8, Canada.Received: 27 September 2018 Accepted: 25 April 2019References1. Rosenbaum L, Lamas D. Residents’ duty hours – towards an empiricalnarrative. NEJM. 2012;367(21):2044–9.2. Lockley SW, et al. Effect of reducing interns’ weekly work hours on sleepand attentional failures. NEJM. 2004;351:1829–37.3. Pattani R, et al. Resident duty hours in Canada: past, present, and future.CMAJ. 2014;186(10):761–5.4. Landrigan CP, et al. Effect of reducing interns’ work hours on seriousmedical errors in intensive care units. NEJM. 2004;351:1638–48.5. Veasey S, et al. Sleep loss and fatigue in residency. JAMA. 2002;288(9):1116–24.6. Bilimoria K, et al. National Cluster-Randomized Trial of duty-hour flexibility insurgical training. NEJM. 2016;374:713–27.7. Desai S, et al. Education outcomes in a duty-hour flexibility trial in internalmedicine. NEJM. 2018;378:1494–508.8. Dussault C, et al. 16-hour call duty schedules: the Quebec experience. BMCMed Educ. 2014;14(Suppl 1):S10.9. Andre D, et al. The development of the SenseWear armband, arevolutionary energy assessment device to assess physical activity andlifestyle: Body Media, Inc.10. Fruin M, Rankin J. Validity of a multi-sensor armband in estimating rest andexercise energy expenditure. Med Sci Sports Exerc. 2004;36(6):1063–9.11. Miwa H, et al. Roll-over detection and sleep quality measurement using awearable sensor. Conf Proc IEEE Eng Med Biol Soc. 2007:1507–10.12. Amirian I, et al. Working night shifts affects surgeons’ biological rhythm. AmJ Surg. 2015;210(2):389–95.13. Rose M, et al. Effects of call on sleep and mood in internal medicineresidents. Behav Sleep Med. 2008;6(2):75–88.14. Morhardt D, et al. Determining resident sleep during and after call withcommercial sleep monitoring devices. Urology. 2017;106:39–44.15. Ko J, et al. Call schedule and sleep patterns of urology residents followingthe 2011 ACGME reforms. Urol Pract. 2016;3(2):147–52.16. Piercy KL, et al. The physical activity guidelines for Americans. JAMA. 2018;320(19):2020–8.17. Owen N, et al. Sedentary behavior: emerging evidence for a new health risk.Mayo Clin Proc. 2010;85(12):1138–41.18. Diaz KM, et al. Potential effects of replacing sedentary time with shortsedentary bouts or physical activity on mortality: a National Cohort Study.Am J Epidemiol. 2018:kwy271.Sy et al. Journal of Occupational Medicine and Toxicology           (2019) 14:14 Page 5 of 5