- Email: [email protected]
Impact of Resident Overnight Duty Hour Changes on Obstetrical Outcomes: A Population-Based Cohort Study
EDUCATION
Impact of Resident Overnight Duty Hour Changes on Obstetrical Outcomes: A Population-Based Cohort Study Brian J. Liu, MD, MSc;1 Michael Ordon, MD, MSc;2,3 Janet Bodley, MD, MSc;1 Grace Liu, MD; MSc;1 Jamie Kroft, MD, MSc1 1
Division of Minimally Invasive Gynecologic Surgery, Department of Obstetrics and Gynaecology, Sunnybrook Health
Sciences Centre, Toronto, ON 2
B.J. Liu
Division of Urology, Department of Surgery, St. Michael’s Hospital, Toronto, ON
3
The Institute for Clinical Evaluative Sciences, Toronto, ON
Abstract
Résumé
Objective: To determine whether obstetrical patient outcomes have changed following the introduction of restricted resident work hours.
Objectif : Déterminer si la restriction des heures de travail des résidents a eu un effet sur les résultats de santé des patientes en obstétrique.
Methods: A population-based retrospective cohort study of the effects of restricted duty hours for residents in July 2013 at three academic hospitals in Toronto, ON using linked health care databases. The study included 6763 deliveries in the 2 years preexposure and 5548 deliveries in the 2 years post-exposure.
Méthodologie : Nous avons interrogé des bases de données de soins de santé liées pour mener une étude de cohorte rétrospective basée sur la population portant sur les effets de la restriction des heures de travail des résidents en juillet 2013 dans trois hôpitaux universitaires de Toronto (Ontario). Nous nous sommes penchés sur 6763 accouchements survenus dans les deux ans précédant la restriction et sur 5548 accouchements survenus dans les deux ans suivant celle-ci.
Results: The primary outcome, planned prior to data collection, was a composite index of 29 maternal/fetal outcomes including maternal transfusion/postpartum hemorrhage, maternal infection, fetal mortality, NICU admissions, and surgical/obstetrical complications. There were seven secondary outcomes analysed: NICU admissions; neonatal death; maternal transfusion or postpartum hemorrhage; maternal infection; and three composite measures. A generalized estimating equation model, clustered by institution, was utilized to assess for differences post-intervention. We found no significant differences in baseline demographics between groups. After the implementation of duty hour restrictions, no significant difference was seen in the primary outcome. However, an increased incidence of composite maternal surgical/obstetrical outcomes (OR 1.191; 95% CI 1.037–1.367, P = 0.013) and transfusion/postpartum hemorrhage (OR 1.232; 95% CI 1.074–1.413, P = 0.003) was found. There were no significant differences in other secondary outcomes. Conclusion: Since the implementation of resident duty hour restrictions, there was no overall change in patient outcomes. However, there was an increase in surgical/obstetrical complications and transfusion/postpartum hemorrhage. This suggests that duty hour restrictions may not be beneficial to patient outcomes. It highlights the need to further investigate the clinical impact of a change in resident duty hours. Key Words: Obstetrics, education, quality improvement, pregnancy complications Corresponding Author: Dr. Jamie Kroft, Sunnybrook Health Sciences Centre, Toronto, ON. [email protected] Competing interests: The authors declare that they have no competing interests. Received on December 7, 2017
Résultats : Le résultat primaire à l’étude, choisi avant la collecte de données, était un indice composé de 29 issues maternelles et fœtales, dont la transfusion maternelle et l’hémorragie de la délivrance, l’infection maternelle, la mortalité fœtale, l’admission à l’UNSI et les complications chirurgicales et obstétricales. Sept résultats secondaires ont également été étudiés : l’admission à l’UNSI, le décès néonatal, la transfusion maternelle et l’hémorragie de la délivrance, l’infection maternelle et trois mesures composées. Un modèle d’équation d’estimation généralisée regroupant les données par établissement a été utilisé pour évaluer les différences post-intervention. Aucune différence significative n’a été observée entre les groupes quant aux caractéristiques démographiques de référence. Nous n’avons observé aucun changement significatif du résultat primaire après la restriction des heures de travail, mais avons constaté une augmentation de la mesure composée pour les complications chirurgicales et obstétricales maternelles (RC : 1191; IC à 95 % : 1037–1367; P = 0,013) et des transfusions et des hémorragies de la délivrance (RC : 1232; IC à 95 % : 1074–1413; P = 0,003). Aucune différence significative n’a été observée quant aux autres résultats secondaires. Conclusion : La restriction des heures de travail des résidents n’a pas eu d’effet sur les résultats de santé des patientes. Une augmentation des complications chirurgicales et obstétricales ainsi que des transfusions et des hémorragies de la délivrance a toutefois été constatée. Ces résultats laissent penser que la restriction des heures de travail pourrait ne pas être avantageuse sur le plan des résultats de santé des patientes, et mettent en évidence le besoin d’étudier davantage les effets cliniques des changements apportés aux heures de travail des résidents.
■■ JOGC ■■ 2018 • 1
EDUCATION
Copyright © 2018 The Society of Obstetricians and Gynaecologists of Canada/La Société des obstétriciens et gynécologues du Canada. Published by Elsevier Inc. All rights reserved.
J Obstet Gynaecol Can 2018;■■(■■):■■–■■ https://doi.org/10.1016/j.jogc.2018.02.001
INTRODUCTION
M
edical residency programs have historically imposed long shifts on residents, a tradition that has raised concerns about patient and resident safety. Evidence shows that acute sleep deprivation can lead to negative effects on mood, cognition, and psychomotor function,1–3 and can have similar effects as alcohol consumption.4,5 The United States Accreditation Council of Graduate Medical Education imposed duty hour restrictions in 2011.6 Following these changes, a pan-Canadian steering committee released suggestions to limit resident work hours in 2013.7 Nevertheless, these changes are controversial, as there are conflicting data regarding the effects of this new system, in the context of both patient outcomes and the resident experience.6,8–14 The obstetrics program at the University of Toronto has recently moved away from 24-hour shifts and adopted duty hour restrictions in the form of a night float (NF) system to improve patient outcomes and resident satisfaction. In this system, residents typically work for 12–14 hours overnight for four to five consecutive nights in a row. Upon completion of their NF block, they typically have one rest day before returning to a day schedule. The primary aim of this study is to determine whether the new system of restricted resident work hours has influenced Toronto obstetrical patient outcomes. MATERIALS AND METHODS
We conducted a population-based, retrospective cohort study in Toronto, ON using health care administrative databases. Patient level data was linked using encoded identifiers and analyzed at the Institute for Clinical Evaluative Sciences. The study was approved by the Research Ethics Boards at Sunnybrook Health Sciences Centre (Project Identification Number 507–2015) and the Institute for Clinical Evaluative Sciences. Reporting of this study follows guidelines set for observational studies.15 Three health care databases were used: the Ontario Health Insurance Plan (OHIP), the Canadian Institute for Health Information-Discharge Abstract Database (CIHI-DAD) and the Mom-Baby Database (MOMBABY). These databases
2 • ■■ JOGC ■■ 2018
are routinely used for research purposes, and their quality has been previously demonstrated.16 In the province of Ontario, OHIP is a single payer universal health care insurance plan. The OHIP dataset contains all of the claims paid for by OHIP from July 1991 onward. The CIHIDAD is a national database of all admissions to acute care institutions. The quality of CIHI-DAD for coding accuracy has been demonstrated with re-abstraction studies.17 The MOMBABY is a dataset linking the admission records of delivering mothers and their newborns. The study included all obstetrical patients who underwent an overnight delivery of any kind between July 2011 and June 2015 at one of three academic hospitals that have Obstetrical residents in Toronto. The residents rotate between all three sites. These academic centres were Sunnybrook Health Sciences Centre, Mount Sinai Hospital, and St. Michael’s Hospital. No significant changes in resident staffing occurred between the three sites either before or after the intervention, except duty hour restrictions. The preintervention cohort comprised of patients who delivered between midnight and 7:00 AM before resident duty hour restrictions took effect (July 2011–June 2013). The delivery time was chosen because we postulated that fatigue would be highest at this time and deliveries between midnight and 7:00 AM have a unique OHIP fee code, allowing them to be identified from the cohort. Shifts for residents at all three sites during this time period lasted 24 hours. The postintervention cohort included patients who delivered between midnight and 7:00 AM after the implementation of the NF system (July 2013–June 2015). NF at the three hospitals involved 12–14 hour shifts for four to five consecutive nights. Baseline patient characteristics recorded for all patients in the study cohort included: maternal age, maternal Aggregated Diagnosis Groups score (a standardized comorbidity index),18 parity, multiple gestation, Caesarean delivery, preterm labour, placenta previa, and post dates (>40 weeks GA). Patients were excluded if they had invalid Institute for Clinical Evaluative Sciences key numbers (an anonymized patient identifier), missing mother’s or baby’s sex (because this data could be inaccurate or incomplete), and if the delivering physician was not an obstetrician (because family physicians and midwives do not have residents that work under the night float system). Deliveries that did not occur between midnight and 7:00 AM at the three academic hospitals listed above were also excluded to ensure that we only accounted for deliveries that occurred during duty hour restricted NF shifts. The primary outcome was a composite index of 29 maternal and fetal outcomes measured up to 30 days postpartum
Impact of Resident Overnight Duty Hour Changes on Obstetrical Outcomes: A Population-Based Cohort Study
(see online Appendices 1, 2, and 3). The seven secondary outcomes included: a composite of 18 maternal outcomes (see online Appendix 1); maternal transfusion or postpartum hemorrhage (PPH); maternal infection; a composite of 11 surgical/obstetrical complications that were a subset of maternal outcomes (see online Appendix 2); a composite of 11 fetal outcomes (see online Appendix 3); neonatal death within 30 days, and NICU admissions.
Table 1. Baseline demographics of study population Characteristics Mother’s characteristics
N = 6739
N = 5208
Mean (±SD)
32.7 (4.94)
32.9 (4.86)
Median (IQR)
33 (30–36)
33 (30–36)
49 (0.7)
37 (0.7)
Mother’s age at delivery (years)
ADG score, n (%) 0–1
The baseline characteristics between the pre-intervention and post-intervention cohorts were compared using standardized differences. This metric describes differences between group means relative to the pooled standard deviation; differences greater than 0.10 reflect the potential for meaningful imbalance.19 Multivariate regression analysis was used to determine if there was an association between timing of delivery (before or after the introduction of NF) and the primary and secondary outcomes. A generalized estimating equation model was used, clustered by institution, to control for the location of delivery. The models were adjusted for patient baseline characteristics (see online Appendix 4). All analyses were completed with SAS version 9.4 (SAS Institute Inc., Cary, NC). RESULTS
The pre-intervention cohort consisted of 6763 deliveries, and the post-intervention cohort 5548 deliveries. In the preintervention cohort, 1848 deliveries occurred at Sunnybrook Health Sciences Centre, 3295 occurred at Mount Sinai Hospital, and 1620 occurred at St. Michael’s Hospital. In the post-intervention cohort, 1587 deliveries occurred at Sunnybrook Health Sciences Centre, 2750 occurred at Mount Sinai Hospital, and 1211 occurred at St. Michael’s Hospital. There were no significant differences in baseline patient demographics between groups (Table 1). There was no significant difference between groups for the primary outcome, a composite index of all maternal and fetal outcomes, with an OR of 0.968 (95% CI 0.933–1.004, P = 0.082). Maternal outcomes included complications such as maternal mortality, transfusion or PPH, infection, reoperation, third- or fourth-degree lacerations, retained placenta, uterine inversion, and disruption of obstetrical/ surgical wounds. Fetal outcomes included fetal mortality, NICU admissions, intracranial hemorrhage or scalp laceration, birth injuries, and asphyxia. In terms of the secondary outcomes, there was no significant difference between groups for the composite of all maternal outcomes (OR 1.026; 95% CI 0.882–1.193, P = 0.741), maternal infection (OR 0.950; 95% CI 0.673–1.339, P = 0.766), composite of all fetal
Pre-intervention Post-intervention
2–3
664 (9.9)
443 (8.5)
4–5
1496 (22.2)
1064 (20.4)
6–9
3297 (48.9)
2647 (50.8)
10 +
1233 (18.3)
1017 (19.5)
History of diabetes, n (%) Yes
187 (2.8)
145 (2.8)
No
6552 (97.2)
5063 (97.2)
Yes
127 (1.9)
112 (2.2)
No
6612 (98.1)
5096 (97.8)
Yes
2416 (35.9)
1688 (32.4)
No
4323 (64.1)
3520 (67.6)
N = 6995
N = 5730
Female
3401 (48.6)
2844 (49.6)
Male
3594 (51.4)
2886 (50.4)
2011
1805 (25.8)
0 (0.0)
2012
3470 (49.6)
0 (0.0)
2013
1720 (24.6)
1748 (30.5)
2014
0 (0.0)
3465 (60.5)
History of hypertension, n (%)
Parity, n (%)
Baby’s characteristics Baby’s sex, n (%)
Baby’s year of birth (years), n (%)
2015
0 (0.0)
517 (9)
N = 6763
N = 5548
Sunnybrook Health Sciences Centre
1848 (27.3)
1587 (28.6)
Mount Sinai Hospital
3295 (48.7)
2750 (49.6)
St. Michael’s Hospital
1620 (24)
1211 (21.8)
Yes
233 (3.4)
183 (3.3)
No
6530 (96.6)
5365 (96.7)
Yes
1774 (26.2)
1386 (25)
No
4989 (73.8)
4162 (75)
Delivery characteristics Hospital delivered, n (%)
Multiple births, n (%)
Caesarean delivery, n (%)
Preterm labour, n (%) Yes
656 (9.7)
547 (9.9)
No
6107 (90.3)
5001 (90.1) Continued
■■ JOGC ■■ 2018 • 3
EDUCATION
DISCUSSION
Table 1. Continued Characteristics
Pre-intervention
Post-intervention
Yes
734 (10.9)
660 (11.9)
No
6029 (89.1)
4888 (88.1)
Yes
463 (6.8)
441 (7.9)
No
6300 (93.2)
5107 (92.1)
Although sleep deprivation may have negative effects on resident physicians, our population-based, retrospective cohort study showed no significant difference in most outcomes before and after the implementation of a resident NF system. There was, however, an increased incidence of maternal transfusion/PPH and an increase in the composite of all surgical/obstetrical complications after duty hour restrictions. These findings question the validity of one of the main benefits of duty hour restrictions: improved patient outcomes.
Premature rupture of membranes, n (%)
Preeclampsia, n (%)
Placenta previa, n (%) Yes
51 (0.8)
22 (0.4)
No
6712 (99.2)
5526 (99.6)
Post-dates, n (%) Yes
728 (10.8)
552 (9.9)
No
6035 (89.2)
4996 (90.1)
Yes
11 (0.2)
9 (0.2)
No
6752 (99.8)
5539 (99.8)
Stillbirth, n (%)
ADG: Aggregated Diagnosis Groups; IQR: interquartile range.
outcomes (OR 0.905; 95% CI 0.747–1.096, P = 0.305), fetal mortality (OR 0.847; 95% CI 0.504–1.423, P = 0.531), and NICU admissions (OR 0.910; 95% CI 0.747–1.109, P = 0.351). However, after resident work hour restrictions, there was a significant increased risk of maternal transfusion or postpartum hemorrhage (OR 1.232; 95% CI 1.074– 1.413, P = 0.003) and for the composite of all surgical/ obstetrical complications (OR 1.191; 95% CI 1.037–1.367, P = 0.013). See Table 2 for regression analyses with covariates that were included and Table 3 for the unadjusted results.
This study is a large population based study including over 12 000 deliveries over a four-year timespan. It addresses the widely-debated issue of the safety of duty hour restrictions in a specialty where studies have been limited. In the US, there are limited data that mortality and morbidity in obstetrics did not significantly change after duty hour restrictions.6,13 On the other hand, a 2017 study showed that there was a modest decrease in obstetrical complications after duty hour restrictions, but only during the first two months of the resident academic year, referred to as the “July effect.”14 Obstetrics is unique in that resident workload on the labour floor is usually consistent throughout a 24hour period compared with other medical and surgical specialties where clinical responsibilities may lighten overnight. The seemingly neutral, or possibly negative effects of work hour limits may be attributed to several factors. By decreasing a resident’s shift length, patient care handovers inevitably
Table 2. Adjusted GEE model results Post-intervention OR
CI
P
0.968
0.933–1.004
0.082
Neonatal deathb
0.847
0.504–1.423
0.53
Composite of all fetal outcomesa
0.905
0.747–1.096
0.31
Composite of all maternal outcomesc
1.026
0.882–1.193
0.74
Primary outcome Composite of all maternal and fetal outcomesa Secondary outcomes
a
0.910
0.747–1.109
0.35
Maternal transfusion or postpartum hemorrhagec
1.232
1.074–1.413
0.003
Maternal infectiond
0.949
0.673–1.339
0.77
Composite of all surgical/obstetrical complicationsc
1.191
1.037–1.367
0.013
NICU admissions
GEE: generalized estimating equation. NOTE: All outcomes were measured up to 30 days post delivery. a
Covariates included: mother’s age, mother’s ADG score, parity, multiple gestation.
b
Covariates included: maternal age, maternal ADG score, parity, multiple gestation, preterm labour, Caesarean delivery.
c
Covariates included: mother’s age, mother’s ADG score, parity, multiple gestation, Caesarean delivery, preterm labour, placenta previa, post dates.
d
Covariates included: mother’s age, mother’s ADG score, parity, multiple gestation, preterm labour, post dates.
4 • ■■ JOGC ■■ 2018
Impact of Resident Overnight Duty Hour Changes on Obstetrical Outcomes: A Population-Based Cohort Study
Table 3. Unadjusted results Pre-Intervention rate per 1000 deliveries
Post-Intervention rate per 1000 deliveries
P value
443.32
447.29
0.653
Primary outcome Composite of all maternal and fetal outcomes Secondary outcomes Neonatal death
10.01
8.55
0.396
Composite of all fetal outcomes
170.26
159.34
0.099
Composite of all maternal outcomes
341.42
354
0.144
NICU admissions
167.55
157.42
0.124
Maternal transfusion or postpartum hemorrhage
51.60
64.53
0.002
Maternal infection
22.18
21.45
0.783
Composite of all surgical/obstetrical outcomes
65.80
79.67
0.003
All outcomes were measured up to 30 days post-delivery.
increase. It has been established that handovers are a significant source of medical errors.20 In fact, a resident’s discontinuity of care due to handovers has been shown to increase preventable adverse events.21,22 Studies have discovered that with each additional handover, patient information is omitted or inaccuracies are added which can affect patient outcomes.22–24 Another possible explanation for why limiting resident work hours has not improved patient outcomes is that the overall fatigue levels have remained similar. This is likely due to the structure of NF, in which the residents worked for 12–14 hours for four or five consecutive nights in a row, took one day off for recovery, and then returned to a day schedule. A recent survey of internal medicine residents in a similar NF system revealed that, although they experienced reduced fatigue in the early morning, they felt worsened long-term fatigue compared with the traditional 26-hour system.25 This increased fatigue over time may counteract any short-term beneficial effects that would be gained from duty hour restrictions. In comparison, other health care providers such as registered nurses also work under similar schedules, although normally with fewer consecutive nights or more time off work in between blocks of shifts. This study is limited by the fact that it is an observational study and cannot infer causation. As well, the three academic sites in Toronto may have had slight variations in the number of consecutive nights that residents worked, or how frequently residents were on night float, which could not be controlled for in this study. Finally, one could argue that patient outcomes may not have been influenced substantially because staff obstetricians were present at all deliveries. Staff obstetricians at the three centres included generally work 24-hour shifts and that remained
the same throughout the study period. This may have neutralized some effects to patient care; however, generally, the resident physicians manage most patient care overnight leading up to the delivery and act as the primary surgeon/ attendant during the delivery. Therefore, certainly the residents can substantially impact patient outcomes. As well, the outcomes that were significantly impacted during the study (surgical/obstetrical outcomes and transfusion and PPH) are both factors that are likely more dependent on resident management given they act as primary surgeon/ attendant. Certainly, further study on the impact of changes to duty hours for staff obstetricians needs to be examined as well.
CONCLUSION
We found that the implementation of the night float system for Obstetrical residents in Toronto did not cause a change in overall patient outcomes, but there was an increased incidence of transfusion and PPH and surgical/obstetrical complications. These results suggest that duty hour restrictions may not be as beneficial as initially thought, and adds to the controversial data from other specialties. It certainly highlights the need to further investigate the clinical impact a change in resident duty hours has, as well as the effect on resident well-being and education, to determine the best strategy to adopt moving forward. Acknowledgements
This study was funded by the Sunnybrook Education Advisory Council Education Research Unit Education Research and Scholarship Grant. Special thanks to Rafik Saskin and Eliane Kim at the Institute for Clinical Evaluative Sciences. This work has been presented as an oral presentation
■■ JOGC ■■ 2018 • 5
EDUCATION
at the American Congress of Obstetricians and Gynecologists Annual Clinical Meeting, May 8, 2017. SUPPLEMENTARY DATA
Supplementary data related to this article can be found online at https://doi.org/10.1016/j.jogc.2018.02.001. REFERENCES 1. Olson E, Drage L, Auger R. Sleep deprivation, physician performance, and patient safety. Chest 2009;136:1389–96.
12. Scally C, Ryan A, Thumma J, et al. Early impact of the 2011 ACGME duty hour regulations on surgical outcomes. Surgery 2015;158:1453–61. 13. Bailit J, Blanchard M. The effect of house staff working hours on the quality of obstetric and gynecologic care. Obstet Gynecol 2004;103:613– 6. 14. Peltan ID, Brown CE, Burke AK, et al. The July Effect on Maternal Peripartum Complications Before and After Resident Duty Hour Reform: A Population-Based Retrospective Cohort Study. Am J Perinatol 2017;34:818–25. 15. Von Elm E, Altman DG, Egger M, et al. The strengthening the reporting of observational studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. J Clin Epidemiol 2008;61:344.
2. Phillbert I. Sleep loss and performance in residents and nonphysicians: a meta-analytic examination. Sleep 2005;1392–402.
16. Williams J, Young WA. Summary of Studies on the Quality of Health Care Administrative Databases in Canada. 339–345 (Canadian Medical Association); 1996.
3. Wilhelm B, Widmann A, Durst W, et al. Objective and quantitative analysis of daytime sleepiness in physicians after night duties. Int J Psychophysiol 2009;72:307–13.
17. Canadian Institute for Health Information. CIHI Data Quality Study of the 2009–2010 Discharge Abstract Database. (CIHI, Ottawa, ON); 2012.
4. Arnedt J, Owens J, Crouch M, et al. Neurobehavioral performance of residents after heavy night call vs after alcohol ingestion. JAMA 2005;294:1025–33. 5. Dawson D, Reid K. Fatigue, alcohol and performance impairment. Nature 1997;388:235. 6. Rajaram R, Chung JW, Cohen ME, et al. Association of the 2011 ACGME resident duty hour reform with postoperative patient outcomes in surgical specialties. J Am Coll Surg 2015;221:748–57. 7. National Steering Committee on Resident Duty Hours. Fatigue, Risk and Excellence: Towards a Pan-Canadian Consensus on Resident Duty Hours. (The Royal College of Physicians and Surgeons of Canada, Ottawa, Ontario); 2013. 8. Ahmed N, Devitt KS, Keshet I, et al. A systematic review of the effects of resident duty hour restrictions in surgery: impact on resident wellness, training, and patient outcomes. Ann Surg 2014;259:1041–53. 9. Philibert I, Nasca T, Brigham T, et al. Duty-hour limits and patient care and resident outcomes: can high-quality studies offer insight into complex relationships? Annu Rev Med 2013;64:467–83.
18. The Johns Hopkins University Bloomberg School of Public Health. The Johns Hopkins ACG Case-Mix Documentation & Application Manual Version 5.0. Johns Hopkins University; 2001. 19. Austin P. Using the standardized difference to compare the prevalance of a binary variable between two groups in observational research. Commun Stat Simul Comput 2009;38:1228–34. 20. Institute of Medicine (US) Committee on Optimizing Graduate Medical Trainee (Resident) Hours and Work Schedule to Improve Patient Safety. Resident Duty Hours: Enhancing Sleep, Supervision, and Safety. Washington DC: National Academies Press (US); 2009. 21. Laine C, Goldman L, Soukup J, Hayes J. The impact of a regulation restricting medical house staff working hours on the quality of patient care. JAMA 1993;269:374–8. 22. Petersen L, Brennan T, O’Neil A, et al. Does housestaff discontinuity of care increase the risk for preventable adverse events? Ann Intern Med 1994;121:866–72. 23. Arora V, Kao J, Lovinger D. Medication discrepancies in resident signouts and their potential to harm. J Gen Intern Med 2007;22:1751–5.
10. Jamal M, Wong S, Whalen T. Effects of the reduction of surgical residents’ work hours and implications for surgical residency programs: a narrative review. BMC Med Educ 2014;14:S14.
24. Horwitz L, Krumholz H, Green M, et al. Transfers of patient care between house staff on internal medicine wards: A national survey. Arch Intern Med 2006;166:1173–7.
11. Jamal M, Doi SA, Rousseau M, et al. Systematic review and meta-analysis of the effect of North American working hours restrictions on mortality and morbidity in surgical patients. Br J Surg 2012;99:336–44.
25. Mathew R, Gundy S, Ulic D, et al. A Reduced Duty Hours Model for Senior Internal Medicine Residents: A Qualitative Analysis of Residents’ Experiences and Perceptions. Acad Med 2016;91:1284–92.
6 • ■■ JOGC ■■ 2018
Impact of Resident Overnight Duty Hour Changes on Obstetrical Outcomes: A Population-Based Cohort Study Brian J. Liu, MD, MSc;1 Michael Ordon, MD, MSc;2,3 Janet Bodley, MD, MSc;1 Grace Liu, MD; MSc;1 Jamie Kroft, MD, MSc1 1
Division of Minimally Invasive Gynecologic Surgery, Department of Obstetrics and Gynaecology, Sunnybrook Health
Sciences Centre, Toronto, ON 2
B.J. Liu
Division of Urology, Department of Surgery, St. Michael’s Hospital, Toronto, ON
3
The Institute for Clinical Evaluative Sciences, Toronto, ON
Abstract
Résumé
Objective: To determine whether obstetrical patient outcomes have changed following the introduction of restricted resident work hours.
Objectif : Déterminer si la restriction des heures de travail des résidents a eu un effet sur les résultats de santé des patientes en obstétrique.
Methods: A population-based retrospective cohort study of the effects of restricted duty hours for residents in July 2013 at three academic hospitals in Toronto, ON using linked health care databases. The study included 6763 deliveries in the 2 years preexposure and 5548 deliveries in the 2 years post-exposure.
Méthodologie : Nous avons interrogé des bases de données de soins de santé liées pour mener une étude de cohorte rétrospective basée sur la population portant sur les effets de la restriction des heures de travail des résidents en juillet 2013 dans trois hôpitaux universitaires de Toronto (Ontario). Nous nous sommes penchés sur 6763 accouchements survenus dans les deux ans précédant la restriction et sur 5548 accouchements survenus dans les deux ans suivant celle-ci.
Results: The primary outcome, planned prior to data collection, was a composite index of 29 maternal/fetal outcomes including maternal transfusion/postpartum hemorrhage, maternal infection, fetal mortality, NICU admissions, and surgical/obstetrical complications. There were seven secondary outcomes analysed: NICU admissions; neonatal death; maternal transfusion or postpartum hemorrhage; maternal infection; and three composite measures. A generalized estimating equation model, clustered by institution, was utilized to assess for differences post-intervention. We found no significant differences in baseline demographics between groups. After the implementation of duty hour restrictions, no significant difference was seen in the primary outcome. However, an increased incidence of composite maternal surgical/obstetrical outcomes (OR 1.191; 95% CI 1.037–1.367, P = 0.013) and transfusion/postpartum hemorrhage (OR 1.232; 95% CI 1.074–1.413, P = 0.003) was found. There were no significant differences in other secondary outcomes. Conclusion: Since the implementation of resident duty hour restrictions, there was no overall change in patient outcomes. However, there was an increase in surgical/obstetrical complications and transfusion/postpartum hemorrhage. This suggests that duty hour restrictions may not be beneficial to patient outcomes. It highlights the need to further investigate the clinical impact of a change in resident duty hours. Key Words: Obstetrics, education, quality improvement, pregnancy complications Corresponding Author: Dr. Jamie Kroft, Sunnybrook Health Sciences Centre, Toronto, ON. [email protected] Competing interests: The authors declare that they have no competing interests. Received on December 7, 2017
Résultats : Le résultat primaire à l’étude, choisi avant la collecte de données, était un indice composé de 29 issues maternelles et fœtales, dont la transfusion maternelle et l’hémorragie de la délivrance, l’infection maternelle, la mortalité fœtale, l’admission à l’UNSI et les complications chirurgicales et obstétricales. Sept résultats secondaires ont également été étudiés : l’admission à l’UNSI, le décès néonatal, la transfusion maternelle et l’hémorragie de la délivrance, l’infection maternelle et trois mesures composées. Un modèle d’équation d’estimation généralisée regroupant les données par établissement a été utilisé pour évaluer les différences post-intervention. Aucune différence significative n’a été observée entre les groupes quant aux caractéristiques démographiques de référence. Nous n’avons observé aucun changement significatif du résultat primaire après la restriction des heures de travail, mais avons constaté une augmentation de la mesure composée pour les complications chirurgicales et obstétricales maternelles (RC : 1191; IC à 95 % : 1037–1367; P = 0,013) et des transfusions et des hémorragies de la délivrance (RC : 1232; IC à 95 % : 1074–1413; P = 0,003). Aucune différence significative n’a été observée quant aux autres résultats secondaires. Conclusion : La restriction des heures de travail des résidents n’a pas eu d’effet sur les résultats de santé des patientes. Une augmentation des complications chirurgicales et obstétricales ainsi que des transfusions et des hémorragies de la délivrance a toutefois été constatée. Ces résultats laissent penser que la restriction des heures de travail pourrait ne pas être avantageuse sur le plan des résultats de santé des patientes, et mettent en évidence le besoin d’étudier davantage les effets cliniques des changements apportés aux heures de travail des résidents.
■■ JOGC ■■ 2018 • 1
EDUCATION
Copyright © 2018 The Society of Obstetricians and Gynaecologists of Canada/La Société des obstétriciens et gynécologues du Canada. Published by Elsevier Inc. All rights reserved.
J Obstet Gynaecol Can 2018;■■(■■):■■–■■ https://doi.org/10.1016/j.jogc.2018.02.001
INTRODUCTION
M
edical residency programs have historically imposed long shifts on residents, a tradition that has raised concerns about patient and resident safety. Evidence shows that acute sleep deprivation can lead to negative effects on mood, cognition, and psychomotor function,1–3 and can have similar effects as alcohol consumption.4,5 The United States Accreditation Council of Graduate Medical Education imposed duty hour restrictions in 2011.6 Following these changes, a pan-Canadian steering committee released suggestions to limit resident work hours in 2013.7 Nevertheless, these changes are controversial, as there are conflicting data regarding the effects of this new system, in the context of both patient outcomes and the resident experience.6,8–14 The obstetrics program at the University of Toronto has recently moved away from 24-hour shifts and adopted duty hour restrictions in the form of a night float (NF) system to improve patient outcomes and resident satisfaction. In this system, residents typically work for 12–14 hours overnight for four to five consecutive nights in a row. Upon completion of their NF block, they typically have one rest day before returning to a day schedule. The primary aim of this study is to determine whether the new system of restricted resident work hours has influenced Toronto obstetrical patient outcomes. MATERIALS AND METHODS
We conducted a population-based, retrospective cohort study in Toronto, ON using health care administrative databases. Patient level data was linked using encoded identifiers and analyzed at the Institute for Clinical Evaluative Sciences. The study was approved by the Research Ethics Boards at Sunnybrook Health Sciences Centre (Project Identification Number 507–2015) and the Institute for Clinical Evaluative Sciences. Reporting of this study follows guidelines set for observational studies.15 Three health care databases were used: the Ontario Health Insurance Plan (OHIP), the Canadian Institute for Health Information-Discharge Abstract Database (CIHI-DAD) and the Mom-Baby Database (MOMBABY). These databases
2 • ■■ JOGC ■■ 2018
are routinely used for research purposes, and their quality has been previously demonstrated.16 In the province of Ontario, OHIP is a single payer universal health care insurance plan. The OHIP dataset contains all of the claims paid for by OHIP from July 1991 onward. The CIHIDAD is a national database of all admissions to acute care institutions. The quality of CIHI-DAD for coding accuracy has been demonstrated with re-abstraction studies.17 The MOMBABY is a dataset linking the admission records of delivering mothers and their newborns. The study included all obstetrical patients who underwent an overnight delivery of any kind between July 2011 and June 2015 at one of three academic hospitals that have Obstetrical residents in Toronto. The residents rotate between all three sites. These academic centres were Sunnybrook Health Sciences Centre, Mount Sinai Hospital, and St. Michael’s Hospital. No significant changes in resident staffing occurred between the three sites either before or after the intervention, except duty hour restrictions. The preintervention cohort comprised of patients who delivered between midnight and 7:00 AM before resident duty hour restrictions took effect (July 2011–June 2013). The delivery time was chosen because we postulated that fatigue would be highest at this time and deliveries between midnight and 7:00 AM have a unique OHIP fee code, allowing them to be identified from the cohort. Shifts for residents at all three sites during this time period lasted 24 hours. The postintervention cohort included patients who delivered between midnight and 7:00 AM after the implementation of the NF system (July 2013–June 2015). NF at the three hospitals involved 12–14 hour shifts for four to five consecutive nights. Baseline patient characteristics recorded for all patients in the study cohort included: maternal age, maternal Aggregated Diagnosis Groups score (a standardized comorbidity index),18 parity, multiple gestation, Caesarean delivery, preterm labour, placenta previa, and post dates (>40 weeks GA). Patients were excluded if they had invalid Institute for Clinical Evaluative Sciences key numbers (an anonymized patient identifier), missing mother’s or baby’s sex (because this data could be inaccurate or incomplete), and if the delivering physician was not an obstetrician (because family physicians and midwives do not have residents that work under the night float system). Deliveries that did not occur between midnight and 7:00 AM at the three academic hospitals listed above were also excluded to ensure that we only accounted for deliveries that occurred during duty hour restricted NF shifts. The primary outcome was a composite index of 29 maternal and fetal outcomes measured up to 30 days postpartum
Impact of Resident Overnight Duty Hour Changes on Obstetrical Outcomes: A Population-Based Cohort Study
(see online Appendices 1, 2, and 3). The seven secondary outcomes included: a composite of 18 maternal outcomes (see online Appendix 1); maternal transfusion or postpartum hemorrhage (PPH); maternal infection; a composite of 11 surgical/obstetrical complications that were a subset of maternal outcomes (see online Appendix 2); a composite of 11 fetal outcomes (see online Appendix 3); neonatal death within 30 days, and NICU admissions.
Table 1. Baseline demographics of study population Characteristics Mother’s characteristics
N = 6739
N = 5208
Mean (±SD)
32.7 (4.94)
32.9 (4.86)
Median (IQR)
33 (30–36)
33 (30–36)
49 (0.7)
37 (0.7)
Mother’s age at delivery (years)
ADG score, n (%) 0–1
The baseline characteristics between the pre-intervention and post-intervention cohorts were compared using standardized differences. This metric describes differences between group means relative to the pooled standard deviation; differences greater than 0.10 reflect the potential for meaningful imbalance.19 Multivariate regression analysis was used to determine if there was an association between timing of delivery (before or after the introduction of NF) and the primary and secondary outcomes. A generalized estimating equation model was used, clustered by institution, to control for the location of delivery. The models were adjusted for patient baseline characteristics (see online Appendix 4). All analyses were completed with SAS version 9.4 (SAS Institute Inc., Cary, NC). RESULTS
The pre-intervention cohort consisted of 6763 deliveries, and the post-intervention cohort 5548 deliveries. In the preintervention cohort, 1848 deliveries occurred at Sunnybrook Health Sciences Centre, 3295 occurred at Mount Sinai Hospital, and 1620 occurred at St. Michael’s Hospital. In the post-intervention cohort, 1587 deliveries occurred at Sunnybrook Health Sciences Centre, 2750 occurred at Mount Sinai Hospital, and 1211 occurred at St. Michael’s Hospital. There were no significant differences in baseline patient demographics between groups (Table 1). There was no significant difference between groups for the primary outcome, a composite index of all maternal and fetal outcomes, with an OR of 0.968 (95% CI 0.933–1.004, P = 0.082). Maternal outcomes included complications such as maternal mortality, transfusion or PPH, infection, reoperation, third- or fourth-degree lacerations, retained placenta, uterine inversion, and disruption of obstetrical/ surgical wounds. Fetal outcomes included fetal mortality, NICU admissions, intracranial hemorrhage or scalp laceration, birth injuries, and asphyxia. In terms of the secondary outcomes, there was no significant difference between groups for the composite of all maternal outcomes (OR 1.026; 95% CI 0.882–1.193, P = 0.741), maternal infection (OR 0.950; 95% CI 0.673–1.339, P = 0.766), composite of all fetal
Pre-intervention Post-intervention
2–3
664 (9.9)
443 (8.5)
4–5
1496 (22.2)
1064 (20.4)
6–9
3297 (48.9)
2647 (50.8)
10 +
1233 (18.3)
1017 (19.5)
History of diabetes, n (%) Yes
187 (2.8)
145 (2.8)
No
6552 (97.2)
5063 (97.2)
Yes
127 (1.9)
112 (2.2)
No
6612 (98.1)
5096 (97.8)
Yes
2416 (35.9)
1688 (32.4)
No
4323 (64.1)
3520 (67.6)
N = 6995
N = 5730
Female
3401 (48.6)
2844 (49.6)
Male
3594 (51.4)
2886 (50.4)
2011
1805 (25.8)
0 (0.0)
2012
3470 (49.6)
0 (0.0)
2013
1720 (24.6)
1748 (30.5)
2014
0 (0.0)
3465 (60.5)
History of hypertension, n (%)
Parity, n (%)
Baby’s characteristics Baby’s sex, n (%)
Baby’s year of birth (years), n (%)
2015
0 (0.0)
517 (9)
N = 6763
N = 5548
Sunnybrook Health Sciences Centre
1848 (27.3)
1587 (28.6)
Mount Sinai Hospital
3295 (48.7)
2750 (49.6)
St. Michael’s Hospital
1620 (24)
1211 (21.8)
Yes
233 (3.4)
183 (3.3)
No
6530 (96.6)
5365 (96.7)
Yes
1774 (26.2)
1386 (25)
No
4989 (73.8)
4162 (75)
Delivery characteristics Hospital delivered, n (%)
Multiple births, n (%)
Caesarean delivery, n (%)
Preterm labour, n (%) Yes
656 (9.7)
547 (9.9)
No
6107 (90.3)
5001 (90.1) Continued
■■ JOGC ■■ 2018 • 3
EDUCATION
DISCUSSION
Table 1. Continued Characteristics
Pre-intervention
Post-intervention
Yes
734 (10.9)
660 (11.9)
No
6029 (89.1)
4888 (88.1)
Yes
463 (6.8)
441 (7.9)
No
6300 (93.2)
5107 (92.1)
Although sleep deprivation may have negative effects on resident physicians, our population-based, retrospective cohort study showed no significant difference in most outcomes before and after the implementation of a resident NF system. There was, however, an increased incidence of maternal transfusion/PPH and an increase in the composite of all surgical/obstetrical complications after duty hour restrictions. These findings question the validity of one of the main benefits of duty hour restrictions: improved patient outcomes.
Premature rupture of membranes, n (%)
Preeclampsia, n (%)
Placenta previa, n (%) Yes
51 (0.8)
22 (0.4)
No
6712 (99.2)
5526 (99.6)
Post-dates, n (%) Yes
728 (10.8)
552 (9.9)
No
6035 (89.2)
4996 (90.1)
Yes
11 (0.2)
9 (0.2)
No
6752 (99.8)
5539 (99.8)
Stillbirth, n (%)
ADG: Aggregated Diagnosis Groups; IQR: interquartile range.
outcomes (OR 0.905; 95% CI 0.747–1.096, P = 0.305), fetal mortality (OR 0.847; 95% CI 0.504–1.423, P = 0.531), and NICU admissions (OR 0.910; 95% CI 0.747–1.109, P = 0.351). However, after resident work hour restrictions, there was a significant increased risk of maternal transfusion or postpartum hemorrhage (OR 1.232; 95% CI 1.074– 1.413, P = 0.003) and for the composite of all surgical/ obstetrical complications (OR 1.191; 95% CI 1.037–1.367, P = 0.013). See Table 2 for regression analyses with covariates that were included and Table 3 for the unadjusted results.
This study is a large population based study including over 12 000 deliveries over a four-year timespan. It addresses the widely-debated issue of the safety of duty hour restrictions in a specialty where studies have been limited. In the US, there are limited data that mortality and morbidity in obstetrics did not significantly change after duty hour restrictions.6,13 On the other hand, a 2017 study showed that there was a modest decrease in obstetrical complications after duty hour restrictions, but only during the first two months of the resident academic year, referred to as the “July effect.”14 Obstetrics is unique in that resident workload on the labour floor is usually consistent throughout a 24hour period compared with other medical and surgical specialties where clinical responsibilities may lighten overnight. The seemingly neutral, or possibly negative effects of work hour limits may be attributed to several factors. By decreasing a resident’s shift length, patient care handovers inevitably
Table 2. Adjusted GEE model results Post-intervention OR
CI
P
0.968
0.933–1.004
0.082
Neonatal deathb
0.847
0.504–1.423
0.53
Composite of all fetal outcomesa
0.905
0.747–1.096
0.31
Composite of all maternal outcomesc
1.026
0.882–1.193
0.74
Primary outcome Composite of all maternal and fetal outcomesa Secondary outcomes
a
0.910
0.747–1.109
0.35
Maternal transfusion or postpartum hemorrhagec
1.232
1.074–1.413
0.003
Maternal infectiond
0.949
0.673–1.339
0.77
Composite of all surgical/obstetrical complicationsc
1.191
1.037–1.367
0.013
NICU admissions
GEE: generalized estimating equation. NOTE: All outcomes were measured up to 30 days post delivery. a
Covariates included: mother’s age, mother’s ADG score, parity, multiple gestation.
b
Covariates included: maternal age, maternal ADG score, parity, multiple gestation, preterm labour, Caesarean delivery.
c
Covariates included: mother’s age, mother’s ADG score, parity, multiple gestation, Caesarean delivery, preterm labour, placenta previa, post dates.
d
Covariates included: mother’s age, mother’s ADG score, parity, multiple gestation, preterm labour, post dates.
4 • ■■ JOGC ■■ 2018
Impact of Resident Overnight Duty Hour Changes on Obstetrical Outcomes: A Population-Based Cohort Study
Table 3. Unadjusted results Pre-Intervention rate per 1000 deliveries
Post-Intervention rate per 1000 deliveries
P value
443.32
447.29
0.653
Primary outcome Composite of all maternal and fetal outcomes Secondary outcomes Neonatal death
10.01
8.55
0.396
Composite of all fetal outcomes
170.26
159.34
0.099
Composite of all maternal outcomes
341.42
354
0.144
NICU admissions
167.55
157.42
0.124
Maternal transfusion or postpartum hemorrhage
51.60
64.53
0.002
Maternal infection
22.18
21.45
0.783
Composite of all surgical/obstetrical outcomes
65.80
79.67
0.003
All outcomes were measured up to 30 days post-delivery.
increase. It has been established that handovers are a significant source of medical errors.20 In fact, a resident’s discontinuity of care due to handovers has been shown to increase preventable adverse events.21,22 Studies have discovered that with each additional handover, patient information is omitted or inaccuracies are added which can affect patient outcomes.22–24 Another possible explanation for why limiting resident work hours has not improved patient outcomes is that the overall fatigue levels have remained similar. This is likely due to the structure of NF, in which the residents worked for 12–14 hours for four or five consecutive nights in a row, took one day off for recovery, and then returned to a day schedule. A recent survey of internal medicine residents in a similar NF system revealed that, although they experienced reduced fatigue in the early morning, they felt worsened long-term fatigue compared with the traditional 26-hour system.25 This increased fatigue over time may counteract any short-term beneficial effects that would be gained from duty hour restrictions. In comparison, other health care providers such as registered nurses also work under similar schedules, although normally with fewer consecutive nights or more time off work in between blocks of shifts. This study is limited by the fact that it is an observational study and cannot infer causation. As well, the three academic sites in Toronto may have had slight variations in the number of consecutive nights that residents worked, or how frequently residents were on night float, which could not be controlled for in this study. Finally, one could argue that patient outcomes may not have been influenced substantially because staff obstetricians were present at all deliveries. Staff obstetricians at the three centres included generally work 24-hour shifts and that remained
the same throughout the study period. This may have neutralized some effects to patient care; however, generally, the resident physicians manage most patient care overnight leading up to the delivery and act as the primary surgeon/ attendant during the delivery. Therefore, certainly the residents can substantially impact patient outcomes. As well, the outcomes that were significantly impacted during the study (surgical/obstetrical outcomes and transfusion and PPH) are both factors that are likely more dependent on resident management given they act as primary surgeon/ attendant. Certainly, further study on the impact of changes to duty hours for staff obstetricians needs to be examined as well.
CONCLUSION
We found that the implementation of the night float system for Obstetrical residents in Toronto did not cause a change in overall patient outcomes, but there was an increased incidence of transfusion and PPH and surgical/obstetrical complications. These results suggest that duty hour restrictions may not be as beneficial as initially thought, and adds to the controversial data from other specialties. It certainly highlights the need to further investigate the clinical impact a change in resident duty hours has, as well as the effect on resident well-being and education, to determine the best strategy to adopt moving forward. Acknowledgements
This study was funded by the Sunnybrook Education Advisory Council Education Research Unit Education Research and Scholarship Grant. Special thanks to Rafik Saskin and Eliane Kim at the Institute for Clinical Evaluative Sciences. This work has been presented as an oral presentation
■■ JOGC ■■ 2018 • 5
EDUCATION
at the American Congress of Obstetricians and Gynecologists Annual Clinical Meeting, May 8, 2017. SUPPLEMENTARY DATA
Supplementary data related to this article can be found online at https://doi.org/10.1016/j.jogc.2018.02.001. REFERENCES 1. Olson E, Drage L, Auger R. Sleep deprivation, physician performance, and patient safety. Chest 2009;136:1389–96.
12. Scally C, Ryan A, Thumma J, et al. Early impact of the 2011 ACGME duty hour regulations on surgical outcomes. Surgery 2015;158:1453–61. 13. Bailit J, Blanchard M. The effect of house staff working hours on the quality of obstetric and gynecologic care. Obstet Gynecol 2004;103:613– 6. 14. Peltan ID, Brown CE, Burke AK, et al. The July Effect on Maternal Peripartum Complications Before and After Resident Duty Hour Reform: A Population-Based Retrospective Cohort Study. Am J Perinatol 2017;34:818–25. 15. Von Elm E, Altman DG, Egger M, et al. The strengthening the reporting of observational studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. J Clin Epidemiol 2008;61:344.
2. Phillbert I. Sleep loss and performance in residents and nonphysicians: a meta-analytic examination. Sleep 2005;1392–402.
16. Williams J, Young WA. Summary of Studies on the Quality of Health Care Administrative Databases in Canada. 339–345 (Canadian Medical Association); 1996.
3. Wilhelm B, Widmann A, Durst W, et al. Objective and quantitative analysis of daytime sleepiness in physicians after night duties. Int J Psychophysiol 2009;72:307–13.
17. Canadian Institute for Health Information. CIHI Data Quality Study of the 2009–2010 Discharge Abstract Database. (CIHI, Ottawa, ON); 2012.
4. Arnedt J, Owens J, Crouch M, et al. Neurobehavioral performance of residents after heavy night call vs after alcohol ingestion. JAMA 2005;294:1025–33. 5. Dawson D, Reid K. Fatigue, alcohol and performance impairment. Nature 1997;388:235. 6. Rajaram R, Chung JW, Cohen ME, et al. Association of the 2011 ACGME resident duty hour reform with postoperative patient outcomes in surgical specialties. J Am Coll Surg 2015;221:748–57. 7. National Steering Committee on Resident Duty Hours. Fatigue, Risk and Excellence: Towards a Pan-Canadian Consensus on Resident Duty Hours. (The Royal College of Physicians and Surgeons of Canada, Ottawa, Ontario); 2013. 8. Ahmed N, Devitt KS, Keshet I, et al. A systematic review of the effects of resident duty hour restrictions in surgery: impact on resident wellness, training, and patient outcomes. Ann Surg 2014;259:1041–53. 9. Philibert I, Nasca T, Brigham T, et al. Duty-hour limits and patient care and resident outcomes: can high-quality studies offer insight into complex relationships? Annu Rev Med 2013;64:467–83.
18. The Johns Hopkins University Bloomberg School of Public Health. The Johns Hopkins ACG Case-Mix Documentation & Application Manual Version 5.0. Johns Hopkins University; 2001. 19. Austin P. Using the standardized difference to compare the prevalance of a binary variable between two groups in observational research. Commun Stat Simul Comput 2009;38:1228–34. 20. Institute of Medicine (US) Committee on Optimizing Graduate Medical Trainee (Resident) Hours and Work Schedule to Improve Patient Safety. Resident Duty Hours: Enhancing Sleep, Supervision, and Safety. Washington DC: National Academies Press (US); 2009. 21. Laine C, Goldman L, Soukup J, Hayes J. The impact of a regulation restricting medical house staff working hours on the quality of patient care. JAMA 1993;269:374–8. 22. Petersen L, Brennan T, O’Neil A, et al. Does housestaff discontinuity of care increase the risk for preventable adverse events? Ann Intern Med 1994;121:866–72. 23. Arora V, Kao J, Lovinger D. Medication discrepancies in resident signouts and their potential to harm. J Gen Intern Med 2007;22:1751–5.
10. Jamal M, Wong S, Whalen T. Effects of the reduction of surgical residents’ work hours and implications for surgical residency programs: a narrative review. BMC Med Educ 2014;14:S14.
24. Horwitz L, Krumholz H, Green M, et al. Transfers of patient care between house staff on internal medicine wards: A national survey. Arch Intern Med 2006;166:1173–7.
11. Jamal M, Doi SA, Rousseau M, et al. Systematic review and meta-analysis of the effect of North American working hours restrictions on mortality and morbidity in surgical patients. Br J Surg 2012;99:336–44.
25. Mathew R, Gundy S, Ulic D, et al. A Reduced Duty Hours Model for Senior Internal Medicine Residents: A Qualitative Analysis of Residents’ Experiences and Perceptions. Acad Med 2016;91:1284–92.
6 • ■■ JOGC ■■ 2018