Barber, Emma L. MD; Eisenberg, David L. MD; Grobman, William A. MD, MBA
Physician practice patterns and decision-making can be altered by nonmedical influences with resulting effects on patient care.1–4 For example, alterations of physician behavior based on financial incentives have been suggested by multiple studies. Specifically in obstetrics, women who are within a fee-for-service reimbursement system have been shown to have higher cesarean delivery rates when compared with those who are within health management organizations or systems with capitated payments.5,6 However, nonmedical influences may not only be financial. Accordingly, the context of work, such as the time of the day, the day of the week, or length of shift, may affect the behavior of health care providers. For example, some studies have revealed an increased risk-adjusted cesarean delivery rate and operative vaginal delivery rate during the evening hours and on Fridays.7–10 Furthermore, working more than 12 hours in a given shift has been associated with increased medical errors.11
For resident physicians, nationally mandated work hour restrictions of 80 hours per week were implemented in 2003.12 To comply with this mandate, many residency programs have replaced traditional night call schedules with a night-float system.13 In a traditional obstetric call system, the provider works a 36-hour shift (ie, a complete workday followed by working overnight and again the next day). In contrast, in a night-float system, there are distinct day and night teams, each working approximately 12-hour shifts. Night-float residents have no daytime responsibilities and no members of the day team stay overnight for call. Attending physicians in some obstetric practices also have implemented a night-float model in which one physician from the group has nighttime, but no daytime, responsibilities for a week. It remains unknown whether this change is associated with differences in physician behavior or patient outcomes.
Our hypothesis was that a night-float system, which guarantees that a well-rested physician is dedicated to the labor floor each evening and has no subsequent daytime clinical responsibilities, would be associated with differences in obstetric management, such as a lesser tendency to schedule labor inductions and potentially differences in patient outcomes. Correspondingly, this longitudinal cohort study was designed to estimate whether the change to a night-float system among a group of physicians at a single institution was associated with changes in practice patterns as well as obstetric outcomes for women admitted for delivery.
MATERIALS AND METHODS
Medical records were reviewed for all deliveries performed by two physician groups between October 1, 2002 and March 31, 2003. One of these groups, designated the study group, changed to a night-float schedule midway through the study period (ie, January 1, 2003); the other group (the control group) used a traditional call schedule throughout the study period. Thus, time period A refers to October 1–December 31, 2002, when both groups were using a traditional call schedule, and time period B refers to January 1–March 31, 2003, after the study group was using a night-float schedule but the control group remained with a traditional call schedule.
Both the study group and control group were private practice general obstetrics groups and each employed six physicians during the study period. No patients cared for by midwives were included in this study. Both groups delivered patients only at Northwestern Memorial Hospital, a tertiary care urban academic medical center. House staff were present to assist with the management of all patients and an additional attending obstetrician, not from either group, was present on labor and delivery for general oversight.
In this study, physicians working within a traditional call schedule were responsible for laboring patients for approximately one night of the week between the hours of 5:00 PM and 7:00 AM, and they continued to work during daytime hours during the rest of the week. In the night-float schedule, conversely, one physician covers all labor management and deliveries of the group's patients between 7:00 PM and 7:00 AM every night for 1 week. The night-float physician has no daytime responsibilities during the week when they are working at night, and when they are off shift, a partner from the same group covers the patients during the day. When on-call in both the night-float system and the traditional call system, physicians are responsible for all of their patients on the labor floor and for answering patient phone calls. Patients of the group presenting to the emergency department with gynecologic symptoms or a gynecologic emergency requiring surgery are also the responsibility of the on-call physician.
Patients were included in the study if they had a singleton gestation and delivered at Northwestern Memorial Hospital and were cared for by one of these two physician groups during the defined study period. The medical records of eligible women were reviewed and data were abstracted for demographic characteristics, intrapartum events, maternal outcomes, and neonatal outcomes. Abstracted data, stratified by study group and study period, were compared using the Student t test for comparing normally distributed continuous variables, the Mann-Whitney U test for comparing continuous variables that were not normally distributed, χ2 analysis for comparing categorical variables, and Fisher exact analysis for comparison of outcomes with a cell frequency of less than five. A power analysis revealed that 206 women in each group were required, with an alpha of 0.05 and a beta of 0.2, to demonstrate that a change in the type of call was associated with a 35% reduction in labor induction from a baseline frequency of 36%. All tests were two-tailed and P<.05 was used to define statistical significance. SAS 9.2 software was used for statistical analysis. The Institutional Review Board of Northwestern University approved this study.
The study group delivered 208 patients meeting inclusion criteria during the study period and the control group delivered 228 patients. Patient demographic characteristics stratified by patient group and study period are presented in Table 1. Patient demographic characteristics were similar between time period A and time period B for the study group. The majority of characteristics for patients of the control group also were unchanged between the two time periods, although the frequency of nulliparity was higher for this group during time period A. Also, in both groups, each individual physician delivered a similar proportion of the group's patients in time period A and time period B (data not shown).
Regardless of the time period, the practice patterns, such as the use of labor induction, of the control group remained unchanged. This finding is in contrast to the changes that were associated with the switch to the night-float schedule among physicians in the study group (Table 2). Specifically, in time period B after the change to the night-float schedule, the frequency of induction of labor and scheduled cesarean deliveries decreased significantly. This decrease was coupled with a corresponding increase in the percentage of patients admitted in spontaneous labor. Moreover, when patients did have labor induced, it was significantly less likely to be initiated between 4:00 AM and 1:00 PM, a time period during which women may have labor initiated if a daytime delivery is sought. The change to a night-float call schedule also was associated with changes in the management of spontaneous labor (Table 3). Once a night-float schedule had begun, physicians were more likely to use oxytocin augmentation and were more likely to begin oxytocin augmentation between 7:00 PM and 7:00 AM. Despite these changes, time from admission to delivery remained similar during both time periods for both groups.
Certain events and outcomes at the time of delivery also were noted to be different after the change to a night-float schedule (Table 4). The use of episiotomy decreased after the night-float system was implemented, as did the incidence of third-degree or fourth-degree perineal lacerations. Physicians in the night-float system were less likely to manually remove the placenta and less likely to deliver a newborn with an umbilical cord arterial pH of less than 7.10. There were no changes in these outcomes between time periods A and B for the control group. Other outcomes, such as the frequency of cesarean delivery and postpartum hemorrhage, were no different for either group based on the time period of delivery.
In this study, several differences were seen in association with implementation of a night-float system for labor and delivery call. These differences were with respect to both care processes (such as use and timing of labor induction) and with respect to outcomes (such as perineal laceration). These differences did not appear to be attributable to differences in the type of women who received care, because demographic characteristics for patients of the study group were similar both before and after the change in scheduling. The changes also could not be attributable to differences in the physicians who were providing care, because there was no change in personnel, and individual physicians within the group participated in a similar number of deliveries both before and after the change to the night shift. Last, it does not appear that these differences were a reflection of other secular trends that were occurring simultaneously at the same institution. Another group delivering during the same time period, but who did not have a change in the type of their call schedule, did not demonstrate any process or outcome changes with respect to the women whom they delivered.
The exact reasons for these differences remain uncertain, although it is at least plausible that they are related to the context of the work schedule. Physicians participating in a night-float schedule have their evening hours reserved for work and have no need to remain awake during the day. Thus, for example, physicians may feel less inclined to schedule a labor induction in such a way to maximize daytime delivery or more inclined to begin augmentation during the night, even though that may increase the likelihood that a patient will deliver during the night. Other studies have suggested that the context of a work schedule can affect labor management. In one study, presence of an attending physician in-house at night has been shown to be associated with a decreased cesarean delivery rate.14,15
Much of the literature surrounding call schedule changes and patient outcomes has not focused on attending physicians, but rather has focused on resident physicians and the shift to an 80-hour work week. In some of these studies, decreased work hours have been associated with decreased mortality in trauma patients,16 decreased mortality in cardiac patients,17 decreased medical errors, decreased physician motor vehicle crashes, and decreased percutaneous injuries.18,19 Specifically addressing a traditional 36-hour call schedule for attending physicians, one study found an increased risk of complications among surgical procedures performed on the day after a night call during which the physician had slept less than 6 hours.20 Another study compared an intensivist continuous work schedule (day shift for 7 consecutive days) to an interrupted schedule (rotating schedule with a different intensivist each day) and found no difference in patient outcomes, but there was increased physician satisfaction and less burnout for physicians in the interrupted group.21 These studies have shown that work schedules can be associated with physician behavior, care delivery, and patient outcomes, although in nonobstetric disciplines.
There are several strengths in the design of this study that should support the reliability of the findings. One method to evaluate the effect that practice environment has on outcomes is to perform a cross-sectional study, such that those patients cared for in one type of system are compared with patients from another type of system. However, the potential for bias in such a study is great, because it remains unknown whether any observed differences are attributable to the system itself or the different types of people who happen to be within each system. The longitudinal design of this study precludes the potential for this type of bias; the physicians during each time period were identical. Although a longitudinal design does raise the concern that changes in practice could be related to other changes in the environment (eg, changes in medical evidence, changes in the professional liability climate) over time, as already noted, this study included a comparison group to elucidate whether other changes, unrelated to the schedule change, existed. The possibility of an unrelated change in practice also is decreased by the short study period. This short study period also was beneficial in that it limited the potential for other changes that could affect processes and outcomes, such as changes in patient characteristics or changes in the personnel within each physician group. The study period, which occurred several years ago, was specifically chosen given that this was a time when one group changed to a night-float schedule and was unaffected by personnel changes.
Although the short time period conveyed several strengths to the study, the weakness introduced by this design is that the number of patients studied was relatively small, thus reducing the power to detect differences in other end points. Additionally, this study examines a single private generalist practice of six physicians at a large, urban, academic medical center with 24-hour resident staffing and the results may not be generalizable to different patient populations, locations, physicians, or physician groups. Also, it should be noted that we cannot know that the changes that were found in association with a change in the call schedule were causally related to the exposure; this is an observational study, and as such, causality cannot be known. Last, even if changes in care were attributable to the change in call schedule, this study cannot, by itself, answer the question of which patients received optimal care or failed to receive “good” care.
Long-standing efforts have been made to improve the quality and safety of patient care within obstetrics. Many of these efforts have involved the study of prophylactic or therapeutic interventions, although relatively fewer have involved the study of how obstetric health services can be best-structured such that the interventions that we do have can be optimally applied. The ever-increasing desire to ensure patient safety and physician satisfaction emphasizes the need to understand and further evaluate which health services factors optimize obstetric outcomes.
1. Cromwell J, Mitchell JB. Physician-induced demand for surgery. J Health Econ 1986;5:293–313.
2. Fuchs VR. The supply of surgeons and the demand for operations. J Hum Resour 1978;13(Suppl):35–56.
3. Gruber J, Owings M. Physician financial incentives and cesarean section delivery. Rand J Econ 1996;27:99–123.
4. Léonard C, Stordeur S, Roberfroid D. Association between physician density and health care consumption: a systematic review of the evidence. Health Policy 2009;91:121–34.
5. Lipkind HS, Duzyj C, Rosenberg TJ, Funai EF, Chavkin W, Chiasson MA. Disparities in cesarean delivery rates and associated adverse neonatal outcomes in New York City hospitals. Obstet Gynecol 2009;113:1239–47.
6. Spetz J, Smith MW, Ennis SF. Physician incentives and the timing of cesarean sections: evidence from California. Med Care 2001;39:536–50.
7. Peled Y, Melamed N, Chen R, Pardo J, Ben-Shitrit G, Yogev Y. The effect of time of day on outcome of unscheduled cesarean deliveries. J Matern Fetal Neonatal Med 2011;24:1051–4.
8. Brown HS 3rd. Physician demand for leisure: implications for cesarean section rates. J Health Econ 1996;15:233–42.
9. Curtin SC, Park MM. Trends in the attendant, place, and timing of births, and in the use of obstetric interventions: United States, 1989–97. Natl Vital Stat Rep 1999;47:1–12.
10. Goldstick O, Weissman A, Drugan A. The circadian rhythm of “urgent” operative deliveries. Isr Med Assoc J 2003;5:564–6.
11. Rogers AE, Hwang WT, Scott LD, Aiken LH, Dinges DF. The working hours of hospital staff nurses and patient safety. Health Aff (Millwood) 2004;23:202–12.
12. Philibert I, Barach P. Residents' hours of work. BMJ 2002;325:1184–5.
13. Fletcher KE, Saint S, Mangrulkar RS. Balancing continuity of care with residents' limited work hours: defining the implications. Acad Med 2005;80:39–43.
14. Klasko SK, Cummings RV, Balducci J, DeFulvio JD, Reed JF 3rd. The impact of mandated in-hospital coverage on primary cesarean delivery rates in a large nonuniversity teaching hospital. Am J Obstet Gynecol 1995;172:637–42.
15. Poma PA. Effect of departmental policies on cesarean delivery rates: a community hospital experience. Obstet Gynecol 1998;91:1013–8.
16. Morrison CA, Wyatt MM, Carrick MM. Impact of the 80-hour work week on mortality and morbidity in trauma patients: an analysis of the National Trauma Data Bank. J Surg Res 2009;154:157–62.
17. Gopaldas RR, Huh J, Bakaeen FG, Wang XL, Coselli JS, LeMaire SA, et al.. The impact of resident work-hour restrictions on outcomes of cardiac operations. J Surg Res 2009;157:268–74.
18. 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;296:1055–62.
19. Reed DA, Fletcher KE, Arora VM. Systematic review: association of shift length, protected sleep time, and night float with patient care, residents' health, and education. Ann Intern Med 2010;153:829–42.
20. Rothschild JM, Keohane CA, Rogers S, Gardner R, Lipsitz SR, Salzberg CA, et al.. Risks of complications by attending physicians after performing nighttime procedures. JAMA 2009;302:1565–72.
21. Ali NA, Wolf KM, Hammersley J, Hoffmann SP, O'Brien JM Jr, Phillips GS, et al.. Continuity of Care in Intensive Care Units: A Cluster-Randomized Trial of Intensivist Staffing. Am J Respir Crit Care Med 2011 [Epub ahead of print].