Salim A, Teixeira PGR, Chan L, Oncel D, Inaba K, Brown C, Rhee P, Berne TV. Impact of the 80-Hour Workweek on Patient Care at a Level I Trauma Center. Arch Surg. 2007;142(8):708–714. doi:10.1001/archsurg.142.8.708
The 80-hour workweek limitation for surgical residents is associated with an increase in mortality and complication rates among adult trauma surgical patients.
Retrospective cohort study.
Academic level I trauma center.
Trauma patients admitted before and after the 80-hour workweek limitation.
We compared death and complication rates for adult trauma patients admitted during a 24-month period before (2001-2003) and a 24-month period after (2004-2006) implementation of the 80-hour workweek at our institution. Relative risk and its 95% confidence intervals were examined.
Main Outcome Measures
Patient care outcomes included preventable and nonpreventable complications and deaths.
The patient populations from the 2 time periods were clinically similar. No significant differences were found in the total and the preventable death rates. The time period after the 80-hour workweek mandate had a significantly higher total complication rate (5.64% vs 7.28%; relative risk, 1.29; 95% confidence interval, 1.15-1.45; P < .001), preventable complication rate (0.89% vs 1.28%; relative risk, 1.43; 95% confidence interval, 1.06-1.91; P = .02), and nonpreventable complication rate (4.75% vs 5.81%; relative risk, 1.22; 95% confidence interval, 1.08-1.39; P = .002).
Although there was no difference in deaths between the 2 time periods, there was a significant increase in total, preventable, and nonpreventable complications. This increase in complication rate may be due, in part, to the new 80-hour workweek policy.
As a result of the Institute of Medicine's eye-opening report To Err Is Human: Building a Safer Health System, the concept of medical errors has been an issue in the minds of patients, politicians, and the national media.1 Although many factors contribute to medical errors, the concept of fatigue among medical providers has been a key focus. A number of studies suggest that fatigue has a detrimental effect on the technical performance of surgical residents during simulated laparoscopic models.2- 4 Similarly, fatigue has been associated with substantial decreases in performance on neurocognitive and simulated clinical tests for other physicians.5- 9 In response to these studies and to growing public pressure, the Accreditation Council for Graduate Medical Education (ACGME) mandated that all US residency programs limit work hours for trainees starting July 1, 2003.10 The rationale was that a decrease in residents' work hours would translate into less-fatigued residents and improved patient outcomes. The ACGME continues to release yearly summaries of achievements after the nationwide implementation of work-hour restrictions related to compliance and organization.11 Patient care outcomes are never addressed. Presently, 2 years after the nationwide implementation and 18 years after the first state-run program of duty-hour restrictions, the effect of this change on patient safety is still unknown.
The objective of this study was to determine if residency work-hour limitations had an effect on the overall incidence of complications and deaths at a busy level I trauma center that relies heavily on resident physician involvement. We paid particular attention to the incidence of preventable complications and deaths. Our hypothesis was that the new 80-hour workweek would be associated with an increase in preventable complications and deaths.
The division of trauma and critical care at the Los Angeles County + University of Southern California Medical Center implemented a work hour–restricted schedule on July 1, 2003, according to the ACGME guidelines. Specifically, these guidelines called for duty-hour limits of 80 hours per week averaged across 4 weeks; 1 in 7 days free from all educational and clinical activities; a 10-hour work-free period after a 24-hour shift; and no more than a 6-hour period for continuity of care after a 24-hour call shift. Patient care outcomes for 2 years prior to implementation of the restricted schedule (July 1, 2001, to June 30, 2003) were compared with 2 years after implementation (July 1, 2004, to June 30, 2006). The year the new schedule was implemented was not included in the evaluation because we felt that it took about 1 year for the services to get used to the new schedule.
The trauma service at Los Angeles County + University of Southern California Medical Center is made up of 3 separate teams with 3 residents per team (1 senior-level resident, 1 junior-level resident, and 1 intern). The call schedule has always been 1 in 3 days. Prior to institution of the limited work schedule, all teams were responsible for their own admissions. This meant that postcall teams often stayed late to continue to provide care for their patients. Patients were passed on infrequently, and postcall residents often came back to the hospital to address any patient deterioration. Residents worked a mean of 100 to 120 hours per week with no set days off. The new work hour–restricted schedule emphasizes liberal patient transfers among physicians. The on-call team assumes the care for all patients on the 3 trauma teams, and postcall residents never come back to the hospital for patient care issues. Resident attendance at educational conferences is not required and physician assistants have become an integral component of the trauma teams. With the new schedule, residents work between 60 and 90 hours each week, depending on the number of calls per week.
Patient care outcomes included complications and deaths. Both were recorded on a daily basis by the house staff and independently verified by a trauma registry nurse. All complications and deaths were then comprehensively reviewed at a mandatory weekly morbidity and mortality conference. All attendings, fellows, and residents in the division of trauma and critical care as well as those services that may have been directly involved in each case, such as neurosurgery, are present at the morbidity and mortality conferences. Complications were presented and classified by consensus according to preventability by the trauma attendings and fellows. An attending trauma surgeon documented each complication, its root causes, outcomes, and loop closure recommendations. All deaths were further reviewed by the Combined Trauma Death Review Committee. The Combined Trauma Death Review Committee is a multidisciplinary committee (trauma surgery, critical care, emergency medicine, neurosurgery, nursing, and forensic pathology), chaired by the coroner, that is responsible for evaluating all deaths. After review of clinical and autopsy data, the deaths were classified by consensus according to 3 categories of preventability (nonpreventable, potentially preventable, and preventable). This classification system of preventability was outlined by MacKenzie12 in 1999. A potentially preventable death must meet 3 criteria: (1) the injury must be survivable, (2) the delivery of care was suboptimal, and (3) the error in care must be directly or indirectly implicated in the death of the patient. Preventable deaths are frankly defined as those in which an error clearly led to the patient's death. This process of reporting, reviewing, and classification of complications and deaths have not changed during the study time period. Because both preventable and potentially preventable events were thought to be equally important as outcome measures, they were combined and considered to be preventable.
All Combined Trauma Death Review Committee and morbidity and mortality reports from July 1, 2001, to June 30, 2005, were reviewed. Clinical data, including patients' demographics, injury history, injury severity, course in hospital, procedures performed, and circumstances surrounding the complications and deaths, were retrieved from the trauma registry and individual medical records. Preventable complications were further classified into 5 functional categories (airway, postoperative complications, delay in diagnosis/treatment, catheters/drains/tubes, and operative complications).
We compared the patients' demographic and clinical characteristics between the 24-month period before (2001-2003) and a 24-month period after (2004-2006) implementation of the 80-hour workweek using the χ2 test for evaluating the difference between proportions, and the 2-sample t test or Mann-Whitney test wherever appropriate for analyzing the difference between the means of continuous variables. We derived the relative risk for each outcome measure and its 95% confidence interval using the time period before the 80-hour restriction as a reference. Statistical significance was set at P < .05. This study was approved by our institutional review board. The need for informed consent was waived.
There were 8939 trauma patients admitted in the time period before the restriction, compared with 7915 trauma admissions in the period after the restriction. Table 1 compares the demographics and injury severity of the patients admitted during the 2 time periods. Overall, the demographic and injury characteristics were similar. Although the time period after the 80-hour restriction had a statistically significantly lower age (mean [SD], 35.1 years [18 years] vs 35.7 years [18 years]; P = .047), the higher percentage of blunt injury (75.7.2% vs 74.2%; P = .02) and longer hospital (6.5 vs 6.3 days; P < .001) and intensive care unit (1.4 vs 1.0 days; P = .004) lengths of stay were not clinically significant. Table 2 compares major procedures performed during the 2 time periods. There were significantly more sternotomies performed in the time period before the 80-hour workweek restriction.
Table 3 compares the complication and death rates of the 2 time periods. There were no differences in the total death rate and the preventable death rate between the 2 time periods. However, the time period after the 80-hour workweek limitation was associated with a significantly higher total complication rate (5.64% vs 7.28%; relative risk, 1.29; 95% confidence interval, 1.15-1.45; P < .001), preventable complication rate (0.89% vs 1.28%; relative risk, 1.43; 95% confidence interval, 1.06-1.91; P = .02), and nonpreventable complication rate (4.75% vs 5.81%; relative risk, 1.22; 95% confidence interval, 1.08-1.39; P = .002). When we examined the specific type of preventable complication (Table 4), we found that the category of airway/intubation had a significant difference in complication rate (0.17% vs 0.39%; P = .005). There were twice as many missed injuries in the time period after the 80-hour workweek restriction. However, this did not reach statistical significance (0.09% vs 0.20%; P = .05).
In response to growing public pressure, the ACGME mandated that all US training programs limit work hours starting July 1, 2003.10 This limit of 80 hours was first introduced by the Bell Commission in New York in 1989 and was subsequently adopted by the ACGME.13 This was an arbitrary time limit with no evidence to support its use. Nevertheless, advocates for this policy change argued that long work hours resulted in fatigue, increased medical errors, and increased mood changes, all at the expense of patient care.14 Critics claimed that the loss of continuity of care owing to shorter work hours and more resident transitions and patient transfers would actually result in worse patient care.15 In the present study, when compared with a 2-year time period before the institution of the 80-hour workweek, a 2-year period after the restriction was associated with significantly higher total, preventable, and nonpreventable complication rates. When the specific type of preventable complication was examined, there were twice as many missed injuries in the 80-hour workweek time period. There was no difference in mortality between the 2 time periods.
As previously mentioned, New York state instituted an 80-hour resident workweek in 1989 as a result of the Bell Commission and code 405.13 Despite more than 18 years of experience with resident work limitations in this state, there have been surprisingly few studies that have addressed patient outcomes. Howard et al15 performed a retrospective analysis using a New York state database to assess the effect of work-hour restrictions on the in-house mortality of congestive heart failure, acute myocardial infarction, and pneumonia patients. They found that the new work-hour restriction had no effect on patient outcomes. In a retrospective cohort study on a single general internal medicine service, Laine et al16 found that although the new work-hour limitations had no effect on in-house mortality, they were associated with delayed test ordering and increased in-hospital complications. Daigler et al17 performed a retrospective cohort study assessing the work-hour effects for interns in pediatrics in a New York hospital. They found no difference in mortality, morbidity, or unexpected intensive care unit transfers. In one of the only studies that demonstrated a beneficial effect of the work-hour limitations, Gottlieb et al18 performed a prospective time-series study in an internal medicine program at a New York veterans' hospital. They found a decrease in length of stay, number of laboratory tests ordered, and medication errors after institution of the schedule change. There have been a number of survey studies that have all reached similar conclusions; despite an improvement in sleep and quality of life, there is a perceived negative effect on patient outcomes owing to a loss of continuity of care.14,19
With the national implementation of the reduced work-hour schedule in 2003, a number of studies have attempted to measure its effect on patient outcomes. Bailit and Blanchard20 found that work-hour restrictions have shown minimal improvement in quality of care on an obstetrics and gynecological service. Kaafarani et al21 used National Surgical Quality Improvement Program data and found no difference in postoperative morbidity and mortality in surgical patients after implementation of the 80-hour workweek. Landrigan et al,22 in a prospective randomized study comparing serious medical errors made by interns working according to a traditional or an intervention schedule that eliminated extended work shifts found a significant increase in serious medical errors during the traditional schedule. This is the best-designed study to date in terms of evaluating the effect of a change in work hours. However, because the study was not designed to detect an effect on preventable adverse events, no conclusion can be made with respect to its effect on patient outcomes.23 In addition, because a sophisticated method was used to detect errors, the observed reduction in errors would probably not be apparent in most hospitals.23
Attempts to measure any effect from the change in duty hours have proved to be challenging. Programs have implemented duty-hour restrictions by using various methods, such as a night-float system, workload distribution, or physician extenders. Patient care outcomes have ranged from medication errors to preventable mortality. Consequently, each study is a measure of a single institution's duty-hours policy, and its applicability to other centers is questionable. In addition, there are many confounding variables for which it is difficult to measure and correct. Patient factors, such as age, sex, and severity of injury and illness, can easily be adjusted for. Institutional factors, such as new health care, hospital, and service policies that may have taken effect, are more difficult to control for. Changes in attending staff coverage at a particular institution may affect outcomes positively or negatively. Increased attending physician supervision has been associated with decreased complication and mortality rates.24 However, measuring the actual amount of supervision has proven to be very difficult. The inability to demonstrate changes in patient outcomes may actually be a consequence of increased attending physician involvement and supervision. Other difficult-to-measure factors, such as resident schedules, service workload, new surgical techniques, and new institutional processes of care, may also influence patient outcomes.21
Despite these factors, we had the perception, just like a number of survey studies,25- 29 that patient care had been negatively impacted by the new duty hours. Probably one of the better measures of assessing an effect on patient outcome is the number of preventable events occurring during a period of time. Therefore, we were not surprised with the significant increase in preventable complications observed during the time period after the 80-hour limitation. However, these findings should be interpreted with caution. Besides the new work-hour restrictions, the observed increase in complications may be due to other factors that were not studied. Since there was a simultaneous increase in nonpreventable complications, other unexpected confounding factors may be present. With no simultaneous change in mortality during the same time period, the increased complication rate could be viewed as a transient phenomenon or as a result of other factors, such as problems with data collection. Even with other confounding factors, a doubling of missed injuries (though not statistically significant) is worrisome. Miscommunication and lapses in continuity of care could be potential byproducts of our liberal transfer policy. Potentially preventable adverse events have been associated with cross-coverage by physicians who are less familiar with patients than the patients' usual physicians.30
Most studies to date have failed to demonstrate any impact of duty-hour restrictions on patient care outcomes. The conflicting results will continue until better-controlled studies that measure relevant patient care outcomes and adjust for the various confounding factors are performed. Because the general public believes that restricting work hours will improve patient safety, the limitation of resident work hours will most likely be permanent.31 Providing optimal patient outcomes by creating a schedule that provides the appropriate balance of resident fatigue and continuity of care is the next necessary step.31 Presently, as our data suggest, optimal patient care outcomes have not yet been realized.
We have found that the change in duty hours at our institution may be associated with a significant increase in preventable and nonpreventable complications. A similar increase in mortality was not observed. Of concern was a doubling of the number of missed injuries after implementation of the 80-hour workweek.
Correspondence: Ali Salim, MD, Los Angeles County + University of Southern California Medical Center Medical Center, 1200 N State St, Room 9900, Los Angeles, CA 90033 (firstname.lastname@example.org).
Accepted for Publication: March 8, 2007.
Author Contributions:Study concept and design: Salim, Teixeira, Brown, Rhee, and Berne. Acquisition of data: Salim and Oncel. Analysis and interpretation of data: Salim, Chan, and Inaba. Drafting of the manuscript: Salim. Critical revision of the manuscript for important intellectual content: Salim, Teixeira, Chan, Inaba, Brown, Rhee, and Berne. Statistical analysis: Chan. Study supervision: Salim, Teixeira, Inaba, Brown, Rhee, and Berne.
Financial Disclosure: None reported.
Previous Presentation: This paper was presented at the 78th Annual Meeting of the Pacific Coast Surgical Association; February 20, 2007; Kohala Coast, Hawaii; and is published after peer review and revision. The discussions that follow this article are based on the originally submitted manuscript and not the revised manuscript.
David Spain, MD, Stanford, California: Congratulations, Dr Salim, on a very nice presentation. I have several questions. With the addition of PAs [physician assistants] to your service, did you stagger your call schedules so you always had at least one member of the team in-house? With the new system, did you develop a formal sign-out policy? How are the handoffs handled? Do you require a face-to-face sign-out, or is this an informal thing done over the phone? Finally, was there any change in severity of injuries during this time period that might account for differences in complications?
Lastly, I would like to point out what can be done with a serious commitment to performance improvement and patient safety. The group at USC [University of Southern California], under the guidance of Dr Demetriades, have truly set the standard in this regard.
Dr Berne: I think all of us cringed when we heard about this idea of surgery residents working 80-hour workweeks and being forced to go home and turn off beepers and all of the other things that have come along with the ACGME restrictions. Dr Salim was right when he said, “It's here to stay; it's not going to change.” The important thing then is to look at what is happening to surgical training and say, “How do we adapt to this? How do we continue to maintain the quality of patient care that is necessary?” Another thing that we didn't quite appreciate when this hit was, “How do we train surgeons?” Serious educational problems occur, for instance, when a patient dehisces near the end of that sacrosanct 6-hour postcall period just as the resident is walking out the hospital door. There is presently no way to have him stay and see what was wrong with the wound closure that caused the dehiscence. How are we going to deal with that kind of loss of experience and that loss of the ongoing responsibility that most of us had?
As was just mentioned, the reality of the situation is that we are stuck with most of this arrangement, so we better look very carefully at it to figure out what it is doing to patient care, what it is doing to residency training, and fix the problems we find. This paper was an effort to do that, but it is hard to do. No one can run a randomized controlled clinical trial on this matter. I don't think the ACGME will let us do that. So we have to deal with impure kinds of data with confounding problems, such as changing faculty and hospital policies, the increasing pressure to push patients out of the hospital faster and faster, and all of the other things that go on that change what we do every day.
The data Dr Salim presented does show that it is possible to pinpoint some clinical areas of concern where we can focus our efforts. We are very happy to know that our mortality rate didn't change. The hospital did some things that I am sure helped prevent problems; they gave us PAs that we didn't have before. We schedule them so that they are around longer postcall than the residents. They actually represent some continuity postcall that would have been lost without some kind of physician extenders. Also in our program, there is a lot of attending staff involvement. Since Dr DeMeester came 15 years ago, we have always had 2 teams, a trauma team and a nontrauma emergency team in-house, and both of those 2 services have in-house, 24-7 attending staffs. Also, those attending staffs work closely with the residents, and I think that may have been what prevented any change in mortality rate.
Regarding our formal handoff policy, we have always had a very detailed 7 AM morning pass-on rounds with a complete discussion of the cases, and the team going off makes rounds with the team coming on. We examined the patients together and passed on information. That has also changed a little bit in that we now do the same thing around 7 PM in the evening every day to be sure that everything is being followed up and that all of the details that were brought up in the morning pass-on rounds have been taken care of.
The other question Dr Spain asked was about the possibility of some drift in the severity of injury. We do know that the average ISS [injury severity score] rates and average mortalities were essentially the same.
I hope that others will take the opportunity to look at what is going on with their own patient care and investigate some other pieces of this 80-hour workweek puzzle. The question is should there be some greater flexibility put into the ACGME restrictions. It's a “one size fits all” plan right now. Maybe some changes are possible, for instance, so that if the patient I mentioned earlier has just eviscerated as the resident who did the operation is about to walk out of the hospital, he could stay around for a few more hours to see what actually went wrong. Then he could take a couple of hours off some other time. But right now, because the system is rigid, it is harmful to resident education.
Henry M. Cryer, MD, Los Angeles, California: Dr Salim, a great presentation, and Dr Berne, a couple of questions for you. First of all, it's a very important topic and, as Dr Spain alluded to, the trauma community has been leading the way with performance improvement requirements and the mechanisms to do it, but that doesn't mean we have perfected it yet. If you look at the NTDB [National Trauma Data Bank], there is extremely wide variability in the number of complications that are reported from various trauma centers across the country. What I am getting at here is, how do you identify the complications, because there is a big difference between concurrent review where you go over them every day from the night before and retrospective review, which is how most people have done it traditionally over the years. My first question for you is, how do you know that in your 2 time periods you just didn't get better at finding the complications? You had more PAs, you had perhaps a little different infrastructure that changed when you went to the 80-hour workweek, and you had to do things a little differently. So on identifying complications, are you doing it concurrently or by retrospective review, and did that change in any way between the 2 time periods?
Additionally, almost all preventable complications are not the fault of a single person or team. They are part of a structural failure or system failure, and that would include, for instance, missed injuries. What's happening with the 80-hour workweek in the department of radiology? Are the CT [computed tomography] findings being reported to your team accurately? Those are the 2 things that struck me about the paper. Great paper and timely, and I think it is important for all of us.
Dr Berne: Yes, the way we do the complications is that each service is responsible for identifying their own. It is largely done by the faculty on each of the 3 admitting teams in discussion with the residents each week before our trauma M&M [mortality and morbidity] conference. That hasn't changed at all.
In regard to missed injuries and the possibility that we missed fractures and other important findings on CT scans, we have actually experienced a big improvement toward having real-time readings, including middle-of-the-night readings. If anything, we should not have had as many missed injuries based on the radiological services.
Thomas R. Russell, MD, Chicago, Illinois: I want to congratulate you and Dr Salim on this paper. In my position at the American College of Surgeons, I have heard many opinions about the 80-hour workweek from different surgeons, but I hear very few facts.
There is no going back. As many of you know, the ACGME is undergoing change now because the director, David Leach, is stepping down in a few months. Many members of the surgical community are concerned about ACGME's policymaking capacity, because it is really more focused on primary care. We have some opportunities to talk to them about the flexibility in the portfolios that they need for residency review programs, which are being discussed. Nonetheless, there is no going back. The 80-hour workweek, I think, is here to stay. But perhaps we could add some flexibility.
Having visited LA County Hospital, my question is, how can you relate unintended extubations and missed injuries to something as simple as the 80-hour workweek? It seems like it would be kind of a stretch. But, I really congratulate you on trying to get some facts about this very difficult issue.
Dr Berne: Yes, we’ve talked about that to see if we could figure out the explanation, and it's not easy. There have been changes in the ICU [intensive care unit]. They are such things as waking patients up every day to avoid over-sedation and much greater use of spontaneous breathing trials. It is possible such changes may have led to more unintended extubations and not be due to resident-hour problems. On the other hand, the residents have always been the watchdogs of sedation levels and made most decisions about extubation. It is possible that with more frequent changes of responsibility each time the new team coming on doesn't really have the same understanding of exactly how to manage the sedation of the patients that they have been passed. This might lead to the patient being too agitated and to unintended extubations.
Unfortunately, we don't know what the injuries are that were missed. They are likely to be relatively minor things like fractured scapulas or bone injuries that are minor but showed up later on in a delayed fashion on subsequent x-rays. Those are the kinds of things that we have to look at and improve.
James J. Peck, MD, Portland, Oregon: Tom, I want to congratulate you on this paper and Dr Salim's excellent presentation. Two questions: on pass-on rounds, is there a written or computer-generated form that is passed from one team to the next, or is it purely verbal communication? Second, who decides who is passed on, who is kept? During my residency at LAC + USC, there were cowboys that never passed on a operative case to another team. Is it the attending man or the resident in charge who decides?
Dr Berne: Right now there is no computerized system to do this on trauma. The emergency nontrauma service has one and trauma is interested in adapting it. There are forms that are filled out. Every patient is discussed. The attending writes a note on all of the patients at the time of pass-on rounds.
The philosophy of “I won't pass anything on” used to be the macho thing to do, and there were surgery residents who almost never passed anything on. That has been completely changed. The rule now is that if there is any question about anything, you pass the patient on, and somebody is assigned to follow the patient and do what has to be done.
Kenneth Waxman, MD, Santa Barbara, California: Accidental extubation may be a pretty good surrogate of quality of critical care. While there are multiple reasons for accidental extubation, one of the most common is that physicians don't extubate their patient soon enough. One hypothesis to explain your data might be that patients were left intubated longer because the postcall team might not have been as aggressive in getting patients extubated. I am wondering if you have looked or could look at duration of intubation.
Dr Berne: Yes, we could do that and probably should to focus on this particular issue. I agree with that, and I believe that the lack of vesting in moving patients along needs to be watched carefully and could well be a factor. But on the other hand, we have instituted a number of the ICU protocols mentioned before, which should have led to earlier and safer extubations. So the reason for more unplanned extubations remains unclear.
Bruce E. Stabile, MD, Torrance, California: I want to congratulate Dr Salim and coworkers on this study. I think it brings some important data to an area that has been somewhat of a black hole. We at Harbor-UCLA have also studied this issue of trauma complications before and after the institution of the 80-hour workweek, and clearly we are having some diversity of findings here.
I have 2 questions. In the presentation it was mentioned that there has been a liberal policy of pass-ons, but I am wondering if there has been any substantive change in the policy regarding pass-ons from before and after the 80-hour workweek change?
The second question is somewhat heretical because your difference was statistically significant, but I wonder if this was just statistical noise in the system, as the absolute change in complication rate was only 0.4%. Now, admittedly, the numbers included in this study were large and the statistics were significant, but an absolute change of 0.4% in the preventable complication rate does not strike me as a very dramatic change, so I question whether this is simply noise in the system. Along this line, have you looked at 2 time periods before the 80-hour workweek and wouldn't you possibly find the same degree of variance among any 2-year time comparisons?
Dr Berne: Bruce, I don't remember any specific policy changes about pass-on rounds that were ever formalized, but the whole philosophy of passing on changed significantly with the 80-hour workweek simply because it had to. House staff just couldn't hang around forever to watch their patients the way they did before. They used to want to stay and see what happened to an interesting patient. For instance, seeing what evolved when a patient who had pelvic fracture bleeding and was going to have angiography. Residents used to stay to be present in the angio suite and observe what happened to their patients. They managed by taking in-house catnaps along the way. Therefore, before work-hour rules, it would sometimes happen that that kind of patient was not passed on but was just kept by the original team. That just can't happen now. So there clearly was a change.
In regard to the “noise” question, yes it's a small difference. However, even though we are not sure how clinically relevant these differences are, we believe that we should dig deeper into what showed “significance” to make sure we know what was happening.
Roger E. Alberty, MD, Portland: From the information available, it would appear that the sentinel event precipitating the 80-hour week had everything to do with handoffs and nothing to do with fatigue. When the residency review committee came around after institution of this 80-hour workweek, I asked the reviewer, were there any plans to follow-up on how this program does? What's the impact? He said no. He said you have to understand, this is a political decision and they are not going to follow-up on this. I think Dr Russell's comments that this is here to stay is absolutely correct. I think it is something we have to deal with, and I think this is a very important paper. It is going to force me to go back home and look at how we do business on handoffs.