Work-hour restrictions came about in part to ensure that residents would be sufficiently rested so that they could optimally learn from the patients they encountered. The ensuing angst of schedule changes and new patterns of resident case coverage raised concerns about a decline in case experience and thus the quality of the resident's experience. Early assessments of the impact of work-hour restrictions in New York State suggested that quality of life for residents, scores on standardized in-service examinations, and surgical caseload did not change.1 In a recent subjective survey, educators felt that work hours have a negative effect on resident education: 63% reported that overall resident education is worse and that resident surgical volume has diminished.2 Objective national data on resident experience before and after the imposition of work-hour restrictions, compared with national practice trends, should clarify this question. The objective of this study was to evaluate resident experience in obstetric and gynecologic procedures after implementation of resident work-hour restrictions.
MATERIALS AND METHODS
Data on resident technical experience from the 3 years before the implementation of the rules (July 1, 2003) and from a 3-year period beginning 4 years after implementation (ending June 30, 2009) were evaluated. Median resident experience data (median number of cases per resident) in the role of “surgeon” were obtained from both published and unpublished compilations by the Resident Review Committee in Obstetrics and Gynecology of the Accreditation Council for Graduate Medical Education. These statistics document the median resident technical experience including both obstetric and gynecologic procedures over the 4 years of their training. Core obstetric procedures studied were spontaneous, cesarean, and operative vaginal deliveries. Gynecologic procedures included abdominal and vaginal hysterectomies, laparoscopic procedures (other than sterilization), laparotomy (for other than hysterectomy), and operative hysteroscopy. Three-year averages of the medians were compiled along with measures of the rate of change for both the individual 3-year windows and for the change between the two time periods. Percentage comparisons are against the baseline values (before work hours) except where noted.
For comparison, additional data regarding national procedural counts, rates, and trends for the same procedures (where available) were obtained to assess simultaneous changes in practice patterns. Sources for these data included the Centers for Disease Control and Prevention, National Center for Health Statistics, United States Department of Health and Human Services, Agency for Healthcare Research and Quality (Healthcare Cost and Utilization Project), and the American College of Obstetricians and Gynecologists Resource Center. For gynecologic data, numbers were abstracted based on the Current Procedural Terminology codes used by the Resident Review Committee in defining the categories reported in their data. Data were obtained beginning no later than 1997 and continued to the most recent published data. National data were occasionally not available for years more recent than 2006 or 2007, and in these cases, projections based on linear regression models of the available data were used to extrapolate the information forward. Ten-year rates of change through the projected 2008 data were compiled in a similar manner to those used for the resident experience data. For comparisons, national trend data (ending December 31) are compared with the data for resident graduates for the subsequent June 30. That is, national data for calendar 2008 are compared against 2009 resident graduates, and so forth.
Whereas resident experience data represent information on the experience of between 1,119 and 1,173 residents (depending on the year), the national median values (“surgeon” cases per resident) reported represent data sets of three points, one point for each time period. Hence, no attempts at statistical comparisons have been made, and only descriptive statistics and comparisons for trends against national data are supplied.
The median resident obstetric experience for the periods ending June 30, 2001–2003, and June 30, 2007–2009, are shown in Table 1. Overall, there was a 6.9% decline (compared with baseline average) in the number of spontaneous deliveries per resident reported in the 3-year period under work-hour restrictions compared with before their implementation, although there has been a 7.1% rise (18 cases) within the most recent 3-year period. The net result was a decline of 19 cases for the respective 3-year averages (282 compared with 263). There was a decline in the number of operative vaginal deliveries between these two time periods (17 cases per resident, 39.1% decline compared with baseline average). These declines in experience with spontaneous and operative deliveries were offset by a rise in the number of cesarean deliveries per resident (55 cases per resident, 33.0% change), resulting in a decline in total obstetric experience of only 19 cases per resident (3.8% difference). As a result of the decline in spontaneous deliveries and the rise in cesarean deliveries reported by the residents, their cesarean delivery rates (percentage) rose by 28.2%, resulting in their cesarean delivery rate rising by 9.5 points to a total of 43.3% of resident deliveries as “surgeon.”
National obstetric statistics for the period beginning in 1998 and the most current available, projected to 2008, are shown in Table 2. National births rose over the 10-year period by 9.4%, although the birth rate (per 1,000) declined by 4.3%. The cesarean delivery rate rose by 60.7% over the period, resulting in a rate of 31.8% in 2007 and projected rate of 34.0% for 2008. When the rate of resident cesarean delivery was compared with the trends in the national data, the rise in resident experience parallels the national trends and the most recent 3-year trend leads the recent national trends (3-year linear regression slope: resident experience 1.52, national rate 1.33). During the 10-year comparison period, the rate of operative vaginal deliveries declined by 55.2% to 7.3% in 2005, projected to a rate of 5.3% for 2008. The decline in resident operative vaginal delivery rates seen during the 2007–2009 period matched that for the national data (3-year linear regression slope: resident experience −0.36, national rate −0.37). This is in contrast with the much more rapid decline in resident operative delivery experience seen during the 2001–2003 period when there was a greater than threefold difference (3-year linear regression slope: resident experience −0.95, national rate −0.30).
Resident experiences in the reference gynecologic procedures for the periods ending June 30, 2001–2003, and June 30, 2007–2009, are shown in Table 3. For 2009 residency graduates, vaginal hysterectomy and laparoscopically assisted hysterectomy were reported separately, but these have been combined as vaginal hysterectomy for comparison. Resident experience with laparotomy for cases other than hysterectomy was not reported for 2009 residency graduates. For this particular data, a 2-year average was used. National statistics for gynecologic procedures from 1998 to the most recent available, with projections to 2008, are shown in Table 4.
Three-year average of median resident “surgeon” experience with abdominal hysterectomies dropped by 8.3 cases (9.9%) after the implementation of work-hour rules, although there was a difference of only 4.8 cases between the values for 2003 and 2007. Median abdominal hysterectomy rates underwent a further 14.0% decline during the years 2007–2009 (11 cases) with the largest proportion occurring between 2008 and 2009 (7.9 cases). These changes are similar to national changes in abdominal hysterectomy rates during the initial time period. Both the resident trend during the second time period and the overall change between times reflect declines not seen in the overall abdominal hysterectomy count or rate (9.9% decline for residents compared with 13.3% decline for the number of procedures and 21.5% decline for the rate per 10,000 female population).
Resident experience with vaginal hysterectomy (including laparoscopically assisted cases) stayed relatively unchanged over both time periods, with a change of only 3.3 additional cases (10.5%) for the 3-year average experience after work-hour restrictions, although most of this was accounted for by a jump of more than six cases in 2009. During the period from 2000 to 2008, the national rate of vaginal hysterectomy (per 10,000 female population) declined by 22.4%. Laparoscopically assisted procedures rose during this period by 44.0% and the rate by 39.9%. Compared with the national data, resident experience is rising more rapidly than the national rate for these procedures. Because of the magnitude of the decline in abdominal cases, median resident experience with all types of hysterectomy showed a 2.9% decline (3.3 cases) between comparative time periods.
In contrast with hysterectomy experience, resident experience with laparotomy, laparoscopy, and hysteroscopy all increased after the implementation of work-hour rules. Median laparotomy experience through the 2008 graduating class rose to a 2-year average of 54.6 cases compared with a 3-year average of 41.7 (31.0% increase over baseline) before the work-rule changes. Hysteroscopic experience showed a similar increase, with median cases rising by 23 cases (62.7 compared with 39.7, a 58.1% increase) between the two time periods, although some of this change may be due to a relaxation in the definition of hysteroscopic experience from strictly “operative” cases during 2001–2003. The largest increase in gynecologic experience occurred in laparoscopic cases, where the average of resident median cases rose by 37.3 cases (94 compared with 56.7, a 65.9% increase). This increase in laparoscopic experience outstrips the change in national data, which has shown only a 13.9% growth over the past 10 years.
Any evaluation of resident experience is, perforce, limited by issues of reporting validity and drifting definitions. Where definitions have changed, they have been indicated and it is reasonable to assume that the rates of resident under- and over-reporting are generally similar from year to year and from time period to time period. Despite this, the nature of these data makes the drawing of any unequivocal conclusions problematic and scientifically questionable. In addition, because of the sometimes small numbers involved, such an evaluation can also be subverted into a form of “statistical lying,” when naked percentages are used to prove a bias for or against a particular change or result. As a result, the best that one can hope for from a study such as this is a catalyst for further discussion and study. When viewed in absolute terms, there has been little change in the reported median resident experience in core procedures.
Intuitively, any reduction in the amount of time that residents are available for educational experiences might result in a reduction in technical (operative) experience. Some care must be taken in the interpretation of this study because of this very logical underpinning. Unlike many studies where there is a null hypothesis (eg, the results of two treatments do not differ), this study posits an if-then question. Inherent is the assumption that if hour restrictions reduce resident experience (P) then there will be a decline in the number of cases performed by them (Q), expressed logically as if P then Q (P → Q). With this structure it is a logical flaw (affirming the consequent) to say that if we document a change in experience (Q) that it is due to the hour restrictions (P). We can, however, under the logical principal of modus tollens, say that if no change in resident experience is found (¬Q), then no reduction in resident experience has occurred because of hour restrictions (¬P). (That is: if P then Q, not Q therefore not P, written in logical notation P → Q, ¬Q |- ¬P, where |- is the logical assertion.)
When viewed objectively, between the time periods 2001–2003 and 2007–2009 there have been few substantive negative changes in resident case experience at a national level. The number of vaginal deliveries declined slightly between periods (but shows a current rising trend), but few would argue that the magnitude of the decline might degrade the already robust educational experience of the residents. The median number of cesarean deliveries performed as “surgeon” per resident actually rose between these two time periods—a change that would seem to be neither educationally necessary nor the result of less time available for educational experiences. Although experience with cesarean deliveries increased, there was, however, a dramatic decline in resident experience with forceps- and vacuum-assisted (operative) vaginal deliveries; these changes are not driven by the hour restrictions but rather are an extension of a trend begun in 2001–2003 and reflective of the national trend away from these procedures. Speculation about the sufficiency of the resultant resident experience to result in clinical competency is beyond the scope of this evaluation.
The use of resident cesarean delivery and operative delivery rates (percentage of deliveries) for comparison with national rates could be criticized because of the educational selection bias that should result in higher resident rates than seen in normal practice settings. (Residents attend procedures with educational value while other providers may handle less specialized events, such as routine vaginal deliveries.) Whereas this bias precludes direct comparison of rates, it should not affect trends over time, which should be sensitive to changing practice patterns, or to the time available for education—work hours, the subject of interest here.
Surgical experience in core gynecologic procedures (except for abdominal hysterectomy) all showed increases in the median number of cases reported. It is unclear why such a change might occur in a setting presumed to allow for fewer hours of possible exposure in the operating room. Unknown are the effects of increased efficiency, a greater priority given to resident education over service, the impact of altered rotations structures such as night-float systems, or faulty perceptions of time committed before the implementation of hours restrictions.
Overall, published data on the impact of mandated work-hour restrictions are somewhat conflicting. Recent data suggest that the reduction in resident work hours may result in improved patient safety, but this has been at the cost of increased staff time, increased cost, and reduced faculty job satisfaction.3–5 Other studies have been unable to document significant changes in 30-day readmission rate, in hospital mortality, patient's length of stay, or residents' performance on in-training examinations.6 In contrast to the findings of this study, some studies have documented an apparent reduction in case experience in other fields of training.7 What is emerging is that both the great hopes and the great fears surrounding resident work-hour restrictions have not come to pass.
1. Kelly A, Marks F, Westhoff C, Rosen M. The effect of the New York State restrictions on resident work hours. Obstet Gynecol 1991;78(3 pt 1):468–73.
2. Espey E, Ogburn T, Puscheck E. Impact of duty hour limitations on resident and student education in obstetrics and gynecology. J Reprod Med 2007;52:345–8.
3. Privette AR, Shackford SR, Osler T, Ratliff J, Sartorelli K, Hebert JC. Implementation of resident work hour restrictions is associated with a reduction in mortality and provider-related complications on the surgical service: a concurrent analysis of 14,610 patients. Ann Surg 2009;250:316–21.
4. Payette M, Chatterjee A, Weeks WB. Cost and workforce implications of subjecting all physicians to aviation industry work-hour restrictions. Am J Surg 2009;197:820–5; discussion 826–7.
5. Vanderveen K, Chen M, Scherer L. Effects of resident duty-hours restrictions on surgical and nonsurgical teaching faculty. Arch Surg 2007;142:759–64; discussion 764–6.
6. Shonka DC Jr, Ghanem TA, Hubbard MA, Barker DA, Kesser BW. Four years of accreditation council of graduate medical education duty hour regulations: have they made a difference? Laryngoscope 2009;119:635–9.
7. Connors RC, Doty JR, Bull DA, May HT, Fullerton DA, Robbins RC. Effect of work-hour restriction on operative experience in cardiothoracic surgical residency training. J Thorac Cardiovasc Surg 2009;137:710–3.