Advertisement

Effect of Student Duty Hours Policy on Teaching and Satisfaction of 3rd Year Medical Students

      In July 2003, the Accreditation Council for Graduate Medical Education (ACGME) mandated standard duty hour restrictions for all resident physicians.

      Accreditation Council for Graduate Medical Education. Resident Duty Hours and the Working Environment. Available at: http://www.acgme.org/acWebsite/dutyHours/dh_Lang703.pdf. Accessed July 17, 2005.

      This step was fueled by concerns of resident sleep deprivation and the impact of excessive fatigue on physician health and patient safety.
      • Berkoff K.
      • Rusin W.
      Pediatric house staff’s psychological response to call duty.
      • Grunebaum A.
      • Minkoff H.
      • Blake D.
      Pregnancy among obstetricians: a comparison of births before, during, and after residency.
      • Klebanoff M.A.
      • Shiono P.H.
      • Rhoads G.G.
      Outcomes of pregnancy in a national sample of resident physicians.
      • Marcus C.L.
      • Loughlin G.M.
      Effect of sleep deprivation on driving safety in housestaff.
      • Steele M.T.
      • Ma O.J.
      • Watson W.A.
      • et al.
      The occupational risk of motor vehicle collisions for emergency medicine residents.
      • Veasey S.
      • Rosen R.
      • Barzansky B.
      • et al.
      Sleep loss and fatigue in residency training: a reappraisal.
      Although this policy change is generally regarded as an advance in graduate medical education, there are still many unintended consequences of these restrictions. One concern expressed by medical educators is the impact of these restrictions on the teaching and training of medical students. Although the Liaison Committee on Medical Education (LCME) has endorsed monitoring medical student duty hours and recommends specifically that medical student hours not exceed residents’ (a distinct possibility), suggestions for how to implement student duty hours are absent.

      Liason Committee on Medical Education. Educational Program for the MD Degree, ED-38. Available from: http://www.lcme.org/standard.htm. Accessed July 18, 2005.

      • Chen J.
      • Kamath R.
      • Lam A.
      • et al.
      How will changes in resident work hours affect medical students?.
      The effect of ACGME duty hours for residents and resident educational satisfaction has been extensively evaluated, but less attention has been focused on the effect of ACGME duty hours on students and student satisfaction.
      • Chen J.
      • Kamath R.
      • Lam A.
      • et al.
      How will changes in resident work hours affect medical students?.
      • Lin G.A.
      • Beck D.C.
      • Garbutt J.M.
      Residents’ perceptions of the effects of work hour limitations at a large teaching hospital.
      • Mathis B.R.
      • Diers T.
      • Hornung R.
      • et al.
      Implementing duty-hour restrictions without diminishing patient care or education: can it be done?.
      • Vidyarthi A.R.
      • Katz P.P.
      • Wall S.D.
      • et al.
      Impact of reduced duty hours on residents’ educational satisfaction at the University of California, San Francisco.
      • Jagsi R.
      • Shapiro J.
      • Weinstein D.F.
      Perceived impact of resident work hour limitations on medical student clerkships: a survey study.
      • Kogan J.R.
      • Bellini L.M.
      • Shea J.A.
      The impact of resident duty hour reform in a medicine core clerkship.
      • White C.B.
      • Haftel H.M.
      • Purkiss J.A.
      • et al.
      Multidimensional effects of the 80-hour work week at the University of Michigan Medical School.
      Kogan et al studied medical students on their third-year medicine clerkship before and after ACGME resident duty hour reform and found no difference between the proportion of time for direct or indirect patient care, education activities, educational value of activities, and time spent in the hospital.
      • Kogan J.R.
      • Bellini L.M.
      • Shea J.A.
      The impact of resident duty hour reform in a medicine core clerkship.
      Jagsi et al similarly found that resident work hour restrictions did not significantly affect student educational satisfaction.
      • Jagsi R.
      • Shapiro J.
      • Weinstein D.F.
      Perceived impact of resident work hour limitations on medical student clerkships: a survey study.
      However, White et al found negative effects of resident work hour restrictions on third-year medical student experiences, especially on surgery-oriented clerkships (such as obstetrics-gynecology and surgery).
      • White C.B.
      • Haftel H.M.
      • Purkiss J.A.
      • et al.
      Multidimensional effects of the 80-hour work week at the University of Michigan Medical School.
      PubMed search results returned no studies that have examined the effect of a policy restricting medical student duty hours on medical student teaching processes and satisfaction.
      The formation of a student duty hour policy (Appendix) arose from assessment of the medical student role with respect to the advent of ACGME resident duty hours. A guiding principle behind this policy is that students are not hospital employees, and their primary responsibility is to their own learning, which includes attending structured educational activities outside of the clinical wards, such as directed reading about patients, attending preceptor groups, and preparing for examinations. Thus, while a resident may have the opportunity to go home postcall and sleep, students often have educational responsibilities that continue beyond time spent in the hospital. Students also are being evaluated on their performance. Although additional learning might take place if students spend the entire call night with their team, the overall educational experience must be considered. Because new patients are not routinely admitted to the on-call teams after midnight, but instead are admitted by a night float resident, students would not be admitting and working a patient de novo, but instead helping the resident continue the work-up of patients admitted earlier that night and responding to active complaints of current inpatients on other teams (such as cross-coverage). It is also possible that a particularly interesting case, procedure, or surgery might be missed if the student was not required to stay the entire night. However, the added educational benefit from such patients is uncertain and perhaps not substantially diminished if the teaching moment occurred during the subsequent day rounds.
      At the University of Chicago Pritzker School of Medicine, these concerns were brought to the Clinical Curriculum Review Committee (CCRC), which is composed of administration representatives, elected student representatives, and clerkship directors. A special working group was formed to consider student duty hours. The initial proposal by this workgroup—that students would depart on-call by midnight—was opposed due to the concerns of missed learning opportunities overnight. One suggestion was to create an optional rule to allow students to leave if they wanted. However, students felt that this “option” would introduce variability and peer pressure.
      From a legal perspective, in Illinois, the Illinois Medical Practice Act requires a medical student’s practice to occur “under the direct, on premises supervision of a physician who is licensed to practice medicine in all its branches in Illinois and who is a member of the faculty of an accredited medical or osteopathic college.”
      Given this mandate, at the Pritzker School of Medicine, it was important to restructure medical student duty hours to ensure effective onsite supervision. The new policy, which took effect in July 2004 (Appendix), requires students to receive patients by 10 pm and be finished with clinical responsibilities by midnight. The adopted policy, one of the first such policies for medical students in the nation, was then circulated to clerkship directors, program directors, all faculty, and third- and fourth-year medical students.
      • Beller P.C.
      Your Intern [Medical Student] Today Is Both Sleepy and Bored Feel Better?.
      At the 1-year follow-up to the implementation, perceived effect among faculty and student representatives to CCRC was positive. The aim of this study is to assess the effect of the Pritzker School of Medicine student duty hour policy on student teaching and satisfaction in the core third-year inpatient medicine rotation.

      Methods

      From June 2002 to July 2005, third-year medical students at the Pritzker School of Medicine were surveyed after completion of their inpatient general medicine rotation, which is required in the third-year internal medicine core clerkship. From June 2002 to July 2003, medical students rotated with residents on inpatient general medical services that were not subject to resident or student duty hour restrictions. In July 2003, residents who rotated on inpatient general medicine were subject to a limit of 80 hours per week and no more than 30 hours consecutively.

      Accreditation Council for Graduate Medical Education. Resident Duty Hours and the Working Environment. Available at: http://www.acgme.org/acWebsite/dutyHours/dh_Lang703.pdf. Accessed July 17, 2005.

      To meet these requirements, internal medicine call schedules were structured every fourth night, with resident shifts from 8 am on the on-call day to 2 pm on the postcall day. In addition, a night float service was created to admit new patients between midnight and 7 am.
      • Arora V.
      • Dunphy C.
      • Chang V.Y.
      • et al.
      The effects of on-duty napping on intern sleep time and fatigue.
      Medical students continued to be assigned to teams (1 attending, 1 resident, and 2 interns) with no instruction regarding their duty hour requirements.
      In July 2004, the Pritzker School of Medicine instituted the student duty hour policy (Appendix). Thus, 3 cohorts of students with 3 different experiences on inpatient general medicine services were surveyed: cohort 1 included students before ACGME duty hour restrictions (July 2002 to June 2003); cohort 2 included students after ACGME duty hour restrictions only (July 2003 to June 2004); and cohort 3 included students after ACGME duty hour restrictions, and implementation of the institutional student duty hours policy (July 2004 to June 2005). Because of the inherent differences between these 3 cohorts and conditions, and to fully assess the effect of the implementation of a student duty hours policy on teaching and student satisfaction, cohort 3 was first compared with cohort 1 and then to cohort 2. Although not a primary aim of this study, the effect of ACGME duty hour restrictions on teaching and student satisfaction were also examined by comparing cohorts 2 and 1. The University of Chicago Institutional Review Board approved this study.

       Data Collection

      Students were asked to complete a 32-item end-of-clerkship survey, adapted from prior studies, which assessed teaching processes and satisfaction with their inpatient general medicine rotation.
      • Wachter R.M.
      • Katz P.
      • Showstack J.
      • et al.
      Reorganizing an academic medical service: impact on cost, quality, patient satisfaction, and education.
      Using open-ended responses, students were asked to detail teaching processes, such as the approximate number of times that month they were observed performing a history and physical examination, the number of hours per week they were taught by their attending or resident, both didactic and at the bedside, and the number of times they received feedback about their work. Students were also asked to rate their overall satisfaction with their general medicine rotation using a 5-point Likert scale. Using this same scale, students also rated their satisfaction with specific components of their rotation that mapped to teaching (eg, clinical topics) and their attending or resident (eg, availability or relationship).

       Data Analysis

      T-tests were used to discern the effect of student duty hours on the continuous responses to open-ended items regarding teaching processes. Because the responses to the open-ended questions regarding the number of times a student was observed performing an interview or physical examination were mostly zeros, a binary outcome variable was generated by assigning a “1” to all students who reported they were observed performing an interview or physical examination at least once and a “0” to those students reporting they were not observed. The authors acknowledge that student satisfaction with the inpatient general medicine rotation is skewed; the majority of students report “Somewhat Satisfied” or “Very Satisfied.” In addition, prior work demonstrates that high levels of satisfaction lead to a sustained career interest in general medicine.
      • Arora V.
      • Wetterneck T.B.
      • Schnipper J.L.
      • et al.
      Effect of the inpatient general medicine rotation on student pursuit of a generalist career.
      Because of this trend, focus shifted to how the policy change would affect students reporting the highest level of satisfaction (“Very Satisfied”), a meaningful education-related outcome. Therefore, responses to the Likert scale items on satisfaction were dichotomized to generate binary outcome variables that represent a high degree of satisfaction with a particular element by assigning responses of “Very Satisfied” a “1” and all those less than “Very Satisfied” as “0.”
      • Chung P.
      • Morrison J.
      • Jin L.
      • et al.
      Resident satisfaction on an academic hospitalist service: time to teach.
      Chi-squared tests were used to test the effect of student duty hours on satisfaction with certain elements of the clerkship. In addition, although it was not the primary research question, Likert scores were compared between cohorts using Wilcoxon tests. No significant differences were found between cohorts on any of the variables. For the sake of brevity and because these results did not contradict the primary analyses, only the results from dichotomizing the Likert scores are reported.
      All statistical tests were performed using Intercooled Stata 8.0 (Stata Corp., College Station, Tex), with statistical significance defined as P <.05.

      Results

      Of the 335 third-year medical students who rotated on general medicine services from July 2002 to June 2005, 213 (64%) completed the survey. Response rates per cohort were as follows: Cohort 1—70/113 (62%); Cohort 2—68/116 (59%); Cohort 3—75/106 (71%). When comparing teaching processes between students with student duty hours (cohort 3) and students without duty hours (either cohort 2 or cohort 1), the number of hours per week spent with the resident or intern fell markedly. Students with restricted student duty hours reported spending, on average, roughly 30 hours per week less with their resident or intern when compared with students before resident duty hour restrictions (cohort 1) [−28.9 (−33.2 to −24.6)] and 23 hours per week less compared with students after resident duty hour restrictions without a student duty hours policy (cohort 2) [−23.8 (−27.8 to −19.8)] (Table 1). Despite these substantial decreases in overall time spent with the resident with restricted student duty hours, no change was observed in the number of hours residents were teaching, either didactic-style or at the bedside. However, students with a student duty hours policy were much less likely to report they were observed by their resident performing a physical examination at least once (24%) when compared with either cohort 1 (40.6%, P <.03) or cohort 2 (42.7%, P <.02). Of note, little change was reported in time spent with the attending or teaching by the attending in a didactic setting with fairly consistent reported means across all 3 academic years. Interestingly, when compared with cohort 1 and after the implementation of a student duty hours policy, students were more likely to report an additional hour of bedside teaching by their attending. There were no changes in any of these teaching processes reported by students with the introduction of ACGME duty hour restrictions (cohort 2 vs cohort 1).
      Table 1Effect of Student Duty Hours on Teaching Processes
      Teaching ProcessesC1
      Cohort 1 (C1): 3rd year medical students who rotated on inpatient general medicine during 2002-2003 (pre-resident duty hours). Cohort 2 (C2): 3rd year medical students who rotated on inpatient general medicine during 2003-2004 (post resident duty hours). Cohort 3 (C3): 3rd year medical students who rotated on inpatient general medicine during 2004-2005 (post medical student duty hour).
      C2
      Cohort 1 (C1): 3rd year medical students who rotated on inpatient general medicine during 2002-2003 (pre-resident duty hours). Cohort 2 (C2): 3rd year medical students who rotated on inpatient general medicine during 2003-2004 (post resident duty hours). Cohort 3 (C3): 3rd year medical students who rotated on inpatient general medicine during 2004-2005 (post medical student duty hour).
      C3
      Cohort 1 (C1): 3rd year medical students who rotated on inpatient general medicine during 2002-2003 (pre-resident duty hours). Cohort 2 (C2): 3rd year medical students who rotated on inpatient general medicine during 2003-2004 (post resident duty hours). Cohort 3 (C3): 3rd year medical students who rotated on inpatient general medicine during 2004-2005 (post medical student duty hour).
      Difference C3-C1 (95% CI)Difference C3-C2 (95% CI)
      Conferences missed per week2.32.82.80.5 (−0.2-1.2)0.03 (−0.7-0.7)
      Attending
       Hours/week overall14.616.415.81.2 (−0.4-2.7)−0.7 (−0.4-2.7)
       Hours/week didactic teaching33.03.10.09 (−0.7-0.9)0.1 (−0.7-0.9)
       Hours/week bedside teaching3.23.94.41.2 (0.06-2.3)
      Because of non-normal distribution with responses skewed towards zero, responses dichotomized by assigning “1” for those students that reported that they were observed in these skills at least once and “0” for those students that reported that they were not observed. Chi-squared tests used to test differences. Confidence intervals generated from 2 sample tests of proportions.
      0.4 (−0.8-1.7)
       Received feedback2.21.61.9−0.3 (−1.2-0.6)0.4 (−0.2-0.9)
       Observed performing interview
      Because of non-normal distribution with responses skewed towards zero, responses dichotomized by assigning “1” for those students that reported that they were observed in these skills at least once and “0” for those students that reported that they were not observed. Chi-squared tests used to test differences. Confidence intervals generated from 2 sample tests of proportions.
      24.625.421.3−3.3 (−1.7-1.0)−4.0 (−18.0-9.9)
       Observed performing physical examination
      Because of non-normal distribution with responses skewed towards zero, responses dichotomized by assigning “1” for those students that reported that they were observed in these skills at least once and “0” for those students that reported that they were not observed. Chi-squared tests used to test differences. Confidence intervals generated from 2 sample tests of proportions.
      39.138.840.00.09 (−15.1-16.8)1.2 (−14.9-17.3)
      Resident
       Hours/week with overall51.846.722.9−28.9 (−33.2-−24.6)
      P value <.05.
      −23.8 (−27.8-−19.8)
      P value <.05.
       Hours/week didactic teaching3.24.03.2−0.02 (−0.9-0.9)−0.9 (−1.9-0.2)
       Hours/week bedside teaching4.75.55.50.8 (−0.8-2.4)0.01 (−1.8-1.8)
       Received feedback
      Students asked to estimate hours per week spent with attending or resident overall, receiving didactic teaching, or bedside teaching and the number of times received feedback from the attending or resident in that month.
      3.92.53.7−0.3 (−1.7-1.2)1.2 (0.3-2.1)
      Because of non-normal distribution with responses skewed towards zero, responses dichotomized by assigning “1” for those students that reported that they were observed in these skills at least once and “0” for those students that reported that they were not observed. Chi-squared tests used to test differences. Confidence intervals generated from 2 sample tests of proportions.
       Observed performing interview
      Because of non-normal distribution with responses skewed towards zero, responses dichotomized by assigning “1” for those students that reported that they were observed in these skills at least once and “0” for those students that reported that they were not observed. Chi-squared tests used to test differences. Confidence intervals generated from 2 sample tests of proportions.
      30.438.226.7−3.8 (−18.5-11.0)−11.6 (−26.9-3.7)
       Observed performing physical examination
      Because of non-normal distribution with responses skewed towards zero, responses dichotomized by assigning “1” for those students that reported that they were observed in these skills at least once and “0” for those students that reported that they were not observed. Chi-squared tests used to test differences. Confidence intervals generated from 2 sample tests of proportions.
      40.642.724.0−16.6 (−31.7-−1.4)
      P value <.05.
      −18.6 (−33.9-−3.4)
      P value <.05.
      low asterisk Cohort 1 (C1): 3rd year medical students who rotated on inpatient general medicine during 2002-2003 (pre-resident duty hours). Cohort 2 (C2): 3rd year medical students who rotated on inpatient general medicine during 2003-2004 (post resident duty hours). Cohort 3 (C3): 3rd year medical students who rotated on inpatient general medicine during 2004-2005 (post medical student duty hour).
      Students asked to estimate hours per week spent with attending or resident overall, receiving didactic teaching, or bedside teaching and the number of times received feedback from the attending or resident in that month.
      Because of non-normal distribution with responses skewed towards zero, responses dichotomized by assigning “1” for those students that reported that they were observed in these skills at least once and “0” for those students that reported that they were not observed. Chi-squared tests used to test differences. Confidence intervals generated from 2 sample tests of proportions.
      § P value <.05.
      Despite these changes in teaching processes with the introduction of a student duty hours policy, there was little change in student satisfaction, either overall or with various elements of the rotation (Table 2). Overall student satisfaction remained high, with over 60% of students reporting that they were “Very Satisfied” with their inpatient general medicine rotation. Furthermore, despite a substantial reduction in the number of hours spent with their resident, there were no differences in student satisfaction with the availability of, relationship with, clinical excellence of, or emphasis on education by their resident; the majority of students reported a high degree of satisfaction with these characteristics of their residents.
      Table 2Effect of Student Duty Hours on Student Satisfaction with Rotation
      Outcomes (percent very satisfied)
      Likert scale responses for all items except for patient care items ranged from 1 (Very Dissatisfied) to 5 (Very Satisfied). Values reported are the percentage of students rating 5 (Very Satisfied).
      C1C2C3Difference C3-C1 (95% CI)Difference C3-C2 (95 % CI)
      Overall
       Overall satisfaction61.447.860.3−1.2 (−17.2-14.8)12.5 (−3.9-28.9)
      Teaching
       Presence of structured curriculum56.655.243.8−12.9 (−27.1-1.3)
      P value <.05; reported from chi-squared tests after dichotomizing Likert scale responses by assigning a “1” for those students that rated 5 (Very Satisfied) and “0” for those students that response less than 5 (Very Satisfied).
      −11.4 (−25.6-2.7)
       Teaching about clinical topics50.039.750.70.7 (−15.7-17.1)11.0 (−5.3-27.2)
       Teaching about basic science topics17.710.314.7−3.0 (−15.1-9.1)4.4 (−6.4-15.2)
       Emphasis on evidence-based medicine44.148.552.07.9 (−8.5-24.2)3.5 (−12.9-19.9)
       Teaching about cost-effectiveness10.57.417.36.9 (−4.4-18.2)10.0 (−0.6-20.6)
      P value <.05; reported from chi-squared tests after dichotomizing Likert scale responses by assigning a “1” for those students that rated 5 (Very Satisfied) and “0” for those students that response less than 5 (Very Satisfied).
       Teaching about managed care6.04.46.70.7 (−7.3-8.7)2.3 (−5.2-9.7)
      Attending characteristics
       Quality of attending rounds42.730.940.0−2.6 (−18.8-13.5)9.1 (−6.5-24.7)
       Clinical excellence61.870.669.37.6 (−8.0-23.1)−1.3 (−16.3-13.8)
       Availability57.451.566.79.3 (−6.6-25.2)15.2 (−0.8-31.2)
      P value <.05; reported from chi-squared tests after dichotomizing Likert scale responses by assigning a “1” for those students that rated 5 (Very Satisfied) and “0” for those students that response less than 5 (Very Satisfied).
       Relationship54.455.258.74.3 (−12.0-20.5)3.4 (−12.9-19.8)
       Emphasis on education48.554.457.38.8 (−7.5-25.1)2.9 (−13.4-19.2)
       Provision of timely feedback45.638.249.33.7 (−12.6-20.1)11.1 (−5.1-27.3)
      Resident characteristics
       Clinical excellence75.066.276.01.0 (−13.1-15.1)9.8 (−5.0-24.7)
       Availability80.672.173.3−7.3 (−21.0-6.5)1.3 (−13.4-15.9)
       Relationship76.564.765.3−11.1 (−25.9-3.6)0.6 (−15.0-16.3)
       Emphasis on education60.350.057.3−3.0 (−19.1-13.2)7.3 (−9.0-23.7)
       Provision of timely feedback52.947.153.30.4 (−16.0-16.8)6.3 (−10.1-22.7)
      Patient care
       Patient volume
      Likert scale responses ranged from 1 (Too Little) to 5 (Too Much) with 3 (Just Right). Values reported are the percentages of students rating 3 (Just Right). Confidence intervals generated from 2 sample tests of proportions.
      82.688.184.01.4 (−10.8-13.6)−4.1 (−15.4-7.3)
       Supervision of your clinical work
      Likert scale responses ranged from 1 (Too Little) to 5 (Too Much) with 3 (Just Right). Values reported are the percentages of students rating 3 (Just Right). Confidence intervals generated from 2 sample tests of proportions.
      81.480.978.7−2.8 (−15.8-10.2)−2.2(−15.4-10.9)
      low asterisk Likert scale responses for all items except for patient care items ranged from 1 (Very Dissatisfied) to 5 (Very Satisfied). Values reported are the percentage of students rating 5 (Very Satisfied).
      Likert scale responses ranged from 1 (Too Little) to 5 (Too Much) with 3 (Just Right). Values reported are the percentages of students rating 3 (Just Right). Confidence intervals generated from 2 sample tests of proportions.
      P value <.05; reported from chi-squared tests after dichotomizing Likert scale responses by assigning a “1” for those students that rated 5 (Very Satisfied) and “0” for those students that response less than 5 (Very Satisfied).

      Discussion

      This article reports the results of the first-known study of the effect of an institutional policy restricting third-year medical student duty hours during their required clerkships on students’ reports of teaching processes and satisfaction with a core rotation. After the institution of a student duty hours policy, students spent substantially less time with their residents. However, this decrease in overall time with the resident did not change the amount of teaching received—overall, didactic, or at the bedside—from their residents. However, students were less likely to be observed by their resident performing a physical examination with the institution of a student duty hours policy. Despite these changes, there were no differences in satisfaction with the rotation overall or with characteristics related to the attending, resident, or teaching methods. Objective differences in student performance in the wake of these duty hour changes were examined by reviewing the average score on the National Board of Medical Examiners (NBME) medicine shelf examination that is taken by all Pritzker School of Medicine third-year medical students at the end of their medicine clerkship. After the implementation of the student duty hours policy, the average shelf examination scores for cohort 3 were relatively stable, with only 1% difference from cohort 2. Interestingly, performance decreased on the shelf examination after the implementation of resident duty hours, with a 16% decrease between cohort 1 and cohort 2. Upon further examination, cohort 1 consistently performed well on national standardized tests, raising the possibility of cohort effects.
      Several implications of this study exist. Most importantly, although a student duty hours policy did not change overall satisfaction with the rotation or satisfaction with major elements of the general medicine rotation, the effects of the substantial reduction of time spent with residents on a core inpatient rotation must be fully explored. Students did not report any differences in formal teaching by their residents, which suggests that students may spend a lot of time on-call with their residents, not engaged in formal teaching. Some of this time may represent student participation in routine clinical care activities with their residents, such as looking up laboratory results or ordering tests. However, this time likely includes valuable educational opportunities such as residents observing students performing physical examinations. Given that approximately one half of students reported that they were observed performing these core skills by an attending, the loss of residents as teachers and observers of this skill is potentially harmful. In addition, this decrease in student time with their residents may also have consequences on acquisition of advanced clinical or practical insight regarding residency training. Although student performance did not appear to change on the NBME shelf examination with the institution of student duty hours, the most sensitive educational outcomes are not being measured, and further work may be needed to truly delineate the consequences of students spending less time with their residents.
      There are several limitations of this study. First, the response rate of 64% is less than ideal. Possible selection bias exists if only students who were satisfied with their experience answered the survey, which is also further complicated by the fact that most students were very satisfied with their rotation and with characteristics of the rotation. In addition, because 3 different cohorts of students were surveyed during 3 different time periods, none of the students experienced more than one of the conditions (baseline, post-ACGME duty hours, poststudent duty hours). There could be inherent differences between these 3 cohorts of students, leading to the differences observed in perception of teaching processes and satisfaction with their inpatient general medicine clerkship. For example, cohorts may have differed in the perceived time and education they received from their attendings or residents, or the perceived impact of additional rest on their educational experience. From another question of this survey, there was no difference in the interest expressed in a career in general internal medicine after the clerkship. However, this assessment does not exclude other inherent differences from potentially confounding these results. Regardless, student responses to certain items, such as teaching processes regarding their attending, remained remarkably consistent during these 3 years.

      Conclusion

      A policy to restrict student duty hours resulted in substantial decreases in student time spent with residents. However, this decrease did not appear to affect the time spent by the resident teaching or satisfaction with the inpatient general medicine rotation.

      Acknowledgments

      This study was supported by the Agency for Healthcare Research and Quality grant R01 10597 A Multicenter Trial of Academic Hospitalists. The authors are grateful to Veronica Tirado and Katie Chiu for their excellent research assistance. They also thank Dr. Halina Brukner and the members of the Student Duty Hours Workgroup: Joel Schwab, MD, Loretto Glynn, MD, Adam Cifu, MD, Sandy Cook, PhD, and student representatives Wil Van Cleave, Geoff Oxnard, John Meier, and Adam Holdt.

      Appendix 1.

      Figure thumbnail gr1

      References

      1. Accreditation Council for Graduate Medical Education. Resident Duty Hours and the Working Environment. Available at: http://www.acgme.org/acWebsite/dutyHours/dh_Lang703.pdf. Accessed July 17, 2005.

        • Berkoff K.
        • Rusin W.
        Pediatric house staff’s psychological response to call duty.
        J Dev Behav Pediatr. 1991; 12: 6-10
        • Grunebaum A.
        • Minkoff H.
        • Blake D.
        Pregnancy among obstetricians: a comparison of births before, during, and after residency.
        Am J Obstet Gynecol. 1987; 157: 79-83
        • Klebanoff M.A.
        • Shiono P.H.
        • Rhoads G.G.
        Outcomes of pregnancy in a national sample of resident physicians.
        N Engl J Med. 1990; 323: 1040-1045
      2. Kohn L.T. Corrigan J.M. Donaldson M.S. To Err is Human: Building a Safer Health System. Institute of Medicine National Academy Press, Washington, DC2000
        • Marcus C.L.
        • Loughlin G.M.
        Effect of sleep deprivation on driving safety in housestaff.
        Sleep. 1996; 19: 763-766
        • Steele M.T.
        • Ma O.J.
        • Watson W.A.
        • et al.
        The occupational risk of motor vehicle collisions for emergency medicine residents.
        Acad Emerg Med. 1999; 6: 1050-1053
        • Veasey S.
        • Rosen R.
        • Barzansky B.
        • et al.
        Sleep loss and fatigue in residency training: a reappraisal.
        JAMA. 2002; 288: 1116-1124
      3. Liason Committee on Medical Education. Educational Program for the MD Degree, ED-38. Available from: http://www.lcme.org/standard.htm. Accessed July 18, 2005.

        • Chen J.
        • Kamath R.
        • Lam A.
        • et al.
        How will changes in resident work hours affect medical students?.
        Curr Surg. 2003; 60 (discussion 270): 268-270
        • Lin G.A.
        • Beck D.C.
        • Garbutt J.M.
        Residents’ perceptions of the effects of work hour limitations at a large teaching hospital.
        Acad Med. 2006; 81: 63-67
        • Mathis B.R.
        • Diers T.
        • Hornung R.
        • et al.
        Implementing duty-hour restrictions without diminishing patient care or education: can it be done?.
        Acad Med. 2006; 81: 68-75
        • Vidyarthi A.R.
        • Katz P.P.
        • Wall S.D.
        • et al.
        Impact of reduced duty hours on residents’ educational satisfaction at the University of California, San Francisco.
        Acad Med. 2006; 81: 76-81
        • Jagsi R.
        • Shapiro J.
        • Weinstein D.F.
        Perceived impact of resident work hour limitations on medical student clerkships: a survey study.
        Acad Med. 2005; 80: 752-757
        • Kogan J.R.
        • Bellini L.M.
        • Shea J.A.
        The impact of resident duty hour reform in a medicine core clerkship.
        Acad Med. 2004; 79: S58-S61
        • White C.B.
        • Haftel H.M.
        • Purkiss J.A.
        • et al.
        Multidimensional effects of the 80-hour work week at the University of Michigan Medical School.
        Acad Med. 2006; 81: 57-62
      4. Medical Practice Act of 1987. 1987 (225 ILCS 60)
        • Beller P.C.
        Your Intern [Medical Student] Today Is Both Sleepy and Bored.
        The New York Times. 2005 (Education. September 14)
        • Arora V.
        • Dunphy C.
        • Chang V.Y.
        • et al.
        The effects of on-duty napping on intern sleep time and fatigue.
        Ann Intern Med. 2006; 144: 792-798
        • Wachter R.M.
        • Katz P.
        • Showstack J.
        • et al.
        Reorganizing an academic medical service: impact on cost, quality, patient satisfaction, and education.
        JAMA. 1998; 279: 1560-1565
        • Arora V.
        • Wetterneck T.B.
        • Schnipper J.L.
        • et al.
        Effect of the inpatient general medicine rotation on student pursuit of a generalist career.
        J Gen Intern Med. 2006; 21: 471-475
        • Chung P.
        • Morrison J.
        • Jin L.
        • et al.
        Resident satisfaction on an academic hospitalist service: time to teach.
        Am J Med. 2002; 112: 597-601