Pediatric Residents’ Perspectives on Reducing Work Hours and Lengthening Residency: A National Survey
OBJECTIVE: In 2011, the Accreditation Council on Graduate Medical Education increased restrictions on resident duty-hours. Additional changes have been considered, including greater work-hours restrictions and lengthening residency. Program directors tend to oppose further restrictions; however, residents’ views are unclear. We sought to determine whether residents support these proposals, and if so why.
METHODS: We surveyed US pediatric residents from a probability sample of 58 residency programs. We used multivariate logistic regression to determine predictors of support for (1) a 56-hour workweek and (2) the addition of 1 year to residency to achieve a 56-hour week.
RESULTS: Fifty-seven percent of sampled residents participated (n = 1469). Forty-one percent of respondents supported a 56-hour week, with 28% neutral and 31% opposed. Twenty-three percent of all residents would be willing to lengthen training to reduce hours. The primary predictors of support for a 56-hour week were beliefs that it would improve education (odds ratio [OR] 8.6, P < .001) and quality of life (OR 8.7, P < .001); those who believed patient care would suffer were less likely to support it (OR 0.10, P < .001). Believing in benefits to education without decrement to patient care also predicted support for a 56-hour-week/4-year program.
CONCLUSIONS: Pediatric residents who support further reductions in work-hours believe reductions have positive effects on patient care, education, and quality of life. Most would not lengthen training to reduce hours, but a minority prefers this schedule. If evidence mounts showing that reducing work-hours benefits education and patient care, pediatric residents’ support for the additional year may grow.
- ACGME —
- Accreditation Council on Graduate Medical Education
- M/LWP —
- married/living with partner
- NIH —
- National Institutes of Health
- OR —
- odds ratio
What's Known on This Subject:
In 2011, the Accreditation Council on Graduate Medical Education increased restrictions on resident duty-hours. Further changes have been considered, including greater work-hour restrictions and lengthening residency. Residents’ views about these policies are unclear.
What This Study Adds:
This is the first systematic, national inquiry into resident opinions on reduced work-hours and longer residency. More pediatric residents support than oppose reduced hours, and a minority would add a year to residency to achieve them.
The Accreditation Council for Graduate Medical Education (ACGME) released new resident work-hours regulations in July 2011, after an Institute of Medicine report linking physician fatigue with medical errors.1,2 The new regulations call for shorter shifts, limited to 16 consecutive hours of work for interns and 28 consecutive hours for higher-level residents.2 US residency program directors indicate concern about implementing these requirements, noting the tension between curtailing shifts and maintaining service and training needs within the existing structure of residency.3
Even as shorter shifts are put in place, total weekly work-hours remain unchanged under the 2011 ACGME limits; interns and other residents continue to be allowed to work up to 80 hours per week (averaged over 4 weeks; 88 hours in some surgical programs). An alternative approach to managing physician fatigue is reducing overall hours worked per week. This is the approach taken by most industrialized nations worldwide (Table 1). Reducing shift length and overall hours worked per week has been found in most studies to be associated with reduced medical error and improvements in physician health and safety.1,4-13 Trainees report improved well-being and decreased stress and burnout in the wake of work-hour reductions.6,8,9 Adoption of this approach in the United States has been inhibited in part by fears that reduction in work-hours will correspond to less effective training.3,9
Lengthening residency by an amount fully compensating for work-hour reductions is a proposed solution that allows for the benefits of reduced hours, including preventing fatigue-related errors, while maintaining total time spent in training.14 There is evidence that the same amount of instruction over more time enhances learning.15,16 Lengthening training is a common strategy used to accommodate parental leave during residency, but it may not be acceptable to the general population of residents.17,18
There has been little systematic inquiry into residents’ views of prolonging residency or reducing work-hours below the 80-hour limit. In this study, we surveyed a nationally representative sample of pediatric residents to investigate their support of further work-hour reductions and to assess whether they would be willing to add a year to residency to reduce work-hours. We also investigated predictors of supporting this trade-off, seeking to understand what might motivate support for this policy. This information can inform a residency redesign that addresses workforce needs and maintains pediatrics as an attractive specialty choice.
We anticipated that lengthening residency would be unpopular overall, but hypothesized that residents who experience negative emotional effects of extended work-hours6,9,19,20 (eg, burnout) would be more likely to support reduced hours and a longer residency. We also hypothesized that those individuals with specific motivations for limiting the duration of residency (eg, high debt burden, plan to pursue additional training) would prefer to maintain the status quo.
Study Design and Population
We conducted a cross-sectional survey of US pediatric residents from a proportionately stratified probability cluster sample of residents. We identified 191 US pediatric residency programs through the ACGME’s Graduate Medical Education Directory 2008–2009. We classified them by the Association of Pediatric Program Directors regional designations (www.appd.org), condensed to Northeast, South, Midwest, and West. We also dichotomized programs by size, with large programs defined as exceeding the median size (>39 residents). As a result, we had 8 strata from which to sample, cross-tabulated by size and region. Next, we randomly sampled 74 programs and asked their program directors to administer our questionnaire. These programs were distributed throughout each of the 8 strata, proportionate to the size of the strata. That is, if a stratum had a small population, then fewer programs were sampled from it. If a program director declined to participate, we contacted the next randomly selected program within the stratum. Fifty-eight programs ultimately distributed surveys to their residents. We attempted to survey all pediatric residents within each program.
After we sampled a program, we contacted the program director to enroll it and solicit help with survey administration. Program directors received $20. As an incentive for residents, the program with the highest response rate received $200 for a party for all pediatric residents.
Program directors (or their representatives) administered an anonymous paper survey during a regularly scheduled conference. Residents sealed their completed survey in an opaque envelope. An online survey was also available; electronic responses were accessible only to the researchers.
The Children’s Hospital Boston Institutional Review Board approved this study.
We developed the questionnaire based on a literature review and on interviews with 10 key stakeholders, who discussed their experience as or with residents related to work-hours, career plans, and emotional health. We then pretested our draft questionnaire with 10 local pediatric residents, who helped us to clarify our wording.
We framed questions about work-hour reduction as follows: “I support reducing all pediatric residents’ work-hours from 80 hours per week to 56 hours per week” (yes, no, not sure). A separate item asked all residents: “To reduce work-hour limits from 80 to 56, I would be willing to change my current schedule to add a year to residency” (5-point Likert scale rated from Disagree Strongly to Agree Strongly). The 2 questions were independent, so a resident could convey opposition to work-hour reduction and support for the trade-off. We selected 56 hours because it was the limit in the United Kingdom during the survey period (without lengthened training) and was discussed as a policy alternative in the United States.21
Descriptive Measures and Predictor Variables
We included in our models factors that we hypothesized to be related to the dependent variable in 4 categories: program-level characteristics, resident emotional experiences, resident beliefs about the impact of work-hour reduction, and demographics. Program-level characteristics were region, size, and research focus (with the use of the top quartile of National Institutes of Health [NIH] grant dollars22 as a proxy measure). Resident emotional experiences were assessed by using previously validated scales for burnout and depression. We assessed burnout, a psychological construct for long-term exhaustion and diminished interest, by using the Maslach Burnout Inventory.23 Burnout rates have been reported between 18% and 82% among medical trainees.24 We assessed depression, which is also common among residents,6,25–27 by using the Harvard National Depression Screen, defining positive as a score ≥9. The Harvard National Depression Screen has high sensitivity (95%) and specificity (94%) for detecting major depression.28 Resident beliefs about the impact of work-hour reduction on patient care, education, and quality of life were assessed by asking to what extent the resident agreed with a series of statements by using a 5-point Likert scale. Responses to the following representative statements from each domain were dichotomized for simplification (strongly agree to agree versus neutral to strongly disagree) in the multivariate model: “Patient care would suffer”; “I would learn more pediatrics”; “My quality of life would improve.” Finally, we included these demographic variables: postgraduate year, gender, family relationships (married/living with partner [M/LWP] and parenting, M/LWP not parenting, not M/LWP and parenting, not M/LWP and not parenting), career plans (general versus subspecialty pediatrics), and household indebtedness (<$50K, $50–150K, $150–250K, >$250K).
Descriptive measures are reported as percentage of population who responded favorably (agree or strongly agree on a 5-point Likert scale). We assessed differences in proportions of residents who endorsed specific beliefs by their support of work-hour policy by using χ2 tests. To determine predictors of supporting a 56-hour workweek, both with and without an additional year of residency, we performed multivariate logistic regressions by use of the above measures. We considered program- and resident-level characteristics separately, and then in combined analysis. We used poststratified inverse proportional sampling weights to account for variable program participation (by region and size), and variable nonresponse by resident program, gender, and program year. Relevant population data were available through the American Board of Pediatrics. We used these weights for both descriptive statistics and regression analyses. For each variable, we analyzed missing values as a distinct category. We performed all analyses by using Stata 11.1 survey commands.29
Characteristics of Responding Programs
Among participating programs (n = 58), 43.1% were large, 32.7% were in the Northeast, 32.8% were in the South, 28.6% were in the Midwest, and 6.8% were in the West (Table 2).
Characteristics of Responding Residents
Enrolled programs had 2555 residents in postgraduate years 1 to 3. We received 1469 responses from postgraduate year 1 to 3 residents (57.5%). Characteristics of responding residents and population data are presented in Table 3.
Residents’ Attitudes Toward Work-hour Reduction
Forty-one percent of residents supported reducing work-hours to 56 hours, 27.7% were unsure, and 31.4% did not support it (Table 4).
When asked about the impact of this reduction, most reported that their quality of life (87.7%) and stress level (80.3%) would improve; 27.3% felt the change would prevent them from becoming a competent physician. Although 49.7% of residents believed they would be better able to take care of patients, 42.2% believed that patient care would suffer (Table 4). Residents who supported work-hour reduction were much more likely to predict positive effects on patient care, education, and quality of life than those who were neutral or opposed to the policy (Table 5).
In multivariate analysis (Table 6), the primary predictors of support for work-hour reduction were beliefs about positive policy impact on education and quality of life with no decrement to patient care. We found some regional variation in policy support, with programs in the South less likely than those in the Northeast to support reduced hours (odds ratio [OR] 0.57, P = .03). When program- and resident-level characteristics were analyzed separately, results were unchanged (data not shown).
Willingness to Add a Year to Residency
Twenty-three percent of sampled residents (including those who did not support work-hour reduction alone) would add a year to residency to reduce hours, 9.3% were neutral, and 67.3% were opposed (Table 4). Of the residents who supported or were neutral toward the 56-hour week, 57.4% were willing to add a year of training. Residents who supported the additional year of residency were much more likely to predict positive effects on patient care, education, and quality of life than those who opposed or were neutral toward the policy (Table 5).
In combined multivariate analysis (Table 6), belief in the positive impact of the 56-hour/4-year policy on education predicted support (OR 1.71, P < .001). Residents who believe that patient care would suffer were less likely to endorse this policy (OR 0.45, P < .001). Belief that work-hour reduction would improve quality of life did not translate into support for the additional training year. Support did vary by postgraduate year, with third-year residents more likely to support the policy (OR 1.65, P = .004). Single parents were also more likely to support the policy than residents who were partnered without children (OR 6.10, P = .03). Burnout also predicted willingness to reduce work-hours and lengthen residency (OR 1.46, P = .02).
When considering program- and respondent-level characteristics separately, programs with a strong research focus (OR 0.79, P = .04) and individuals with subspecialty career plans (OR 0.74, P = .04) were less inclined to support lengthening training (data not shown). Otherwise, effect sizes and SEs were virtually identical.
This study found that more pediatric residents supported a 56-hour workweek than opposed it, with many undecided. Although the majority of residents did not support lengthening residency, 23% of all sampled residents would lengthen residency training by 1 year to create a 56-hour workweek. Support for reducing hours with and without an additional year of residency appears to be driven primarily by beliefs in positive policy impact on patient care and education. Lengthening residency with a shorter workweek has particular appeal for residents who experience burnout or who plan not to pursue subspecialty training. To our knowledge, this is the first multicenter study investigating how residents view a 56-hour workweek/4-year residency as a proposed solution to improve patient care, medical training, and physician wellness.
Work-hour reduction remains controversial. Concerns include potential for increasing the cost of medical care and degradation of training quality.3,9,30,31 Whether cost would in fact increase or training quality decrease remains uncertain. Lengthening residency may offset the potential cost of work-hour reduction by maintaining a workforce that is less expensive than its substitutes (eg, nurse practitioners), and would add additional training time. Regarding training quality, 1 review identified 14 studies of the educational impact of reducing work-hours; of these, 4 showed improvements and 9 showed no change.12 There is little literature about the educational impact of lengthening residency. Educational studies suggest that lengthening the duration of learning and increasing opportunity for repetition would only improve knowledge acquisition and retention.15,16 Demonstrations of this principle in medical training exist, but are isolated to specific content areas.16,32 One small study of pediatric residents who deliberately pursued a longer, slower residency showed no difference in in-training examination scores or board passage rates compared with coresidents in the same program on a traditional schedule.33
Lengthening residency is often assumed to be an unpalatable option for residents themselves, who would defer for another year a full physician salary and career. Concerns to date have been largely speculative. The largest inquiry into resident opinion about current ACGME guidelines used a convenience sample with a 22% response rate.18 Attitudes about prolonged time in training were provided by only half of respondents, limiting the ability to draw conclusions. In a small, single-center study from Colorado, 16% of internal medicine residents preferred to or were undecided about whether to extend residency to 4 or 5 years to reduce work-hours to 60 per week; burned-out residents were more likely to prefer to extend residency.34 Two small studies of residency programs that piloted an additional year in training noted that attitudes about specialty choice remained positive, and later applications to subspecialty training did not suffer.35,36 Furthermore, length of training period did not have a significant impact on specialty choice in a large study of US medical students.37
We extended the existing work by using a large, nationally representative sample of pediatric residents to examine the desirability of proposed schedule changes. We investigated the program and resident-level characteristics, including emotional experience, beliefs about policy effects, and demographics, that underlie opinion. This study shows that the primary predictors of schedule preference are residents’ beliefs about the impact of reduced work-hours on their patients, medical education, and quality of life. It is not clear whether these beliefs stem from personal experience, knowledge of evidence in the literature, or local culture. Whatever the source of their belief, residents who felt they would be better able to deliver safe, effective patient care and learn medicine by working a reduced-hours schedule were more likely to accept the longer residency. Given the size of this group, offering flexibility in training schedule and length either within or among residency programs may help to improve patient care and trainee education.17,33
Residents experiencing burnout were more likely to support a reduced hours/longer residency. This mirrored the result in the Colorado study. In our sample, 45% of residents met criteria for burnout, similar to national averages.24 Burned-out residents did not prefer work-hour reduction without the additional year of training, possibly reflecting greater ambivalence that this proposal alone would provide adequate education and patient care. This group is likely to experience improvements in burnout with further work-hour reductions; some evidence suggests that reductions in medical errors may follow.6,8,9,11,19,20,24
Although these are good arguments for a longer, slower residency, this schedule has potential drawbacks. In our study, this is highlighted by the fact that the majority of residents did not support it. We believe this reflects desire to forge ahead with a planned career. Our analysis shows, for example, that subspecialty career plans were associated with lower likelihood of supporting this policy. Similarly, when examining program characteristics, research focus predicted lower likelihood of support. The similarity of these measures likely reduced statistical power in analysis combining both program- and resident-level characteristics. Those planning a subspecialty career typically have at least 3 more years of training ahead; it is intuitive that they would not favor adding another. In other specialties with longer residencies, proportional lengthening may present additional challenges in this regard. In pediatrics, this finding is of particular concern when considering lengthening residency, because workforce predictions suggest a need for more specialists.38,39 Any proposed increase in residency length should address this challenge, perhaps with early graduation for specialists or broadening availability of nonstandard pathways. Several nonstandard pathways reduce clinical training in general pediatrics to allow for research training. In the Accelerated Research Pathway, trainees complete 2 years of residency and then begin an extended fellowship. The Integrated Research Pathway incorporates blocks of research time during a 3-year general pediatrics residency.
We had anticipated that residents with high debt might also oppose an additional year of training. Debt level was not, however, related to schedule preference in our analysis. It may be that household debt alone could not capture the complexity of the role of finances in career decisions, or that finances are not primary motivators of career decisions among pediatric residents.40,41 The relationship between household finances and career decisions warrants further study in this population, given its high and increasing debt-to-income ratio and the potential impact of proposed scheduling changes.
This study has several limitations. First, as a cross-sectional survey, the ability to draw causal inferences is limited. In addition, although response rates were high for surveys of residents,42,43 and weights were used to carefully balance participants with known characteristics of the population, the possibility of some remaining nonresponse bias related to attitudes toward the survey’s topic exists. We attempted to minimize this bias by creating an administration protocol whereby program directors did not advertise the survey content beyond requiring conference attendance. Finally, our policy proposals and pediatric residents’ opinions may not generalize across specialties.
Pediatric residents who support further reductions in work-hours believe reductions have positive effects on patient care, education, and quality of life. Most pediatric residents would not lengthen training to reduce weekly hours, but a minority prefers this schedule. From the pediatric residents’ perspective, if a 4-year/56-hour residency is considered, it should make accommodations such as early graduation for those pursuing fellowship. If evidence mounts showing that reducing work-hours benefits education and patient care, pediatric residents’ support for the additional year may grow.
We thank the Association for Pediatric Program Directors and their associates who helped with data collection.
- Accepted March 20, 2012.
- Address correspondence to Mary Beth Gordon, MD, Division of General Pediatrics, Children’s Hospital Boston, 300 Longwood Ave, Boston, MA 02115. E-mail:
Dr Gordon had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
FINANCIAL DISCLOSURE: Dr Landrigan has reviewed cases evaluating the effects of sleep deprivation on provider performance for Sachs Waldman, and has received honoraria from multiple institutions for lecturing on sleep deprivation, patient care handoffs, and patient safety; the other authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: Supported by the Association for Pediatric Program Directors Special Projects Program, the Frederick H. Lovejoy Jr. MD Housestaff Research and Education Fund, and the Harvard Pediatric Health Services Research Fellowship, HRSA T32 HP10018.
- Institute of Medicine
- ↵Accreditation Council on Graduate Medical Education. Common program requirements. 2010. Available at: www.acgme-2010standards.org/pdf/Common_Program_Requirements_07012011.pdf. Accessed July 20, 2011
- Cappuccio FP,
- Bakewell A,
- Taggart FM,
- et al.,
- Warwick EWTD Working Group
- Landrigan CP,
- Fahrenkopf AM,
- Lewin D,
- et al
- Moonesinghe SR,
- Lowery J,
- Shahi N,
- Millen A,
- Beard JD
- ↵Fleisher GR. American Pediatric Society 2011 Presidential Address: Earthquakes, tectonic shifts in graduate medical education, and the role of the APS in finding solid ground. Pediatric Res. 2011; 70(6):647–651
- Dempster FN,
- Farris R
- Bjork RA
- ↵ACGME. Shaping the future. In: Proceedings of Accreditation Council for Graduate Medical Education: Annual Educational Conference; March 5–8, 2009; Grapevine, TX
- ↵National Institutes of Health Research Portfolio Online Reporting Tools (RePORT). Medical schools ranking tables 2005. Available at: http://report.nih.gov/award/rank/MedSchool_Departments.cfm?Department=PEDIATRICS. Accessed February 15, 2011
- Maslach C,
- Jackson SE,
- Leiter MP
- Fahrenkopf AM,
- Sectish TC,
- Barger LK,
- et al
- ↵Stata Statistical Software: Release 11 [computer program]. College Station, TX: StataCorp LP; 2009
- American Academy of Pediatrics Committee on Pediatric Workforce
- Freed GL,
- Dunham KM,
- Research Advisory Committee of the American Board of Pediatrics
- Grava-Gubins I,
- Scott S
- Copyright © 2012 by the American Academy of Pediatrics