Advertisement
Mayo Clinic Proceedings Home

Physician Satisfaction and Burnout at Different Career Stages

      Abstract

      Objective

      To explore the work lives, professional satisfaction, and burnout of US physicians by career stage and differences across sexes, specialties, and practice setting.

      Participants and Methods

      We conducted a cross-sectional study that involved a large sample of US physicians from all specialty disciplines in June 2011. The survey included the Maslach Burnout Inventory and items that explored professional life and career satisfaction. Physicians who had been in practice 10 years or less, 11 to 20 years, and 21 years or more were considered to be in early, middle, and late career, respectively.

      Results

      Early career physicians had the lowest satisfaction with overall career choice (being a physician), the highest frequency of work-home conflicts, and the highest rates of depersonalization (all P<.001). Physicians in middle career worked more hours, took more overnight calls, had the lowest satisfaction with their specialty choice and their work-life balance, and had the highest rates of emotional exhaustion and burnout (all P<.001). Middle career physicians were most likely to plan to leave the practice of medicine for reasons other than retirement in the next 24 months (4.8%, 12.5%, and 5.2% for early, middle, and late career, respectively). The challenges of middle career were observed in both men and women and across specialties and practice types.

      Conclusion

      Burnout, satisfaction, and other professional challenges for physicians vary by career stage. Middle career appears to be a particularly challenging time for physicians. Efforts to promote career satisfaction, reduce burnout, and facilitate retention need to be expanded beyond early career interventions and may need to be tailored by career stage.

      Abbreviations and Acronyms:

      DP (depersonalization), EE (emotional exhaustion)
      Understanding factors that affect physician career satisfaction is important because it is associated with quality of care,
      • Melville A.
      Job satisfaction in general practice: implications for prescribing.
      • Linn L.S.
      • Yager J.
      • Cope D.
      • Leake B.
      Health status, job satisfaction, job stress, and life satisfaction among academic and clinical faculty.
      • Katz A.
      Better outcome means more job satisfaction: pilot project in Winnipeg and Halifax to enhance physician-patient communication.
      patient satisfaction,
      • Linn L.S.
      • Brook R.H.
      • Clark V.A.
      • Davies A.R.
      • Fink A.
      • Kosecoff J.
      Physician and patient satisfaction as factors related to the organization of internal medicine group practices.
      • Haas J.S.
      • Cook E.F.
      • Puopolo A.L.
      • Burstin H.R.
      • Cleary P.D.
      • Brennan T.A.
      Is professional satisfaction of general internists associated with patient satisfaction.
      and patient adherence to medical treatments.
      • DiMatteo M.R.
      • Sherbourne C.D.
      • Hays R.D.
      • et al.
      Physicians' characteristics influence patients' adherence to medical treatment: results from the Medical Outcomes Study.
      Furthermore, dissatisfied physicians are more likely to reduce their clinical work hours, leave their current practice, or retire early—all of which disrupt patient-physician relationships and affect access to medical care.
      • Williams E.S.
      • Konrad T.R.
      • Scheckler W.E.
      • et al.
      Understanding physicians' intentions to withdraw from practice: the role of job satisfaction, job stress, mental and physical health.
      Dissatisfied physicians are also at higher risk for professional burnout,
      • Shanafelt T.D.
      • Balch C.M.
      • Bechamps G.J.
      • et al.
      Burnout and career satisfaction among American surgeons.
      which is a potential barrier to successful health care reform.
      • Dyrbye L.N.
      • Shanafelt T.D.
      Physician burnout: a potential threat to successful health care reform.
      Although most physicians are satisfied with their careers,
      • Mechanic D.
      Physician discontent: challenges and opportunities.
      career satisfaction varies by specialty, income, region, and age.
      • Leigh J.P.
      • Kravitz R.L.
      • Schembri M.
      • Samuels S.J.
      • Mobley S.
      Physician career satisfaction across specialties.
      Data from the 1996-1997 Community Tracking Physician Survey suggest that there is a U-shaped relationship between age and career satisfaction, with younger physicians and older physicians having greatest career satisfaction and middle career physicians having lowest satisfaction.
      • Leigh J.P.
      • Kravitz R.L.
      • Schembri M.
      • Samuels S.J.
      • Mobley S.
      Physician career satisfaction across specialties.
      Nearly a quarter of today’s physicians are in the challenging middle career stage.
      American Medical Association
      Physician Characteristics and Distribution in the U.S.
      We conducted a study to (1) further evaluate the work lives of US physicians by career stage, (2) explore the relationship between career stage and multiple dimensions of professional satisfaction and burnout, and (3) examine whether these relationships by career stage persist across sexes, specialties, and practice settings.

      Participants and Methods

      Study Population

      As previously described,
      • Shanafelt T.D.
      • Boone S.
      • Tan L.
      • et al.
      Burnout and satisfaction with work-life balance among U.S. physicians relative to the general U.S. population.
      in June 2011 we selected a diverse sample of 89,831 physicians from all specialty disciplines from the American Medical Association Physician Masterfile. These physicians were sent an e-mail inviting them to participate in an anonymous, voluntary study. In accordance with established survey methods,

      American Association for Public Opinion Research. Standard definitions. Final disposition of case codes and outcome rates for surveys. 2011. http://www.aapor.org/AM/Template.cfm?Section=Standard_Definitions2&Template=/CM/ContentDisplay.cfm&ContentID=3156. Accessed May 4, 2012.

      the 27,276 physicians who opened at least one invitation e-mail were considered to have received an invitation to participate in the study.

      Study Measures

      The survey included items that inquired about demographic characteristics, hours worked per week, frequency of overnight call, specialty area, years in practice (beyond residency and fellowship), primary practice setting, burnout, work-home conflicts, and career satisfaction. We collapsed specialties into primary care (ie, general pediatrics, general internal medicine, and family medicine), surgical specialties, medicine and pediatric subspecialties, and other (ie, anesthesiology, dermatology, emergency medicine, radiology, neurology, pathology, physical medicine and rehabilitation, psychiatry, and other).

      Burnout

      Burnout was measured using the Maslach Burnout Inventory, which is considered the criterion standard tool for measuring burnout.
      • Maslach C.
      • Jackson S.E.
      • Leiter M.P.
      Maslach Burnout Inventory Manual.
      • Rafferty J.P.
      • Lemkau J.P.
      • Purdy R.R.
      • Rudisill J.R.
      Validity of the Maslach Burnout Inventory for family practice physicians.
      • Lee R.T.
      • Ashforth B.E.
      A meta-analytic examination of the correlates of the three dimensions of job burnout.
      • Leiter M.
      • Durup J.
      The discriminant validity of burnout and depression: a confirmatory factor analytic study.
      Burnout encompasses 3 domains (emotional exhaustion [EE], depersonalization [DP], and low sense of personal accomplishment), which have been confirmed in factor analyses.
      • Maslach C.
      • Jackson S.E.
      • Leiter M.P.
      Maslach Burnout Inventory Manual.
      Reliability evidence is supported by a Cronbach coefficient α of 0.90 for EE, 0.79 for DP, and 0.71 for personal accomplishment in large population samples.
      • Maslach C.
      • Jackson S.E.
      • Leiter M.P.
      Maslach Burnout Inventory Manual.
      Scores within individual burnout domains were categorized into low, intermediate, and high scores using established cutoffs.
      • Maslach C.
      • Jackson S.E.
      • Leiter M.P.
      Maslach Burnout Inventory Manual.
      Because high scores on either the EE (≥27) or DP (≥10) scales can distinguish clinically burned out from non–burned out individuals
      • Schaufeli W.
      • Bakker A.
      • Hoogduin K.
      • Schaap C.
      • Kladler A.
      On the clinical validity of the Maslach Burnout Inventory and the Burnout Measure.
      and because many studies have identified high levels of either EE or DP as the underpinning of burnout in physicians,
      • Thomas N.K.
      Resident burnout.
      • Shanafelt T.D.
      • Bradley K.A.
      • Wipf J.E.
      • Back A.L.
      Burnout and self-reported patient care in an internal medicine residency program.
      • Rosen I.M.
      • Gimotty P.A.
      • Shea J.A.
      • Bellini L.M.
      Evolution of sleep quantity, sleep deprivation, mood disturbances, empathy, and burnout among interns.
      we considered physicians with a high EE or DP score as having at least one manifestation of professional burnout.
      • Maslach C.
      • Jackson S.E.
      • Leiter M.P.
      Maslach Burnout Inventory Manual.

      Work-Home Conflict

      Consistent with our prior approach,
      • Dyrbye L.N.
      • Shanafelt T.D.
      • Balch C.
      • Satele D.
      • Freischlag J.
      Relationship between work-home conflicts and burnout among American surgeons: a comparison by sex.
      • Dyrbye L.N.
      • Freischlag J.
      • Kaups K.A.
      • et al.
      Work-home conflicts have a substantial impact on career decisions that affect the adequacy of the surgical workforce.
      physicians were asked whether they had experienced a conflict between work (clinical or administrative) and personal responsibilities within the past 3 weeks and how the most recent work-home conflict was resolved (ie, in favor of work responsibilities, in favor of personal responsibilities, or in a manner that met both responsibilities).

      Career Satisfaction

      The survey included comprehensive measures of professional satisfaction in 6 different dimensions, including career satisfaction, specialty satisfaction, satisfaction with work-life balance, intent to reduce clinical work hours, intent to leave current practice, and whether they would recommend medicine as a career option to their children (only asked of those with children). On the basis of similar measures from previous physician surveys to assess career satisfaction,
      • Shanafelt T.D.
      • Balch C.M.
      • Bechamps G.J.
      • et al.
      Burnout and career satisfaction among American surgeons.
      • Shanafelt T.D.
      • Bradley K.A.
      • Wipf J.E.
      • Back A.L.
      Burnout and self-reported patient care in an internal medicine residency program.
      • Kuerer H.M.
      • Eberlein T.J.
      • Pollock R.E.
      • et al.
      Career satisfaction, practice patterns and burnout among surgical oncologists: report on the quality of life of members of the Society of Surgical Oncology.
      • Frank E.
      • McMurray J.E.
      • Linzer M.
      • Elon L.
      Career satisfaction of US women physicians.
      • Lemkau J.
      • Rafferty J.
      • Gordon Jr., R.
      Burnout and career-choice regret among family practice physicians in early practice.
      • Goitein L.
      • Shanafelt T.D.
      • Wipf J.E.
      • Slatore C.G.
      • Back A.L.
      The effects of work-hour limitations on resident well-being, patient care, and education in an internal medicine residency program.
      physicians were asked if given the opportunity to revisit their career choice whether they would choose to become a physician again (career choice). Another item asked if given the opportunity to revisit their specialty choice whether they would choose the same specialty again (specialty choice). Response options included “definitely not,” “probably not,” “not sure, neutral,” “probably,” and “definitely yes.” Responses of “probably” or “definitely yes” were considered to indicate greater career satisfaction.
      Satisfaction with work-life balance was ascertained by asking physicians their level of agreement with the statement that their work schedule leaves them enough time for their personal or family life. Physicians who answered “agree” or “strongly agree” were considered to be satisfied with their work-life balance.
      Physicians were asked the likelihood that (1) they would reduce their work hours devoted to clinical care during the next 12 months or (2) leave their current practice within 2 years (none, slight, moderate, likely, or definite).
      • Dyrbye L.N.
      • Freischlag J.
      • Kaups K.A.
      • et al.
      Work-home conflicts have a substantial impact on career decisions that affect the adequacy of the surgical workforce.
      Physicians who indicated a moderate or higher likelihood of reducing their clinical work hours were asked to indicate their primary reason for doing so (ie, frustration with Medicare and insurance reimbursement issues, desire to spend more time with family, decreasing reimbursement for clinical care, desire to pursue administrative or leadership opportunities, desire to pursue research or medical education opportunities, and other). Similarly, physicians who indicated a moderate or higher likelihood of leaving their current practice were asked what they would do if they left their current practice (ie, look for a different practice opportunity and continue to work as a physician, look for a different job in medicine and no longer work as a physician, leave the practice of medicine altogether to pursue a different career, retire, and other).

      Career Stage

      We classified a priori physicians who had been out of training and in practice for 10 years or less as early career, 11 to 20 years as middle career, and 21 or more years as late career. We opted to use years in practice rather than age because of variation in training lengths of residencies or fellowships and to accommodate those who had a career before medicine. Our definitions are generally consistent with those of the US Office of Personnel Management.

      United States Office of Personnel Management. Career Patterns: A 21st Century Approach to Attracting Talent. http://www.opm.gov/hcaaf_resource_center/careerpatterns/CPGuideV1.pdf. Accessed October 1, 2012.

      Statistical Analyses

      We used standard descriptive summary statistics and applied the Kruskal-Wallis test (continuous variables) or χ2 test (categorical variables) to explore for associations among variables. We performed basic forward-stepping logistic model with backwards stepping
      • Greene W.H.
      Econometric Analysis.
      to confirm the findings to evaluate associations of the independent variables with burnout, career satisfaction, specialty satisfaction, and work-life balance satisfaction. The independent variables used in the burnout model included demographic characteristics (ie, age, sex, children, and relationship status [married or partnered vs single or widowed as the referent]), personal work characteristics (ie, years in practice [0-10, 11-20, ≥21], work hours, nights on call, and specialty [primary care, surgical, medicine and pediatric specialties, or other], practice setting [private practice, academic medical center, veterans’ hospital, or active military practice]), recent work-home conflict, and how the most recent work-home conflict was resolved. The independent variables used in the career, specialty, and work-life balance satisfaction models also included burnout. All tests were 2-sided, with type I error rates of .05. All analyses were performed using SAS statistical software, version 9 (SAS Institute Inc).

      Results

      As previously reported,
      • Shanafelt T.D.
      • Boone S.
      • Tan L.
      • et al.
      Burnout and satisfaction with work-life balance among U.S. physicians relative to the general U.S. population.
      7288 of the 27,276 physicians who received an invitation to participate (26.3% cooperation rate) completed surveys. Participating physicians had similar demographic characteristics as the 814,022 US physicians listed in the Physician Masterfile, although participants were slightly older and further removed from medical school graduation. Early and later responders (a measure of response bias) were statistically similar with respect to age, sex, and specialty.
      Of the 7288 physicians who completed the survey, 1583 (22.2%) were early career, 1634 (22.9%) were middle career, and 3906 (54.8%) were late career; 165 did not indicate their years in practice and were excluded from the analysis. Demographic and practice characteristics of responding physicians by career stage are summarized in Table 1. Practice characteristics, work-home conflicts, burnout, and career satisfaction by career stage and sex can be found in the Supplemental Appendix (available online at http://www.mayoclinicproceedings.org).
      Table 1Demographic Characteristics, Practice Characteristics, and Work-Home Conflicts by Career Stage
      Percentages account for missing values not included in the table.
      CharacteristicCareer stage, No. (%)
      Data are presented as No. (percentage) of study participants unless otherwise indicated.
      P value
      Early (0-10 y) (n=1583)Middle (11-20 y) (n=1634)Late (≥21 y) (n=3906)
      Sex<.001
       Male858 (54.2)1055 (64.6)3208 (82.1)
       Female725 (45.8)579 (35.4)698 (17.9)
      Age (y), median384961<.001
      Children<.001
       Yes1120 (70.8)1375 (84.2)3520 (90.1)
       No463 (29.2)258 (15.8)386 (9.9)
      Specialty<.001
       Surgical360 (23.1)416 (26.2)1035 (27.2)
       Primary care
      Primary care specialties include general internal medicine, general practice, family medicine, and general pediatrics.
      410 (26.3)350 (22.1)817 (21.5)
       Internal medicine and pediatric subspecialty243 (15.6)272 (17.1)714 (18.8)
       Other546 (35.0)549 (34.6)1238 (32.5)
      Primary practice setting<.001
       Private practice801 (51.9)962 (60.2)2235 (58.9)
       Academic medical center425 (27.6)320 (20.0)716 (18.9)
       Veterans’ hospital36 (2.3)33 (2.1)112 (3.0)
       Active military practice27 (1.8)23 (1.4)14 (0.4)
       Not in practice or retired5 (0.3)5 (0.3)68 (1.8)
       Other248 (16.1)255 (16.0)649 (17.1)
      Hours worked per week, mean (SD)52.5 (16.2)54.6 (16.0)50.2 (16.5)<.001
      Nights on call per week, mean (SD)1.9 (2.1)2.5 (2.4)2.3 (2.5)<.001
      Experienced work-home conflict in past 3 wk<.001
       Yes921 (58.5)839 (51.7)1377 (35.5)
       No654 (41.5)783 (48.3)2502 (64.5)
      How most recent conflict resolved<.001
       Resolved in favor of work552 (36.7)533 (34.4)887 (24.3)
       Resolved in favor of personal191 (12.7)176 (11.4)355 (9.7)
       Able to resolve to meet both763 (50.7)840 (54.2)2405 (65.9)
      a Percentages account for missing values not included in the table.
      b Data are presented as No. (percentage) of study participants unless otherwise indicated.
      c Primary care specialties include general internal medicine, general practice, family medicine, and general pediatrics.

      Work Hours, Call, and Work-Home Conflicts

      Middle career physicians reported working more hours and taking more nights on call than early or late career physicians (both P<.001; Table 1). Despite working fewer hours, early career physicians were more likely to have experienced a recent work-home conflict and were least likely to have been able to resolve the conflict in a manner that allowed both home and work responsibilities to be met (both P<.001).

      Burnout

      Statistically significant differences were found in the prevalence of high EE, high DP, and overall burnout by career stage (all P<.001; Table 2). Middle career physicians were more likely to have high EE and be burned out than early or late career physicians. In contrast, high DP was greatest among early career physicians and decreased incrementally in prevalence at middle and late career. The higher prevalence of burnout among middle career physicians persisted when analyzed by sex (Figure, A), across specialties (Figure, B), and by practice setting (Figure, C). Table 3 indicates that higher rates of burnout among middle career physicians persisted after adjusting for a variety of personal and professional factors. Interestingly, when we repeated multivariate analysis, including those who had retired or were not in practice, we found that being retired or not in practice remained independently associated with burnout (odds ratio, 1.93; 95% CI, 1.03-3.59), suggesting burnout may have contributed to the decision to leave practice in these individuals.
      Table 2Burnout and Career Satisfaction by Career Stage
      VariableCareer stage, No. (%)P value
      Early (0-10 y) (n=1583)Middle (11-20 y) (n=1634)Late (≥21 y) (n=3906)
      Burnout indices
       High emotional exhaustion629 (39.9)763 (47.0)1289 (33.4)<.001
       High depersonalization569 (36.2)557 (34.4)933 (24.2)<.001
       Burned out
      High score on emotional exhaustion and/or depersonalization scale (see “Methods”).
      798 (50.5)876 (53.9)1566 (40.4)<.001
      Dimensions of professional satisfaction
       Satisfied with work-life balance<.001
      Yes695 (44.0)638 (39.3)2101 (54.0)
      No883 (56.0)986 (60.7)1791 (46.0)
       Satisfied with choice to become a physician<.001
      Yes986 (62.4)1037 (63.9)2955 (76.1)
      No594 (37.6)585 (36.1)928 (23.9)
       Satisfied with specialty choice<.001
      Yes1098 (69.4)1075 (66.3)2847 (73.4)
      No483 (30.6)547 (33.7)1033 (26.6)
      Recommend career in medicine to children
      Only asked of those with children.
      .002
       Yes594 (53.3)711 (51.9)1982 (57.0)
       No520 (46.7)659 (48.1)1494 (43.0)
      Moderate or greater likelihood of reducing clinical hours within next 12 mo<.001
       Yes178 (20.8)219 (20.9)1004 (31.6)
       No677 (79.2)831 (79.1)2176 (68.4)
      Primary reason for considering reducing clinical work hours
      Only asked of physicians who indicated a moderate or higher likelihood of reducing hours.
      <.001
       Missing/No/Slight chance of reducing hours122913012784
       Frustration w/Medicare and insurance issues25 (7.1)62 (18.6)170 (15.2)
       Spend more time with family190 (53.7)109 (32.7)327 (29.1)
       Declining reimbursement for clinical care29 (8.2)32 (9.6)108 (9.6)
       Pursue admin/leadership opportunities29 (8.2)47 (14.1)106 (9.4)
       Pursue research/education opportunities42 (11.9)15 (4.5)43 (3.8)
       Other39 (11.0)68 (20.4)368 (32.8)
      Moderate or greater likelihood of leaving current practice within 2 y<.001
       Yes539 (34.1)405 (24.9)1541 (39.6)
       No1041 (65.9)1222 (75.1)2350 (60.1)
      What the physician would do if he/she left the current practice
      Only asked of physicians who indicated a moderate or higher likelihood of leaving current practice.
      <.001
       Missing/No/slight chance of leaving current practice104812342393
       Look for diff practice and continue to work as423 (79.1)202 (50.5)355 (23.5)
       Look for diff job in medicine and leave medicine41 (7.7)75 (18.8)167 (11.0)
       Leave medicine altogether to pursue different26 (4.9)56 (14.0)81 (5.4)
       Retire4 (0.7)22 (5.5)736 (48.6)
       Other41 (7.7)45 (11.3)174 (11.5)
      a High score on emotional exhaustion and/or depersonalization scale (see “Methods”).
      b Only asked of those with children.
      c Only asked of physicians who indicated a moderate or higher likelihood of reducing hours.
      d Only asked of physicians who indicated a moderate or higher likelihood of leaving current practice.
      Figure thumbnail gr1
      FigurePrevalence of burnout among middle career physicians compared with early or late career physicians according to sex (A), specialty area (B), and practice setting (C), with differences statistically significant for all variables (all P≤.01) except for the veterans’ hospital settings (P=.59).
      Table 3Multivariable Analysis for Burnout
      The independent variables used in the model included demographic characteristics (ie, age, sex, children, and relationship status [married or partnered vs single or widowed]), personal work characteristics (ie, years in practice [0-10, 11-20, or ≥21], work hours, nights on call, specialty [primary care, surgical, internal medicine and pediatric specialties, or other], practice setting [private practice, academic medical center, veterans’ hospital, or active military practice]), recent work-home conflict, and how the most recent work-home conflict was resolved.
      VariableOdds ratio (95% CI)P value
      Early career (0-10 y)
      Referent was middle career.
      0.75 (0.64-0.89)<.001
      Late career (≥21 y)
      Referent was middle career.
      0.74 (0.65-0.85)<.001
      Have children0.63 (0.54-0.73)<.001
      Hours worked per week (for each additional hour)1.02 (1.01-1.02)<.001
      Nights on call per week (for each additional night)1.03 (1.002-1.05).03
      Surgical specialty
      Referent was primary care.
      0.71 (0.60-0.83)<.001
      Internal medicine or pediatrics subspecialty
      Referent was primary care.
      0.72 (0.60-0.86)<.001
      Academic practice
      Referent was private practice.
      0.63 (0.55-0.73)<.001
      Recent work-home conflict2.47 (2.20-2.77)<.001
      Most recent work-home conflict resolved in favor of personal responsibility
      Referent was resolved in a manner that allowed meeting both work and personal responsibilities.
      1.29 (1.07-1.54).006
      Most recent work-home conflict resolved in favor of work responsibility
      Referent was resolved in a manner that allowed meeting both work and personal responsibilities.
      2.13 (1.87-2.43)<.001
      a The independent variables used in the model included demographic characteristics (ie, age, sex, children, and relationship status [married or partnered vs single or widowed]), personal work characteristics (ie, years in practice [0-10, 11-20, or ≥21], work hours, nights on call, specialty [primary care, surgical, internal medicine and pediatric specialties, or other], practice setting [private practice, academic medical center, veterans’ hospital, or active military practice]), recent work-home conflict, and how the most recent work-home conflict was resolved.
      b Referent was middle career.
      c Referent was primary care.
      d Referent was private practice.
      e Referent was resolved in a manner that allowed meeting both work and personal responsibilities.

      Career Satisfaction

      Satisfaction with overall career choice (being a physician) was lowest among early career physicians, higher in middle career, and highest in late career (P<.001; Table 2). Satisfaction with overall career choice was lowest among early career physicians for both sexes (P<.001) and for those who worked in private practice (P<.001), academia (P<.001), and veterans’ hospitals (P=.09). Career satisfaction was lowest among early career physicians for primary care physicians and surgeons but lowest among middle career physicians for internal medicine and pediatric subspecialty physicians and other specialists (all P<.001).
      Middle career physicians had the lowest satisfaction with specialty choice and work-life balance (both P<.001) and were the least likely to recommend medicine as a career option to their children (P=.002). Satisfaction with specialty choice was lowest among middle career physicians for both women (P=.004) and men (P=.0015), across all specialties (all P<.001 except for primary care [P=.08] and subspecialty internal medicine and pediatric practice [P=.98]), and within all practice settings (all P<.01, except for private practice [P=.0009], academic medical centers [P=.0001, and veterans’ hospitals [P>.99]). After adjustment for a variety of personal and professional factors, physicians in late career had greater odds of being satisfied with their career than those in middle career (Table 4), but the relationship between career stage and specialty satisfaction was no longer statistically significant (data not given). Physicians in early and late career had greater odds of being satisfied with their work-life balance than middle career physicians after adjusting for personal and professional factors.
      Table 4Multivariable Analysis for Career Satisfaction and Satisfaction With Work-Life Balance
      The independent variables used in the model included demographic characteristics (ie, age, sex, children, and relationship status [married or partnered vs single or widowed]), personal work characteristics (ie, years in practice [0-10, 11-20, or ≥21], work hours, nights on call, specialty [primary care, surgical, medicine and pediatric specialties, or other], practice setting [private practice, academic medical center, veterans’ hospital, active military practice]), recent work-home conflict, how the most recent work-home conflict was resolved, and burnout.
      VariableOdds ratio (95% CI)P value
      Career satisfaction
       Late career (≥21 y)
      Referent was middle career.
      1.56 (1.35-1.81)<.001
       Have children1.33 (1.13-1.57)<.001
       Academic practice
      Referent was private practice.
      1.63 (1.39-1.92)<.001
       Recent work-home conflict0.78 (0.68-0.89)<.001
       Most recent work-home conflict resolved in favor of personal responsibility
      Referent was resolved in a manner that allowed meeting both work and personal responsibilities.
      0.79 (0.65-0.97).02
       Most recent work-home conflict resolved in favor of work responsibility
      Referent was resolved in a manner that allowed meeting both work and personal responsibilities.
      0.79 (0.69-0.91)<.001
       Burned out0.24 (0.21-0.28)<.001
      Satisfaction with work-life balance
       Male (vs female)1.37 (1.18-1.60)<.001
       Married or partnered (vs single or widowed)1.33 (1.08-1.63).007
       Early career (0-10 y)
      Referent was middle career.
      1.33 (1.10-1.60).003
       Late career (≥21 y)
      Referent was middle career.
      1.24 (1.06-1.46).007
       Hours worked per week (for each additional hour)0.95 (0.94-0.95)<.001
       Nights on call per week (for each additional night)0.94 (0.92-0.97)<.001
       Recent work-home conflict0.34 (0.30-0.38)<.001
       Most recent work-home conflict resolved in favor of work
      Referent was resolved in a manner that allowed meeting both work and personal responsibilities.
      0.43 (0.37-0.50)<.001
       Burned out0.33 (0.29-0.37)<.001
      a The independent variables used in the model included demographic characteristics (ie, age, sex, children, and relationship status [married or partnered vs single or widowed]), personal work characteristics (ie, years in practice [0-10, 11-20, or ≥21], work hours, nights on call, specialty [primary care, surgical, medicine and pediatric specialties, or other], practice setting [private practice, academic medical center, veterans’ hospital, active military practice]), recent work-home conflict, how the most recent work-home conflict was resolved, and burnout.
      b Referent was middle career.
      c Referent was private practice.
      d Referent was resolved in a manner that allowed meeting both work and personal responsibilities.

      Intent to Reduce Clinical Work Hours or Leave Medicine

      Late career physicians were the most likely to report that they intended to reduce their clinical hours within the next 12 months, whereas similar rates were reported among early and middle career physicians (P<.001; Table 2). The reasons for planning to reduce clinical hours, however, varied by career stage. Younger physicians were most likely to report doing so to spend more time with family or to pursue research- or medical education–related work, whereas middle career physicians were most likely to report doing so out of frustration with Medicare or insurance reimbursement.
      Middle career physicians were the least likely to be seriously considering leaving their current practice within the next 2 years (P<.001; Table 2). Early career physicians who intended to leave their current practice were the most likely to be planning to relocate to a new practice to work as a physician. Middle career physicians were more likely to be contemplating leaving to pursue a different career in medicine (but no longer practicing clinical medicine) or intending to leave medicine altogether for another career. As expected, late career physicians were the most likely to report they were leaving their practice to retire.

      Discussion

      In this national sample of US physicians, we identified several notable differences in the work lives, burnout, and career satisfaction of physicians by career stage. Physicians early in their career had the lowest career satisfaction, greatest rates of work-home conflicts, more difficulty resolving work-home conflicts in a manner that allowed both work and home responsibilities to be met, and greater DP. The prevalence of these issues decreased among middle career physicians and was lowest among late career physicians. In contrast, middle career physicians worked more hours, took more call duty, reported the lowest specialty satisfaction, were more dissatisfied with work-life balance, and struggled more with EE and burnout than their early or late career colleagues. Late career physicians were generally the most satisfied and had the lowest rates of distress.
      One notable finding of this analysis is the distinct challenges experienced by middle career physicians. Long hours and frequent call duty likely contribute to worse work-life balance, greater EE and burnout, and more professional dissatisfaction. Middle career physicians were more likely to be planning to leave their current practice out of frustration and to pursue a career that involved no direct patient care or was outside medicine altogether. These findings are particularly concerning because middle career is often the most productive time of a physician’s career in terms of the amount of patient care provided.
      US Department of Health and Human Services
      Physician Workforce: Projections and Research Into Current Issues Affecting Supply and Demand.
      Although leaving the practice of medicine may have personal benefits for the individual physician, from a societal perspective it amplifies the physician workforce shortage
      US Department of Health and Human Services
      Physician Supply and Demand: Projections to 2020.
      and may create access problems in many specialties and smaller communities. In addition, organizational costs for a hospital or practice group to replace a physician can be staggering, with some studies suggesting that replacing a single physician can cost $115,000 to $587,000, depending on specialty.
      • Schloss E.P.
      • Flanagan D.M.
      • Culler C.L.
      • Wright A.L.
      Some hidden costs of faculty turnover in clinical departments in one academic medical center.
      Practice turnover also leads to disruptions to patient care that may adversely affect quality metrics.
      Our study also illustrates some unique challenges for early career physicians. Even though early career physicians worked fewer hours and took less call duty, they were more likely to experience work-home conflicts. Because early career physicians overall are younger, their children may also be younger, which may increase the likelihood of work-home conflicts. Also, a greater proportion of early career physicians in our cohort were women, and previous data suggest that women experience more work-home conflicts than men.
      • Dyrbye L.N.
      • Shanafelt T.D.
      • Balch C.
      • Satele D.
      • Freischlag J.
      Relationship between work-home conflicts and burnout among American surgeons: a comparison by sex.
      When work-home conflicts occur, early career physicians were the least likely to report being able to resolve the conflict in a manner that allowed both home and work responsibilities to be met. Whether this is due to personal characteristics, a lack of experience, or greater practice-related restrictions (eg, more junior members in a practice often have less autonomy and flexibility than their more senior colleagues) cannot be determined from this study. The experience of work-home conflicts and how they are ultimately resolved when they occur have been reported to influence career decisions, career satisfaction, and burnout.
      • Dyrbye L.N.
      • Shanafelt T.D.
      • Balch C.
      • Satele D.
      • Freischlag J.
      Relationship between work-home conflicts and burnout among American surgeons: a comparison by sex.
      • Dyrbye L.N.
      • Freischlag J.
      • Kaups K.A.
      • et al.
      Work-home conflicts have a substantial impact on career decisions that affect the adequacy of the surgical workforce.
      Our findings on the challenges of middle career are consistent with the Community Tracking Study of physicians conducted in the 1990s
      • Leigh J.P.
      • Kravitz R.L.
      • Schembri M.
      • Samuels S.J.
      • Mobley S.
      Physician career satisfaction across specialties.
      and with studies conducted in other fields.
      • Clark A.
      • Oswald A.
      • Wan P.
      Is job satisfaction u-shaped in age?.
      Given that the nadir in satisfaction during middle career is not a new phenomenon, it appears unlikely that the findings of our study are due to generational effects alone. Middle and late career physicians who have experienced previous health care systems have to adapt to ongoing changes in the work environment. That phenomenon is likely to be as true in the past, because the health care system has been undergoing tremendous structural changes for decades (eg, health maintenance organizations in the 1980s and early 1990s), as in the future, when the changes during the next several decades may be even more profound as the nation reforms the health care system. Those in middle and late career will always be adapting to changes that are, to the earlier career physician, the only way of doing things they have experienced. The differences in work hours, satisfaction with work-life balance, and other factors by career stage likely also contribute to some of the observed differences. Other unmeasured aspects not unique to the practice of medicine, such as unrealized career expectations, new roles at home or work, and psychosocial development (ie, Erikson’s stage of middle adulthood—generativity vs stagnation
      • Erikson E.H.
      Identity and the Life Cycle.
      ) may also have a role, particularly at the individual level. However, the data presented represent the average experience of a large sample of physicians and are unlikely to be affected by personal factors that influence only a small subset of physicians. In contrast, greater challenge with a given issue (such as work-life balance) at a given career stage likely implies there are common factors at that stage (eg, age of children) that lead, on average, to more challenge for individuals at that point in their career. It has also been postulated that the recovery in career satisfaction among later career physicians is due to self-selection among older physicians and exit from practice of those who are least satisfied.
      • Leigh J.P.
      • Kravitz R.L.
      • Schembri M.
      • Samuels S.J.
      • Mobley S.
      Physician career satisfaction across specialties.
      This theory is supported by our finding that physicians not currently in practice or retired had a more than 2-fold greater odds of burnout on multivariate analysis after controlling for other personal and work-related factors. Our analysis also identifies some factors independent of career stage that relate to burnout and career satisfaction. For example, individuals in academic practice had lower burnout and greater career satisfaction than those in private practice. Work-home conflict (and how it is resolved) is also a powerful driver of both burnout and career satisfaction independent of career stage.
      This study has a number of limitations. First, the design is cross-sectional; thus, conclusions about causation and the potential direction of the effects seen cannot be made. Second, we relied on self-reported intent to reduce clinical hours or leave current practice rather than actual behavior. Such intentions, however, are necessary antecedents of staying or leaving behavior and are the greatest predictor of actual voluntary turnover.
      • Alexander J.A.
      • Lichtenstein R.
      • Joo O.H.H.
      • Ullman E.
      A causal model of voluntary turnover among nursing personnel in long-term psychiatric settings.
      • Bluedorn A.C.
      The theories of turnover: causes, effects, and meaning.
      Third, our response rate of 26.7% raises the possibility of response bias. Because substantial differences between responding and nonresponding physicians have not been found in previous studies,
      • Kellerman S.
      • Herold J.
      Physician response to surveys: a review of the literature.
      existence of such extreme bias is unlikely. The similarity between the demographic characteristics of early and late career responding physicians (a standard approach to evaluate for response bias) further suggests that responders were representative of US physicians. Fourth, the work-related factors we assessed are not comprehensive. For example, in this study we did not ask physicians to report their income or their primary method of compensation. Income is an important part of overall career satisfaction.
      • Leigh J.P.
      • Kravitz R.L.
      • Schembri M.
      • Samuels S.J.
      • Mobley S.
      Physician career satisfaction across specialties.
      • Landon B.E.
      • Reschovsky J.
      • Blumenthal D.
      Changes in career satisfaction among primary care and specialist physicians, 1997-2001.
      However, in our previous study of approximately 8000 US surgeons, the primary method of compensation (salaried, incentive-based pay, or mix) was not independently related to satisfaction with overall career choice (being a physician) or specialty choice (being a surgeon).
      • Shanafelt T.D.
      • Balch C.M.
      • Bechamps G.J.
      • et al.
      Burnout and career satisfaction among American surgeons.
      Nonetheless, physician income is linked to practice setting, specialty, and work hours, and adjustment for these variables likely accounts for a large proportion of physician income. Other unexplored factors in these domains may also affect career satisfaction at various career stages.
      Strengths of our study include obtaining a large sample of physicians from the Physician Masterfile, which is a complete registry of all US physicians (regardless of American Medical Association membership). Participating physicians were similar to US physicians. Furthermore, we used a validated instrument to assess burnout and well-established items to explore career satisfaction across multiple dimensions.

      Conclusion

      Physicians face different challenges at different stages of their career. Middle career is a particularly challenging career stage characterized by high work effort, a higher prevalence of burnout, and greater career dissatisfaction regardless of sex, practice setting, and specialty. Strategies to improve the work experience of middle career physicians are needed to maintain productivity, minimize turnover, and improve patient care.

      Supplemental Online Material

      References

        • Melville A.
        Job satisfaction in general practice: implications for prescribing.
        Soc Sci Med Med Psychol Med Sociol. 1980; 14A: 495-499
        • Linn L.S.
        • Yager J.
        • Cope D.
        • Leake B.
        Health status, job satisfaction, job stress, and life satisfaction among academic and clinical faculty.
        JAMA. 1985; 254: 2775-2782
        • Katz A.
        Better outcome means more job satisfaction: pilot project in Winnipeg and Halifax to enhance physician-patient communication.
        Can Fam Physician. 1999; 45: 218-220
        • Linn L.S.
        • Brook R.H.
        • Clark V.A.
        • Davies A.R.
        • Fink A.
        • Kosecoff J.
        Physician and patient satisfaction as factors related to the organization of internal medicine group practices.
        Med Care. 1985; 23: 1171-1178
        • Haas J.S.
        • Cook E.F.
        • Puopolo A.L.
        • Burstin H.R.
        • Cleary P.D.
        • Brennan T.A.
        Is professional satisfaction of general internists associated with patient satisfaction.
        J Gen Intern Med. 2000; 15: 122-128
        • DiMatteo M.R.
        • Sherbourne C.D.
        • Hays R.D.
        • et al.
        Physicians' characteristics influence patients' adherence to medical treatment: results from the Medical Outcomes Study.
        Health Psychol. 1993; 12: 93-102
        • Williams E.S.
        • Konrad T.R.
        • Scheckler W.E.
        • et al.
        Understanding physicians' intentions to withdraw from practice: the role of job satisfaction, job stress, mental and physical health.
        Health Care Manage Rev. 2001; 26: 7-19
        • Shanafelt T.D.
        • Balch C.M.
        • Bechamps G.J.
        • et al.
        Burnout and career satisfaction among American surgeons.
        Ann Surg. 2009; 250: 463-471
        • Dyrbye L.N.
        • Shanafelt T.D.
        Physician burnout: a potential threat to successful health care reform.
        JAMA. 2011; 305: 2009-2010
        • Mechanic D.
        Physician discontent: challenges and opportunities.
        JAMA. 2003; 290: 941-946
        • Leigh J.P.
        • Kravitz R.L.
        • Schembri M.
        • Samuels S.J.
        • Mobley S.
        Physician career satisfaction across specialties.
        Arch Intern Med. 2002; 162: 1577-1584
        • American Medical Association
        Physician Characteristics and Distribution in the U.S.
        American Medical Association, Chicago, IL2012
        • Shanafelt T.D.
        • Boone S.
        • Tan L.
        • et al.
        Burnout and satisfaction with work-life balance among U.S. physicians relative to the general U.S. population.
        Arch Intern Med. 2012; 172: 1377-1385
      1. American Association for Public Opinion Research. Standard definitions. Final disposition of case codes and outcome rates for surveys. 2011. http://www.aapor.org/AM/Template.cfm?Section=Standard_Definitions2&Template=/CM/ContentDisplay.cfm&ContentID=3156. Accessed May 4, 2012.

        • Maslach C.
        • Jackson S.E.
        • Leiter M.P.
        Maslach Burnout Inventory Manual.
        3rd ed. Consulting Psychologists Press, Palo Alto, CA1996
        • Rafferty J.P.
        • Lemkau J.P.
        • Purdy R.R.
        • Rudisill J.R.
        Validity of the Maslach Burnout Inventory for family practice physicians.
        J Clin Psychol. 1986; 42: 488-492
        • Lee R.T.
        • Ashforth B.E.
        A meta-analytic examination of the correlates of the three dimensions of job burnout.
        J Appl Psychol. 1996; 81: 123-133
        • Leiter M.
        • Durup J.
        The discriminant validity of burnout and depression: a confirmatory factor analytic study.
        Anxiety Stress Coping. 1994; 7: 357-373
        • Schaufeli W.
        • Bakker A.
        • Hoogduin K.
        • Schaap C.
        • Kladler A.
        On the clinical validity of the Maslach Burnout Inventory and the Burnout Measure.
        Psychol Health. 2001; 16: 565-582
        • Thomas N.K.
        Resident burnout.
        JAMA. 2004; 292: 2880-2889
        • Shanafelt T.D.
        • Bradley K.A.
        • Wipf J.E.
        • Back A.L.
        Burnout and self-reported patient care in an internal medicine residency program.
        Ann Intern Med. 2002; 136: 358-367
        • Rosen I.M.
        • Gimotty P.A.
        • Shea J.A.
        • Bellini L.M.
        Evolution of sleep quantity, sleep deprivation, mood disturbances, empathy, and burnout among interns.
        Acad Med. 2006; 81: 82-85
        • Dyrbye L.N.
        • Shanafelt T.D.
        • Balch C.
        • Satele D.
        • Freischlag J.
        Relationship between work-home conflicts and burnout among American surgeons: a comparison by sex.
        Arch Surg. 2011; 146: 211-217
        • Dyrbye L.N.
        • Freischlag J.
        • Kaups K.A.
        • et al.
        Work-home conflicts have a substantial impact on career decisions that affect the adequacy of the surgical workforce.
        Arch Surg. 2012; 147: 933-939
        • Kuerer H.M.
        • Eberlein T.J.
        • Pollock R.E.
        • et al.
        Career satisfaction, practice patterns and burnout among surgical oncologists: report on the quality of life of members of the Society of Surgical Oncology.
        Ann Surg Oncol. 2007; 14: 3043-3053
        • Frank E.
        • McMurray J.E.
        • Linzer M.
        • Elon L.
        Career satisfaction of US women physicians.
        Arch Intern Med. 1999; 159: 1417-1426
        • Lemkau J.
        • Rafferty J.
        • Gordon Jr., R.
        Burnout and career-choice regret among family practice physicians in early practice.
        Fam Pract Res J. 1994; 14: 213-222
        • Goitein L.
        • Shanafelt T.D.
        • Wipf J.E.
        • Slatore C.G.
        • Back A.L.
        The effects of work-hour limitations on resident well-being, patient care, and education in an internal medicine residency program.
        Arch Intern Med. 2005; 165: 2601-2606
      2. United States Office of Personnel Management. Career Patterns: A 21st Century Approach to Attracting Talent. http://www.opm.gov/hcaaf_resource_center/careerpatterns/CPGuideV1.pdf. Accessed October 1, 2012.

        • Greene W.H.
        Econometric Analysis.
        7th ed. MacMillan, New York, NY1993
        • US Department of Health and Human Services
        Physician Workforce: Projections and Research Into Current Issues Affecting Supply and Demand.
        US Dept of Health and Human Services, Rockville, MD2008
        • US Department of Health and Human Services
        Physician Supply and Demand: Projections to 2020.
        US Dept of Health and Human Services, Rockville, MD2006
        • Schloss E.P.
        • Flanagan D.M.
        • Culler C.L.
        • Wright A.L.
        Some hidden costs of faculty turnover in clinical departments in one academic medical center.
        Acad Med. 2009; 84: 32-36
        • Clark A.
        • Oswald A.
        • Wan P.
        Is job satisfaction u-shaped in age?.
        J Occup Organ Psychol. 1996; 69: 57-81
        • Erikson E.H.
        Identity and the Life Cycle.
        WW Norton & Co, New York, NY1980
        • Alexander J.A.
        • Lichtenstein R.
        • Joo O.H.H.
        • Ullman E.
        A causal model of voluntary turnover among nursing personnel in long-term psychiatric settings.
        Res Nurs Health. 1998; 21: 415-427
        • Bluedorn A.C.
        The theories of turnover: causes, effects, and meaning.
        Res Soc Organizations. 1982; 1: 75-128
        • Kellerman S.
        • Herold J.
        Physician response to surveys: a review of the literature.
        Am J Prev Med. 2001; 20: 61-71
        • Landon B.E.
        • Reschovsky J.
        • Blumenthal D.
        Changes in career satisfaction among primary care and specialist physicians, 1997-2001.
        JAMA. 2003; 289: 442-449