Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Prediction of well-being and insight into work-life integration among physicians using machine learning approach

  • Masahiro Nishi ,

    Roles Data curation, Formal analysis, Investigation, Writing – original draft, Writing – review & editing

    nishim@koto.kpu-m.ac.jp

    Affiliations Department of Cardiovascular Medicine, Graduate School of Medical Science, Kyoto Prefectural University of Medicine, Kyoto, Japan, Cardiovascular Branch, National Heart, Lung and Blood Institute, National Institutes of Health, Bethesda, Maryland, United States of America

    ⨯
  • Michiyo Yamano,

    Roles Investigation, Writing – review & editing

    Affiliation Department of Cardiovascular Medicine, Graduate School of Medical Science, Kyoto Prefectural University of Medicine, Kyoto, Japan

    ⨯
  • Satoaki Matoba

    Roles Conceptualization, Supervision, Writing – review & editing

    Affiliation Department of Cardiovascular Medicine, Graduate School of Medical Science, Kyoto Prefectural University of Medicine, Kyoto, Japan

    ⨯

Abstract

There has been increasing interest in examining physician well-being and its predictive factors. However, few studies have revealed the characteristics associated with physician well-being and work-life integration using a machine learning approach. To investigate predictive factors of well-being and obtain insights into work-life integration, the survey was conducted by letter mail in a sample of Japanese physicians. A total of 422 responses were collected from 846 physicians. The mean age was 47.9 years, males constituted 83.3% of the physicians, and 88.6% were considered to be well. The most accurate machine learning model showed a mean area under the curve of 0.72. The mean permutation importance of career satisfaction, work hours per week, existence of family support, gender, and existence of power harassment were 0.057, 0.022, 0.009, 0.01, and 0.006, respectively. Using a machine learning model, physician well-being could be predicted. It seems to be influenced by multiple factors, such as career satisfaction, work hours per week, family support, gender, and power harassment. Career satisfaction has the highest impact, while long work hours have a negative effect on well-being. These findings support the need for organizational interventions to promote physician well-being and improve the quality of medical care.

Citation: Nishi M, Yamano M, Matoba S (2021) Prediction of well-being and insight into work-life integration among physicians using machine learning approach. PLoS ONE 16(7): e0254795. https://doi.org/10.1371/journal.pone.0254795

Editor: Vincenzo De Luca, University of Toronto, CANADA

Received: December 18, 2020; Accepted: July 3, 2021; Published: July 15, 2021

Copyright: © 2021 Nishi et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All relevant data are within the manuscript and its Supporting Information file.

Funding: The authors received no specific funding for this work.

Competing interests: The authors have declared that no competing interests exist.

Introduction

Physician well-being extends beyond individual issues and has a significant impact on healthcare systems and patient care. Unwell physicians could contribute to low quality of care, lapses in professionalism, and jeopardizing patient safety [1–6]. Demographic characteristics such as gender, age, and relationship status are reported to be associated with physician well-being [7]. Physician well-being is not merely the absence of distress, but also comprises being challenged, thriving, and achieving success in various kinds of personal and social aspects [8]. It is a complex and multifactorial issue.

Problems with work-life integration and burnout are common among physicians [9–11] and physician assistants [12]. As current medical practice demands more time, physicians are more likely to be fatigued, stressed, or burned out [13–15]. Physicians specializing in emergency medicine and primary care have a higher risk of burnout [16–20]. Among Japanese physicians working in stroke care units, heavy workload, short sleep duration, little experience, and low mental quality of life were determined to be the risk factors for burnout [21]. A systematic review shows that burnout or stress in physicians is predominantly associated with workplace-related factors, such as work demands and poor work environment, rather than non-modifiable and non-workplace-related factors [22].

Limited research has been conducted to identify the factors determining well-being and work-life integration among physicians, despite the abundant reports about physician burnout. To our knowledge, there are no reports using a machine learning approach to study physician well-being. To bridge this gap in the literature, a survey was conducted among 846 physicians who belonged to a union of the Department of Second Internal Medicine at Kyoto Prefectural University of Medicine in Japan. In this study, we reveal that the machine learning approach is feasible for predicting physician well-being, and variables such as career satisfaction and work hours per week are important predictive factors for physician well-being.

Methods

A letter mail survey was conducted among 846 physicians, including cardiologists, nephrologists, pulmonologists, hematologists, gastroenterologists, and family doctors (S1 Fig) in a union of the Department of Second Internal Medicine at Kyoto Prefectural University of Medicine in Japan, which was established in 1921. The participants responded to the survey by mail. They volunteered to participate, informed consent was obtained by written form before filling out the survey, and all responses remained anonymous. The survey was conducted between March 1 and 31, 2019. The data analysis took place from November 1, 2019, to March 30, 2020. A total of 422 responses were collected, yielding a participation rate of 49.9%. The characteristics of total survey subjects and survey respondents were shown in S2 Fig. Of the respondents, 63 with no work hours were excluded because they were considered to have retired. The protocol for the present study was approved by the committee of the graduates’ association of Kyoto Prefectural University of Medicine. The entire protocol of the present study was designed in accordance with the Declaration of Helsinki.

Study measures

The primary independent variable in the present study was well-being, which was assessed on a scale of 1 to 10. Physicians were considered to be well when they reported a high score of 6 to 10 and unwell when they reported a low score of 1 to 5. Age, gender, relationship status, work, work style, work hours per week, career satisfaction, family support, sexual harassment, power harassment, equality at home, and equality at the workplace were the 12 variables considered (S1 Appendix). The variables were used as categorical values (well-being, gender, relationship status, work, work style, career satisfaction, family support, sexual harassment, power harassment, equality at home, and equality at the workplace) or numeric values (age and work hours per week).

Primary analysis

An unwell state of well-being was set as the target of prediction modeling. An ensemble model of machine learning, comprising elastic net (ENET), average (AVG), median (MED), and generalized linear model (GLM) blender, was used for the dual classification of well-being (S2 Appendix). To avoid overfitting, 10-fold cross validation and 10 seeds randomization were performed. The performance of each model was evaluated and compared using the validation data set (10% of sample). AUC was used as an indicator of the model accuracy. Permutation importance was calculated for variable impact on target, according to previous studies [23,24]. Mutual information was calculated for each variable. Low-impact variables were excluded at each round when the minimum permutation importance was lower than 0; consequently, a total of 5 rounds were conducted. Partial dependence was calculated to evaluate the effect of each variable on the target. Analysis and modeling were performed using Python 3.7.4, with DataRobot 2.21.3 being deployed.

Results

Sample characteristics

Data on participant characteristics associated with well-being and work-life integrity were available for 359 physicians (S3 Appendix), of whom 88.6% [318 of 359] were found to be well; their mean [SD] age was 47.9 [14.8] years; 83.3% [299 of 359] of them were male; 64.9% [233 of 359] were hospital workers and 91.9% [330 of 359] were full time workers. Their mean [SD] work hours per week were 45.5 [17.6]. A majority of them, 90.8% [326 of 359], were married; 37.6% [135 of 359] had family support; 48.7% [175 of 359] enjoyed equality at work; and 41.8% [150 of 359] enjoyed equality at home. However, 74.4% [267 of 359] of the physicians had suffered power harassment and 52.9% [190 of 359] had experienced sexual harassment. Despite this, the majority, 72.1% [259 of 359], enjoyed career satisfaction (Table 1). Work hours per week and career satisfaction were significantly different between unwell and well state (S4 Appendix).

thumbnail
Download:
Table 1. Participant characteristics associated with well-being and work–life integration.

https://doi.org/10.1371/journal.pone.0254795.t001

Machine-learning models and variable importance

The mean [SD] area under the curve (AUC) was 0.70 [0.02] in round 1, where all 12 variables were used for prediction modeling. In round 2, excluding work and work style, the mean [SD] AUC was 0.70 [0.02]. In round 3, excluding equality at work and relationship status, the mean [SD] AUC was 0.70 [0.02]. In round 4, excluding the existence of sexual harassment and equality at home, the mean [SD] AUC was 0.69 [0.02]. Excluding age, the model in round 5 with the remaining five covariates showed the highest accuracy: mean [SD] AUC 0.72 [0.02] (Table 2). Therefore, the remaining five variables were further examined. Several regression models were also evaluated to predict well-being index (1–10); however, those accuracy were not sufficient compared to binary classification models (S5 Appendix). The mean [SD] permutation importance of career satisfaction, work hours per week, family support, gender, and power harassment were 0.057 [0.012], 0.022 [0.009], 0.009 [0.003], 0.01 [0.005], and 0.006 [0.006], respectively. Among them, career satisfaction had the highest impact, with work hours per week also showing a high impact on well-being (Fig 1A). The network between factors associated with well-being was depicted according to the mutual information and the permutation importance of each variable using Cytoscape 3.6.1 (Fig 1B) [25]. The partial dependence of each variable is shown in Fig 2. The difference in the mean unwell rate was 16% between participants with or without career satisfaction. Long work hours per week were correlated with the unwell rate. The difference in the mean unwell rate was 3.5% between participants with or without family support. The difference in the mean unwell rate was 5.6% between males and females. Consistent with low permutation importance, the partial dependence of power harassment seemed indistinct due to its large error. To exclude the influence of work style as a possible cofounder, the data excluding part-time worker was also analyzed, then permutation importance for each variable was calculated (S3 Fig). The work hours per week showed similar importance on well-being to all data analysis.

thumbnail
Download:
Fig 1. Variable impact on physician well-being.

(A) Permutation importance on physician well-being was calculated for selected covariates: career satisfaction, work hours per week, family support, gender, and power harassment. Boxplots represents median, interquartile range, and extreme values. + indicates mean value. (B) Network between factors associated with well-being. The network between well-being and each variable is depicted in blue; the edge width and node size is determined according to the permutation importance of each variable. The network between variables is depicted in gray; the edge width is determined based on the mutual information between the two variables.

https://doi.org/10.1371/journal.pone.0254795.g001

thumbnail
Download:
Fig 2. Effect of variable on physician well-being.

Partial dependence on physician unwell state was shown for selected covariates: career satisfaction, work hours per week, family support, gender, and power harassment. Boxplots represents median, interquartile range, and extreme values. + indicates mean value. Unwell rate indicates predicted probability of unwell state.

https://doi.org/10.1371/journal.pone.0254795.g002

thumbnail
Download:
Table 2. Model accuracy of each round and random seed.

https://doi.org/10.1371/journal.pone.0254795.t002

Discussion

Among Japanese physicians in this study, well-being was predicted by ensemble models of machine learning, and the determinants of work-life integration were evaluated. Model accuracy using AUC was sufficient to predict well-being by selecting five important covariates: career satisfaction, work hours per week, family support, gender, and power harassment. In particular, career satisfaction and work hours per week were high-impact predictive factors for well-being.

This study provided insights into the association between physician well-being and work-life integration. Career satisfaction was the highest-impact predictive factor for well-being. Physicians with burnout are more likely to experience not only career dissatisfaction [7,14,26] but also career choice regret among resident physicians [27]. Long work hours per week had a significant impact on well-being. Those who frequently work long shifts suffer negative consequences in terms of healthcare performance [28,29]. In addition, they have an increased risk of percutaneous injuries and motor vehicle accidents [30,31]. Moreover, physicians without family support had a higher partial dependence on the unwell rate despite its low permutation importance. Various kinds of harassment prevail in medical training [32]. Sexual harassment was excluded as a covariate because the permutation impact was low. Power harassment is a form of workplace harassment in which the someone in a greater position uses the power to harass or bully a victim who is lower on the office hierarchy. Power harassment was a low-impact predictive factor for well-being, with unexpected results of partial dependence. Although this survey included an array of doctors with varying specializations, specialty was not evaluated. There are also different types of hospital workers from the business management to work shift staff. This study showed no impact of such work differences between health practitioners and hospital workers in terms of well-being.

Demographic characteristics such as age, gender, and relationship status were evaluated in association with physician well-being. This survey was conducted only among Japanese physicians, although race/ethnicity is an independent predictive factor of physician burnout [33–35]. Age and relationship status were excluded as covariates because the permutation importance was low. Young trainee physicians are at high-risk for burnout and stress compared to older physicians [36–38]. Gender was a predictive factor, but it had a relatively low impact on well-being. Male physicians showed a higher partial dependence on the unwell rate than did female physicians. Another report revealed that female residents have a higher risk of burnout than do male residents [27]. In addition, female physicians are more likely to experience work–home conflicts than their male counterparts [9]. Moreover, many female physicians are subject to various hurdles such as male-dominant structures of medical society, unconscious discrimination, or unfair evaluation by medical directors [39,40]. Thus, consistent with a previous report [22], demographic and non-work place factors such as gender, age, and relationship status had a relatively lower impact on physician well-being in this study.

This study has several limitations. First, this survey was conducted in a single center for Japanese physicians. Second, the limited number of participants might have caused the study to be underpowered. The observational findings will need to be evaluated in nationwide survey. Third, some important factors associated with well-being, such as specialty, race, and income of physicians, were not assessed in this survey. Fourth, the evaluation measurement for subjective indicators might be difficult to reproduce in future studies. Fifth, the participation rate of 49.9% was not high. Sixth, the previous emotional state or medical history of participants, such as depression or posttraumatic stress disorder, was not investigated. A comparative study before and after COVID19 pandemic will be interesting.

In conclusion, physician well-being was predicted using a machine learning model. Career satisfaction, work hours per week, family support, gender, and power harassment were predictive factors for well-being. Career satisfaction had a significant impact, and long work hours per week had a negative effect on well-being. Further studies are needed to better understand these results and ensure improvements in physician well-being, work–life integration, and medical care quality.

Supporting information

S1 Fig. Specialty of total survey subjects.

Distribution of each speciality is indicated by n (%).

https://doi.org/10.1371/journal.pone.0254795.s001

(PPTX)

S2 Fig. Characteristics of total survey subjects and survey respondents.

A, Gender distribution. Distribution for each gender is indicated by n (%). B, Age distribution. Age for each gender is indicated by mean [SD].

https://doi.org/10.1371/journal.pone.0254795.s002

(PPTX)

S3 Fig. Permutation importance of each variable for the data excluding part-time worker.

Ensemble models were used for the dual classification of well-being. Mean area under the curve (AUC) was used as an indicator of the model accuracy. Boxplots represent median, min,and max value.

https://doi.org/10.1371/journal.pone.0254795.s003

(PPTX)

S1 Appendix. Survey for physician well-being.

https://doi.org/10.1371/journal.pone.0254795.s004

(DOCX)

S2 Appendix. Algorithm of machine learning model.

https://doi.org/10.1371/journal.pone.0254795.s005

(DOCX)

S3 Appendix. All data set of the survey.

https://doi.org/10.1371/journal.pone.0254795.s006

(XLSX)

S4 Appendix. Participant characteristics for unwell and well state.

https://doi.org/10.1371/journal.pone.0254795.s007

(DOCX)

S5 Appendix. Accuracy of regression models for physician well-being.

https://doi.org/10.1371/journal.pone.0254795.s008

(DOCX)

Acknowledgments

The authors thank all the participants for completing the survey as well as Shinpei Ijichi and Isao Sugawara (DataRobot Japan) and Toshiya Mitsuhashi (NS Solutions) for machine learning modeling.

References

  1. 1. Wallace JE, Lemaire JB, Ghali WA. Physician wellness: a missing quality indicator. Lancet. 2009;374(9702):1714–21. Epub 2009/11/17. pmid:19914516.
  2. 2. West CP, Tan AD, Habermann TM, Sloan JA, Shanafelt TD. Association of resident fatigue and distress with perceived medical errors. JAMA. 2009;302(12):1294–300. Epub 2009/09/24. pmid:19773564.
  3. 3. Dyrbye LN, Massie FS Jr., Eacker A, Harper W, Power D, Durning SJ, et al. Relationship between burnout and professional conduct and attitudes among US medical students. JAMA. 2010;304(11):1173–80. Epub 2010/09/16. pmid:20841530.
  4. 4. Dewa CS, Loong D, Bonato S, Trojanowski L, Rea M. The relationship between resident burnout and safety-related and acceptability-related quality of healthcare: a systematic literature review. BMC Med Educ. 2017;17(1):195. Epub 2017/11/11. pmid:29121895; PubMed Central PMCID: PMC5680598.
  5. 5. De Stefano C, Philippon AL, Krastinova E, Hausfater P, Riou B, Adnet F, et al. Effect of emergency physician burnout on patient waiting times. Intern Emerg Med. 2018;13(3):421–8. Epub 2017/07/06. pmid:28677043.
  6. 6. Shanafelt TD, Balch CM, Bechamps G, Russell T, Dyrbye L, Satele D, et al. Burnout and medical errors among American surgeons. Ann Surg. 2010;251(6):995–1000. Epub 2009/11/26. pmid:19934755.
  7. 7. Shanafelt TD, West CP, Sinsky C, Trockel M, Tutty M, Satele DV, et al. Changes in Burnout and Satisfaction With Work-Life Integration in Physicians and the General US Working Population Between 2011 and 2017. Mayo Clin Proc. 2019;94(9):1681–94. Epub 2019/02/26. pmid:30803733.
  8. 8. Eckleberry-Hunt J, Kirkpatrick H, Taku K, Hunt R, Vasappa R. Relation Between Physicians’ Work Lives and Happiness. South Med J. 2016;109(4):207–12. Epub 2016/04/05. pmid:27043800.
  9. 9. Dyrbye LN, Shanafelt TD, Balch CM, Satele D, Sloan J, Freischlag J. Relationship between work-home conflicts and burnout among American surgeons: a comparison by sex. Arch Surg. 2011;146(2):211–7. Epub 2011/02/23. pmid:21339435.
  10. 10. Dyrbye LN, Sotile W, Boone S, West CP, Tan L, Satele D, et al. A survey of U.S. physicians and their partners regarding the impact of work-home conflict. J Gen Intern Med. 2014;29(1):155–61. Epub 2013/09/18. pmid:24043567; PubMed Central PMCID: PMC3889954.
  11. 11. Sinsky CA, Dyrbye LN, West CP, Satele D, Tutty M, Shanafelt TD. Professional Satisfaction and the Career Plans of US Physicians. Mayo Clin Proc. 2017;92(11):1625–35. Epub 2017/11/06. pmid:29101932.
  12. 12. Dyrbye LN, West CP, Halasy M, O’Laughlin DJ, Satele D, Shanafelt T. Burnout and satisfaction with work-life integration among PAs relative to other workers. JAAPA. 2020;33(5):35–44. Epub 2020/04/30. pmid:32345947.
  13. 13. Wright AA, Katz IT. Beyond Burnout—Redesigning Care to Restore Meaning and Sanity for Physicians. N Engl J Med. 2018;378(4):309–11. Epub 2018/01/25. pmid:29365301.
  14. 14. Dyrbye LN, Shanafelt TD. Physician burnout: a potential threat to successful health care reform. JAMA. 2011;305(19):2009–10. Epub 2011/05/19. pmid:21586718.
  15. 15. Thomas LR, Ripp JA, West CP. Charter on Physician Well-being. JAMA. 2018;319(15):1541–2. Epub 2018/03/30. pmid:29596592.
  16. 16. Shanafelt TD, Boone S, Tan L, Dyrbye LN, Sotile W, Satele D, et al. Burnout and satisfaction with work-life balance among US physicians relative to the general US population. Arch Intern Med. 2012;172(18):1377–85. Epub 2012/08/23. pmid:22911330.
  17. 17. Michel JB, Sangha DM, Erwin JP, 3rd. Burnout Among Cardiologists. Am J Cardiol. 2017;119(6):938–40. Epub 2017/01/10. pmid:28065490.
  18. 18. Panagioti M, Geraghty K, Johnson J. How to prevent burnout in cardiologists? A review of the current evidence, gaps, and future directions. Trends Cardiovasc Med. 2018;28(1):1–7. Epub 2017/07/19. pmid:28716506.
  19. 19. Lall M, Gaeta T, Chung A, Dehon E, Malcolm W, Ross A, et al. Assessment of Physician Well-being, Part One: Burnout and Other Negative States. Western Journal of Emergency Medicine. 2019;20(2):278–90. pmid:30881548
  20. 20. Kuhn G, Goldberg R, Compton S. Tolerance for uncertainty, burnout, and satisfaction with the career of emergency medicine. Ann Emerg Med. 2009;54(1):106–13 e6. Epub 2009/02/10. pmid:19201058.
  21. 21. Nishimura K, Nakamura F, Takegami M, Fukuhara S, Nakagawara J, Ogasawara K, et al. Cross-sectional survey of workload and burnout among Japanese physicians working in stroke care: the nationwide survey of acute stroke care capacity for proper designation of comprehensive stroke center in Japan (J-ASPECT) study. Circ Cardiovasc Qual Outcomes. 2014;7(3):414–22. Epub 2014/05/16. pmid:24823957.
  22. 22. Zhou AY, Panagioti M, Esmail A, Agius R, Van Tongeren M, Bower P. Factors Associated With Burnout and Stress in Trainee Physicians: A Systematic Review and Meta-analysis. JAMA Netw Open. 2020;3(8):e2013761. Epub 2020/08/19. pmid:32809031; PubMed Central PMCID: PMC7435345.
  23. 23. Fisher A, Rudin C, Dominici F. All Models are Wrong, but Many are Useful: Learning a Variable’s Importance by Studying an Entire Class of Prediction Models Simultaneously. Journal of Machine Learning Research. 2019;20(177):1–81.
  24. 24. Altmann A, Tolosi L, Sander O, Lengauer T. Permutation importance: a corrected feature importance measure. Bioinformatics. 2010;26(10):1340–7. Epub 2010/04/14. pmid:20385727.
  25. 25. Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 2003;13(11):2498–504. Epub 2003/11/05. pmid:14597658; PubMed Central PMCID: PMC403769.
  26. 26. Shanafelt TD, Balch CM, Bechamps GJ, Russell T, Dyrbye L, Satele D, et al. Burnout and career satisfaction among American surgeons. Ann Surg. 2009;250(3):463–71. Epub 2009/09/05. pmid:19730177.
  27. 27. Dyrbye LN, Burke SE, Hardeman RR, Herrin J, Wittlin NM, Yeazel M, et al. Association of Clinical Specialty With Symptoms of Burnout and Career Choice Regret Among US Resident Physicians. JAMA. 2018;320(11):1114–30. Epub 2018/11/14. pmid:30422299; PubMed Central PMCID: PMC6233627.
  28. 28. Landrigan CP, Rothschild JM, Cronin JW, Kaushal R, Burdick E, Katz JT, et al. Effect of reducing interns’ work hours on serious medical errors in intensive care units. N Engl J Med. 2004;351(18):1838–48. Epub 2004/10/29. pmid:15509817.
  29. 29. Umehara K, Ohya Y, Kawakami N, Tsutsumi A, Fujimura M. Association of work-related factors with psychosocial job stressors and psychosomatic symptoms among Japanese pediatricians. J Occup Health. 2007;49(6):467–81. Epub 2007/12/14. pmid:18075207.
  30. 30. Ayas NT, Barger LK, Cade BE, Hashimoto DM, Rosner B, Cronin JW, et al. Extended work duration and the risk of self-reported percutaneous injuries in interns. JAMA. 2006;296(9):1055–62. Epub 2006/09/07. pmid:16954484.
  31. 31. Barger LK, Cade BE, Ayas NT, Cronin JW, Rosner B, Speizer FE, et al. Extended work shifts and the risk of motor vehicle crashes among interns. N Engl J Med. 2005;352(2):125–34. Epub 2005/01/14. pmid:15647575.
  32. 32. Fnais N, Soobiah C, Chen MH, Lillie E, Perrier L, Tashkhandi M, et al. Harassment and discrimination in medical training: a systematic review and meta-analysis. Acad Med. 2014;89(5):817–27. Epub 2014/03/29. pmid:24667512.
  33. 33. Garcia LC, Shanafelt TD, West CP, Sinsky CA, Trockel MT, Nedelec L, et al. Burnout, Depression, Career Satisfaction, and Work-Life Integration by Physician Race/Ethnicity. JAMA Netw Open. 2020;3(8):e2012762. Epub 2020/08/09. pmid:32766802; PubMed Central PMCID: PMC7414389.
  34. 34. Afzal KI, Khan FM, Mulla Z, Akins R, Ledger E, Giordano FL. Primary language and cultural background as factors in resident burnout in medical specialties: a study in a bilingual US city. South Med J. 2010;103(7):607–15. Epub 2010/06/10. pmid:20531049.
  35. 35. Meredith LS, Schmidt Hackbarth N, Darling J, Rodriguez HP, Stockdale SE, Cordasco KM, et al. Emotional exhaustion in primary care during early implementation of the VA’s medical home transformation: Patient-aligned Care Team (PACT). Med Care. 2015;53(3):253–60. Epub 2015/02/13. pmid:25675403.
  36. 36. Peisah C, Latif E, Wilhelm K, Williams B. Secrets to psychological success: why older doctors might have lower psychological distress and burnout than younger doctors. Aging Ment Health. 2009;13(2):300–7. Epub 2009/04/07. pmid:19347697.
  37. 37. Stucky ER, Dresselhaus TR, Dollarhide A, Shively M, Maynard G, Jain S, et al. Intern to attending: assessing stress among physicians. Acad Med. 2009;84(2):251–7. Epub 2009/01/29. pmid:19174680.
  38. 38. Dyrbye LN, West CP, Satele D, Boone S, Tan L, Sloan J, et al. Burnout among U.S. medical students, residents, and early career physicians relative to the general U.S. population. Acad Med. 2014;89(3):443–51. Epub 2014/01/23. pmid:24448053.
  39. 39. Villanueva AM, Kaye D, Abdelhak SS, Morahan PS. Comparing selection criteria of residency directors and physicians’ employers. Acad Med. 1995;70(4):261–71. Epub 1995/04/01. pmid:7718057.
  40. 40. Dumelow C, Griffiths S. We all need a good wife to support us. J Manag Med. 1995;9(1):50–7. Epub 1994/12/09. pmid:10142779.