THE PRACTICE OF EMERGENCY MEDICINE/BRIEF RESEARCH REPORT
Attrition From Emergency Medicine Clinical Practice in the United States Adit A. Ginde, MD, MPH, Ashley F. Sullivan, MS, MPH, Carlos A. Camargo Jr., MD, DrPH From the Department of Emergency Medicine, University of Colorado Denver School of Medicine, Aurora, CO (Ginde); and the Department of Emergency Medicine, Massachusetts General Hospital, Harvard Medical School, Boston, MA (Sullivan, Camargo).
Study objective: We estimate the annual attrition from emergency medicine clinical practice. Methods: We performed a cross-sectional analysis of the American Medical Association’s 2008 Physician Masterfile, which includes data on all physicians who have ever obtained a medical license in at least 1 US state. We restricted the analysis to physicians who completed emergency medicine residency training or who obtained emergency medicine board certification. We defined attrition as not being active in emergency medicine clinical practice. Attrition was reported as cumulative and annualized rates, with stratification by years since training graduation. Death rates were estimated from life tables for the US population. Results: Of the 30,864 emergency medicine–trained or emergency medicine board-certified physicians, 26,826 (87%) remain active in emergency medicine clinical practice. Overall, type of attrition was 45% to non– emergency medicine clinical practice, 22% retired, 14% administration, and 10% research/teaching. Immediate attrition (⬍2 years since training graduation) was 6.5%. The cumulative attrition rates from 2 to 15 years postgraduation were stable (5% to 9%) and thereafter were progressively higher, with 18% having left emergency medicine clinical practice at 20 years postgraduation and 25% at 30 years postgraduation. Annualized attrition rates were highest for the first 5 years postgraduation and after 40 years postgraduation; between 5 and 40 years, the rates remained low (⬍1%). The overall annual attrition rate from emergency medicine clinical practice, including estimated death rate, was approximately 1.7%. Conclusion: Despite the high stress and demands of emergency medicine, overall attrition remains low and compares favorably with that of other medical specialties. These data have positive implications for the emergency physician workforce and are important for accurate estimation of and planning for emergency physician workforce needs. [Ann Emerg Med. 2010;56:166-171.] Please see page 167 for the Editor’s Capsule Summary of this article. 0196-0644/$-see front matter Copyright © 2009 by the American College of Emergency Physicians. doi:10.1016/j.annemergmed.2009.11.002
INTRODUCTION Background For decades, the physical and emotional demands of emergency medicine clinical practice have raised concerns about physicians’ ability to practice in the emergency department (ED) setting for an entire career.1-5 Shift work, unpredictable practice environment, high acuity, and sometimes difficult patients contribute to physical and mental stress of emergency physicians.6-12 The consequences of high physician burnout include low morale, deterioration in quality of care, job turnover, and ultimately attrition from clinical practice.13,14 Previous studies, which used limited samples, yielded conflicting results, with 1% to 12% annual attrition rates reported.1,2,6-10,15-17 Although the American Medical Association reported annual attrition rates of 2% to 3% for all physicians,18,19 accurate and comprehensive estimates for emergency physicians are lacking. 166 Annals of Emergency Medicine
Importance We recently reported that 69% of emergency physicians in 2008 were emergency medicine trained or emergency medicine board certified.20 Workforce estimation and planning rely on accurate understanding of the demand for services, current workforce, supply of new emergency medicine residency graduates, willingness of new graduates to work in underserved areas, and attrition from clinical practice.18,21 Accurate estimation of future emergency physician workforce supply will guide health policymakers in their efforts to advance access to high-quality emergency care. Goals of This Investigation We sought to estimate rates and type of attrition from emergency medicine clinical practice. Because of high stress and burnout among emergency physicians, our hypothesis was that Volume , . : August
Ginde, Sullivan & Camargo
Editor’s Capsule Summary
What is already known on this topic There are too few emergency physicians to meet current and near-future needs. The rate of emergency physicians ceasing clinical practice is an unknown but important factor in addressing this deficit. What question this study addressed How often do practicing emergency physicians leave practice and when? What this study adds to our knowledge According to the American Medical Association’s master file of self-reported data, 87% of emergency physicians remain in active clinical practice. Most who leave emergency department clinical practice move to a nonclinical practice (45%) or retire from medicine (22%), and attrition within 2 years of starting practice (6.5%). How this might change clinical practice This will not change practice but suggests that attrition is not the main reason for the shortage and that solutions must focus on developing new providers. attrition from clinical practice would be higher for emergency physicians than the 2% to 3% reported for all physicians.18,19
MATERIALS AND METHODS Study Design This study was a cross-sectional analysis of the 2008 American Medical Association Physician Masterfile, which contains data on all 940,000 US allopathic and osteopathic physicians and has been in use since 1906. Use of these data for analysis of the emergency physician workforce has been previously described.20 Briefly, physician data are entered in the Masterfile during initial medical license application and updated during subsequent renewals and by periodic survey. Membership in the American Medical Association is not required for inclusion. We purchased the commercially available deidentified database on November 7, 2008, from Medical Marketing Service, Inc. (Wood Dale, IL). This study was approved as an exempt protocol. Selection of Participants We included all emergency medicine–trained or emergency medicine board-certified physicians in our analysis. Emergency medicine training was defined as completion of emergency medicine residency, combined internal medicine– emergency medicine Volume , . : August
Emergency Medicine Attrition residency, or pediatric emergency medicine fellowship program. Years since training was based on the graduation year from the most recently completed Accreditation Council for Graduate Medical Education–accredited training program. Emergency medicine board certification was defined as successful completion of emergency medicine or pediatric emergency medicine examination administered by the American Board of Emergency Medicine or the American Board of Pediatrics. Since 1993, completion of an Accreditation Council for Graduate Medical Education–accredited emergency medicine residency program was required for eligibility for emergency medicine board certification by the American Board of Emergency Medicine. For non-emergency-medicine-trained physicians, we classified their training according to the most recently completed program (if more than one residency program was completed). Methods of Measurement Physicians are required to self-designate one primary specialty and could optionally designate a secondary specialty. Additionally, their primary type of practice is required of a complete application. Emergency medicine was defined as selfdesignated primary or secondary specialty of emergency medicine, internal medicine– emergency medicine, pediatric emergency medicine, pediatric emergency medicine– emergency medicine, or pediatrics– emergency medicine. We identified physicians as not clinically active by using the following primary type of practice: retired, administration, teaching, research, and not active for other reasons. Using these variables, we classified emergency physicians as active in emergency medicine clinical practice or no longer in emergency medicine clinical practice, with the latter representing attrition. Attrition was further classified as active in non– emergency medicine clinical practice, retired, administration, teaching/research, and other. Physicians who died were identified by report from state license boards and regular searches of the Social Security Death Index. Deceased physicians are removed from the Masterfile quarterly, and thus we were unable to include death in the primary attrition analysis. Primary Data Analysis We performed statistical analysis with Stata 10.1 (StataCorp, College Station, TX) and summarized data with descriptive statistics. We calculated cumulative attrition rates as proportion of emergency medicine–trained or emergency medicine boardcertified physicians who are no longer active in emergency medicine clinical practice, stratified by years since training graduation. We then estimated annualized attrition rates and summed these estimates, weighted by number of emergency medicine–trained or emergency medicine board-certified physicians per years since training graduation to obtain the combined estimate of attrition (not including death). This method uses time since training graduation in these crosssectional data to approximate a longitudinal cohort (eg, attrition Annals of Emergency Medicine 167
Emergency Medicine Attrition rates of physicians in earlier years were used to represent future patterns of more recent graduates). Because deceased physicians are removed from the Masterfile, we could not determine vital status of actual emergency physicians. However, we used the most recent version (2004) of age- and sex-stratified life tables for the US population, according to national vital statistics,22 to estimate annual death rates for emergency physicians. We applied the annual probability of death, stratified by age and sex, and summed estimates weighted by number of emergency medicine residency–trained or emergency medicine board-certified physicians active in emergency medicine clinical practice in a given category. This estimate was adjusted to 80% of its total value to account for the lower occupation-specific mortality rates among physicians, as previously reported23 and used by the Association of American Medical Colleges for workforce estimation.18 This calculated annual death rate was added to the previously calculated combined attrition rate (excluding death) to estimate the overall emergency physician attrition rate for direct comparison with estimates from other specialties, which included death in the attrition rate.18,19
RESULTS Of the 30,864 emergency medicine–trained or emergency medicine board-certified physicians, 26,826 (87%) remain active in emergency medicine clinical practice. The 4,038 emergency physicians who left emergency medicine clinical practice were older (median age 55 versus 44 years), were more likely to be men (81% versus 75%), were more likely an international medical graduate (14% versus 7%), and had longer duration since training graduation (median 23 versus 10 years) compared with those who remained. Additionally, those who left emergency medicine clinical practice were less likely to be emergency medicine trained (49% versus 80%) and less likely emergency medicine board-certified (79% versus 83%) compared with those who remained. Overall, the types of attrition were 1,820 (45%) to non-emergency-medicine clinical practice, 892 (22%) retired, 577 (14%) administration, 410 (10%) research/teaching, and 339 (8%) other. The most common specialties of those in non-emergency-medicine clinical practice were internal medicine (457 [25%]) and family/ general medicine (388 [21%]). Characteristics stratified by practice status and years since training graduation are displayed in the Table. Most attrition among emergency physicians within 30 years of training graduation was to non-emergency-medicine clinical practice, whereas those greater than 30 years since training graduation left emergency medicine clinical practice for retirement. The differences in emergency medicine training and emergency medicine board certification rates between those who left and remained in emergency medicine clinical practice were mostly attributable to those who were 2 to 30 years from training graduation. The cumulative attrition rates stratified by years since training graduation are displayed in Figure 1. Overall, 168 Annals of Emergency Medicine
Ginde, Sullivan & Camargo immediate attrition (⬍2 years since training graduation) was 6.5%. The cumulative attrition rates from 2 to 15 years postgraduation were stable (5% to 9%). Thereafter, cumulative attrition rates were progressively higher, with 18% having left emergency medicine clinical practice at 20 years postgraduation and 25% at 30 years postgraduation. More than 50% remained active in emergency medicine clinical practice until after 41 years postgraduation. We annualized these attrition rates by years since training graduation, as shown in Figure 2. The annualized attrition rates were highest for the first 5 years postgraduation and after 40 years postgraduation; between 5 and 40 years, the rates remained low (⬍1%). Applying these annualized attrition rates stratified by years since training graduation to actual numbers of emergency medicine–trained or emergency medicine boardcertified physicians, the annual attrition rate (excluding death) was 1.3%. According to occupation-adjusted life table analysis of emergency medicine–trained or emergency medicine boardcertified physicians active in emergency medicine clinical practice, the annual death rate for emergency physicians was an estimated 0.4%. Thus, the total estimated emergency physician attrition rate from emergency medicine clinical practice was approximately 1.7% per year.
LIMITATIONS This study has several potential limitations. Although the Masterfile is updated weekly, individual emergency physician data are typically updated at state medical license renewal, which occurs every 1 to 3 years. This creates a time lag between actual and recorded practice status and would likely cause a slight underestimation of attrition rates. Additionally, attrition was defined by primary type of practice—it is possible that physicians labeled as attrition work part time in emergency medicine clinical practice. Because the Masterfile is linked to only to the American Board of Medical Specialties, we could not include emergency medicine board certification by other organizations (eg, the approximately 1,700 diplomates of the American Osteopathic Board of Emergency Medicine). Although the American Medical Association seeks to achieve a high degree of data accuracy through standard data elements at medical license renewal supplemented by periodic surveys, we cannot be entirely sure that all data elements are consistently requested by every state licensing board at each licensure. Calculation of rates and inferences from temporal trends in the attrition prevalence by years since training graduation is limited by the cross-sectional analysis. For example, our approach assumed that the hazard function was the same for all strata of time since graduation; thus, if physicians who entered practice over time had different patterns of attrition, this study design would not detect it. Additionally, as the specialty matures and a higher proportion of total emergency physicians are older and in later practice years, the attrition rates, particularly from retirement and death, will likely increase. However, the present results reflect a current and complete sample of emergency physicians, and a large, representative Volume , . : August
Ginde, Sullivan & Camargo
Emergency Medicine Attrition
Table. Characteristics of emergency medicine–trained or emergency medicine board-certified physicians by years since training graduation and status of emergency medicine clinical practice or attrition. <2 Years
Characteristics Demographics Age, y (median [interquartile range]) Female sex International medical graduate Type of attrition Retired Non-emergency-medicine clinical practice Administration Teaching/research Other Training Residency Emergency medicine Family medicine Internal medicine Pediatrics General surgery Internship only Other residency None Board certification Emergency medicine Other specialty None
Emergency Medicine Practice, nⴝ2,715 33 (31–35) 996 (37) 160 (6)
— — — — —
2–30 Years Attrition, nⴝ153
Emergency Medicine Practice, nⴝ22,277
35 (33–40)
44 (38–52)
53 (35) 7 (5)
5,535 (25) 1,325 (6)
0 147 (96) 0 3 (2) 3 (2)
>30 Years Attrition, nⴝ2,789 52 (44–57) 629 (23) 323 (12)
Emergency Medicine Practice, nⴝ1,834 64 (61–67) 135 (7) 348 (19)
Attrition, nⴝ1,096 71 (64–79) 81 (7) 268 (24)
— —
230 (8) 1,491 (53)
— —
662 (60) 182 (17)
— — —
442 (16) 367 (13) 259 (9)
— — —
135 (12) 40 (4) 77 (7)
2,715 (100) 0 0 0 0 0 0 0
153 (100) 0 0 0 0 0 0 0
18,556 (83) 534 (2) 1,566 (7) 170 (⬍1) 625 (3) 462 (2) 113 (⬍1) 251 (1)
1,768 (63) 87 (3) 486 (17) 90 (3) 139 (5) 128 (5) 31 (1) 60 (2)
69 (4) 84 (5) 454 (25) 49 (3) 328 (18) 538 (29) 50 (3) 262 (14)
60 (5) 49 (4) 272 (25) 43 (4) 221 (20) 347 (32) 31 (3) 73 (7)
535 (20) 168 (6) 53 (35)
59 (39) 41 (27) 53 (35)
19,698 (89) 465 (2) 1,851 (8)
2,042 (73) 462 (17) 285 (10)
1,556 (99) 7 (⬍1) 7 (⬍1)
1,072 (98) 9 (⬍1) 15 (1)
Data are No. (%) except where indicated. –, Not applicable.
Figure 1. Cumulative attrition rates stratified by years since training graduation.
prospective cohort would be technically challenging and require decades to collect necessary data.
DISCUSSION In this national study of attrition from emergency medicine clinical practice, we found that the annual attrition Volume , . : August
Figure 2. Annualized attrition rates stratified by years since training graduation.
rate was lower (1.7%) than the 2% to 3% rate reported for other specialties.18,19 Indeed, most (75%) physicians who were emergency medicine trained or emergency medicine board-certified have remained in emergency medicine clinical practice for the duration of a 30-year career. Our reported attrition rate is similar to that of previous smaller studies of Annals of Emergency Medicine 169
Emergency Medicine Attrition emergency physician longevity1,10,15-17 but much lower than that of initial studies that raised serious concerns about durability of the specialty.2,6-9 These higher attrition estimates were largely based on data from emergency physicians who practiced before the availability or acceptance of emergency medicine residency training, which may explain the difference. Previous studies have reported high levels of occupational stress and burnout among emergency physicians;6-12 however, although this would appear to threaten longevity, emergency physicians still report high levels of career satisfaction.11,12,15 For example, the American Board of Emergency Medicine Longitudinal Study of Emergency Physicians reported that 65% of emergency physicians had high levels of career satisfaction, and 77% stated that emergency medicine met or exceeded their career expectations.11 These results compare favorably with those of other medical specialties, in which 30% to 40% would not choose the same career again.24 Factors that contribute to high levels of career satisfaction in emergency medicine include patient variety, proficient use of skills, job excitement, member of an effective team, high salary, and opportunities for teaching and leadership.9-11 Thus, despite a stressful environment, most emergency physicians are highly satisfied with their career choices and consequently have low rates of attrition. The timing of attrition also argues for the health of the specialty. Although a small (6.5%) early attrition is expected among recent emergency medicine residency graduates,1 the attrition rates then stabilize for the anticipated 30-year practice life of these physicians. The majority (96%) of early attrition is composed of non-emergency-medicine clinical practice (eg, urgent care, primary care, international medicine), similar to that of other specialties.18 When the 1.7% annual attrition rate is applied to the 26,826 emergency physicians active in emergency medicine clinical practice, we anticipate that approximately 450 emergency physicians will leave clinical practice the following year. However, there are approximately 1,350 new emergency medicine residency graduates each year to replace those who leave emergency medicine clinical practice.20,25 Although it will take decades before there is a completely emergency medicine–trained, emergency medicine board-certified workforce26—and the distribution of emergency physicians to rural areas27 and increasing overall demand28 will remain challenging—the favorable attrition figures are welcome news. In summary, most emergency medicine–trained or emergency medicine board-certified physicians remain active in emergency medicine clinical practice. Despite the high stress and demands of the specialty, overall attrition remains low and compares favorably with that of other medical specialties. These data have positive implications for the emergency physician workforce and are important for accurate estimation of and planning for emergency physician workforce needs. 170 Annals of Emergency Medicine
Ginde, Sullivan & Camargo
Supervising editor: Donald M. Yealy, MD Author contributions: AAG, AFS, and CAC conceived and designed the study. AAG provided statistical advice, acquired the data, and performed the analysis. All authors contributed to data interpretation, and AAG drafted the article. All authors contributed substantially to article revision and approved the final version. AAG takes responsibility for the paper as a whole. Funding and support: By Annals policy, all authors are required to disclose any and all commercial, financial, and other relationships in any way related to the subject of this article that might create any potential conflict of interest. The authors have stated that no such relationships exist. See the Manuscript Submission Agreement in this issue for examples of specific conflicts covered by this statement. Publication dates: Received for publication September 3, 2009. Revisions received September 29, 2009, and October 15, 2009. Accepted for publication November 4, 2009. Available online December 24, 2009. Presented at the Society for Academic Emergency Medicine annual meeting, May 2009, New Orleans, LA. Reprints not available from the authors. Address for correspondence: Adit A. Ginde, MD, MPH, Department of Emergency Medicine, University of Colorado Denver School of Medicine, 12401 E 17th Ave, B-215, Aurora, CO 80045; 720-848-6777, fax 720-848-7374; E-mail
[email protected]. REFERENCES 1. Anwar RA. A longitudinal study of residency-trained emergency physicians. Ann Emerg Med. 1983;12:20-24. 2. Leitzell JD, Riggs LM. Emergency medicine: two points of view. N Engl J Med. 1981;304:477-483. 3. Wasserberger J, Ordog JG. Is emergency medicine built to self destruct by 1992? Ann Emerg Med. 1986;15:603-604. 4. Keller KL, Koenig WJ. Sources of stress and satisfaction in emergency practice. J Emerg Med. 1989;7:293-299. 5. Dwyer BJ. Surviving the 10-year ache: emergency medicine burnout. Emerg Med Rep. 1991;23(suppl):S1-S8. 6. Murphy JG, Jacobson S. Satisfaction with practices: emergency physicians versus internists. Ann Emerg Med. 1987;16:277-283. 7. Keller KL, Koenig WJ. Management of stress and prevention of burnout in emergency physicians. Ann Emerg Med. 1989;18: 1157-1161. 8. Gallery ME, Whitley TW, Klonis LK, et al. A study of occupational stress and depression among emergency physicians. Ann Emerg Med. 1992;21:58-64. 9. Doan-Wiggins L, Zun L, Cooper MA, et al. Practice satisfaction, occupational stress, and attrition of emergency physicians. Acad Emerg Med. 1995;2:556-563. 10. Goldberg R, Ross RW, Chan L, et al. Burnout and its correlates in emergency physicians: four years’ experience with a wellness booth. Acad Emerg Med. 1996;3:1156-1164. 11. Cydulka RK, Korte R. Career satisfaction in emergency medicine: the ABEM Longitudinal Study of Emergency Physicians. Ann Emerg Med. 2008;51:714-722. 12. Kuhn G, Goldberg R, Comptom S. Tolerance for uncertainty, burnout, and satisfaction with the career of emergency medicine. Ann Emerg Med. 2009;54:106-113.
Volume , . : August
Ginde, Sullivan & Camargo
Emergency Medicine Attrition
13. Shanafelt T, Sloan J, Habermann T. The well-being of physicians. Am J Med. 2003;114:513-517. 14. Spickard A Jr, Gabbe SG, Christensen JF. Mid-career burnout in generalist and specialist physicians. JAMA. 2002;288:14471450. 15. Hall KN, Wakeman MA, Levy RC, et al. Factors associated with career longevity in residency-trained emergency physicians. Ann Emerg Med. 1992;21:291-297. 16. Hall KN, Wakeman MA. Residency-trained emergency physicians: their demographics, practice evolution, and attrition from emergency medicine. J Emerg Med. 1999;17:7-15. 17. Counselman FL, Marco CA, Patrick VC, et al. A study of the workforce in emergency medicine: 2007. Am J Emerg Med. 2009; 27:691-700. 18. Association of American Medical Colleges Center for Workforce Studies. The complexities of physician supply and demand: projections through 2025. Available at: http://www.aamc.org/ workforce. Accessed August 31, 2009. 19. Kletke PR, Marder WD, Silberger AB. The Demographics of Physicians Supply: Trends and Projections. Chicago, IL: American Medical Association; 1987. 20. Ginde AA, Sullivan AF, Camargo CA Jr. National study of the emergency physician workforce, 2008. Ann Emerg Med. 2009;54: 349-359.
21. Kirsch DG, Vernon DJ. Confronting the complexity of the physician workforce equation. JAMA. 2008;299:2680-2682. 22. Arias E. United States life tables, 2004. Natl Vital Stat Rep. 2007;56:1-39. 23. Johnson NJ, Sorlie PD, Backlund E. The impact of specific occupation on mortality in the US National Longitudinal Mortality Study. Demography. 1999;36:355-367. 24. Zuger A. Dissatisfaction with medical practice. N Engl J Med. 2004;350:69-75. 25. American Board of Emergency Medicine. Examination and diplomate statistics. American Board of Emergency Medicine Web site. Available at: http://www.abem.org/PUBLIC/portal/alias__Rainbow/ lang__en-US/tabID__3373/DesktopDefault.aspx. Accessed August 31, 2009. 26. Camargo CA Jr, Ginde AA, Singer AH, et al. Assessment of emergency physician workforce needs in the United States, 2005. Acad Emerg Med. 2008;15:1317-1320. 27. Handel DA, Hedges JR; SAEM IOM Task Force. Improving rural access to emergency physicians. Acad Emerg Med. 2007;14:562565. 28. Pitts SR, Niska RW, Xu J, et al. National Hospital Ambulatory Medical Care Survey: 2006 emergency department summary. Natl Health Stat Rep. 2008;7:1-38.
Did you know? Podcasts are available for almost every article in Annals.
Visit http://www.annemergmed.com/content/podcast to find out more.
Volume , . : August
Annals of Emergency Medicine 171