US Graduate Medical Education Changing Dynamics Sarah E. Brotherton, PhD; Frank A. Simon, MD; Sylvia I. Etzel The National Graduate Medical Education (GME) Census, jointly administered by the American Medical Association and the Association of American Medical Colleges, completed its second year as an online survey of all GME programs accredited by the Accreditation Council for Graduate Medical Education as well as combined specialty programs approved by their corresponding medical specialty boards. Continuing trends include the growing number of subspecialty programs, which increased by 65 since last year to 3822, and a smaller number of specialty programs, which decreased by 25 to 4203. There were corresponding shifts in the number of residents training in them. The number of graduates from osteopathic medical schools in allopathic GME has increased from 3288 in 1996-1997 to 4658 in 2001-2002, an increase of 42%. Overall, the total number of residents in GME has declined slightly, to an estimated 96 410 this year. The number of residents in graduate year 1 (GY1) positions (entry positions in GME, for which prior training is not required) has also decreased, although the proportion of residents in GY1 positions who have not had prior GME has increased. The number of international medical graduates (IMGs) in GY1 positions has decreased from a high of 6727 in 1999-2000 to 5898 in 2001-2002, a decrease of 12.3%. However, the number who have entered GME immediately on graduation has grown from 310 in 1996-1997 to 936 in 2001-2002. Most of these IMGs were citizens or permanent residents of the United States and attended medical schools in the Caribbean. New race and ethnicity questions, which parallel those of the US Census, continue to reveal disparities in the racial/ethnic distribution of the training physician population relative to the US population. Despite continued interest in resident work hours, there was no decrease in reported hours worked between 2000-2001 and 2001-2002. Physicians training in surgical specialties are reported to work the most hours per week, while internal medicine and pediatrics allow for the most consecutive time on duty, at 30 hours or more. Specialties and subspecialties with limited on-call schedules, such as forensic psychiatry and pediatric pathology, have, on average, the lowest number of duty hours. JAMA. 2002;288:1073-1078 This year's report of the National Graduate Medical Education (GME) Census comes amid anecdotal stories of local physician shortages,1 a forecasted (and controversial) physician shortage predicted for the year 2020,2-4 concerns about immigration restrictions and how they might affect access to physicians,5 an ever-increasing body of literature substantiating differences in health care treatment and outcomes by racial and ethnic groups,6 and public interest in and proposed legislation to regulate resident physician working hours.7 We considered these issues in our analyses. METHODS This was the second year that the joint American Medical Association (AMA) and Association of American Medical Colleges (AAMC) National GME Census was administered through GME Track, an Internet-based AAMC product, continuing an annual event of more than 75 years and contributing to the AMA's Graduate Medical Education Database. We surveyed the active Accreditation Council for Graduate Medical Education (ACGME)–accredited and combined programs in academic year 2001-2002 about the residents active in the programs, as well as those who had graduated. Program directors were asked to reconfirm or modify the training status of those residents who were present in their programs the previous year (year in program, number of years in GME training in a different specialty, start date in the program, expected or actual graduation date), add any new residents to their program that were not already in the GME Database (approximately 19 000 new residents are added to the Database via the National Resident Matching Program and the AAMC's follow-up report of medical schools), and to confirm or add to demographic information including sex, birth date, country of origin, medical school, medical school graduation date, visa or citizenship status, race, and ethnicity. Program directors were also asked questions about their programs and were instructed to edit prior information about their programs. These questions address the education and work environment of the program, as well as employment policies and benefits. Much of the program survey responses are then included in FREIDA Online,8 which is a Web-based public information source on GME programs used by medical students and residents. Of the 8025 programs surveyed, there were 654 programs without any enrolled residents that did not approve their program survey (we use the term "approve" to signify that an electronic version of the survey was submitted or a paper version was returned). Of the remaining programs, 6559 (89.0%) approved their program survey and updated the status of at least 1 resident, who, according to the GME Database, appeared to be active in their program. Another 220 (3.0%) approved only the program survey, and 178 (2.4%) updated information on residents in the program, but did not approve the program survey. There were 414 programs (5.2%) that neither approved the program survey nor confirmed the status of any apparently active resident. Of the 6885 programs with apparently active residents, 634 programs (9.2%) did not update the status of any resident, 391 (5.7%) confirmed the status of some but not all residents (average percentage confirmed among this group was 87.1%), and 5860 programs (85.1%) confirmed the active status of all their enrolled residents, accounting for 93.6% of all apparently active residents. In total, 90.1% of all the residents in the GME Database had their status confirmed (eg, active, graduated, withdrawn) by their program director. Residents for whom we did not receive new information were advanced into the next year of training (6192 [6.4%]), or graduated (4075 [11.8%] of those who, by expected graduation dates, appeared to have graduated), based on the number of years of training and expected graduation dates. RESULTS The National GME Census counted 96 410 resident physicians as of December 31, 2001, enrolled in ACGME-accredited and combined specialty GME programs during academic year 2001-2002. In Table 1, we present 6 years of trend data on the number of residents in training by type of medical school. While most of the numbers are fairly stable, the number of residents who are graduates of osteopathic medical schools (DOs) continues to increase; there are now a total of 42% more DO physicians training in allopathic programs than there were 6 years ago, even though there has been a slight decrease (1.7%) in the number of residents overall. (See Appendix II for additional data on all residents. [http://jama.ama-assn.org/issues/v288n9/rfull/japp2-1.html]) The numbers of specialty and subspecialty programs continue to decrease and increase, respectively, as shown in Table 2. There are now 3.4% fewer specialty programs than there were in 1996-1997. Over this time period, 2 new specialty areas were added and 1 was discontinued, all among the combined specialties (eg, pediatrics/medical genetics). Compared with 6 years ago, there are now 2001 fewer residents training in specialty programs, a 2.3% decrease. Conversely, there are now 11.2% more subspecialty programs than there were in 1996-1997. There was a gain of 19 subspecialty areas for which there are ACGME-accredited GME programs available, with a loss of 1 subspecialty area during the same time period. The number of residents training in subspecialty programs has grown by 355, a 2.8% increase. Residents training in the 19 new subspecialty areas now comprise 5.4% of all subspecialty residents. Graduate Year 1 Residents Table 1 presents the total number of residents, the number of residents in graduate year 1 (GY1) positions (entry-level GME positions that do not require prior GME, but also do not preclude prior GME), and the number of residents in GY1 positions who have not had prior GME. The decrease in the number of GY1 residents without prior GME in 2000-2001 was confounded by respondent confusion over a change in questionnaire formatting for that survey year, which was subsequently rectified. Nonetheless, there has been a decrease of 9.7% in the total number of GY1 residents over the past 2 years, matched by a smaller decrease (4.8%) in the number of GY1 residents without prior GME. The decline in the number of GY1 residents since 1999-2000 is principally due to an 8.9% decrease in the number of graduates of US allopathic medical schools (USMDs) in GY1 positions, down from 17 213 to 15 684 in 2001-2002 (data not shown). Furthermore, the number of international medical graduates (IMGs) in GY1 positions has decreased from a high of 6727 in 1999-2000 to 5898 in 2001-2002, a decrease of 12.3% (data not shown). The number of GY1 residents with prior GME has been decreasing both numerically and proportionally. Two years ago the number of GY1 residents who already had some training in GME was 3178; this year the number is 1780, a decrease of nearly 44%. There was a parallel proportional increase in the number of GY1 residents who have not had prior GME, increasing from 87.5% of GY1 residents in 1999-2000 to 92.3% this year. We further analyzed the subset of residents who entered a GY1 position the same year they graduated from medical school. Table 3 presents data from the past 6 years of the total number of GY1 residents, the number of USMD and IMG residents who entered GME immediately on medical school graduation, and the number of IMGs who were citizens or permanent US residents at the time they entered GME. The number of USMD residents has been stable, while the number of IMGs entering GME immediately on graduation has tripled, from 310 to 936. The number of IMGs in this group who are citizens or permanent residents (72.3% and 7.3%, respectively, over the 6 years) has increased by 140% over the same time period. Table 4 presents the locations of the medical schools from which these IMGs graduated. A great majority (74.4%) graduated from medical schools in the Caribbean. The large increase in the number of graduates from other countries outside the Caribbean in 2001-2002 is made up primarily of graduates from medical schools in Mexico. During the years 1996-1997 through 2000-2001, an average of 8 IMGs directly entered GME immediately on graduation from a Mexican medical school. In 2001-2002, there were 58 IMGs who graduated in 2001 from Mexican medical schools; almost all of them completed a fifth pathway program. This is a reflection of more fifth pathway program positions being made available recently (Susan A. Kline, MD, Executive Vice Dean for Academic Affairs, New York Medical College, oral communication, June 7, 2002). A countervailing trend is a decrease in the number of GY1 IMG residents with J-1 or J-2 visas. International medical graduates with J visas made up 37.6% of GY1 residents with known citizenship status in 1997-1998, a proportion that has decreased steadily. This year, IMGs with J visas accounted for only 23.6% of GY1 residents with known citizenship status (data not shown). Race/Ethnicity This year the National GME Census asked for race and ethnicity following the dictates of the US Office of Management and Budget Statistical Policy Directive 15 and the US Census Bureau.9 The category "Asian/Pacific Islander" was divided into 2 categories: "Asian" and "Native Hawaiian/Pacific Islander." In our previous surveys, Hispanic ethnicity had been a category within the race question. This year the question was asked separately, and therefore respondents were asked to provide both Hispanic ethnicity and race for residents. Table 5 presents the race and Hispanic ethnicity of USMD residents without prior GME in GY1 positions, as well as the race and Hispanic ethnicity of the US population overall. It appears that respondents to our survey continued to use the Hispanic ethnicity option as an exclusive race category, as further analysis reveals that the majority of residents of Hispanic ethnicity (85.0%) were missing a race designation. We hope that the "other/unknown" category will be used less frequently as respondents become more familiar with the dual-part question. However, as Table 5 demonstrates, the racial and ethnic distribution of new entrants to the US physician workforce continues not to reflect the general population and is consistent with our results from previous years.10 (Tables 5 and 8 in Appendix II provide additional statistics on race and Hispanic ethnicity of all residents.) Residency Work Environment Continued scrutiny over resident working conditions, most notably in the form of legislation before Congress, led us to once again look at the numbers of scheduled hours of work reported by program directors. We compared what was reported for academic year 2001-2002 with the previous year and found virtually no differences. Overall, for 91.5% of programs that provided data both years, the reported average number of hours residents worked in their first year of training in a program (PY1) did not change. Similarly, there was no difference for 93.1% of programs in their reporting of the maximum consecutive hours residents were allowed to work, ie, the number of hours consecutively worked at which point a resident must go off duty. Excluding programs that reported a difference of more than 20 hours between the 2 years, we found that the average number of hours per week was reported as 55.9 in 2000-2001, compared with 55.8 this year. Maximum consecutive hours worked decreased from 23.2 in 2000-2001 to 23.1. Averages for specialties at the extreme ends of the distribution for these 2 measures are presented in Table 6. Surgical specialties have the most average work hours per week, while psychiatry and pathology subspecialties tend to have the fewest hours reported. The 5 surgical specialties in Table 6 have consistently reported the most hours over the past several years.10 There were 8 specialties with an average maximum consecutive hours of 30 or more; 4 of those specialties account for 991 GME programs (12.3% of the total) and 40% of all PY1 residents. COMMENT Over the past few years there has been a slight decline in the number of residents training in US GME overall. There has not, however, been a decline in the number of programs to train these residents; there are now 238 more programs overall than there were in 1996-1997, a 3.1% increase. This growth has been completely due to new subspecialty programs, as the number of specialty programs decreased by 148, while there are 386 more subspecialty programs than there were 6 years ago. The growth in subspecialty areas, and the concordant growth in programs, has been noted elsewhere.11 What is new with respect to the growth of subspecialization is that it continues to occur in the face of an apparent decline in the total number of residents. The number of students graduating from US allopathic medical schools has not changed appreciably during this time period.12 The decrease in the number of USMDs in GY1 positions over the last several years has resulted in a near match between the number of graduates and the number of USMD GY1 residents. With the recent decrease in the number of GY1 residents with prior GME, it appears that fewer residents are retraining or switching specialties than what was experienced in the past. The historically stable proportion of residents in GY1 positions who were apparently retraining or switching specialties (approximately 12% for several years) was one consideration in the Council on Graduate Medical Education's decision to revisit its recommendation that the number of GY1 positions be equivalent to 110% of the number of graduates of US allopathic and osteopathic schools.13 The widely anticipated physician surplus did not materialize in 2000.2 A new controversial economic model has been developed that forecasts an upcoming national physician shortage.2-4 The observed decrease in the number of residents overall and in GY1 residents may be a short-term anomaly, and we will continue to follow these trends. The increase in the number of US citizens and permanent residents seeking medical education abroad and returning to the US for GME (while the number of USMDs holds constant) lends support to those who have suggested that there should be an expansion of medical opportunities within the US.14 Furthermore, the declining number of residents training in the US with J visas may in turn reduce the number of physicians seeking J-1 visa waivers, exacerbating current shortages in medically underserved areas.15 These 2 trends also merit further monitoring. The unchanged racial and ethnic distribution of physicians in training continues not to reflect that of the US population. This should continue to concern those who propose, as one method to reduce health disparities between majority and minority populations, increasing access of patients to health care professionals of similar background, including race and ethnicity.6 Despite recent public scrutiny of the working conditions of residents in GME,7 we do not find evidence of reductions in the total number of weekly work hours or in the number of consecutive work hours. Recently, the ACGME,16 the AAMC,17 and the AMA,18 have developed guidelines and recommendations to address the issue of resident duty hours. While there are variations among the guidelines established by each organization, all impose an initial limit on duty hours at 80 hours per week with call no more frequent than every third night and at least one 24-hour period off every 7 days. Consecutive hours of call would be limited to 24 hours with the provision of additional time for follow-up and the transfer of care. This is specifically limited to an additional 6 hours by the ACGME and the AMA. These policies reflect concerns about fatigue and sleep deprivation and their effect on the quality of education, the delivery of health care, and patient safety, and also for the quality of life and safety of residents. These reductions in work hours must be achieved within a hospital environment struggling with a nursing shortage and tight financial margins. The design of residency educational programs and the provision of care in hospitals may have to be restructured to manage these competing demands. As with all survey data, limitations to our results include issues concerning self-reporting, and in the case of portions of the resident data, second-hand reporting. We rely on program directors or their designees to supply us with updated information about their programs; our methods for cleaning the data include contacting programs with outlying data and making corrections or removing obviously incorrect values. Program directors are asked to both provide new information on residents and to confirm information already present on residents, much of which comes to the GME Database indirectly from the residents themselves via data transfers from the National Resident Matching Program. As much of this information is to the benefit of the program (including an accurate listing of information on FREIDA Online as well as in the annual Graduate Medical Education Directory) and to the resident (the AMA provides GME training credentials verification services), we believe program directors respond to the GME Census in a thoughtful manner. Nonetheless, as with all self-reported data, there may be biases in reporting by program directors that we are not aware of and cannot control for. Strengths include the collaborative data resources of several organizations and the high response rate of program directors, which allows us to confirm and edit already present data, and to add to the database overall. By virtue of the longitudinal nature of the National GME Census, a major strength of our study is the ability to not only provide a snapshot of the present status of GME, but also to make comparisons to the past and to identify trends. Author/Article Information Author Affiliations: Division of Undergraduate and Graduate Medical Education Policy and Standards, American Medical Association, Chicago, Ill. Corresponding Author and Reprints: Sarah E. Brotherton, PhD, Division of Undergraduate and Graduate Medical Education Policy and Standards, American Medical Association, 515 N State St, Chicago, IL 60610 (e-mail: sarah_brotherton@ama-assn.org). Author Contributions: Study concept and design: Brotherton. Acquisition of data: Brotherton, Etzel. Analysis and interpretation of data: Brotherton, Simon. Drafting of the manuscript: Brotherton. Critical revision of the manuscript for important intellectual content: Simon, Etzel. Statistical expertise: Brotherton. Administrative, technical, or material support: Etzel. Study supervision: Brotherton, Simon. Acknowledgment: We wish to thank the Department of Data Collection at the American Medical Association and the Division of Health Care Affairs at the Association of American Medical Colleges for administering the National GME Census. REFERENCES 1. Booth W. Las Vegas trauma center closes as doctors quit. The Washington Post. July 4, 2002:A02. 2. Cooper RA, Getzen TE, McKee HJ, Laud P. Economic and demographic trends signal an impending physician shortage. Health Aff (Millwood). 2002;21:140-154. MEDLINE 3. Weiner JP. A shortage of physicians or a surplus of assumptions? Health Aff (Millwood). 2002;21:160-162. MEDLINE 4. Snyderman R, Sheldon GF, Bischoff TA. Gauging supply and demand: the challenging quest to predict the future physician workforce. Health Aff (Millwood). 2002;21:167-168. MEDLINE 5. Romney L. Visa cut threatens rural clinics. The Los Angeles Times. July 7, 2002. Available at: http://www.latimes.com/news/local/la-me-docs7jul07005049.story. Accessed July 14, 2002. 6. Smedley B, Stith AY, Nelson A. Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care. National Academy Press – book online. Available at: http://www.nap.edu/books/030908265X/html/. Accessed June 11, 2002. 7. Boodman SG. Waking up to the problem of fatigue among medical interns. Los Angeles Times. April 16, 2001:S1. 8. FREIDA Online. Available at: http://www.ama-assn.org/ama/pub/category/2997.html. Accessed July 25, 2002. 9. Greico EM, Cassidy RC. Overview of Race and Hispanic Origin. Washington, DC: US Census Bureau; March 2001. Census 2000 Brief. 10. Brotherton SE, Simon FA, Etzel SI. US graduate medical education, 2000-2001. JAMA. 2001;286:1056-1060. ABSTRACT | FULL TEXT | PDF | MEDLINE 11. Donini-Lenhoff FG, Hedrick HL. Growth of specialization in graduate medical education. JAMA. 2000:284;1284-1289. MEDLINE 12. AAMC Data Book: Statistical Information Related to Medical Schools and Teaching Hospitals. Washington, DC: Association of American Medical Colleges; 2002. 13. Council on Graduate Medical Education. COGME Physician Workforce Policies: Recent Developments and Remaining Challenges in Meeting National Goals: Fourteenth Report. Rockville, Md: US Dept of Health and Human Services; 1999. 14. Mullan F. The case for more US medical students. N Engl J Med. 2000;343:213-217. MEDLINE 15. Baer LD, Konrad TR, Slifkin RT. If Fewer International Medical Graduates Were Allowed in the US, Who Might Replace Them in Rural Areas? Chapel Hill: North Carolina Rural Health Research and Policy Analysis Center; 2001. Working Paper No. 71. 16. Accreditation Council for Graduate Medical Education. Proposed ACGME Common Program Requirements for Resident Duty Hours. Available at: http://www.acgme.org/new/ProgramDutyHours.pdf. Accessed July 15, 2002. 17. Association of American Medical Colleges. AAMC Policy Guidance on Graduate Medical Education. Available at: http://www.aamc.org/hlthcare/gmepolicy/gmepolicy.pdf. Accessed July 15, 2002. 18. American Medical Association. Council on Medical Education Report 9: Resident Physician Working Conditions. Available at: http://www.ama-assn.org/ama/pub/category/2959.html. Accessed July 15, 2002. |
I M G I, Inc. [ Home | About Us | Advertisers ]
|