Charles, Anthony G. MD, MPH; Walker, Elizabeth G. MSPH; Poley, Stephanie T.; Sheldon, George F. MD; Ricketts, Thomas C. PhD; Meyer, Anthony A. MD, PhD
While the U.S. population has grown, the number of surgeons in the United States has not kept pace. The number of medical school graduates matriculating into residency training in the United States has changed very little since 1980, and, likewise, the number of residents annually completing their general surgery residency has held steady at around 1,000 residents per year.1 In 2009, according to the American Board of Surgery, only 909 chief residents became board-certified general surgeons (personal communication with American Board of Surgery, February 2010). Further, data from the National Resident Matching Program reveal no appreciable increase in the number of available categorical general surgery positions (remaining steady at between 1,057 and 1,077) between 2006 and 2010, even though 99% of the available positions are filled.2 General surgery is one of the few specialties that consistently attracts more applicants than there are positions available.
In 1994, Kwakwa and Jonasson3 estimated that the number of surgeons needed to ensure access to high-quality and timely care was 7.1 per 100,000 people. Later, these authors increased the estimate to 7.53 per 100,000 people for the year 2001.4 In that year (2001), according to these numbers, the United States—with 17,243 clinically active, nonfederally employed general surgeons, a ratio of 6.4 general surgeons per 100,000—was already facing a shortage of general surgeons.5 Since then, the U.S. population has grown without a proportionate increase in the number of surgeons. This declining supply of general surgeons per population over the past 10 to 20 years is partially attributable to the lack of growth in medical school enrollments parallel to population growth.6 In addition, the 1997 passage of the Balanced Budget Act (BBA) capped federal funding of residency training positions at 1996 levels, so few additional residency positions have been added since that time. Salsberg and colleagues7 examined the number of residents in training before and after the BBA and found a net increase of only 1,672 new entrants (7.6%) across all specialties from 2002 to 2007.
Further, researchers anticipate additional declines, likely exacerbated by continued growth of the elderly population, in the general-surgeon-to-population ratio.8 Because older individuals undergo surgical procedures more often, growth in this segment of the population creates a larger caseload for practicing general surgeons.9 Additional decreases in general surgery capacity occur because more surgeons trained in general surgery narrow their practice by either subspecializing or not covering general surgeries.
Rural areas are particularly vulnerable to insufficient numbers of surgeons. Those communities have significantly fewer general surgeons per 100,000 people. In fact, some estimate that in small rural areas there are currently only 4.67 surgeons per 100,000 people.10 A dramatic decline in the number of rural general surgeons in the United States since the early 1980s has further precipitated the crisis in rural general surgery.10 In rural areas, general surgeons are essential members of the local health care system, performing emergency operations, supporting the trauma care system, backing up primary care providers, and contributing to the financial viability of small hospitals.11,12 Surgical services are vital to the continued operation of small hospitals, and the loss of surgical care may have ripple effects for the health care community and economy.13
One logical way to increase the size of the general surgery workforce is to increase the number of physicians trained within existing general surgery residency programs; however, the capacity of existing programs to train additional surgeons is unknown. Thus, we examined the expansion capacity of Accreditation Council for Graduate Medical Education (ACGME)-accredited general surgery residencies.
In the fall of 2009, we conducted a Web-based survey of the directors and coordinators of the then 246 ACGME-accredited general surgery residency programs. We identified programs using publicly available data on the ACGME Web site. After an initial e-mail that included the questionnaire, inviting program directors and coordinators to participate, we sent up to two additional e-mails to encourage completion of the survey. Survey instructions encouraged program staff to complete those questions that did not require direct responses by the program director. We offered responders no incentives for participating in the survey, and we guaranteed each program director's anonymity by reporting the data in aggregate form. The 14-item questionnaire requested information about the characteristics of each program including the demographic profile of current surgery residents, institutional case volume, resident funding sources, subspecialization of graduating residents, and attrition. We defined attrition as “the transfer of a resident from a categorical general surgery position to a different specialty within graduate medical education.” On the basis of our assumption that surgeons in some fields are still likely to perform general surgery, we included the following surgical fields in our definition of “general surgery”: breast, colorectal, trauma/critical care, endocrine, hepatobiliary and pancreas, minimally invasive/laparoscopic, surgical oncology, and transplant. We defined “specialty surgery” as specialties not likely to include general surgery cases: cardiothoracic, hand, pediatric, plastic, vascular, and other.
The Web-based questionnaire also asked the program directors specifically (i.e., not staff) to share their opinions regarding the current state of the surgery workforce and the capacity for expansion within their residency program. The full survey is provided as Supplemental Digital Appendix 1, http://links.lww.com/ACADMED/A49.
We obtained some additional information, including the number of approved chief resident (fifth-year resident) positions, about the general surgery residency programs in the fall of 2009 from publicly available information on the ACGME Web site.
For the purposes of our analyses, we defined academic general surgery residency programs as those located at institutions with a medical school, and we defined community programs as those located at institutions without a medical school. We did not consider military programs to be either academic or community programs, and we analyzed them separately.
The survey used the University of North Carolina at Chapel Hill Sheps Center's Integrated Research System, a modular Web-based system for customized research surveys. We stored the data electronically on a password-protected, secure server and analyzed the data using Microsoft Excel (2007; Redmond, Washington) and Stata/SE (version 10.1 for Windows; College Station, Texas). The University of North Carolina at Chapel Hill biomedical institutional review board granted approval for the research.
Of the 246 programs that received the survey, 123 (50%) responded to the questionnaire, though some did not address all questions. Of 123 responding programs, 115 (93%) answered questions that required responses from the program director rather than staff, including questions about additional capacity. The response rate was slightly lower for academic programs than for community programs. Exactly half (50%; 75/150) of all academic general surgery residency programs responded to the survey, whereas 52% (45/86) of community general surgery residency programs responded. Three out of 10 (30%) military residency training programs responded.
The response rate for residency programs varied geographically (we defined regions using the U.S. Census Bureau definitions): The lowest response rate was from programs in the Northeast (43%; 33/76 programs), and the highest response rate was from those in the South (54%; 43/79 programs); the Midwest and West regions had response rates of 51% each (28/55 and 18/35, respectively). The one ACGME-approved general surgery residency program in a U.S. territory (Puerto Rico) also responded to the survey. We learned from the ACGME Web site that all (responding and nonresponding) general surgery residency programs had, on average, 4.6 chief resident positions and were associated with, on average, 3.6 other institutions. We detected no differences between responding and nonresponding programs.
Of the 115 general surgery residency program directors who responded to our survey, 110 (96%) reported that they believe the United States is currently facing a shortage of general surgeons, and 92 (80%) reported that their programs would currently have clinical and operative volume to accommodate an increase in resident complement if funding were not an issue (see Table 1).
Respondents of these 92 programs estimated capacity for an average of 1.9 additional residents per year. Academic programs reported slightly higher excess capacity of 2.0 residents per year compared with 1.7 per annum for community programs. Additional capacity varied by geographic region. Respondents from programs in the Northeast (including two outlier programs reporting capacity for 10 and 30 additional residents, respectively) reported greater expansion capacity than that reported by programs in the South, West, and Midwest. Programs reporting no excess capacity had more current chief resident positions than those reporting expansion capacity: 6.0 positions versus 4.3, respectively.
The rate of categorical residents withdrawing from responding programs over the five-year period (2005–2009) to enter other specialties was 15.3% (452 of 2,942 total categorical residents entering the general surgery programs). Of the 2,490 chief residents graduating from responding programs and declaring a fellowship choice over the five-year period, 1,773 (71%) reportedly entered fellowships immediately after graduation.
Among the responding programs, a higher percentage of graduates of the 55 larger programs (five or more chief resident positions, constituting 45% of the 123 responding programs) entered fellowships compared with those from smaller programs (four or fewer chief resident positions, 68 [55%] of the programs). At larger programs, an average of 75% of graduates reportedly entered fellowships compared with an average of 63% of graduates at smaller residency programs (P = .0027).
Of the 1,753 fellowship-bound residents for whom specific specialties were reported, 906 (52%) entered fellowships in fields we defined as general surgery, and the balance (847 [48%]) entered fellowships in fields we defined as specialty surgery.
There is some capacity to expand the number of residents trained in a majority of existing ACGME-accredited general surgery residency programs. In the absence of sufficient data to determine the true capacity for all programs, we used a crude calculation to estimate the total capacity under an expansion.
We added the average number of chief resident positions currently approved at those programs with expansion capacity (4.3) to their average reported expansion capacity (1.9 residents per year) and weighted this by the percent of responding programs reporting expansion capacity (80%): (4.3 + 1.9) × 0.8 = 4.96. Then, we used the percent of responding programs without expansion capacity (20%) to weight the average number of chief resident positions currently approved in programs without expansion capacity (6.0): 6 × 0.2 = 1.2. This provided an estimate of an average total of 6.16 (i.e., 4.96 + 1.2) chief residents per program per year under an expansion. Across 246 programs, this would result in a total of 1,515 general surgery residents graduating each year, beginning five years after implementing the expansion. This number represents an increase of 378 residents over the existing 1,137 approved chief resident positions in these 246 programs. On the basis of our survey results with regard to fellowships, we expect that approximately 110 (29%) of these 378 additional surgeons will go straight into general surgical practice and that 139 (52%) of the remaining residents will pursue general surgery fellowships. This would realistically result in 249 additional residents entering general surgery practice per year.
A key requirement for expansion of residency training programs is a sufficient patient case volume necessary to competently train additional general surgery residents. The Surgical Council on Resident Education (SCORE), a voluntary consortium of six organizations that focus on resident education in surgery, divides procedures and operations into three levels: (1) “essential and common,” (2) “essential and uncommon,” and (3) “complex.” Of the 114 program directors who answered the relevant question, 99 (87%) reported that the number of essential and common procedures available either meets or exceeds the needs of current residents (see Table 2). The 15 directors (13% of 114) reportedly struggling to meet resident training needs on at least some of the “essential and common” procedures most frequently reported that their programs have deficient numbers of procedures in the upper and lower endoscopy/bronchoscopy category. Of the 114 directors who answered, 82 (72%) reported that the number of “essential and uncommon” procedures available either meets or exceeds the needs of current residents.
Because sufficient case volume is essential to training additional residents, we explored whether programs reporting additional training capacity also reported excess procedures. Of the 92 programs reporting additional training capacity, 91 responded to the questions about availability of common and uncommon procedures. Of these 91 programs, 65 (71%) also said that the number of “essential and common” procedures exceeds training needs for their residents. However, just 25 (27%) of the programs reporting additional training capacity reported that the number of “essential and uncommon” procedures exceeds resident training needs. An additional 48% (44/91) reported that they have “just the right amount” of “essential and uncommon” procedures to meet resident training needs.
Many general surgery residency programs reported applying to the ACGME Residency Review Committee for increases in resident complement in recent years. Of the 123 programs responding to the survey, 68 (55%) applied to the ACGME at least once for an increase in resident complement between 2005 and 2009, and the ACGME granted 60% (52/86) of the requested additional residency positions over this time period.
Of the 108 programs in our survey providing funding source information, the percent of residents supported by Medicare Direct Graduate Medical Education funds ranged from no residents to all residents, and 81 programs (75%) reported that all of their residents were supported by these funds.
One logical solution to ameliorate the declining general surgery workforce is to expand existing ACGME-accredited general surgery programs. However, on the basis of our results, it is unclear whether such expansion would be sufficient to meet future surgical demand given population demographic trends, subspecialization, and attrition from general surgery training. Previous studies, supported by our own findings regarding progressive specialization, suggest that at least 70% of general surgery graduates pursue subspecialty training.14 Since 1984, the number of general surgery program graduates who have chosen to practice as general surgeons has decreased 25%.15 Although some surgeons who pursue subspecialty training still provide general surgery call, some researchers are concerned that a provider who spends the majority of his or her time in specialized elective care in vascular or thoracic diseases may lose skills in the management of emergency general surgery cases and thus produce poor outcomes and/or work unsafely.16 To further complicate general surgery workforce modeling, progressive specialization has now evolved to early specialization via integrated programs, particularly within the specialties of plastic surgery, vascular surgery, and cardiothoracic surgery.
Attrition of the surgical resident is particularly concerning. Attrition rates among general surgery residents range from 17% to 23%.17,18 The enormous changes in surgical technology, sicker patients who typically increase a surgeon's workload, declining reimbursement, government oversight, and other factors may further increase attrition rates.18 In addition, lifestyle has become an important factor in choosing careers in medicine for both men and women. Students now prefer specialties, unlike general surgery, that permit a controllable lifestyle.19 Finally, the objective data on which residency programs select residents have little predictive value for performance or retention. Given that fact, it is not surprising that some residents find themselves ill suited for a surgical residency.20 More careful resident selection and increased attention to mitigating the stresses (high workload, poor reimbursement, increased oversight, lopsided family-work balance) inherent in a surgical residency may lead to improved retention. Without such improvements, however, simply increasing the number of residency positions may have limited impact on the output of practicing general surgeons.
Another issue that complicates assessment of appropriate surgery training capacity is the recent downward trend in physician work hours. Between 1977 and 1997, physician work hours ranged from 54.6 to 55.9 hours per week, but steady declines have resulted in a workweek of an average of 51 hours in 2007.21 Whether reductions in work hours will result in the need for more surgeons to handle surgical demand is unclear, but this downward trend should be carefully considered for any plan to expand the surgery workforce.
A major impediment to any residency training program expansion is the paucity of available funding. One estimate for the cost of training the surgical workforce to meet the needs of the population is $10 billion over the next 20 years.22 Currently, the federal government supports most medical education through Medicare. Our own findings lend credence to this reality: Program directors reported that their programs rely heavily on Medicare funds. In light of limitations on Medicare-funded graduate medical education, some hospitals are moving toward privately financed residency training using funding from clinical practices, research grants, endowments, and professional associations.7,23 Any expansion of residency training programs will likely require pursuing these and other alternative funding sources.
Additionally, any program wishing to expand residency training must meet special criteria established by the ACGME, including a detailed operative case log that demonstrates its ability to provide sufficient operative experience for the existing and proposed new residents. Programs with additional training capacity must request permission from ACGME to expand their programs. ACGME can encourage expansion by streamlining its process for approving such changes.
We believe, based on the ratio of number of applicants to positions available, that the demand for categorical general surgery residency positions is currently robust and that there is currently some capacity to absorb the projected increases in the number of graduates from both current and new U.S. osteopathic and medical-degree-granting medical schools.24
Expanding existing residency programs is a logical and important strategy for improving the surgery workforce; however, this strategy alone may be inadequate to accommodate future demand for surgical services. Additional strategies are likely necessary. One such strategy is to establish new surgery residency programs in hospitals (such as those in the Council of Teaching Hospitals and Health Systems) that do not already support an established surgical residency program but that do have an existing training infrastructure. Another, more controversial way to meet the need for surgeons in the United States is to reduce the length of surgical training, focusing fundamentally on the “essential and common” procedures established by SCORE.
These survey results are subject to several limitations. We achieved only a 50% response rate, and nonresponders may not resemble responders in terms of expansion capacity and average program size, thereby potentially limiting the accuracy of our results. All data are self-reported and depend on the knowledge of the respondent. The assessment of available expansion capacity in our survey derived primarily from the availability of prerequisite clinical and operative volume as the fundamental basis for any expansion, and we did not explicitly address faculty capacity and other requirements. Our study also assumes that all finishing chief resident positions will be filled and result in a graduating surgeon.
Future studies should focus on the financial impediments to expansion of general surgery residency programs and the role of academic health centers, hospitals, and insurance companies in proffering solutions that will ease expansion capacity in the era of funding caps for graduate medical education.
Producing an appropriate surgical workforce to meet future needs in the United States will be challenging. Population changes, such as the expansion of the elderly population, may create increased demand for services, while changes in the retention of general surgeons may simultaneously reduce the available supply. This study highlights one potential strategy for increasing the supply of general surgeons: adding new residency positions within existing programs. Strategies, such as developing new training programs, cultivating new medical education funding streams, and changing the surgical training paradigm, may also be necessary to address current and future deficiencies in the general surgery workforce.
The authors wish to thank the participating general surgery residency program directors and staff for their time in completing the survey.
Financial support for this study was provided by the American College of Surgeons Health Policy Research Institute and the University of North Carolina-Chapel Hill Department of Surgery.
The University of North Carolina at Chapel Hill biomedical institutional review board granted approval for the research.
Portions of the survey results were presented at the American Association of Medical Colleges Physician Workforce Research Conference, May 6–7, 2010, in Alexandria, Virginia.