View Full Report (PDF - 34 pages)
In its Third Report, Improving Access to Health Care Through Physician Workforce Reform: Directions for the 21st Century, COGME made recommendations to address the problems of physician oversupply, increasing specialization, geographic maldistribution, and minority underrepresentation. Recognizing a progressive oversupply of specialists as well as a shortage of generalist physicians, COGME set forth goals for the number and type of physicians entering residency. It was recommended that the number of physicians entering residency be reduced from 140% to 110% of the number of graduates of allopathic and osteopathic medical schools in the United States in 1993 and that the percentage of those graduates who complete training and enter practice as generalists should be increased from the current level of 30% to 50% (ie, the so-called "110:50/50 recommendation"). The Fourth Report issued by COGME provided an update of the Third Report and recommended legislation to achieve these workforce goals though allocation of reduced numbers of residency positions to consortia of medical schools and teaching hospitals. The Sixth Report, Managed Health Care, documented fundamental changes occurring in the United States health care systems and outlined implications for medical education and the physician workforce based on a reassessment of physician supply and requirements. In response to a changing congressional environment, COGME's Seventh Report, Physician Workforce Funding Recommendations for Department of Health and Human Services' Programs, recommended that planned reductions in Medicare funding of graduate medical education (GME) be targeted specifically to reducing the number of first-year residents by reducing GME payments for international medical graduates (IMGs).
The implications of implementing the 110:50/50 recommendation of COGME's Third Report are great. First-year residency positions in the United States would be reduced from 25,000 to approximately 19,600 - a 22% decrease. The number of physicians entering specialties would drop 44%, while the number of generalist physicians would increase by one-third.
Since publication of the Third Report in 1992, the health care delivery system has been changing rapidly as a result of progressive implementation of managed systems of care and competitive medical practice. Consequently, in this report physician supply and requirements are reassessed in the context of a health care system increasingly dominated by a managed care. Methodologies and available analyses for forecasting patterns of physician utilization (ie, requirements) are compared under various assumptions, and the appropriateness of the 110:50/50 recommendations for GME are reassessed. In addition, COGME's position regarding the specialty compositional the workforce and the output of training programs is clarified.
In the 1960s and 1970s, in response to a physician shortage, the number of graduates of medical schools in the United States doubled and the United States government fostered immigration of physicians trained in foreign medical schools. When the shortage eased, the number was not adjusted subsequently. While the output of medical schools in the United States has remained stable for over a decade, the number of IMGs who entered residency training each year almost doubled between 1988 and 1994 - from 3,600 to 6,700. As a result, the number of first-year residency positions filled has increased to 140% of the number of United States medical graduates (USMGs), and the nation's physician-to-population ratio has increased rapidly.
Between 1965 and 1992, the patient care physician-to-population ratio (excluding resident physicians) increased by 65%, from 115 to 190 physicians per 100,000 population, almost entirely in the medical specialties. The specialist physician-to-population ratio increased by 121%, from 56 to 123 specialists per 100,000 population, while the generalist ratio increased only 13%, form 59 to 67 generalists per 100,000 population.
If numbers of those entering GME remain at current levels, the patient care physician-to-population ration will continue to increase until 2010 - an additional 15% over the 1992 level - from 190 to 219 physicians per 100,000 population. Assuming that 70% will continue to enter specialty practice, the specialist-to-population ratio will increase another 23%, from 123 to 152 specialists per 100,000, while the generalist physician-to-population ratio will remain stable at 67 per 100,000.
These projections may actually understate the future total patient care physician supply. Entry into GME in this country is a pathway to entering practice in the United States for both domestic and international medical school graduates. The increases in IMGs could continue. Several new schools of osteopathy are under development, adding to the medical supply. Furthermore, anecdotal reports suggest that the unprecedented demand by United States citizens for medical education is leading to increases in their enrollment in schools outside the United States that are not accredited by the Liaison Committee for Medical Education.
The estimates of future numbers of generalists may also be understated, inasmuch as these projections are based on past patterns of specialty choice. As perceptions of a surplus of specialists become prevalent, graduates of medical schools in the United States appear to be exhibiting increased interest in generalist specialties. According to information derived from the Association of American Medical Colleges Graduation Questionnaire, medical students' interest in generalist careers has increased from its nadir of 14.6% in 1992 to 27.6% in 1995. At the same time, students' interest in the medical subspecialties and hospital-based specialties appears to be declining.
It is unclear what long-term impact this increasing interest in generalist specialties will have on total numbers of physicians entering specialty medicine. In the past, patterns of reimbursement for GME have provided incentives for hospitals to maintain and expand their training programs. Residents provide valued service and are a source of revenue through GME reimbursement. These incentives may result in increasing numbers of IMGs being recruited to fill vacant subspecialty positions, in which case the surplus of specialists would not be moderated.
Congress now is contemplating reductions in the funding of GME. It is COGME's belief that across-the-board reductions may have little impact on the total number of residents entering GME. Targeted reductions that decrease funding in the subspecialties may result in more physicians entering generalist specialties but may not reduce total numbers entering GME. However, targeted reductions in GME reimbursement for IMGs - such as those recommended in COGME's Seventh Report - are most likely to reduce total numbers of physicians.
Five studies have been undertaken to determine projections for physician requirements for the next century. Four of these studies are demand-based methodologies, and one large-scale effort - the Graduate Medical Education National Advisory Council (GMENAC) Study - utilized a needs-based methodology to estimate requirements for practicing physicians. The GMENAC model projected physician need based on the prevalence of illness and estimates by provider panels of physician services required to manage these illnesses. Conversely, the demand based models establish their assumptions on the manner in which medical services are paid (eg, the percentage of capitated managed care versus fee-for-service) and current patterns of utilization. COGME places special emphasis on those demand models that assume increasing domination of the health care system by managed care arrangements. These systems use fewer patient care physicians per 100,000 population and a higher proportion of generalists than do the fee-for-service arrangements that have dominated health care delivery in this nation.
For all five models, requirements for patient care generalist physicians in year 2000 lie within a range of 60-80 generalists per 100,000 population. Specialist requirements in the five models varied from 82-138 specialists per 100,000 population. The differences in the five models lie in the degree to which historic increases in the demand for specialists are assumed to continue in the increasingly competitive managed care setting. Given the widespread consensus that the future health care system will be dominated by manage care (ie, capitated financing with strong utilization controls), COGME believes that ranges of patient care generalists between 60-80 per 100,000 population and specialists between 85-105 per 100,000 population are reasonable estimates of physician utilization in the early 21st century. COGME believes that market forces will at least balance increasing demand for specialty services resulting form new technology. Consequently, increasing demand for specialists is not anticipated.
If physician staffing patterns were at the midpoint of the requirement ranges (165 patient care physicians with 70 generalists and 95 specialists per 100,000 population) the workforce would consist of approximately 42% generalists and 58% specialists. In comparison, in the year 2000, COGME projects 203 patient care physicians with 63 generalist and 140 specialists per 100,000 population - a specialty mix of 31% generalists and 69% specialists.
Comparison of Alternative Supply Scenarios With Requirements
In its Third and Fourth Reports, COGME recommended that the nation reduce the number of physicians entering GME to the number of United States allopathic (MD) and osteopathic (DO) medical school graduates plus 10%, and that at least 50% of these residency graduates enter practice as generalists. This recommendation and various alternative scenarios of numbers of first-year residents and alternative specialty mixes were compared with the aforementioned patient care physician requirement ranges.
If recent patterns of residency education continue with first-year residency positions at a level of 140% of U.S. medical graduates and an output of 30% generalists and 70% specialists, imbalances between supply and requirements will worsen. Generalist supply will remain in the lower portion of the requirement range of 60-80 generalists per 100,000 population. At the same time, the patient care specialist supply will increase to 152 physicians per 100,000 population in 2010, far above COGME's estimated requirement range of 85-105 specialists per 100,000 population. The number of specialists will be 100,000 above the upper level of the range in the year 2000 and will grow to 130,000 above the range by the year 2010. At the some time, the supply of generalists will be 48,000 below the top of the requirement range in the year 2000 and will be approximately 39,000 below the upper end of the range in 2010.
* Assuming current physician output at 140% of United States Medical Graduates
* Assuming current physician output at 140% of United States Medical Graduates
*Assuming current specialty mix of 30% generalists and 70% specialists.
*Assuming current specialty mix of 30% generalists and 70% specialists.
Neither reducing the number of first-year residents nor increasing the generalist output to as high as 60% would alone bring both generalist and specialist supplies within the requirement ranges of a managed care-dominated system by the year 2010. If the number of first-year residency positions is not reduced and the proportion of generalist trainees grows to 50% with the increasing attractiveness of generalist careers and reduction in GME funding of subspecialty training, the number of generalists will exceed 80/100,000 before 2010 and the ratio of specialists will markedly exceed the requirement range. However, a reduction of first-year residents to the number of USMGs plus 10%, in combination with an increase in the proportion of generalists to at least 50% of those educated annually, will minimize the projected specialty surplus while maintaining generalist supply. Under this scenario, in 2010, the specialist physician-to-population ratio will be 134 per 100,000 - 87,000 specialist physicians above the requirement range. At the same time, the generalist physician-to-population ratio will be 77 per 100,000 - 8,000 generalist physicians below the upper level of the requirement range.
The ultimate requirements for generalists and specialists will obviously depend on the configuration of future health care systems. In a competitive environment, that system will be structured through an interaction of factors including cost efficiency mechanisms, consumer desires, and workforce availability. The Council anticipates that nurse practitioners and physician assistants will be utilized increasingly, both in specialty care and in primary care. It also believes that specialists will provide a portion of primary care services for chronically ill patients in managed care systems. At the same time, COGME anticipates that generalists will assume increased roles in coordinating care and providing primary care as managed systems of care proliferate, thus reducing demand for specialists.
It has been suggested that the current generalist supply will be adequate in a system dominated by managed care. Evidence to support this conclusion is drawn from data demonstrating that generalist-to-population ratios in the United States already approximate current generalist staffing levels in many health maintenance organization (HMOs). However, these conclusions, as well as the studies utilized in establishing COGME's estimations of requirements, are derived by projecting physician staffing patterns in local systems of managed care to the nation as a whole. They do not consider the inevitable geographic variation in physician supply. While variation in physician supply across states and regions may be reduced as managed care progressively dominates the health care delivery system, it is not realistic to expect that physicians in the future will be distributed evenly. The ranges of requirements are intended to be broad enough to take into consideration geographic and other local variations.
Current levels of generalist supply have been achieved through public support of generalist training. If generalist training is to be expanded by one-third, as would be the case with implementation of the 110:50/50 recommendation, the educational infrastructure must be maintained and enhanced. Training programs, particularly those serving rural and inner-city areas, should continue to receive training grant support at least for a decade or until managed care efforts have clearly replaced these needs.
In the final analysis, COGME recognizes that the nation's most significant workforce problem is an increasing surplus of physicians, primarily of specialists. In a setting of overall surplus, the issue of optimal requirement ranges becomes moot. The real issue becomes identifying where the system has the capacity to productively employ additional physicians. At present, this country has very limited capacity to absorb additional specialists while still being able to employ many additional generalists productively.
The health care system is in a state of dynamic change. Patterns of delivery may change with time, and current data are incomplete. Ongoing studies of workforce supply and requirements are needed.
Despite the aforementioned uncertainties, current data support a goal that total first-year residency positions be reduced to 110% of 1993 USMGs and that 50% of this reduced number enter practice as generalists. Implementing this recommendation will require fundamental changes in current patterns of GME which should be instituted as rapidly as possible. If this goal is achieved, the nation's physician workforce will more closely correspond to physician requirements early in the next century.