Written by: Tao Wang, Ph.D., Chris Carnahan
Download the full whitepaper here.
The United States has been facing a shortage of physicians, primarily because of the growth and aging of the population and the impending retirements of older physicians. In 2015, 43.2% of active U.S. physicians were age 55 or older [1,2]. In addition, the number of Medicare-funded residency training positions have remained unchanged for two decades, limiting the number of new physicians.  The Association of American Medical Colleges (AAMC) predicts that by 2025, the country will have a shortfall of between 61,700 and 94,700 physicians. The shortage range is 14,900 to 35,600 in primary care and between 37,400 and 60,300 in non–primary care specialties. 
In primary care specialties including Family Medicine, Internal Medicine and Pediatrics, many studies after the year 2000 have shown a severe undersupply of physicians. As for non-primary care specialties, Psychiatry remains an especially worrisome specialty shortage, particularly in child and adolescent Psychiatry. A recent survey suggests that psychiatry’s workforce needs to add 2,800 more physicians to its existing 45,580 and reveals that many areas in the country have no psychiatrists at all.
Psychiatry is not the only such non-primary care specialty. In its 2016 report, AAMC updated its supply and demand data and identified increasing shortages in the surgical specialties, not only General Surgery, Urology, and Ophthalmology but also Vascular Surgery and Neurosurgery. These shortages are more severe in rural areas. This maldistribution of physicians makes shortage particularly acute for some specialties. In addition to those mentioned above, Emergency Medicine, Hospitalist medicine, Endocrinology, Rheumatology, and Urgent Care are also facing shortages.
Furthermore, the United States has fewer physicians per 1,000 people than 23 of the 28 countries reporting data in 2013.  The United States has 2.56 doctors per 1,000 people, which is far behind countries like Austria (4.99), Norway (4.31), Sweden (4.12), and Germany (4.04).
Although today there is a broad consensus that physician demand will grow faster than supply, this has not always been the case. Early studies such as the GMENAC study pointed to exactly the opposite conclusion and resulted in public policy decisions that contributed to the current situation.
To determine the number of physicians the medical profession needed to recruit, hospitals, medical groups, and other healthcare organizations looked at physician-to-population ratios by specialty. When accurate, analysis of physician supply and demand can improve decision making to help align the nation’s physician workforce with its health needs. When inaccurate, however, the studies actually hamper effective decision making.
This white paper will review the research on this topic with particular attention to the methodologies employed, the population-to-physician ratios, and how these have been refined over time.
In 1976, in response to concerns about the rapidly growing supply of physicians, the Graduate Medical Education National Advisory Committee (GMENAC) was established by the Secretary of the Department of Health, Education and Welfare (now the Department of Health and Human Services) to advise the nation on how many physicians were needed to provide medical care in the United States. After analyzing the distribution among specialty physicians and residents and evaluating alternative approaches to ensure an appropriate balance, GMENAC concluded in 1980 that the nation faced a potentially serious surplus and recommended that it limit the number of medical school positions and severely restrict the number of international medical school graduates (IMGs) entering the United States.
The GMENAC study, which contains 107 recommendations and over 800 pages of analysis, was the first detailed, specialty-by-specialty study of physician supply and demand. The study committee consisted of individuals and physicians holding positions in clinical and academic medicine, nursing, law, hospital administration, economics, and insurance and consulted with many individuals and organizations.  The study sought to develop comprehensive data and to forecast physician supply in the United States in 1990.
This study and the subsequent studies led to the long-standing belief that there would be too many physicians, particularly specialists. This view was accepted by many organization such as the Bureau of Health Professions, the Council on Graduate Medical Education (GME) in the 1990s, and many medical organizations including the Association of American Medical Colleges. Not everyone accepted this conclusion, but the GMENAC study provided the dominant view and helped shape public policy.
Essentially, GMENAC’s methodology is “adjusted needs-based.” It used educated guesses to determine the numbers of physicians that would be required to treat the population in 1990. To project physician supply and demand, the GMENAC staff and technical panel on modeling developed three models: the supply model, the graduate medical education model, and the requirements model.
In the supply model, GMENAC considered the data on current numbers of predoctoral and postdoctoral training, foreign medical graduates, graduate medical education positions, and practitioner supply and attrition rates. GMENAC then estimated the future supply of physicians.
The graduate medical education model was derived from analysis of the training histories of 112,610 physicians who graduated from medical school between 1961 and 1975. By applying this model and considering year of training, GMENAC attempted to specify what the supply of GME positions would be at future points in time.
The requirements model was based on physician manpower. Several assumptions were made for this model based on an assessment of total burden of disease and disability in the target year 1990. GMENAC asked panels of experts in different fields of medicine to reach a consensus on norms of care for each health condition and medical procedure and estimate the needs. Then a modeling panel adjusted the needs by taking into consideration economic, social, and behavioral constraints. [Table 1: physician-to-population ratios]
Despite the inaccurate predictions it yielded, due to its comprehensive analysis, GMENAC continues to be one of the most frequently used ratios for physician supply and demand assessment. Today, however, some suggest that at most the specialty projections should be viewed as approximations or as ranges rather than as hard factors or accurate numbers.