COGME’s 16th Report to Congress: Too Many Physicians Could Be Worse Than Wasted

• COGME
• education
• medical
• graduate
• health planning organizations
• physicians, family/workforce
• delivery of health care
• health services research
• health manpower

COGME’s latest and perhaps last report is a serious departure from its previous 15, which have been fairly consistent in calling for a brake on the production of physicians and a health care workforce focused more on primary care.¹ It is a rigorous and thoughtful analysis, but the recommendations suffer from the same problem Olympic competitor Matt Emmons encountered when his last shot in the 50-meter 3-position rifle final competition in the 2004 games was dead-on—unfortunately, he aimed at the wrong target. The COGME’s modeling offers a great deal of precision, testing many scenarios that might affect demand and need, and gives a brief nod to the potential impact of burgeoning nurse practitioner and physician’s assistant workforces. The COGME goes off-target, however, in basing its recommendations on how physicians currently work rather than incorporating the likely changes during the next 15 to 20 years that they modeled in their scenarios. COGME’s recommendations also fail to account for how the nurse practitioner and physician’s assistant workforces will affect health care in the future. The potential of this report for effecting an expansion of the physician workforce is evidenced by the Association of American Medical College’s use of the findings to change its own workforce policies before to its formal release.² The flaws in COGME’s recommendations and potential for diverting precious resources deserve attention before this momentum grows.

Slightly more than a decade ago COGME used the federal Physician Demand Model to project the physician-to-population ratio at 2020.³ This report projected from 240 to 298 physicians per 100,000 people by 2020, and that a little less than one third of practicing physicians would be generalists. These and other studies prepared for COGME also recommended a generalist physician-to-population ratio of about 80 per 100,000.⁴ In 2004, there were 265 physicians, including 92 generalists, for every 100,000 people in the US population (Table 1⇓), and if we count only those in direct patient care (nearly one third of physicians do not spend the majority of their time in direct patient care) for every 100,000 people, there are 212 physicians including 75 generalists. By either measure, the physician workforce in the United States either currently does or is poised to exceed previous goals set by COGME. This workforce has increased by 78% during the last 25 years, a rate nearly 3 times that of the general population. The range of projections for the population of the United States between 2005 and 2020 suggest that population growth will slow, growing only 9% to 15%, while there is no slowing in the production of physicians.⁵ Thus, we question whether a mismatch will exist between physician production and system needs or demands.

Where this report probably goes astray is in its targeting. First, the report projects the work of physicians in isolation. While acknowledging the need to account for nurse practitioners and physician’s assistants, currently numbering about 150,000 and positioned to double by 2020, they don’t actually do it. The absence of reliable data should not preclude assessing the impact of another clinician workforce of 150,000 to 300,000 clinicians, especially when COGME’s estimates of physicians vs demand range from 50,000 to 170,000. COGME has done previous reports on the importance of interdisciplinary education to quality and safety, but this report does not continue that aim.⁶ The contributions of nurse practitioners and physician’s assistants will be felt, not only in primary care, but across the physician workforce, particularly because there are already more physician’s assistants working outside primary care than in it. Nurse practitioners and physician’s assistants could radically alter the number of physicians needed to deliver the more routinely technical aspects of subspecialty care in the next 15 years, such as endoscopic screenings and intervention, much the way that certified nurse anesthetists have filled a niche working with anesthesiologists. With growing recognition of the need to develop better health care teams designed to meet the needs of patients, we cannot afford to model in isolation.^7,8

The second place this report is off target is in its failure to discuss what physicians will do in new models of care. Rather than “shooting” at the right number and mix of doctors, COGME had an opportunity in this report to decide the types of services we want to produce in the future and how we align the physician workforce to participate in delivering them. COGME seeks tighter methods for planning a workforce, but the report does not offer a vision of what this workforce will accomplish. Improving population health is one option, and in this regard there is consistent evidence that a primary-care–focused health system has beneficial effects for the population and individuals. This evidence led the Institute of Medicine to advocate for it nearly a decade ago.⁹ Since that time, the evidence for cost-effective, patient-centered primary care outcomes has continued to grow, and the United States is falling behind its peers around the world, at least in part because our system is oversubspecializing.^10–12 There is good evidence that subspecialists operating within the boundaries of their training do well, but evidence from our own shores is showing that simply having more subspecialists may be hazardous to our health.^11,12 Producing more doctors to do the same things will compete for resources, consume more of our gross domestic product, and likely have little impact on the well-documented disparities in health and health care in the United States. In fact, if this process outcompetes other factors known to reduce disparities in health, such as education and public health, its impact is likely to be harmful. Alluding to the opportunity costs of producing physicians, one work-force study advisor proclaimed, “too many physicians is worse than wasted.”¹³ The authors of the COGME study acknowledge that it may be desirable to reduce unnecessary services and even do sensitivity analyses to model potential reductions in services; however, the COGME does not embrace this idea in its recommendations.

COGME’s 16^th report to Congress is thorough and practical in its analysis and recommendations. As was Matt Emmons’ fateful, final event in the 2004 Olympics, however, it is technically correct but fundamentally not directed toward the right target. It fails to account for exploding nurse practitioner and physician’s assistant workforces combined with a physician workforce that is still growing much faster than the general US population, and it fails to accommodate what physicians actually will do in redesigned health care. If COGME’s goal was to suggest a minor correction in physician work-force production, suitable for past and present practice approaches, it is probably more right than wrong. But COGME seems to have missed an opportunity to take advantage of its denouement to develop projections based on a vision of a physician workforce that could deliver better health care and better health outcomes for everyone with greater efficiency. History has been humbling for previous physician workforce analyses, and such is likely to be true for this COGME report. COGME’s 16^th report should be seen more as another step in advancing the science of workforce estimation and not as a prescription for change or worry. Great care should be taken before implementing additional enlargement of the US physician workforce. An excess of physicians may produce no gains in population health even as the effort to produce them starves education and other policy options that are critical to health—an outcome not only off-target but “worse than wasted.”

• Received for publication February 22, 2005.
• Revision received April 11, 2005.
• Accepted for publication April 11, 2005.

• © 2005 Annals of Family Medicine, Inc.