Policy Matters: People Matters - Challenges in Workforce Analysis to Inform Health Policy

< Back

August 1, 2013 Volume 77, Number 8
Policy Matters: People Matters - Challenges in Workforce Analysis to Inform Health Policy Thomas R. Miller, Ph.D., M.B.A.

The number of physicians in the U.S. increased 195 percent between 1970 and 2010, from 334,000 to more than 985,000. The number of physicians per 100,000 persons increased almost 100 percent, from 161 to 319 during this same 40-year span.1 In addition to population, the demand for physician services in a particular geographic area is affected by the area’s demographic characteristics (e.g., age and gender), health status and socioeconomic characteristics (e.g., education levels, income and insurance coverage). Today, patient preferences also influence the amount and type of physician services provided, especially preferences among the information-savvy subset of the population that conducts research prior to visiting their physician.2 Additionally, changes in practice patterns resulting from technology advances can dramatically influence the demand for physician services. For example, despite a growing population and increasing burden of cardiovascular disease in the U.S., the number of practicing cardiothoracic surgeons fell in 2003 for the first time in 20 years, partly as a result of the increased use of cardiac stents and the corresponding decline in coronary artery bypass grafting operations between 1997 and 2004.3

Understanding the current and potential future demand for physician services is critical in workforce and medical education planning, establishing payment policy, and in considering licensing and scope-of-practice regulations. For anesthesiology, a 2012 article by Schubert and colleagues should be a “must read” for anesthesiologists.4 They provided an excellent assessment of the anesthesia workforce during the past decade and its outlook for the next seven years, concluding:

“Anesthesiologist supply constrained by small graduation growth combined with generation- and gender-based decrements in workforce contribution is unlikely to keep pace with the substantial population and public-policy generated growth in demand for service, even in the face of productivity improvements and innovation.”5

A workforce “out of balance” – either reflecting a shortage or a surplus – has potentially important policy implications. Shortages may encourage the use of workforce substitutes (technology or other types of manpower) or the rationalization of needed services. Surpluses may reflect market prices (physician compensation) in excess of market equilibrium and might encourage policymakers to seek payment reductions. In addition, surpluses may also create “supplier-induced” demand for unnecessary services. This is often the concern with ancillary services such as imaging or diagnostic testing; and there is evidence of supplier-induced demand in general practice.6 In addition to shortages and surpluses, policymakers are concerned about the apparent geographic misdistribution of providers, especially potential shortages in rural and underserved areas.7

Projecting the demand for physicians is part science and part art. It is replete with uncertainty and necessitates assumptions about future health care policy, technology, market dynamics and patient needs. For those interested in reading further on the topic, economist Uwe Reinhardt of Princeton provided an excellent description of the complexity and nuanced methodologies intrinsic in forecasting the supply of or demand for physicians and the related policy implications.8

Certainly it would seem that current workforce supply data are easy to obtain and consistent among sources. However, there are several complicating factors that impact supply estimates. Physicians may be part-time or full-time, and there is variability in work hours and levels of productivity. Some physicians are only involved in patient care activities, while others are employed as researchers, administrators, or educators and provide very limited patient care services.

There are several national sources of workforce data. The American Medical Association (AMA) maintains comprehensive supply data by specialty and practice type. As a result of the Administrative Simplification provisions of the Health Insurance Portability and Accountability Act of 1996 (HIPAA), the Centers for Medicare & Medicaid Services (CMS) provides the national provider identifier (NPI) dataset. Figure 1 illustrates the differences in the estimated number of anesthesiologists per 100,000 persons by state, based on these two national sources. The AMA data are based on physicians who self-designated their practice as “patient care”-related. Based on the AMA data, the number of anesthesiologists (patient care-related, excluding pain medicine) per 100,000 population ranged from 7.0 in Idaho to 16.9 in New Jersey. The NPI dataset does not distinguish the physicians by primary type of practice. In 36 states, the NPI counts exceeded the patient care-related AMA counts. Based on the AMA counts of all anesthesiologists, regardless of practice activity, the AMA data exceeded the NPI counts in 46 states (not shown). Overall, the NPI count of anesthesiologists exceeded the AMA count of anesthesiologists in patient care activity by 11.4 percent and was 23 percent below the AMA count of all anesthesiologists (excluding pain medicine). There are several potential causes of the differences in these data; some are known and some are not.

Number of Anesthesiologists per 100,000 persons by state, ranked high to low based on NPI data.

Figure 1: Number of anesthesiologists per 100,000 persons by state, ranked from high to low based on NPI data. Calculations by ASA’s Health Policy Research department based on data from the National Provider Identifier dataset (April 2013) and AMA data for physicians designating patient care activity (April 2013) rather than administration, research, etc.

ASA’s Health Policy Research department is obtaining available workforce data and will maintain an anesthesia-related workforce database. In addition, ASA has engaged the RAND Corporation to study the current state of the labor market for anesthesiologists and how it may have changed since RAND’s previous study five years ago.9 Since local and state manpower data may be more recent, accurate or detailed than publicly available data from national sources, ASA component societies are encouraged to contact the Director of Health Policy Research to share workforce information and help ASA develop more accurate and complete data on the anesthesia workforce.

Thomas R. Miller, Ph.D., M.B.A.is ASA’s Director of Health Policy Research.

 At the June Academy Health Annual Research Meeting in Baltimore, Lee Fleisher, M.D.* hosted a planning meeting of the new Perioperative Care and Pain Management Interest Group within AcademyHealth (www.academyhealth.org). At the ANESTHESIOLOGY™ 2013 annual meeting, there will be another meeting of this group and other interested researchers. The interest group is looking for investigators interested in surgical and anesthesia outcomes, including critical care and acute and chronic pain management from a diverse group of both clinical specialties and research disciplines. The goals of these meetings are to discuss the best ways to advance this area of research.

*Dr. Fleisher is the Robert D. Dripps Professor and Chair of Anesthesiology and Critical Care at the Perelman School of Medicine, and Senior Fellow in the Leonard Davis Institute of Health Economics at the University of Pennsylvania.


  1. Heisler EJ. Physician supply and the Affordable Care Act. Congressional Research Service 2013, 7-5700. R42029.
  2. Shih YT, Tai-Seale M. Physicians’ perception of demand-induced supply in the information age: A latent class model analysis. Health Economics 2012; 21:252-269.
  3. Grover A, Gorman K, Dall TM, Jonas R, Lytle B, Shem R, Wood D, Kron I. Shortage of cardiothoracic surgeons is likely by 2020. Circulation 2009; 120:488-494.
  4. Schubert A, Eckhout GV, Ngo AL, Tremper KK, Peterson MD. Status of the anesthesia workforce in 2011: Evolution during the last decade and future outlook. Anesth Analg 2012; 115:407-427.
  5. Ibid. Page 407.
  6. Van Dijk CE, Van Den Berg B, Verheil RA, Spreeuwenberg P, Groenewegen PP, De Bakker DH. Moral hazard and supplier-induced demand: Empirical evidence in general practice. Health Economics 2012; 22:340-352.
  7. Rabinowitz HK, Diamond JJ, Markham FW, Santana AJ. The relationship between matriculating medical students’ planned specialties and eventual rural practice outcomes. Academic Medicine 2012; 87:1086-1090.
  8. Reinhardt UE: Health manpower forecasting: the case of physician supply. In Health Services Research: Key to Health Policy. E Ginzberg, Editor. 1991. Harvard University Press: Cambridge, MA. p.234-379.
  9. The previous RAND report can be accessed at http://www.rand.org/pubs/technical_reports/TR688.html.