Assessing Manpower Needs in Gastroenterology and Digestive Endoscopy

Assessing Manpower Needs in Gastroenterology and Digestive Endoscopy: Lessons from the Past and Implications for the Future of Endoscopic Training

Mihir S. Wagh1 and Alan Barkun2

1 University of Colorado, Denver, CO, USA

2 McGill University, McGill University Health Centre, Montreal, QC, Canada

The practice of medicine has always greatly been influenced by evolving technologies that not only assist physicians in delivering optimal patient care but also often directly contribute to shaping subsequent diagnosis and therapy. This is no truer than for gastroenterology and, more particularly, gastrointestinal endoscopy in which there have been great and rapid evolutions from diagnostic into therapeutic endoscopic techniques amidst the emergence of competing technologies. Such converging dynamics have indeed brought our specialty to a critical crossroad. An understanding of both the contributing factors leading to this and of the processes involved in predicting future patterns of health care delivery are now required as a matter of survival for our specialty. This chapter attempts to focus on the latter.

Modeling manpower

Before one can properly assess manpower needs specifically as they relate to endoscopic training, one must turn to the basic tenets of economics—the law of supply and demand. Unfortunately, projections based exclusively on supply and demand can be woefully erroneous as will be pointed out in the upcoming paragraphs. In economic theory, the notion of “manpower” is seemingly synonymous with human capital. Human capital, when broken down to its core, denotes a stock of knowledge and skills that represents the ability to perform labor, and thus create economic value. The current concept of human capital was introduced by Jacob Mincer (1958), and applied by Gary Becker (1962) [1,2]. In its current view, human capital is thought of as a “physical means of production” (e.g., factories). Therefore, similar to making investments in a factory, individuals can make investments in human capital (e.g., training, education, and medical treatment). The effectiveness of the human capital is determined by the “rate of return,” which is often expressed as the amount of economic value produced in relation to the corresponding investments made in human capital.

An alternative approach to assessing manpower relates to its characterization in the context of labor economics. The fundamentals of labor economics can be traced back to the Roy model that is based on “‘self selection” [3]. A more practical interpretation of this theory is that rational beings make optimizing decisions about what markets to participate in. These decisions apply to all aspects of life, including employment, location, education, and marriage. Through his discourse on “Thoughts on the Distribution of Earnings,” Roy describes the following factors as affecting occupational choice: (1) The fundamental distribution of skills and abilities; (2) Correlations among these skills in the population; (3) Technologies for applying these skills; and (4) Consumer tastes that impact on the demand for different types of outputs.

Econometrics is a discipline that develops and applies quantitative or statistical methods to the study and elucidation of economic principles. So what do these economic theories have to do with manpower issues in endoscopic training? Roy was the first to posit that workers have skills in each occupation, but they can only use one skill or the other. The original Roy model is based on the assumption that skills are distributed in a logarithmic rather than normal fashion and that the pursuit of comparative advantage in a free market reduces earnings inequality compared to the earnings distribution that would result if workers were randomly assigned to sectors. Noted econometrician, Heckman points out in “The empirical content of the Roy model” through complex hierarchical and regression analysis that earning densities need to be factored into the “self selection” and geographical preferences for earnings [3]. So if we analyze manpower issues in endoscopic training from a very theoretical and economic‐based approach, earning potential has to be factored into the discussion and equation. The request for training more endoscopists will only be met if there exists a clear need for their services. The next several paragraphs thus address physician needs.

Physician workforce estimates

The ability to correctly predict future demands of professional services bears obvious and significant implications for both the prospective employer and the student contemplating a future career. Predictions are often just that—estimates based on a perspective, or a mathematically derived model. It is instructive to review the very recent past during which forecasting had suggested both a physician and a gastroenterologist surplus that would overwhelm the medical workplace. Such a review will allow us to factor in many nuances that can lead to a better determination of actual manpower needs in endoscopic training. In 1976, the Graduate Medical Education National Advisory Committee (GMENAC) undertook a study of physician supply. It predicted that by 1990 there would be a surplus of 70,000 active physicians (a 13% surplus) doubling to nearly 145,000 by 2000 [4]. In an extremely analytical and somewhat controversial essay in the mid‐1990s, Cooper suggested that “in terms of physicians alone, there is no evidence of major impending national surplus” [5]. He assessed physician supply and demand for the period extending to 2020 from three perspectives: physician utilization in group‐ and staff‐model health maintenance organizations (HMOs), physician distribution, and the future supply of nonphysician clinicians. Cooper postulated that physician supply would initially increase more rapidly than its demand, resulting in a surplus of 31,000 physicians in the year 2000, increasing to 62,000 physicians in 2010, after which the gap would narrow.

In response to the growing perception of “physician glut” that abounded in the mid‐1990s, two studies approached this question using different methodologies to achieve a more accurate means of calculating physician estimates, including the need for specialists. T.P. Weil used four methodologies in parallel to assess whether a predetermined population base had an inadequate, sufficient, or excess number of physicians. Using physician–population ratios per 100,000 persons, he supplied the number of physicians needed for 28 clinical disciplines, adopting a managed care perspective for most analyses. Although the data reported are outdated, his conclusions based on varying inputs and models remain pertinent today. A first table provides a comparison by clinical specialty of the number of physicians for a service area population of 100,000. The number of specialists needed is reportedly based on (1) the then available numbers in the United States; (2) the estimates of need reported in the GMENAC study; (3) the staffing of a large HMO with 2.4 million subscribers; (4) the staffing in the Minneapolis region with its high HMO penetration; and (5) the staffing in Wichita, Kansas, where there was low managed care penetration. Not surprisingly, the number of full‐time equivalents (FTEs) varied from as low as 1.3 to 19, further highlighting the inexact science for predicting physician requirements, including those of subspecialists.

He then proposed the following guidelines be used to determine physician needs: (1) Determine the composition, utilization patterns, and cost per case (adjusted by case mix intensity) of each member of the hospital’s existing medical staff, clinical department, and sections; (2) Individually interview all key members of the hospital’s medical staff and governing board officers; (3) Administer physician questionnaires to gauge needs; (4) Perform physician–population ratio calculations using various models (as shown above); and (5) A composite summary projecting that specialties should be open or closed to meet community, hospital, or subscriber needs. Using this methodology, he calculated the existing FTE equivalent of three gastroenterologists needed per 100,000 persons to be decreased by a factor 0.04 in a community hospital—not an intuitive means of quantifying the number of gastroenterologists needed.

An alternative approach has also been described to determine an adequate supply of physicians in the United States. Also performed in the late 1990s, Greenberg and Cultice appropriately placed the context of their work by claiming that “observers of the health care scene in the United States increasingly support the proposition that there are too many specialist physicians and not enough generalists and that, as a consequence, the nation is suffering in two respects: excessive health care costs for all and inadequate access for some” [6]. These investigators based their projections on the Health Resources and Services Administration’s Bureau of Health Professions (BHPr) model. The BHPr physician requirement model divides patient care in three domains: Population, physician specialty, and care setting. Secondary data were analyzed and put into matrices for use in the mainframe computer‐based model. This model produced a projected number of physicians, by specialty, required to handle the demand of health care in each of various projection years.

This model is unique in that it allows for changes in the demographic composition of the United States, as well as assumptions about the percentage of patients insured by the differing options‐fee‐for‐service, HMOs, and so on versus those who are uninsured, utilization rates, and physician productivity. Despite these permutations, the results of this methodology consistently confirmed that the number of physicians required would decline. Thus, the authors contended that the “Bureau’s demographic utilization model represents improvements over the data‐driven methodologies that rely on staffing ratios and similar supply‐determined bases for estimating requirements. The model’s distinct utility rests in offering national‐level physician specialty requirements forecasts.”

How have these models fared when tasked with estimating requirements for medical specialists? Anderson and colleagues compared three methods for estimating the requirements for otolaryngologists [7]. They identified three basic methods based on (1) the utilization and/or staffing patterns of managed care organizations, (2) current utilization patterns and anticipated changes in demographic and insurance coverage as well as estimates of physician productivity to predict demand for physician services, and (3) a modified Delphi technique that asks a group of physicians in their specialty to suggest the number of individuals with that disease who should see a physician in their specialty, and to estimate the time required to treat the patient. Anderson and colleagues computed national requirements for otolaryngologists in 1994 and 2010 by each of these three methods. The estimates were compared to the number of active otolaryngologists actually (or anticipated to be) providing patient care in that year. The estimated number of otolaryngologists required varied significantly depending on the method used. Wide variations occurred depending on the assumptions used across models. Furthermore, the models were very sensitive to minor alterations in assumptions with disparate resultant conclusions showing surpluses or shortages. They concluded the managed care model had the greatest potential to yield reliable estimates because it reflects actual staffing patterns in institutions that are attempting to use physicians efficiently. The following paragraphs will focus specifically on modeling in gastroenterology and, more specifically, the endoscopic workforce that needs to be trained.

Gastroenterology workforce modeling

The landmark study commissioned by the Gastroenterology Leadership Council (GLC) was the first analyzing the gastroenterology physician workforce [8]. This study compared US gastroenterologist‐to‐population ratios with those of other countries and of major HMOs. Importantly, investigators projected the growth of the supply of gastroenterologists under various scenarios.

Authors abstracted data that examined the current distribution and scope of practice in gastroenterology from two large national data sets: (1) the Health Care Financing Administration (HCFA) Medicare Part B file and (2) the Area Resource File (ARF).

Future supply of gastroenterologists was determined by first obtaining from BHPr the age and sex‐specific physician death rates and retirement/exit rates. They projected their model over a 50‐year period to illustrate the time it would take for some policy options to achieve a steady‐state workforce. Survey of the American Gastroenterology Association (AGA) training programs and the National Study of Internal Medicine Manpower (NaSIMM) was used to determine the training stream, number of positions offered, and matched. They chose to use the NaSIMM estimate of fellowship graduates (n = 490) in the final model. Mathematically, the investigators used the following formula:


A Markov model to determine future supply using a decision tree analysis yielded similar results as those obtained by the methodology described above. Finally, the investigators modeled future supply while adopting five different scenarios: current training levels, or assuming, respectively, 25%, 50%, two‐third reductions in these, or outright elimination of gastroenterology training.

ARF data pointed out that the number of practicing gastroenterologists increased by 500% between 1974 and 1994 while the size of the general population increased by 20%. The percentage of gastroenterologists representing the physician pool increased from 0.8% in 1985 to 1.25% in 1992. The investigators also pointed out that 59% of the approximately 7,500 practicing gastroenterologists were younger than 45 years of age, while only 3% were 65 or older. Taken together, these statistics suggested the existing pool of gastroenterologists in 1996 would remain in practice for the foreseeable future. The period from 1977 to 1994 saw an increase in the number of graduates from 398 to 490. During the same timeframe, the number of 3‐year programs increased from 7 to 45.

This study also elucidated the breakdown of consultative versus procedural services offered by gastroenterologists and nongastroenterologists. Meyer and colleagues found that 60% of the bills generated by gastroenterologists were for office‐based cognitive and consultative services, while only 22% were for procedural services—yet these produced 49% of the income generated by gastroenterologists. Their study also shed further light on the number of “GI” procedures performed by the nongastroenterologists. They found that 21% of the esophageal procedures (1992 Medicare data) were billed by generalists—general practitioners, general pediatricians, family practitioners, and general internists—and a similar percentage was performed by surgeons. Approximately 40% of the colonoscopies were performed by nongastroenterologists. Shockingly, 16% of ERCPs were performed by nongastroenterologists. For all procedures, gastroenterologists performed a higher proportion of therapeutic than diagnostic procedures. These authors also argued that by volume alone, the generalist saw more patients with GI disorders (based on ICD codes) and, hence, the supply for gastroenterologists should be further reduced. The authors then projected the future supply of gastroenterologists in the United States corrected for population growth until 2042. They report that if the current number of training positions remains unaltered, the number of gastroenterologists will double in approximately 28 years; if all fellowship training programs were eliminated, the number of gastroenterologists would decrease by 20% (Figure 34.1). According to the AGA position paper on training and education published in 1989, the suggested need was for 9,000–10,000 gastroenterologists [10]

Only gold members can continue reading. Log In or Register to continue

Stay updated, free articles. Join our Telegram channel

Jul 31, 2022 | Posted by in GASTOINESTINAL SURGERY | Comments Off on Assessing Manpower Needs in Gastroenterology and Digestive Endoscopy

Full access? Get Clinical Tree

Get Clinical Tree app for offline access