The Clinical Endocrinology Workforce: Current Status and Future Projections of Supply and Demand (2024)

Search for other works by this author on:

Article history

Interest in physician workforce issues can be traced to the 1960s when concerns arose regarding both the adequacy of the overall number of physicians and whether they were engaged in primary care or specialty practice. The Graduate Medical Education National Advisory Committee report of 1976 was a landmark in the analysis of the United States' physician workforce. Its goal was to project the overall need for physicians, by specialty, in the year 1990. In a milestone 1986 report, the Council on Graduate Medical Education suggested that there might be too many specialists. These and similar analyses led to the 1993 Health Security Act, which instituted centralized planning and funding of residency positions based on the determination of whether there was a shortage or excess of physicians in a particular specialty.

In recent years, physician workforce analyses have indicated an impending shortage of physicians, particularly specialists (1–4). In 2003 the Council on Graduate Medical Education projected a shortage of approximately 85 000 physicians, mostly specialists, by 2020 unless there was a modest increase in US medical school capacity (5). Indeed, even with an increase in the number of medical school graduates, the American Association of Medical Colleges estimated a shortage of 63 000 physicians by 2015, of which 34 000 are estimated to be specialists (6).

This shift in physician workforce predictions from an excess of specialists to a deficit resulted from changes in two assumptions. First, the 1994 projections incorrectly assumed continued growth of staff model health maintenance organizations with the implication that this would restrain the use of specialists. Second, more recent models assume that continued economic growth will increase demand for specialty care. The premise is that advances in technology provide for an unlimited spectrum of services providing health benefits and use of these services is constrained only by our ability and willingness to pay. With respect to the specialty of endocrinology, diabetes, and metabolism, a 2003 study jointly published by a consortium of professional endocrine societies found that the US demand for endocrinologists exceeded the supply by 15% for the year 1999. The 3623 US endocrinologists treating adults in that year were about 500 too few to accommodate the patient demand; that gap was projected to widen to about 900 adult endocrinologists by the year 2020 when the workforce would be 4500 physicians (4).

The passage of the Patient Protection and Affordable Care Act (PPACA) has refocused attention on the physician workforce in the United States. This act increases health care coverage to millions of people in 2014 and is expected to increase it by substantially more as it is rolled out further during the next several years. Projecting how the PPACA will affect the demand for endocrinology services is part of this analysis, as is the effect of other factors influencing supply and demand that have substantively changed since the previous endocrinology workforce report (4).

Materials and Methods

The Endocrine Society commissioned The Lewin Group, which had previously provided the expertise and models for the 2003 study, to investigate the status of the current endocrine workforce and model the future supply of endocrinologists and the demand for their services using updated assumptions and current data. The Society empaneled a technical expert panel (TEP) to provide institutional and clinical information relevant to the endocrinology physician workforce as well as guidance for this study. The TEP consisted of senior endocrinologists and The Endocrine Society staff who are the authors of this report. The TEP provided their opinion and judgment of the current and future state of the health care market for endocrinology services and the context to help interpret practice pattern data. The TEP high-value perspective was integrated in the model along with objective sources of data. A White Paper describing the complete findings and including the 2012 Endocrinologist Survey questions is available on The Endocrine Society web site.

Data sources of supply and demand projections

The Lewin Group undertook an endocrinologist survey of the current workforce in 2012. This survey comprised 46 questions to assess endocrinologist demographics and clinical and nonclinical practice characteristics (Supplemental Material). It was sent to 1689 board-certified endocrinologists randomly sampled from an American Medical Association master file. A total of 355 survey responses were received, representing a 21% response rate. The survey results provided insight about practice patterns and real-time data that aligned with the data from other sources, including the physician databases that were used as sources for supply and demand calculations.

The model projecting the future supply of endocrinologists incorporated the current active supply adjusted for new entrants into the field and loss of endocrinologists due to the attrition rate. The active supply was derived using the American Medical Association's master file (7) including information on board-certified endocrinologists, which was enhanced using Provider 360. New Entrants were projected from data available from the American Board of Internal Medicine (ABIM), historical trends in the number of fellowship positions, and data from the American Board of Pediatrics (8). The attrition rate was calculated using the retirement rates from the Over 50 Survey from the American Association of Medical Colleges (AAMC) (9) and the mortality rates from the 2010 United States Census.

We projected the active supply of endocrinologists from 2011 through 2025. Supply projections of the endocrine workforce were based on an inventory model framework. The projection started with the number of board-certified endocrinologists in the base year (2011) and added new entrants (physicians completing fellowship training, becoming board certified, and entering the workforce) into the model each year. The attrition aspect of the model consisted of endocrinologists who left the workforce for reasons of emigration, change in professional activity, extended leave, retirement, or death. The attrition rate based on AAMC survey responses regarding intent to retire was not adequate for analysis, so we used data from the Endocrinologist Survey. The active supply in the next year (Y+1) was a function of supply in the current year (Y) plus new entrants minus attrition.

We use three concepts of an endocrinologist to determine the active supply. First, we consider only physicians who are board-certified endocrinologists as endocrinologists. This includes adult endocrinologists and pediatric endocrinologists but, for the purposes of this study, excludes reproductive endocrinologists. Successful completion of an endocrinology fellowship program is a prerequisite for board certification. An estimated 1% of endocrinologists are from an era (before 1972) preceding board certification. They were given grandfathered certification status and are included in this analysis, although many actually took the board examination and passed. Second, of those who are board certified, we consider the number who are clinically active, ie, those who are actively engaged in direct patient care, rather than those engaged in full-time research or in administration, or are inactive as indicated by the 2011 American Medical Association master file. This method adjusts new entrants into the profession to reflect the historical rate at which some of them will not be engaged in clinical care. Third, we consider the notion of a full-time equivalent (FTE) endocrinologist to determine the FTE supply. The FTE endocrinologist is empirically defined by the average number of patient visits that are provided in a year by the typical board-certified endocrinologist engaged in clinical care. This derives from data obtained from the Endocrinologist Survey. By defining an FTE endocrinologist by clinical productivity, an individual endocrinologist may provide more, or less, than one FTE, depending on hourly productivity and the number of hours worked. Average productivity data were compiled for endocrinologists in the work force based on age and gender.

The Demand model is the population or more precisely the population ratio, augmented by estimates of the effect of health care reform and other providers derived from the econometric model. The Demand projections for the model used the following: 1) population projection from US Census projections and The Lewin Group benefit simulation model for impact due to PPACA; 2) Part B Medicare Annual Data file; 3) use rate from the Medical Expenditure Panel Survey (MEPS) 2008–2009; 4) commercial claims data from OptumInsight's deidentified Normative Health Information system; and 5) annual visits to endocrinologist from the Lewin Endocrinologist Survey and the Medical Group Management Association. Using these sources, we calculated use rates for services provided by endocrinologists by patient age, gender, insurance status, and region. We projected the demand for services provided by endocrinologists by multiplying the population projections with the age, gender, and insurance-specific use rate for services provided by endocrinologists. We use FTE endocrinologist to measure demand, in that the number of FTE endocrinologists demanded is a measure of the services that are demanded.

The number of FTE endocrinologists needed to meet the demand is calculated by dividing the projected demand for services provided by FTE endocrinologists with the average annual visits provided.

In these equations, demand for visits in period T+1 equals visits per capita multiplied by the population in T+1. Note that visits per capita varies by age and gender categories and that the actual calculation accounts for shifts in the age/sex distribution over time. These are then converted to FTE endocrinologists by dividing by the number of visits attributed to an FTE endocrinologist.

Results

Current workforce

As of 2011, there were 5496 board-certified adult endocrinologists in the United States. This represents a net increase of 1873 (or ∼52%) compared with 1999 (4). There were 1016 pediatric endocrinologists as of 2011. There has been no previous determination of the pediatric endocrinology workforce. First-year adult endocrinology fellows increased from about 160 in 1999 to 304 by 2013 (10). Thus, taking into account attrition, the 1999 estimate for adult endocrinologist and the 2011 estimate appear consistent. Although the original study baseline prediction of the supply of adult endocrinologists in 2011 was approximately 4600, this assumed that fellowship positions would remain constant at the 1999 level. Using the actual number of adult fellowship positions, the original study accurately forecasted that for 2013 there would be approximately between 5350 and 5500 adult endocrinologists compared with the actual number of 5496.

Approximately 4841 adult endocrinologists and 893 pediatric endocrinologists were engaged in clinical practice in 2011; there were 15.5 board-certified adult endocrinologists in clinical practice per 1 million persons in in the United States. Restricting the patient population to those older than 18 years, there were approximately 21 clinically active board-certified adult endocrinologists per 1 million population older than 18 years, whereas there were 11 clinically active board-certified pediatric endocrinologists per 1 million of the US population aged 18 years or younger. These rates fall in the middle of the distribution of other internal medicine subspecialties.

Approximately 62% of adult endocrinologists are men and 38% are women, whereas for pediatric endocrinologists, 49% are men and 51% are women. The mean age for both adult and pediatric endocrinologists in 2011 was 51 years (Supplemental Figures 1 and 2). In both cases, a higher percentage of men are older than 50 years and are being replaced by women who are younger than 50 years. One implication of these data is that a significant proportion of the workforce will be retiring by 2020.

The mean number of weekly hours and visits per year in direct patient care (inclusive of inpatient care but exclusive of teaching, administration, and continuing education responsibilities) differs based on age and gender and whether it concerns adult or pediatric endocrinologists (Supplemental Tables 1 and 2). Male adult endocrinologists work 42 hours each week, provide 3434 visits per year, and attain peak productivity between the ages of 41 and 45 years. Female adult endocrinologists work fewer hours per week (39), provide fewer patient visits per year (2484), and reach their peak of productivity between the ages of 46 and 50 years. Female pediatric endocrinologists spend more hours per week (31) and provide more visits per year (2158) than their male counterparts (21 hours, 1840 visits). Both reach their peak productivity at ages 46–50 years. The fewer services provided by pediatric endocrinologists reflects their more frequent positions in academic settings in which their time is divided between clinical activities, teaching, and research.

International medical graduates (IMGs) represent 23% of the endocrinology workforce in the United States as determined by the Endocrinologist Survey. The proportion of IMGs in the endocrinology workforce and in the workforce of internal medicine and its subspecialties is likely to increase over time. Indeed, in 2013, 53% of the endocrinology first-year endocrinology fellows were IMGs (11), whereas the percentage of first-year residents in internal medicine who were IMGs was 45% (12). Although precise numbers are not known, most studies suggest that the vast majority of IMGs who become board certified in the United States remain in the country after completion of their training.

The geographical distribution of endocrinologists is uneven (Supplemental Figure 3). As is the case with many other specialists, endocrinologists are more highly concentrated in New England and the Middle Atlantic regions of the country, whereas they are scarce in the South East Central, South West Central, North West Central, and Mountain Central Census regions.

The median income of endocrinologists in 2010 was $211 400, based on the MGMA non-self-reported administrative data. These figures are for endocrinologists employed in group practices for more than 1 year and are similar to those reported in the 2012 Endocrinologist Survey. The survey showed that the mean income for private practice, university faculty, and hospital-based adult endocrinologists was $217 297, $196 154, and $189 063, respectively, and $168 333, $155 000, and $196 875 for pediatric endocrinologists, respectively. The median compensation for endocrinologists is relatively low compared with other, more procedure-based subspecialties, such as noninvasive cardiology and gastroenterology (median incomes of $431 740 and $463 955, respectively) but is similar to the other cognitive subspecialties, such as rheumatology and infectious disease.

Factors affecting supply

The main factors affecting the supply of endocrinologists are the number and fill rate of fellowship positions as well as the attrition of the endocrinologist workforce. The annual entrants into the adult endocrinology workforce from fellowship in 2012 were estimated at 280, as reported by the ABIM in 2011. Since then, the number of adult endocrinology fellows has further increased to 304 as of 2013 (13). Thus, there has been an increase of about nine new adult fellowship positions per year, representing a growth rate of 3.4% per year. First-year fellowship positions in pediatric endocrinology have grown from 41 in 1998 to 94 in 2011 or at a rate of 6.6% per year. The total number of pediatric endocrinology fellows in the 3-year program has increased from 93 in 1998 to 273 in 2012 (14).

Attrition is primarily determined by retirement rates and work satisfaction. Age and health status are the most important factors in the retirement decision, factors over which there is limited influence. Endocrinologists have a high level of satisfaction, which is an important factor in their decision to remain active in medicine. Endocrinologists have similar, or slightly lower, retirement rates relative to other medical specialists and surgical specialists based on data from the AAMC and our Endocrinologist Survey (9). Based on the results of our survey, almost 30% of the Endocrinologist Survey respondents indicated they plan to completely retire between 65 and 69 years of age and 25% between 70 and 74 years of age (Supplemental Figure 4).

For those endocrinologists who originally became board certified in 1990 or after, recertification is required to maintain board certification every 10 years through the ABIM's Maintenance of Certification program. Endocrinologists who were initially board certified prior to 1990 are not required to recertify. However, of those certified in endocrinology between 1990 and 2002, only 79% have recertified by February 2014 (15). The pass rate for first-time takers with valid time-limited endocrinology certificates rates has improved over time from 76% in 2007 to 86% in 2013 (16). The ultimate pass rate is 96%. Because endocrinologists certified prior to 1990 do not require recertification, recertification is required only every 10 years, and the ultimate pass rate is so high, we do not believe that failure to maintain certification is a significant form of attrition from the profession.

Factors affecting demand

Current demand for endocrinologists' services is driven by diabetes care, which represents 46.1% of coded visits. Given the overwhelming number of persons with diabetes, only 15% of all diabetes care is provided by endocrinologists. According to the MEPS database, 85% of diabetes care is provided by nonendocrinologists. Thyroid disorders represent 18.2% of visits to endocrinologists and the remaining 33.7% of visits are distributed among all other endocrine diseases, with no one of them more than 5.5% of visits.

The responses to the Endocrinologist Survey to the question of how many were seeking additional colleagues provides evidence of present excess demand for endocrinologists. Of the 355 respondents, 75 adults and 10 pediatric endocrinologist positions were available. In addition to physicians, endocrine practices are recruiting nonphysician clinicians. Seventy-four percent of private practices and 56% of academic practices indicated they had openings for physician assistants and nurse practitioners. In addition, the survey found that the average wait time for an initial nonurgent consultation visit with an adult endocrinologist was 37 days, which is identical to the mean waiting time found in 1999 despite a 52% growth in the number of adult endocrinologists. In comparison, a phone survey in 2009 of the average waiting time for an initial visit with family medicine, cardiology, and dermatology specialists was 20, 15, and 22 days, respectively (17).

Future demand for endocrinologist is informed by a number of factors, some of which can be reasonably estimated at this time. The impact of others remains less quantifiable. In the former category are current and projected use rates, demographic trends, and insurance status, whereas patient lifestyle trends, regulatory requirements, guidelines for achievement of optimal care, and the availability of new medications and technologies are in the latter category. Because the projected increase in the demand for endocrinologists is based solely on demographic trends, it may underestimate actual demand by not accounting for these difficult to quantify highly relevant factors.

The rates of patient use of endocrinology services increase with age. Use is highest for men and women aged 65–74 years, with 76.5 and 97.4 visits per 1000 persons, respectively, and lowest for ages 25–34 years, at 4.4 and 30.9 visits per 1000 persons, respectively. In addition, demographic trends project a disproportionate growth in the number of people between ages 65 and 85 years. Between 2000 and 2010, the population aged 65 to 84 years grew by about 1 annually. Between 2010 and 2020, this population increase jumps to 3.3% annually. In 2010, there were 37.5 million people aged 65 years or older, constituting about 12.7% of the total population. By 2015, the number of people in the United States over age 65 years will increase to 46.9 million, representing 14.6% of the total population. By 2025, the population aged 65 years or older is expected to number 62.5 million and 17.9% of the population. Because age is a significant factor in the risk for endocrine conditions, including type 2 diabetes and osteoporosis, the demand for the services of endocrinologists will rise proportionally. Expansion of insurance coverage as part of the PPACA will result in increased demand for services, increasing the demand for all physicians because the number of uninsured patients is expected to decline from 50 million to 22 million. Mitigating the effect on endocrinologists is that the largest decline in the number of uninsured will occur in the age cohort of individuals18–24 years of age, who have a relatively low use of their services.

Supply and demand projections

Using 2011–2012 data, we developed a baseline scenario and a sensitivity analysis of factors affecting the future workforce market environment (scenarios 1 and 2) to project the supply and demand for both adult and pediatric endocrinologists from 2011 to 2025 (Table 1). In addition, scenario 3 was created to estimate the number of additional new fellows it would take to close the gap between supply and demand in 5 and 10 years. In each case we included the likely effect of the PPACA. We used FTE endocrinologists as defined above for both estimates.

Adult endocrinologists

There is a significant excess demand in the baseline scenario (Figure 1A), which is closed by 2023–2025 if the number of adult endocrinology fellows continues to increase at the historic rate of 3.4% (Figure 1B). However, should there continue to be an increase in the prevalence of diabetes, the gap will not close and, in fact, continue to widen (Figure 1C) (18). The effect of increasing the supply of adult endocrinology fellows in closing the gap is shown in Figure 1D. New entrants to the workforce must grow at an annual rate of 14%/y to close in 5 years and 5.5%/y, to close the gap in 10 years.

Pediatric endocrinologists

For pediatric endocrinologists, there is excess demand in the short term but the gap closes by 2015, after which supply exceeds demand (Figure 2A). At the current rate of growth of pediatric endocrinology fellows, the excess supply will widen so that by 2025 there will be almost twice as many pediatric endocrinologists as required by the demand curves (Figures 2B). An increase in the prevalence of diabetes (both type 1 and type 2) will have a smaller effect on demand for pediatric endocrinology services, compared with a comparable growth rate in the adult population because the prevalence rate of diabetes is much lower in the pediatric population, approximately 0.26%, compared with the adult population, approximately 11.3%. Therefore, a given percentage growth rate will have a much smaller effect on the pediatric population. Thus, in scenario 2 for pediatric endocrinologists, the demand resulting from the increase in the prevalence of diabetes among population aged 0–18 years is small. This has little effect on the demand curve and has no affect when the supply curve intersects the demand curve (Figure 2C).

Discussion

The Endocrinology Workforce Analysis identifies an increasing demand for the services of adult endocrinologists, which exceeds the available supply of full-time equivalent clinically active adult endocrinologists for at least the next decade. There are several factors at work producing this excess demand. First, the patient population is both growing and aging. Because the prevalence rate of endocrinopathies tends to be higher for older members of the population, the demand for endocrinologists grows by more than simply the growth in the aggregate population. Second, significant numbers of the adult endocrinologist workforce are baby boomers (born between 1946 and 1964). Those physicians who are over age 60 years will leave the workforce or will greatly reduce hours of clinical work over the next decade. Our model of supply incorporates this attrition and changing hours in the projections. Third, there is emerging evidence that those who replace the older cohort may work fewer hours, on average, and see fewer patients.

Our demand model captures the increased use of services because of the changing population demographics. Similarly, our supply model accounts for the effect that changing demography among endocrinologists will have on the number of patient visits provided by the average endocrinologist. At the same time that the baby boom generation will require more care delivered by endocrinologists, baby boomer endocrinologists will be leaving or limiting their practices and will be replaced by physicians who are expected to see fewer patients each year.

For pediatric endocrinologists, some of the same factors are at work, but the effects are attenuated. First, the aging population has little or no effect on the pediatric patient population and therefore is not a source of increasing demand as it is in the case of adult endocrinologists. Moreover, the overall proportion of children in the population is projected to decline, from approximately 27% in 2010 to 26% by 2020, with further declines to 25% by 2035. Second, although the prevalence rate of diabetes is increasing in both the adult and pediatric population, the overall prevalence of diabetes in the pediatric population is small. Therefore, a given percentage increase in the prevalence rate results in a relatively small increase in demand, compared with an equivalent increase in the prevalence rate in the adult population. Nevertheless, we recognize that because most patients with type 1 diabetes are cared for by pediatric endocrinologists, our model may have underestimated the impact that the PPACA will have on demand.

The precise forecasts for both adult and pediatric endocrinologists are subject to uncertainty and could vary widely based on the assumptions entered into the model. The TEP felt that the assumptions in baseline case and in scenarios 1 and 2 were the most likely to occur. In the baseline case, we estimate an excess demand gap of about 1484 FTE adult endocrinologists in 2015. By 2025, this gap will have fallen slightly to 1344 FTEs. For pediatric endocrinologists, our baseline estimate is that there will be excess demand of about 100 FTEs in 2015, but this gap is likely to be approximately closed by 2016. Under the baseline, there may emerge a surplus of FTE pediatric endocrinologists of about 200 by 2025.

There are likely to be adjustments on the supply side related to the excess demand. Basic economics suggest that, in the face of excess demand, earnings in the profession will rise, making it more attractive for physicians to enter the field of adult endocrinology. However, this assumes that the health care market recognizes the contributions that endocrinologists make to the control and prevention of diabetes complications, in particular, and endocrine diseases in general. Relieved of the constraints of the current compensation formulas for endocrine services, this would lead to an expansion of the number of fellowship positions, which are likely to be filled, and may encourage some physicians to delay retirement and others to expand clinical hours. All of these factors may work to reduce the supply-demand gap by expanding supply. Although the demand for health care services overall declined during the 2008–2012 recession, in the longer run demand is likely to expand, not only because of the aging population and increasing prevalence of diabetes but also because of the effects of the PPACA to improve accessibility to care.

In addition, excess demand is likely to encourage more practice efficiencies. When a physician's time will become more valuable, team-based care and task shifting will increase. The use of physician assistants and nurse practitioners, leveraging the endocrinologist's time in management of endocrine diseases, especially diabetes, will become more common. The scope of practice laws and reimbursem*nt policies must obviously adapt to this increase in demand and facilitate these practice efficiencies. Although some market factors may mitigate the level of demand observed over the next decade, the level of excess demand predicted by these models suggests that proactive interventions on the part of the endocrinology community are warranted including, but not limited to, those shown in Table 2.

This study has several strengths. One is the ability to draw on a suite of refined databases that are increasingly reliable and useful to health workforce researchers. Some of that data was used in combination with other data sets. Agreement across multiple data sets results in more reliable results. Finally, with the benefit of hindsight from 2003, the recognition that managed care did not evolve to dominate the health care market, arguably reducing the demand for specialty care, results in a greater demand than was anticipated in the earlier study.

An inherent limitation of this study and all workforce modeling is that these predictions have the appearance of precision but are based on many variables and assumptions. Moreover, the predictions are made without specifying the range of error using statistical methods. Predictive modeling is a statistical estimate or probability of an outcome given a set amount of input data. Newer and more refined data, and modeling techniques that incorporate uncertainty in a more transparent way, may improve future models. An additional limitation in the current paradigm is that there is no well-described, broadly accepted definition of one endocrinologist FTE. Therefore, we defined an FTE endocrinologist empirically, based on what the typical or average endocrinologist actually does, ie, the typical adult endocrinologist provides about 3038 visits per year, whereas the typical pediatric endocrinologist provides about 1964 visits per year. Another limitation is that demand is based on current use measures, and most experts would agree that current use does not achieve the access or benchmark outcome measures most experts would recommend. Finally, some of the data such as salary and access are generated by self-report. The latter may underestimate demand in certain markets where endocrinology physician wait times are 3–6 months.

In summary, this workforce study shows that there is likely to be a significant gap between the demand for and supply of adult endocrinologists for the foreseeable future and thus provides useful data that will enable policy recommendations on how to close this critical gap.

Acknowledgments

The opinions expressed in this paper reflect the personal views of the authors and are not the official views of the United States Army or the Department of Defense.

This study was funded by the Endocrine Society with assistance from the Association of Program Directors in Endocrinology, Diabetes, and Metabolism.

Disclosure Summary: The authors have nothing to declare.

Abbreviations

• AAMC

  American Association of Medical Colleges

• ABIM

  American Board of Internal Medicine

• FTE

  full-time equivalent

• IMG

  international medical graduate

• MEPS

  Medical Expenditure Panel Survey

• PPACA

  Patient Protection and Affordable Care Act

• TEP

  technical expert panel.

References

Copyright © 2014 by the Endocrine Society

Issue Section:

Special features

Email alerts

Citing articles via

More from Oxford Academic