A Longitudinal Analysis of the Pediatric Surgeon Workforce

Abstract

Objective

To describe the trends in the pediatric surgeon workforce during the last 25 years and to provide objective data useful for planning graduate medical education requirements.

Summary Background Data

In 1975, the Study on U.S. Surgical Services (SOSSUS) was published, including a model to survey staffing. A pediatric surgeon workforce study was initiated in conjunction with SOSSUS as a population, supply, and need-based study. The study has been updated every 5 years using the same study model, with the goals of determining the number and distribution of pediatric surgeons in the United States, the number needed and where, and the number of training programs and trainee output required to fill estimated staffing needs. This is the only such longitudinal workforce analysis of a surgical specialty.

Methods

Questionnaires were sent to 100 pediatric surgeons representing the 62 standard metropolitan statistical areas (SMSAs) in the United States with a population of 200,000 or more to verify the names and locations of all active pediatric surgeons and to gain information about the 5-year need for new pediatric surgeons by region. A program was developed to predict the number of pediatric surgeons relative to the total population and the 0-to-17-year-old population in the subsequent 30 years using updated data on the present number and ages of pediatric surgeons, age-specific death and retirement rates, projections of U.S. population by age group, and varying numbers of trainees graduated per year. As each 5-year update was done, previous projections were compared with actual numbers of pediatric surgeons found. The trends during the last 25 years were analyzed and compared and additional information regarding the demographics of practice, trends in reimbursement, and volume and scope of surgery was obtained.

Results

The birth rate has been stable since 1994. The 0-to-17-year-old population has been increasing at 0.65% per year; a 0.64% annual rate is projected to 2040. At present, 661 pediatric surgeons are distributed in every SMSA of 200,000 or more population, with an average age of 45 and an average age of retirement 65. The actual number of pediatric surgeons in each 5-year survey has consistently validated previous projections. Trainee output has increased markedly in the past 10 years. The rate of growth of the pediatric surgeon workforce at present is 50% greater than the forecasted rate of increase in the pediatric age group, and during the past 25 years the rate of growth of the pediatric surgeon workforce has been double that of the pediatric population growth. Nationally, significant changes in reimbursement, volume of surgery, and demographics of practice have occurred.

In 1970, the American Surgical Association and the American College of Surgeons agreed to cosponsor the Study on Surgical Services for the United States (SOSSUS) to evaluate objectively the distribution of surgical services, the problems of staffing, and the interaction of surgery with other fields of medicine, the community, and the law. Its aim was to influence societal plans that would be developed in the United States; it was hoped that a careful and systematic examination of surgery by surgeons would be a constructive influence on the future of American surgery. Dr. George D. Zuidema served as chairman; Dr. Francis D. Moore was vice-chairman. Dr. Moore chaired the subcommittee on surgical manpower, and the senior author of this paper (J.A.O.) served on that committee. SOSSUS published its summary report in 1975. ¹

For the pediatric surgery portion of the SOSSUS study, a methodology was developed to determine the following:

• The number and distribution of pediatric surgeons in the United States

• Within limits, the number needed and where

• The existing and needed number of completely equipped and staffed hospitals

• The number of training programs and trainee output required to fill estimated staffing needs to satisfy societal and academic requirements.

The first analysis of the pediatric surgery workforce, including projected needs, was published in 1976, and the study has been updated every 5 years since. ^2–6 This paper describes the 25-year update of this material and presents a comparison of the trends during this time period. Unfortunately, other surgical specialties have not continued to follow up on workforce evaluation and graduate medical education planning in the same fashion since the original SOSSUS publication. In 1982, the American College of Surgeons initiated the Longitudinal Study of Surgical Residents with yearly updates on the number of residents completing residencies, but information on geographic distribution and estimated requirements is not given. ⁷ Although this is a study of the pediatric surgery workforce, this model may be useful for other specialties; ideally, with the long-term accumulation of longitudinal data, the original aims of SOSSUS to influence national policy may be achieved.

METHODS

Questionnaires were sent to 100 pediatric surgeons, with an approximately equal balance between those in private community practice and those in academic positions. These surgeons were selected to represent the 62 standard metropolitan statistical areas (SMSA) in the United States that have a population of 200,000 or more. In the largest SMSAs, more than one surgeon was surveyed to ensure accuracy. The respondents were asked to verify the names and locations of all practicing pediatric surgeons in their SMSA as provided to them from all available directories and a variety of other sources. Corrections and deletions were made based on whether the surgeons were in active practice or not. No residents or retirees were included, and the respondents were asked to note whether the surgeons were in full-time or part-time practice. The surgeons were also asked to comment on the 5-year need for new pediatric surgeons in their region, either to replace retiring surgeons or to fill new positions based on practice expansion. Both community and academic requirements were assessed.

The methods used were similar to those used in previous reports. ² The statistical software package SAS (SAS Institute, Cary, NC) was used for all computations. A computer program in SAS was written to project total U.S. population and also the 0-to-17-year age group. The middle series estimate of the 1997 census projection was used, the latest available. ⁸ The latter projection of population was used for the 1st, 5th, 10th, and 30th year. Extrapolations were carried out for the 15th and 25th year using the estimated increment rate for these periods. A program was also written to predict the number of pediatric surgeons and their ratio to the total U.S. population and to the 0-to-17-year age group. The number of yearly graduates from training programs was assumed to be constant, as was done in the previous studies and the SOSSUS study. A constant input of 26, 30, or 35 graduate trainees per year between the ages of 30 and 35 years was used for comparison purposes related to total population and the 0-to-17-year age group in the United States. Also included were updated data about the present number and age distribution of pediatric surgeons, and age-specific death and retirement rates. ⁹ The validity of these predictions was checked by projecting the numbers of all board-certified surgical specialists, and this reproduced the projection of all board-certified surgeons as shown in Table 9 in the latter publication. Then, the previous and the current 5-year projections of expected numbers of pediatric surgeons were compared with the actual numbers of pediatric surgeons determined for that interval.

The present number of pediatric surgeons and their locations were determined from the SMSA questionnaires, directories, and personal inquiry. The age distribution was determined from data available from the American College of Surgeons, the area questionnaires, and other sources. The surgeons whose ages were not known were distributed among the age categories as seemed appropriate—for instance, a recent graduate not yet eligible for membership in a national organization was placed in the 30-to-35-year age group.

Data were also collected on the immigration of graduates from Canadian pediatric surgery training programs since 1985, the earliest year for which such data were available. No other immigration of pediatric surgeons was evident. Longitudinal data on U.S. birth rate, number of training programs, and number of trainee graduates were collected. In addition, specific demographic information was collected, including information on physician reimbursement, changes in style of practice, influences of managed care, scope and volume of practice, and competition from other surgical specialties for the same work.

RESULTS

All the questionnaires that were sent out were returned, permitting validation of the number and locations of active pediatric surgeons in the United States. In addition, each informant provided detailed information regarding estimates of need for pediatric surgeons in the next 5 years, including anticipated retirements. The same response rate has been achieved in each 5-year update.

According to the U.S. Census Bureau, projections of total population continue to show a decrease in growth rate. Annual population growth from 1970 to 1980 was 1.09%, from 1980 to 1995 it was 0.97%, and from 1995 to 2000 it was 0.88%; the projection for 2000 to 2040 is 0.75%. The projection for the pediatric population (less than 18 years) for the years 1995 to 2040 is an increase of 0.64% per year (Fig. 1). The Center for Health Statistics of the Centers for Disease Control & Prevention states that the number of live births has remained stable since 1994 at 14.6/1,000. ¹⁰ The Census Bureau states that the rate of immigration will remain stable.

Figure 1. Forecasted growth of the under-18 population. This represents a decrease in the growth rate of the pediatric population from that 20 years ago.

Estimates of the pediatric population indicate slight differences in the expected rate of growth based on subsets of this age group (see Fig. 1). Although the mean growth of the 0-to-17-year age group is 0.64% per year, the expected rate of growth in the 0-to-5-year and 5-to-13-year age groups is 0.62% per year. In the 14-to-17-year age group, which is shared between pediatric surgeons and adult general surgeons, the expected rate of growth is 0.75% per year.

As of January 2000, there were 661 pediatric surgeons in active clinical practice. The average age was 45 years; the average age of graduating trainees was 34 years. The estimated age at retirement or death, according to Moore et al ⁹ and a survey of pediatric surgeons 2 years ago, was 65 years, as it was in the previous studies. There are 32 residency training programs in the United States; 21 of them graduate one resident per year and 11 one resident every other year, for a total of 26.5 graduates per year. In Canada there are six programs, with five graduating one resident each year and one every other year for a total of 5.5 graduates per year. Three and sometimes four come to the United States, a slight decrease compared with previous studies; most of them are U.S. citizens to begin with. There were 14 residency training programs in pediatric surgery in 1975, 17 in 1980, 20 in 1985, 24 in 1990, 32 in 1995, and 38 in 2000 (U.S. and Canadian programs), a growth rate of trainees of almost 25% since 1983. There are no new applications before the Residency Review Committee. ¹¹ The increase in trainee output has paralleled the increase in the number of training programs approved.

An analysis of the geographic distribution of the 661 actively practicing pediatric surgeons indicates that pediatric surgeons are practicing in every SMSA of 200,000 or greater and in some smaller SMSAs as well. State-by-state analysis reveals some differences in density, however. There are no pediatric surgeons in the sparsely populated states of Montana, Idaho, Wyoming, and North Dakota. Efficient transportation systems have been in place for several years to provide tertiary pediatric surgical care in those states. The ratio of pediatric surgeons to the 0-to-17-year population by state ranges from 1/99,000 or less to 1/200,000 or more (Fig. 2). Previously accumulated information on the density of pediatric surgeons per unit of total population in various European countries found a ratio of pediatric surgeons to total population of 1/1 million. ⁶ Although it varies by state, the ratio of pediatric surgeons to population in the United States is approximately 1/500,000 for the total population and 1/100,000 for the 0-to-17-year age group. During the past 25 years, the ratio of pediatric surgeons to the pediatric population has doubled.

Figure 2. Density of pediatric surgeons in the United States based on population. Only the sparsely populated states of Idaho, Montana, Wyoming, and North Dakota do not have pediatric surgeons.

By 1995, it was found that hospital facilities with staffing and resources capable of handling the broad range of complicated pediatric surgical diagnoses existed not only in every SMSA of at least 200,000 population but also in most of those with a population of 100,000. ⁶ At that time, there appeared to be more than enough neonatal centers in the United States.

The 0-to-17-year age group is expected to increase at a rate approximately 85% of the rate of growth of the total population. This primarily reflects a marked expected increase in the number of persons older than 60. Figure 3 shows the estimated number of pediatric surgeons that would result with constant inputs of 26, 30, or 35 trainees per year per 100,000 total population;Figure 4 provides the same information for the 0-to-17-year age group. In previous studies, data were available only for the 0-to-15-year age group, which in many ways may be a better estimation of the population pediatric surgeons are exposed to. The annual growth rate for the 0-to-15-year age range is expected to be 0.52%. Compared with the anticipated 0.64% annual growth rate of the 0-to-17-year population, as displayed in Figure 4, a constant input of approximately 28graduates per year would match the population growth. The forecasted percentages of growth relative to population have been consistent during the past 25 years when calculated either for the 0-to-15-year or 0-to-17-year groups. Currently, 30 or 31 physicians are entering the active practice of pediatric surgery in the United States yearly. During the past 5 years (since the 1997 appointment match), the number of physicians applying for pediatric surgical residencies has been 61 (51% match), 57 (42% match), 72 (42% match), 49 (49% match), and 37 (78% match). ¹²

Figure 3. Predicted numbers of pediatric surgeons relative to total population, taking into account death and retirement, comparing fixed numbers of trainee graduates per year. There are 661 active pediatric surgeons and an input of 30 or 31 surgeons per year.

Figure 4. Same data as in Figure 3 projected for the pediatric population treated by pediatric surgeons. With 30 graduates per year entering practice, the growth rate in the number of surgeons is 1.7 times the population growth rate.

Figures 5 and 6 illustrate the most recent data collected regarding the estimated number of pediatric surgeons needed, and where, during the next 5 years compared with trainee output. The anticipated 5-year need and the apparent demand for new pediatric surgeons vary by region, just as the geographic distribution and density of pediatric surgeons do. In the next 5 years, a mean of 101 pediatric surgeons will be needed, in comparison with a 5-year trainee output of 132.5 (at 26.5 graduates per year), 150 surgeons (at 30 graduates per year), and 175 surgeons (at 35 graduates per year). The current rate of output of trainee graduates is 50% more than the estimated 5-year requirement, taking into account anticipated retirements. Current data from the American College of Surgeons and the American Pediatric Surgical Association list approximately 65 available positions, some which are about to be filled and some of which, by their nature, are not conducive to the classic practice of pediatric surgery. The same clearinghouse registries list 13 practicing pediatric surgeons seeking new positions.

Figure 5. Estimated number of pediatric surgeons needed in the next 5 years and where, considering planned retirements and practice expansion requirements. A mean of 101 surgeons was estimated to be needed.

Figure 6. At the present level of 30 graduates per year, the 5-year growth rate in the number of surgeons exceeds estimates of need, even when retirements and deaths are considered.

Additional demographic information was accumulated along with the questionnaires regarding the volumes and scope of practice. Most respondents reported a marked decrease in reimbursement, which has resulted in the need to do more procedures. For example, surgeons estimated that although recruitment previously depended on the availability of approximately 300 procedures per year to support a new pediatric surgeon, 400 are estimated to be needed now. Most groups around the nation have taken a conservative approach to recruiting, indicating a decreased demand for pediatric surgeons. However, there is still a demand for pediatric surgeons as judged by the clearinghouse data referred to above, although the indication is that the demand is less than in previous surveys, and many available positions are in areas of low population. This is paralleled by the fact that there is a definite trend toward performing surgery in community hospitals in multiple practice sites taking more of the surgeons’ time (as compared with the previous single-practice-site model). This is based on economic need, from the comments made. At least half the surgeons surveyed reported a decrease in the number of procedures, especially simple ones and newborn ones. This was largely the result of managed care organizations steering patients to other surgeons, especially patients with insurance. Also, there is now significant competition from pediatric specialists in urology, otolaryngology, gastroenterology, pulmonology, and interventional radiology for the same pool of patients in their communities. This trend was true for surgeons in community practice as well as in university or children’s hospital practices.

To assess the accuracy of this predictive model (Table 1), the actual number of pediatric surgeons determined to be practicing at each 5-year interval was compared with what had been predicted 5 years previously using the statistical model. The actual number was generally slightly greater than the number predicted, taking into account the number of trainees and death and retirement rates. For example, in 1975, at 20 graduates per year entering the system, it was estimated that there would be 405 pediatric surgeons in 1980; there were 416. These figures indicate the actual number of pediatric surgeons entering the workforce pool each year.

Table 1. NUMBER OF PEDIATRIC SURGEONS VS. PREDICTION

In previous 5-year studies after SOSSUS, the increase in the ratio of pediatric surgeons per unit population could be compared with that of all other surgical specialties, but precise updated information is not available for all the surgical specialties. However, pediatric surgery is increasing at a significantly greater rate than general surgery (which is flat), urology, thoracic surgery, neurosurgery, and orthopedics. The 25-year trend of actual growth of the number of pediatric surgeons relative to the pediatric population is shown in Figure 7, and the trend for estimated growth is similar.

Figure 7. Comparison between the forecasted rate of growth of the pediatric population and that of the number of surgeons with 30 or 31 graduates per year. The marked increase in the number of surgeons can be related to the increase in the number of training programs in the past 10 years in the United States and Canada.

DISCUSSION

Perhaps the best indication that this study is important to pediatric surgeons is that during the past 25 years, essentially 100% of the questionnaires have been completed and returned. When this study was undertaken 25 years ago in conjunction with SOSSUS, the goal was to obtain objective data on a serial basis to develop a way to predict the number of training programs and trainee output required to produce enough, but not too many, surgeons so that children might have access to pediatric surgical services of high quality. This goal is shared by the American Academy of Pediatrics. ¹³ For patients to have access to critical surgical services, there must be enough trained and experienced surgeons, and they must be dispersed over wide geographic areas. There must also be supporting specialists, including pediatric anesthesiologists, as well as properly equipped and staffed hospital facilities. If there are not enough surgeons, then access is limited. If there are too many surgeons, experience is diluted and there is the potential for diminished quality of care.

An additional consideration pointed out by Feil et al ¹⁴ is that the global cost of care increases when there are too many practitioners in a region. Cost is a consideration not only in the latter respect but also in terms of the funding of graduate medical education. As this workforce study has proceeded through the years, it has reinforced the concept that if accurate, objective data can be provided to planning organizations dealing with the accreditation of training programs and government, appropriate national decisions might be possible. If the appropriate decisions are made by the specialty motivated to provide the best quality of care, this process would be better than if it were mandated by government regulation.

Several methodologies for studying workforce are available. The Graduate Medical Education National Advisory Committee published a study estimating the need for various types of specialists based on population trends and projected disease incidence (a so-called adjusted needs-based study). ¹⁵ The Longitudinal Study of Surgical Residents of the American College of Surgeons is a supply-based study. ⁷ The Council on Graduate Medical Education has published a series of reports on the supply of various types of physicians, as well as estimates of need for the sorts of physicians society needs as related to population; this is a combination of a supply, population, and needs-based study. ¹⁶ In separate studies, Weiner ¹⁷ and the Council on Graduate Medical Education ¹⁸ projected workforce requirements on the basis of health maintenance organization staffing patterns as related to population (so-called managed care model of forecasting), but their assumptions that 100% of the population would be insured has not come to pass, and the healthcare environment has been in flux. Various specialty groups have sporadically published data on supply and to some degree estimated need as well. The present workforce study performed during the last 25 years has been designed to be population-based as well as needs, disease, and demand-based to develop, as much as possible, accurate information about both the supply of and the requirement for pediatric surgeons through the years. Further, it has integrated information on the effects of managed care on patterns of practice in various regions of the United States and the impact of technology on practice and the incidence of surgery. Most of the data collected can be related in a reasonable fashion, as will be discussed below.

Data on the demographics of pediatric surgery and population are highly accurate. Estimates of need are less accurate, unless followed over time and correlated with other confirmatory information. This is because estimates of surgeon supply and requirements are based on entrance and exit rates; although this is an accepted forecasting tool, even minor discrepancies in assumptions can create large errors, usually overestimates. This has not been true of this study of the pediatric surgery workforce, because only known data and minimal assumptions have been used, and the previous studies have provided slight underestimates of supply.

The study by Ravitch and Barton ¹⁹ was a disease-based study looking at the known incidence of major, so-called index cases in the United States, concentrating on major congenital malformations and tumors and then estimating the number of surgeons needed to perform a sufficient number of procedures to remain competent. In their study, published in 1974, they estimated that 300 pediatric surgeons would be required to manage the nine index conditions that would be expected to occur in the United States each year, assuming that each surgeon would require 55 index cases yearly to remain competent. When the latter study was performed, the live birth rate was approximately 15.5/1,000 women; currently the birth rate is 14.6/1,000. The latter may be the best indication available regarding disease incidence if one is to attempt a disease-based study; although theoretically this the most appealing approach, it is probably the most difficult and least accurate.

The introduction of technologies such as minimally invasive surgery has increased the incidence of certain procedures, such as cholecystectomy, but the overall incidence of surgery in the pediatric age group has not changed. It is not likely that the introduction of any new technology will have a major impact on the number of surgeons needed if the population to which it is to be applied is either not growing or is growing very slowly. Therefore, the incidence of disease and new technology, at least at present, are not potentially significant confounding variables.

The current trends in population have been operative for the entire 25-year period of this study—that is, the proportion of persons older than 50 years has been steadily increasing, and the rate of growth of the pediatric population has been falling. In 1975 the growth of the pediatric population was approximately 1% per year, whereas it is now 0.64% per year, projected to 2040. The expected rate of growth in the 0-to-5-year and 5-to-13-year groups is 0.62% per year; this is the main age range treated by pediatric surgeons. With stable birth and fertility rates during the past 6 years, there is no reason to believe there will be any substantial change in the estimated rate of growth of the 0-to-17-year age group.

Another important consideration in forecasting workforce requirements in any particular specialty has to do with the economics of providing that care. If one examines ethnic subsets of the birth and fertility rate data, it is evident that the main increases in the pediatric population are in disadvantaged populations with either no health insurance or with Medicaid. ⁸ For example, the Medicaid sector of most pediatric surgery practices represents 30% to 60% of the total. Managed care has added to this burden, raising the question of how many pediatric surgeons can be accommodated based on the economic realities of providing care. The same consideration is true for hospital facilities. Parkerton et al, ²⁰ in their survey of recent graduates, found that starting salaries were reasonable compared with other surgical specialties, but our surveys of all age groups of pediatric surgeons indicated significant decreases in reimbursement and expectations for the future. We made no effort to evaluate this factor, but low expected compensation doubtless influences young people when they consider whether they should spend 2 to 4 extra years in residency training, particularly if they are carrying large debts. ²¹

At present, there are 661 active pediatric surgeons, and 30 or so more will be added to this pool over the next few months, with a handful of anticipated retirements. The average practicing lifetime of pediatric surgeons has been 30 to 32 years during the term of this study. The distribution of pediatric surgeons geographically is broad.

Between 1975 and 2000, the number of training programs increased from 14 to 38 (in the United States and Canada), and the number of surgeons entering practice in those years varied from 12 to 14 in 1975 to 30 or 31 in 2000. ¹¹ By 1995, immigration of foreign-trained surgeons, from other than Canada, was essentially eliminated, and most of those who graduated from Canadian training programs and came to the United States were American citizens to begin with. The estimated need for pediatric surgeons in Canada is approximately 2.5 per year, leaving 3 and occasionally 4 surgeons available for practice in the United States.

During the 25 years of this study, the number of training programs has increased by a factor of 2.7 and the number of graduates by a factor of 2.5. With 30 surgeons entering practice each year, the increase is 1.01% per year, or 1.7 times the population growth. At the same time, there is no prospect that the population growth rate will increase, there is more competition for the volume pool of surgery in the pediatric age group, disease patterns are not changing in any appreciable way, and changes in technology in the past 10 years have not changed the overall requirements for surgical procedures. There is also no evidence that the pediatric population has limited access to pediatric surgical services in any part of the United States, so at least from the standpoint of population-based information, the current rate of growth of pediatric surgeons relative to population may be in excess of the future needs for such surgeons.

Taking into account the serial data accumulated every 5 years since 1975, this predictive model has proven to be remarkably accurate, usually an approximately 5% underestimation. The same sort of analysis can obviously be done for any surgical specialty. Unfortunately, the same degree of accuracy and concordance can never be achieved with estimations of need or demand. However, since 1975, the estimated needs have never been far in excess of the estimated increases in the numbers of practicing pediatric surgeons. The estimated need as well as the demand for pediatric surgeons has clearly been different for different regions of the country.

From comments obtained in this survey and in the survey of new graduates, it is clear that there are changes in the patterns and demographics of pediatric surgical practice in the United States. In the past, graduating pediatric surgeons were recruited to children’s hospitals and university hospitals, but in the past 5 years the trend has been for pediatric surgeons to locate in less and less populous regions and to community hospitals, as also noted by Sheldon ²² and others.

Whereas in the past pediatric surgeons practiced primarily at a single institution or at the most two, the trend in the past 5 years has been for pediatric surgeons to practice at multiple sites, in some cases up to 10, driving long distances and taking much more time than previously. ²⁰ The factors that have influenced this change in practice pattern have been primarily financial and not professionally or quality-based. One must ask whether this trend is a good one from the standpoint of children who need care, because most of these community hospitals do not have the full array of children’s services. It is likely that the multiple practice site model will require more pediatric surgeons than the centralized care model because of time considerations. These trends have not become established throughout the country, so it is difficult to predict their eventual influence. However, it is evident that in geographic areas where the multiple practice site model is developing, an increased hospital demand for pediatric surgeons is evident. Those positions have been more difficult to fill because they are less desirable than the single practice site model, in which it is easier for a surgeon to fulfill his or her responsibilities. Although the demand for pediatric surgeons in some regions is strong, the recent graduates study indicates that since 1994, the likelihood of a recent graduate practicing at a children’s hospital dropped by 50%, and more than two thirds of 1997 graduates were practicing in general hospitals, where the scope of practice was narrower. ²⁰

Another trend that also is influencing the need and demand for pediatric surgeons is competition for pediatric surgical work by adult general surgeons, pediatric urologists, otolaryngologists, gastroenterologists, pulmonologists, and interventional radiologists. These physicians are competing for surgical procedures that were traditionally all performed by pediatric surgeons, so there has been some decrease in the available volume and scope of surgery for pediatric surgeons. Although it is not yet possible to quantify this effect, it was reported from every geographic region in the United States.

For more than 25 years, this study has enabled us to track accurately the workforce trends and socioeconomic considerations in the field of pediatric surgery. The same model is clearly applicable to any surgical specialty. The analysis relates the production and attrition of pediatric surgeons relative to population and region of the country and is qualified by estimates of need and demand. The 25-year trend in terms of need and demand has been followed carefully, which is probably the only way we can balance what would otherwise be a largely subjective picture of workforce need. Initially, production barely met the need, but in at least the past 15 years, production of new surgeons has more than kept pace with need, and there are indications that production is now exceeding need, with the study of recent graduates and physician income data being the best confirmations. ¹⁴

Under most circumstances, market factors influence the number of persons who seek training positions when the future prospects for available jobs are diminished. One must question whether the marked decrease in the number of applicants in the pediatric surgery match this past year is an indication of that or other factors. Other factors that may influence the choice of specialty are physician satisfaction, lifestyle, working hours, and financial incentives.

There is an increasing trend for women to enter the field of pediatric surgery. Women tend to have shorter work lives and are more likely to work part-time because of child-bearing considerations. ^23,24 The long-term effect of these influences is not yet known.

To validate the estimates of need from the questionnaires, as many data sets as possible were accessed. The American Pediatric Surgery Association and the American College of Surgeons have each developed clearinghouse databases that list physicians who are seeking new practice locations as well as available positions, an indication of demand. From these databases, it was evident that 65 open positions, some listed for several years, were available. Based on the questionnaire responses, for the next 5 years, approximately 101 pediatric surgeons will be needed; all these positions can potentially be filled, because in the next 5 years, 150 or more new pediatric surgeons will enter the practice arena, with little attrition during that time. In addition, 13 pediatric surgeons currently in practice are seeking new positions, in most instances for financial reasons. The latter factor cannot be underestimated: it is likely that reimbursement patterns may well cause prospective candidates to reject the field of pediatric surgery when they already owe large amounts of money. ²¹ We already have a ratio of pediatric surgeons to population higher than that of most European countries whose populations are considered to be well served, although they generally use a centralized model of care.

The current regional estimates of need for pediatric surgeons in the United States do not indicate any substantial deficit in numbers that cannot easily be met by even fewer graduates than we have at present. Ravitch and Barton ¹⁹ cautioned that significant decreases in each individual surgeon’s experience in managing complicated pediatric surgical conditions would lead to a decrease in the quality of pediatric surgical care. Holder, ²⁵ in his 1976 workforce report, expressed concern about continuing competency based on adequate technical experience. This factor must be monitored carefully.

The marked increase in the number of training programs in the past several years has appeared to stabilize as those in the field have become aware of marketplace trends. Estimates of the requirements for pediatric surgeons have been elusive, but the 25-year trend indicates that there is no need for more pediatric surgeons than indicated by this study.

Not only pediatric surgery but all surgical specialties clearly need to continue to survey their workforce, its distribution, and the pattern of surgical care they provide. All of the emphasis has appropriately been on continuing assessment and enhancement of quality in our trainees. For many compelling reasons, it is evident that quantity is also becoming a major consideration in influencing quality.

Previous reports have also presented information on the effects of managed care on the practice patterns of pediatric surgery in various regions of the nation. Fewer pediatric surgeons were estimated to be needed wherever there was a high penetration of health maintenance organizations, because many routine pediatric procedures were being performed by general surgeons in community hospitals. It appears that pediatric surgeons are now going to many of these remote practice sites to maintain economic viability, perhaps in some cases at the expense of quality of care. This raises the need to promote a national standard that ensures that children receive surgical, anesthetic, and other aspects of care from qualified pediatric specialists in regional centers of excellence. We also need to ensure that new knowledge will be developed to advance the standard of surgical care, but current evidence suggests that it will be possible to do this at present levels without expanding our training base any further.

The current structure of our boards and residency review committees does not permit any consideration of appropriate numbers of trainees and programs or the needs of society at large. Because of this, numerous professional organizations have suggested that an independent, national workforce planning body be established to coordinate data gathering in a standard fashion, and we agree. ^13,26 Medicare and other groups that pay for graduate medical education will create restrictions if organized surgery does not keep its house in order. It is only by continuing to survey the workforce of each surgical specialty that we will have an opportunity to exert leadership in an appropriate fashion to ensure that professional competence is maintained, quality of care is enhanced, and the needs of society at large are met.

Discussion

Dr. Francis D. Moore, Sr. (Boston, Massachusetts): In a sense this is a follow-up of the SOSSUS Study, an effort made jointly by the College and this Association 25 years ago. In another few years, another such study should begin by taking stock of surgery and surgeons for the entire country.

It is interesting to look at the residences of the authors of this paper: Nashville, Ann Arbor, Toronto, Kansas City, Houston, and Palo Alto. They represent a wonderful and congenial national collaboration. That is what is always needed in this sort of a study.

Some years ago we pointed out that the need for surgeons is determined by the epidemiology of the diseases they treat. A good rule, but it has problems because the diseases surgeons treat change as surgery expands in scope—in this century, as examples, to include the brain, the blood vessels, the heart and transplantation.

In pediatric surgery, there is a very special problem. Pediatric surgeons are operating on the same patients formerly operated on by general surgeons who did not call themselves pediatric surgeons, nor confine their work to children. Pediatric surgeons are in a sense replacing an already active workforce. So we have an interesting question as to what fraction of pediatric surgeons in this country work in teaching centers. Are they a byproduct of the superspecialization of teaching centers associated with children’s hospitals? Or are they scattered throughout the hinterland of American cities and towns and counties?

Because those same children’s diseases are being operated on by general surgeons elsewhere, we have an additional enigma to consider in assessing the surgical workforce. Clearly, pediatric surgeons fill a need. What, precisely, is that need? Is it a need for teaching? Or for special services?

In 1975, the SOSSUS Study suggested that there might be quite a few extra surgeons or excess surgeons by 2000! Would “market forces” control this? It is very difficult to find a surgeon who regards himself or herself as extra or excess. Indeed, surgery has been accused of creating its own need, and in a sense that is true. Think of cardiac surgery, or transplantation, products of surgical research that have created the need for a whole new generation and species of surgeons. Has pediatric surgery done the same?

Finally, I would like to consider several diseases of infants— pyloric stenosis, inguinal hernia, meningocele, and congenital dislocation of the hip—then let us ask the essayists what fraction of those are operated upon by board-certified pediatric surgeons. That is an interesting question, and a difficult answer to get. I do not believe they tried to get that answer, but we would all be very interested to know their guess.

This is the question (in terms of economic analysis) that would be called “market penetration”: To what extent has pediatric surgery taken over the “market” of surgery in children? Do they own it all, or are they just nibbling around the edges?

Presenter Dr. James A. O’Neill, Jr. (Nashville, Tennessee): Dr. Moore, let me first of all congratulate you on your award which came this year. Let me say that my generation of surgeons all dreamed of having the opportunity to interact and train with you. Of course, that happened only to a relatively fortunate few. But many of us do interact with you in national forums, and it was my pleasure to interact with you in the manpower study. I think most here know that Dr. Moore chaired the original workforce study in American surgery, and that was the pacesetting event in that regard.

You made reference to several things, including disease-based studies, which is one other valid form of workforce analysis. I think that we know that the diseases that we see in childhood are the same. Because the preponderance are congenital disorders, while their patterns may have changed a little, the overall spectrum of disease is very much the same. We are able to do a bit more because of the improvements in surgery. So for example, we are able to tackle malignancies more aggressively, we are able to do a bit more for trauma than we had previously been able to do, and so forth. This parallels general surgery. But just like the consideration of technology, if the population to which technology is to be applied is growing slowly or not at all, or the disease-state patterns are not changing remarkably, one would not expect that the incidence of surgery would increase. And indeed we have data on the incidence of surgery, and that has not changed.

I think we know that pediatric surgeons today are doing the majority of things such as pyloric stenosis and inguinal hernias, things which would be construed as general surgery. On the other hand, things which had previously been done by pediatric surgeons which are now relegated to the fields of neurosurgery, cardiothoracic surgery, and urology, and perhaps otolaryngology—those things have very well been taken over by those specialties, as is seen in your own Boston Children’s Hospital. So there has been a very real move towards specialization in that regard.

Economically, this is another issue. I think that the things to be considered there, Dr. Moore, include the consideration that if there are excessive numbers of surgeons, the cost of care will rise. We know that is a very reliable observation. Second, the majority of pediatric surgeons have anywhere from 40% to 60% of their practice as Medicaid. And so there comes the consideration as to how many pediatric surgeons can actually survive economically in this nation because of those reimbursement patterns. That is a very real consideration today as well.

Dr. Olga Jonasson (Chicago, Illinois): In this longitudinal study of the workforce in pediatric surgery, Dr. O’Neill has sounded an alarm for his specialty. There may be too many surgeons for the population of children of this and the next generation. This is not a minor point, for pediatric surgery has had a history of recruiting the very best and brightest to its ranks from among the elite of surgical residency programs and has kept them in one of the longest residency programs in all surgery. No wonder pediatric surgery residents are, on average, 36 years old when they finally emerge from their structured educational programs.

What a supreme waste of time and talent it would be if, after all this, pediatric surgeons find fewer and fewer children who have less and less congenital disease as a result of improved prenatal care and genetic counseling. They will not be able to practice their skills. This study highlights the need for good planning based on the best possible estimates of need for the specialists.

Previous studies, including the SOSSUS report, have not provided the evidence we need to plan the pipeline or the supply. This is because of a fundamental flaw in that needs-assessment component of workforce studies. Who can predict what changes will take place that will influence the need for specialists? For example, did we accurately predict the changes in technology that have impacted so greatly on surgery, the aging of America and the declining birth rate, the transformation of healthcare systems, and the emergence of nonphysician practitioners that are becoming predominant health caregivers in some of the primary specialties?

Because we simply cannot predict the many changes that will influence demand and utilization of health care, Dr. Richard Cooper has proposed this trend model for consideration. Dr. Cooper suggests that we look at supply, for which we have very good data, examine trends of the work we do as influenced by technology, demographics, and other factors such as the evolution of systems and regulations including rationing of health care, and the emergence of nonphysician practitioners.

The Council on Graduate Medical Education and the Bureau of Health Professions are interested in testing this model, which has recently been presented to them. And the American College of Surgeons, in collaboration with health services researchers and the specialties of urology and neurological surgery, is hoping soon to begin an indepth pilot study of the workforce in those specialties as a preliminary to a more general workforce study.

That workforce studies are needed is obvious. What we will do with the results of this study in pediatric surgery or other studies as they are completed is, and will be, a most interesting challenge.

Dr. O’Neill: I think most of you are aware that Dr. Jonasson and her colleagues at the College of Surgeons have been doing the Longitudinal Study of Surgical Residents since 1982. Dr. Jonasson, I did make reference to the fact that technology cannot be expected to influence workforce needs if the population to which it is to be applied is not growing. It turns out that nonphysician practitioners are not an issue in pediatric surgery. As a matter of fact, they help the practice of pediatric surgery and do not compete. Because of the nature of the field, it is not possible to do that.

You are quite right that estimates of supply are highly accurate, and we have shown that, I believe, and that estimates of need or demand may be inaccurate over time. The only way we have gotten around what would otherwise be a very subjective forecast of need—and I think the value of this particular study, is that it has been possible to follow these trends very accurately over the past 25 years, updating it every 5 years. Otherwise, I think we would not be able to make any sort of claim that the data was valid in terms of predicting need.

We are familiar with Dr. Cooper’s model. For the last 10 years, we have been asking, and actually we have worked also with the Bureau of Health Manpower, to coordinate workforce studies across all surgical specialties so that a common methodology would be utilized, and therefore we would have much more data relative to the field of surgery overall. Our effort here has been to keep our own house in order, but I think that it necessarily must occur across the board if surgery is to control what happens to it rather than government.

Dr. Anthony A. Meyer (Chapel Hill, North Carolina): I am going to read some comments that Dr. George Sheldon wanted to make on this paper—he is unfortunately unable to be here—and I have a few questions of my own I would like to toss in at the end. Some of these relate to things you have already said and Dr. Jonasson has already discussed.

The 25-year study or longitudinal review of the pediatric workforce represents the most useful approach to workforce studies, those done by the specialty itself. Dr. O’Neill and coworkers have done us all a service by the methodology and thoughtful analysis included in the paper. Workforce studies are confusing, as they use different methodologies and are not comparable.

During the past year, the Council of Medical Specialty Societies, under a contract with the Council of Graduate Medical Education led by Richard Cooper, as Dr. Jonasson pointed out, the director of Health Policy Institute at the Medical College of Wisconsin, are evaluating specialty workforce studies. Dr. Cooper has also consulted with the American College of Surgeons regarding the surgical workforce analysis.

Dr. Cooper has the best perspective on overall manpower issues and studies. He has collected every workforce study available, and concludes that physician workforce studies largely often miss the issue. The true issues that affect supply and demand in the workforce are geographic distribution and the growth of nonphysician clinicians. For example, a few years ago, Dr. Jonathan Weiner of Johns Hopkins predicted a surplus of 165,000 physicians in the year 2000. Such overcapacity is not discernible today. What is apparent is that everyone is practicing medicine today. There will be more nonphysician clinicians than primary care physicians and more chiropractors and pediatricians by the year 2005. In November 1999, Blue Cross and Blue Shield of Illinois declared that chiropractors were primary care physicians for payment purposes.

Alternative health care workers see more patients than primary care physicians. It is certain that practitioners of shorter educational requirements that provide uncomplicated health services such as hypertension control at lower cost than physicians will be strong competitors for lower technology services.

My two questions: First, when you have enough patients to hire one half-time pediatric surgeon, you don’t have quite enough to have recruit one more pediatric surgeon, how do you go about trying to solve that problem? And second, do you believe that in the Residency Review Committee there will be an attempt to actually decrease the number of pediatric surgical training programs and will that be done without any major consternation?

Dr. O’Neill: I think I have already made reference to nonphysician practitioners, so I won’t cover that any further. I think that the economics of medical practice, Dr. Meyer, pretty well determine whether you are going to recruit another practitioner or not. It doesn’t matter whether it is a pediatric surgeon, plastic surgeon, general surgeon, or whatever. I think that it may depend, too, on whether you are in an academic center, as about 45% of pediatric surgeons are, or whether you are out in the community.

There is no mechanism in the United States through our Residency Review Committees or the American Board of Surgery to control the quantity of any type of practitioner, but rather, the value of exhibiting data such as this is that it allows the specialty itself to police itself, provided the data is accurate and accurately put together. Indeed, that has probably happened, in that there have been no new applications for programs since 1995.

Dr. Robert J. Touloukian (New Haven, Connecticut): The past 20 years have witnessed a dramatic change in the work profile of pediatric surgeons from an emphasis principally on the index case, as first studied in 1975 by Drs. O’Neill and Barton and Ravitch as the basis for manpower needs, to the current time. Now pediatric surgeons, as Dr. O’Neill has correctly pointed out, find themselves increasingly decentralized, traveling to the satellite and back to the home unit to do the research and teaching in what little time remains.

In addition, the young pediatric surgeon finds a true expansion in the area of the nonindex work. This nonindex work includes an increasing demand to care for trauma patients and operate on infants and young children under the age of 5 years with hernia, pyloric stenosis, and even appendicitis, principally driven by younger-generation pediatricians who are demanding this care be given by pediatric surgeons. In some areas, the pediatric surgeon has also assumed additional nonclinical responsibilities, making the need for full-time equivalent (FTE)-based survey essential as an adjunctive study. I would have Dr. O’Neill comment on this.

The questionnaire-based survey in this study becomes the principal vehicle for projecting needs and has proven a reliable projector in the past. Now a paradox seems to exist between the projected 5-year needs based on the current number of trainees produced annually and a demand based on current clearinghouse lists. This is a subject that Dr. O’Neill did not really have time to get into today, but is covered in his manuscript. The latest number reported in the clearinghouse lists 65 available positions.

Dr. O’Neill, in your manuscript, I was not able to determine whether you had the opportunity to crossreference the clearinghouse list with the questionnaire data for the extent of correlation in a given standard metropolitan statistical area. This would be easy to do and provide additional verification of your conclusion that the regional data was reproducible. I would encourage you to do this also.

The decreasing number of the most promising candidates for training positions in pediatric surgery would indicate that we are at an important crossroads. The longitudinal analysis of the pediatric workforce is vitally important for the future of our discipline.

Dr. O’Neill: First of all, in doing this study, we did do it by FTEs. In the area surveys, it was requested as to whether individuals were in full-time practice or part-time practice and research, or whether they were women who were in their childbearing years and therefore were part-time. We have all of that data. I couldn’t display that today. Basically, the base that you saw was based on FTEs.

The 5-year estimates of need—we had previously done 10-year estimates of need, but it really is too much of a changing situation. That is the reason for doing it on a 5-year basis.

The clearinghouse lists come from two locations, the American College of Surgeons—and I am indebted to Dr. Jonasson for providing me with that data—and the American Pediatric Surgical Association. Those two clearinghouse lists were coordinated. They are geographically based. They were coordinated with the data from this study. So the conclusions are based on all of that coordination.

In other words, this is the best one can do in terms of coming up with estimates of need and demand. The only way to verify that is to continue to survey it every 5 years or thereabouts, so that is the reason for this having been done in that fashion.