DUKE UNIVERSIGS LIBRARY Digitized by the Internet Archive in 2021 with funding from Duke University Libraries https://archive.org/details/nhowmanydoctorsdo01 priv HOW MANY DOCTORS DO WE NEED? pet aie How Many Doctors Do We Need? A Policy Agenda for the United States in the 1990s Based on the Tenth Private Sector Conference, 1985 Edited by Duncan Yaggy and Patricia Hodgson Foreword by William G. Anlyan Duke Press Policy Studies Duke University Press Durham, 1986 © 1986 Duke University Press All rights reserved Printed in the United States of America on acid-free paper © Library of Congress Cataloging in Publication Data appear at the end of this book. Contents 11 8 Foreword / WILLIAM G. ANLYAN ix Participants x I Overview of Public and Private Policies Affecting Physician Supply in the U.S. / WILLIAM G. ANLYAN_ I Legislative Perspective / PAUL G. ROGERS 9 Historical Perspective / ROSEMARY A. STEVENS II GMENAC Revisited / ALVIN R. TARLOV 13 The Physician Surplus: Another View / WILLIAM B. SCHWARTZ_ 19 Market Forces and Geographic Distribution of Physicians / JOSEPH P. NEWHOUSE 23 Correcting ‘‘Surpluses” and ‘Shortages’ in Medical Specialties I. The Surgeons / Cc. ROLLINS HANLON 28 Il. The Physicians / ROBERT H. MOSER 30 Life-style Choices and Evolving Practice Patterns / BRITAIN NICHOLSON 34 Market Influences of FMGs / SAMUEL P. ASPER 37 Methodological Problems in Assessing Physician Demand, Need, and Supply: Policy Implications / UWE E. REINHARDT 4I Discussion 46 2 The University’s Role in Establishing Priorities for Medical Education—A Societal Perspective / WILLIAM H. DANFORTH 55 Managing Quality and Quantity in Residency Training Programs / JOHN S. GRAETTINGER 60 The Impact of State Licensing Boards on the Distribution and Quality of Physicians / BRYANT L. GALUSHA 64 The Experience in Great Britain / JOHN LISTER 70 Discussion 80 3 Public and Private Options for Assessing and Managing Physician Manpower Supply / ROBERT G. PETERSDORF 97 Comments I. From the AAMC/ JOHN A. D. COOPER 106 II. From the AMA / JAMES H. SAMMONS 109 III. From a Private Medical School / DANIEL C. TOSTESON II2 IV. From a Public Medical School / M. Roy SCHWARZ _ II5 V. Third-party Payment: Industry / WILLIs GOLDBECK 118 VI. Law / CLARK C. HAVIGHURST 124 VII. State Government / EUGENE S. MAYER 127 Discussion 134 4 Summary I. Alain C.Enthoven 147 II. ArmoldS.Relman 152 Ill. J. Alexander McMahon 155 Index 157 Figures and Tables Figures ite a Io. JUL I2. ee 14. 15. 16. 17. 18. 19. 20. Medicine Evolves into a Science, 1846-1909 2 Post-Flexner to World War II, 1910-44 3 Post—World War II Era: Sputnik and Social Justice, 1945-65 4 Medicare/Medicaid and the Great Society to an Age of Limited Resources, 1965-85 5 Ratio M.D.s/Population 7 Physician Diffusion as Supply Grows: Towns of 5,000, 10,000 and 30,000 Population 23 A Simple Forecasting Equation for Physician Manpower 42 An Economic Definition of a Physician Surplus 44 Ratio of Physicians to Applicants and the Number of Unmatched Applicants 61 Positions Available, by Category, Filled by US. Senior Students 62 Doctor/Population Quotient in Great Britain 72 Career Structure 73 Hospital Staff: England and Wales 74 Numbers of Doctors in Each Grade by Place of Birth: England and Wales 75 Applications and Acceptances for Medical Schools, 1983 78 Growth in Numbers of Junior Doctors and Consultants: Great Britain, 1970-79 79 Specialty Distribution 102 Percentage of Residents by Specialty, 1983 103 Physician/t0,o0o0 Population Ratio for Nonmetropolitan Counties in North Carolina and the United States, 1979-80 129 Change in Physician/Population Ratio by County, 1963-83 130 21. Counties Served by the North Carolina Area Health Education Centers Program 131 Tables 1. Average Number of Office Visits per Year by Gender of Physician: Internal Medicine, United States, 1981 43 2. Differentials in Workload of Men and Women Physicians in Private Practice: France, 1981 43 Foreword For the Tenth Private Sector Conference we chose a topic that troubles physicians, educators, and policymakers: the growing sup- ply of physicians in the United States. We focused on two questions: Is there a surplus of physicians now, or in prospect? If there is, what should be done about it, if anything? The conferees expressed a variety of strong opinions on these questions. All the participants agreed that the growing supply of physicians will have a significant impact on American health care, and on physicians themselves, but they could not agree what the impact will be and whether it will be beneficial or detrimental. There was general agreement that the rapidly increasing influx of new physicians from foreign medical schools posed a serious threat, but a proposal to curtail the flow produced sharp controversy. We agreed on little, but our discussions gave all of us a better understanding of the issues and their significance. And that, after all, is the true purpose of the Private Sector Conference. On behalf of the conferees, I would like to thank the Duke University Medical Center, the Duke Endowment, and the Ameri- can Medical Association for the support that made the conference possible. W.G. Anlyan, M.D. Chancellor for Health Affairs Duke University Participants JOHN E. AFFELDT, M.D., President, Joint Commission on Accredita- tion of Hospitals WILLIAM G. ANLYAN, M.D., Chancellor for Health Affairs, Duke University SAMUEL P. AspER, M.D., President, Education Commission for Foreign Medical Graduates JosePH E Boyte, M.D., President, American Medical Association JoHN W. CoLLoTon, Director and Assistant to the President for Statewide Health Services, University of lowa Hospitals and Clinics JoHN A. D. Cooper, M.D., President, Association of American Medical Colleges WILLIAM H. DANForRTH, M.D. Chancellor, Washington University C. Douc tas EAVENSON, Assistant Director, Employee Benefits and Services, General Motors Corporation RICHARD H. EGDAHL, M.D., Academic Vice President and Direc- tor of the Medical Center, Boston University Paut M. ELLwoop, jr., M.D., President, Interstudy ALAIN C. ENTHOVEN, Ph.D., Mariner S. Eccles Professor of Public and Private Management, Stanford University Harvey V. FINEBERG, M.D., Dean, Harvard School of Public Health ASHLEY H. GALE, JR., Director, Hospital and Child Care Division, The Duke Endowment BRYANT L. GALUSHA, M.D., Executive Vice President, Federation of State Medical Boards Ex1 GINZBERG, Ph.D., Director, Conservation of Human Resources, Columbia University Participants xi WILLIs GOLDBECK, President, Washington Business Group on Health JoHN S. GRAETTINGER, M.D., Associate Dean, Graduate Medical Education, St. Luke’s Medical Center C. RoLtins HANLON, M.D., Director, American College of Surgeons Criark C. HavicHurstT, J.D., School of Law, Duke University JOHN LisTER, M.D., Postgraduate Dean, British Postgraduate Medi- cal Foundation, University of London Bitty G. McCatt, Deputy Executive Director and Secretary, The Duke Endowment J. ALEXANDER McManuon, President, American Hospital Association Marcaret MAHOney, President, The Commonwealth Fund EuGENE S. Mayer, M.D., Associate Dean, AHEC Program Director, University of North Carolina RosBerT H. Moser, M.D., Executive Vice President, American Col- lege of Physicians Tom E. Nessitt, M.D., Urology Associates, Nashville, Tennessee JosepH P NeEwuHouse, Ph.D., Chief, Economic Department, Rand Corporation BRITAIN NICHOLSON, M.D., Boston, Massachusetts Davip J. OTTENSMEYER, M.D., President, Lovelace Medical Foundation ROBERT G. PETERSDORE, M.D., Vice Chancellor and Dean, School of Medicine, University of California Uwe E. REINHARDT, Ph.D., James Madison Professor of Political Economy, Woodrow Wilson School of Public & International Affairs, Princeton University ARNOLD S. RELMAN, M.D., Editor, New England Journal of Medicine B. L. RHODEs, Executive Vice President, Kaiser Foundation Health Plan, Inc. PAuL G. Rocers, Attorney at Law, Hogan & Hartson Law Firm James H. SAMMoNS, M.D., Executive Vice President, American Medical Association Davip SATCHER, M.D., Ph.D., President, Meharry Medical College WILuiAM B. SCHWARTZ, M.D., Vannevar Bush University Professor and Professor of Medicine, Tufts University xii Participants M. Roy SCHWARZ, M.D., Vice President, Medical Education and Science Policy, American Medical Association Jack K. SHELTON, Manager, Employee Insurance Department, Ford Motor Company RosEMARY A. STEVENS, Ph.D., Professor, Department of History and Sociology of Science, University of Pennsylvania ALVIN R. TARLOv, M.D., President, Henry J. Kaiser Family Foundation DANIEL C. TosTEsON, M.D., Dean, Harvard Medical School EDWIN C. WHITEHEAD, Chairman, Whitehead Associates RICHARD S. WILBUR, M.D., Executive Vice President, Council of Medical Specialty Societies JERE WITHERSPOON, Associate Director, Hospital and Child Care Division, The Duke Endowment Rita WroBLEwskI, Ph.D., Director of Medical Affairs, Pfizer, Inc. JOHN IGLEHART, Rapporteur Overview of Public and Private Policies Affecting Physician Supply in the United States WILLIAM G. ANLYAN, M.D. First and foremost, we must define what physicians are all about. I choose to say that the basic mission of physicians is to improve the health care of the people they serve. Physicians’ services are neither a commodity nor a luxury; rather they provide basic needs for the prevention and cure of disease and the alleviation of suffering. In colonial times, most of the physicians in this country were trained in Great Britain, particularly Edinburgh and London. The first controls in medicine came in with the licensing boards in New York in 1760, which were followed by the founding of the first College of Medicine in Philadelphia in 1765, which became the University of Pennsylvania; the founding of Harvard Medical School in 1783; and then Kings College, which became Columbia, in 1792. Interwoven from this point on in this overview of medicine in the United States are the major scientific discoveries and events that have affected the numbers of doctors and the quality of health care. In 1809 the first quasi-elective laparotomy was done in Ken- tucky by Ephraim McDowell. A woman rode on horseback for sixty miles to have a 22.5 pound tumor removed from her abdo- men on a kitchen table without anesthesia. In 1810 Yale Medical School was founded and then came the Jacksonian era of deregulation in the 1830s, leading from four medical schools to thirty by 1840 and to seventy-seven by 1876. During this era of deregulation came the discovery of anesthesia, initially by Crawford Long in 1843, and then the ether dome dem- onstration at Massachusetts General Hospital (MGH) in 1846. Figure 1 shows medical history from this point until 1909. The American Medical Association (AMA) was founded in 1847 to 2 Overview of Public and Private Policies Figure 1. Medicine Evolves into a Science, 1846—1909. 1846 Anesthesia: Long vs. Morton 1847 AMA founded 1864 Specialty societies started; American Ophthalmological Society 1876 AAMC founded 1871-77 Harvard medical reform Pasteur, Lister, Koch 1893 Johns Hopkins 1895 Discovery of x-ray: Dr. Roentgen I9OI Landsteiner 1905 160 medical schools, 5,600 graduates 1907 AMA: CME bring some professional order to medicine. Then the specialty societies evolved, starting with the ophthalmologists in 1864 and the Association of American Medical Colleges (AAMc)} in 1876. At about the same time we had the work of Pasteur, Lister, and Koch leading to antisepsis and bacteriology. Harvard Medical School underwent its reform in the 1870s under President Eliott, over the near-dead body of the Chief of Surgery, Henry Bigelow, at the Mcu. Eliott wanted all students to know how to read and write, and Bigelow thought that that would restrict their admissions pool. In 1893 came another very definite milestone, the founding of Johns Hopkins, the model on which many other schools of medi- cine have been based. In 1895 the discovery of x-ray occurred, and in 1901 Landsteiner performed the first blood transfusion. In 1905 there were 160 medical schools and 5,600 graduates. In 1907 the AMA formed the Council on Medical Education to focus on the problems of medical education within that organization. During the early 1900s medicine evolved into a science as opposed to the previous era of purgare et sangare, which Moliére described as all that physicians could do for their patients: bleed and purge. With the aid of the Carnegie Corporation, Abraham Flexner in 1910 (figure 2) did his study of the quality of medical education, resulting in the closing of many of the second- and third-rate medical colleges of the time. In 1912 the College of Surgeons and the College of Physicians were formed and in 1t915 came the beginning of the specialty boards for examination. By 1920 the number of medical schools William G. Anlyan 3 had shrunk from 160 to 60, and the movement in the country was toward increasing specialization. There were fewer and fewer physi- cians in rural areas and the poor were not sufficiently cared for. Such foundations as the Rockefeller and subsequently the Duke Endowment became concerned and encouraged the formation of new medical schools. In 1924 insulin was discovered, prolonging the lives of subse- quent diabetics. But in 1930 the ama and the aamc began to suggest that we were creating too many medical schools, and in 1934 the president of the aMaA suggested closing half of the schools. Shortly thereafter, in the Aamc setting, Arthur Dean Bevan sug- gested that we cut down on admissions by 5 percent and increase the requirements for licensure. This occurred at the same time that third-party insurance was creeping in and the sulfa drugs were coming into use. In 1937 Weiskotten, heading a statesmanlike commission, sug- gested leaving the numbers of medical students alone but improv- ing their quality. In 1941 the United States entered World War II and needed more doctors, so classes were accelerated with a net gain of perhaps one or one and one-half classes. During the war, penicillin was discovered and some control of microbacterial infec- tions began. Figure 2. Post-Flexner to World War II, 1910-44. eee 1910 Carnegie: Flexner 1912 Colleges 1915 Specialty boards 1920 60 schools 1920-24 Foundations encourage new schools Increasing specialization Decreasing numbers of rural M.D.s Decline in the care of the poor 1924 Insulin 1930 AMA and AAMC: ? too many 1934 AMa (Bierring): close half of the schools 1934-37 Third-party insurance Sulfa drugs 1936 AAMC (Bevan): admissions decline 5%, licensing increases 1937 Weiskotten: quality improves 1941 World War II: increased need for M.D.s, accelerated classes 1944 Penicillin 4 Overview of Public and Private Policies In 1945 (figure 3) the NrH evolved from the adolescent unit that existed at that time, and in 1948 the Hill-Burton Act was enacted, creating rural hospitals. New va hospitals were created in juxtapo- sition to universities. And the Aamc declared that we needed to double medical school enrollment. If somebody had asked me in 1948 what kind of hospital I thought we should build for the next decade, I would have said we need more tuberculosis hospitals and more polio hospitals and convalescent centers. Yet we see from the vantage point of some years later that tuberculosis was coming under control and research in the next few years would lead to the polio vaccine. By 1949—50 the Truman Commission tended to agree with the AAMC about the need for more doctors but the AMa raised the question of whether the data were good enough to support those suggestions. The Ama preferred to leave it to the universities to decide about class size; they were accused of trying to protect the physician’s income at that time and nothing much happened. In the four-year period between 1952-56, the number of foreign medical gradu- ates (FMGs) multiplied fourfold. Figure 3. Post-World War II Era: Sputnik and Social Justice, 1945-65. 1945 NIH 1948 Hill-Burton Act VA-university relationships AAMC: need to double enrollment 1952 Magnuson and Truman: same AMA: ? data and leave to university ? to protect M.D. income FMGs quadruple in four years Tuberculosis declines 1956 Little Rock Polio 1957 Sputnik Heart disease on the rise 1959 Civil Rights movement; increased attempts to solve social problems through government AMA and AAMC agree on need to increase numbers of schools and M.D.s Transplantation 1963 First HPEA Act: construction, loans B.1.G. HPEA revisions: need 50% more M.D.s within a decade 1965 Coggeshall and Millis (aamc) William G. Anlyan 5 Figure 4. Medicare/Medicaid and the Great Society to an Age of Limited Resources, 1965-85. 1965 Medicare/Medicaid 1969 PA.P: increases 10% 1970 Carnegie Commission: 50% increase, and increased federal aid 1971 Health Manpower Act: capitation bonus increases, new schools, convert 2-year schools to 4-year schools 1972 National Hs Corps: geographic maldistribution 1973 Weinberger 1975 70,000 more FMGs: one-third I and R; one-fifth U.S. M.D.s HPEA + U.S. FmMGs (decreased one year later) SOSSUS 1976 AHECS GMENAC (1976-80) 1978 IOM: no increase 1981 Capitation kaput Then came the Little Rock incident in 1956, which completely blurred what happened in Hungary, and in 1957 the Sputnik, fol- lowed by a great movement to increase the civil rights of our citizens and solve social problems through governmental action. Shortly thereafter, both the ama and the aamc agreed upon the need to increase the number of medical schools and the enroll- ment of physicians. Together they persuaded the federal govern- ment to write the first Health Professions Educational Act for the construction of new schools and the expansion of existing schools, for loans, and for basic improvement grants. A year later this was in federal legislation. In 1965 the Coggeshall report supported the expansion of medi- cal education and the Millis Commission of the ama threw its support to the creation of family practice as a specialty. This period brought the evolution of the treatment of heart disease and the beginnings of transplantation of kidneys and other organs in the laboratory. We come to the era of Medicare and Medicaid and the Great Society (figure 4) and its evolution into the age of limited resources. In 1965 the legislation for Medicare/Medicaid was passed. In 1969 the Nixon Administration asked for a physician augmentation program, increasing the enrollment of physicians by 10 percent within nine months. In 1970 the Carnegie Commission suggested 6 Overview of Public and Private Policies a 50 percent increase in the enrollment and an increase in federal aid. In 1971 there was another Health Manpower Act with a capi- tation bonus and clauses to increase medical school size and to convert two-year medical schools to four-year schools. Because of the geographic maldistribution of our population, the National Health Service Corps was also established. In 1973 Caspar Weinberger, then Secretary of the Department of Health, Education, and Welfare, suggested there were too many future physicians in the pipeline. By 1975 there were over 70,000 foreign medical graduates (FMGs) in the system; one-third of the interns and residents were foreign medical graduates and one-fifth of practicing U.S. physicians came from abroad. Shortly thereafter, the renewal of the Health Professions Educational Assistance Act suggested that all schools had to absorb a certain number of Amer- ican FMGs from Guadalajara and the Caribbean Basin, but there was such a hullabaloo that the provision was dropped a year later. In 1975 a study of the surgical services of the United States began looking at the different types of surgeons and whether there was an excess or a shortage. A year later area health education centers, one of the suggestions of the Carnegie Commission, were developed. Between 1976 and 1980 the GMENAC Commission came out with its suggestion that we did indeed have a major surplus of doctors coming down the pipeline. In 1978 the Institute of Medicine suggested maintaining the status quo. By 1981 the capi- tation program had ended; no longer did we get federal support for our medical schools in that form. The physician/population ratio is another way to consider the question of physician excess or shortage. Prior to 1980 the physician/population ratio was highest in 1850 (figure 5), when there was one physician for every 571 people. In 1980 there was one for every 465 people. If you eliminate part-time practicing physicians, there was one for every 520 people. Now let us look at the major forces facing a change of any kind: the aging population will have to be considered in our reeommen- dations, as will the declining birthrate, which is already reflected in the number of applicants to medical schools; the change in the economy from the Great Society of 1965, where more was better and the approach was open-ended, to selection and substi- tuted choices of what the country can do; the unknowns in the scientific breakthroughs that will prevent the polios and cure the tuberculoses; the tremendous ability of our colleagues to produce bigger and better and more complex halfway technologies in trans- William G. Anlyan 7 Figure 5. Ratio M.D.s/Population. M.D.s Population Ratio 1850 40,755 23,261,000 1/571 1860 55,055 31,513,000 1/572 1870 64,414 39,905,000 1/619 1880 85,671 50,262,000 1/587 1890 100,180 63,056,000 1/629 1900 119,749 76,094,000 1/635 1910 135,000 92,407,000 1/684 1920 144,797 106,461,000 1/735 1930 153,803 123,188,000 1/800 1940 175,163 132,122,000 1/754 1950 203,400 151,684,000 1/745 1960 274,833 180,671,000 1/657 1970 348,328 204,879,000 1/588 1980 487,000 226,346,000 1/465 Sources: Historical Statistics of the United States Colonial Times to 1790; Statistical Abstract of the United States, 1984. plantation and heart disease and unknown progress in brain research and in the newer sciences just around the corner; and the increasing number of procedures available to nonsurgeons in the system at the present time. Other major forces are: the increasing number of women in medicine, which has brought a very fine change in sensitivity in medicine; the changing life-styles of our future physicians and younger physicians; the political force of UsFmMGs—we know how strong they are from what happened in the late seventies; the increasing political base of foreign medical graduates; the AMA, the AAMC, the Colleges, and the Boards; the AHA looming as a very major force in graduate medical education, which, to me, is the center of the continuum of continuing medical education today; the states, the federal government, and the business world, which some have dubbed the fourth party at the present time—a very major force; and the third-party insurers who are no longer a pass- through but who are getting into the act of policy decisions. My tentative conclusion in looking at the past is that the United States is a heterogeneous country with an open system of health care. We have a powerful private sector using influence, persuasion, and market economics and not solely government decree. Forty percent of our health care is government controlled; added to that 8 Overview of Public and Private Policies is the va system. Yet the responsiveness of medical care and health care is less dominated by a central government and its politics than is true in other developed nations. The US. health care system is a multisystem. Breakthroughs in biomedical research impact significantly at an accelerating rate. The only major gap in this multisystem is the health care of the uninsured poor. There have been sharp swings of the pendulum regarding whether we have too many doctors or too few. It is obvi- ous that when the aAmc and the AMa row together—particularly if they pull in the foundations and government—the system tends to change. So the questions we face among others are: acknowledging the historic pendulum swings of opinion and the unforeseen break- throughs of biomedical research, do we have a looming major physician surplus? Do we have a continuing uneven specialty and geographic distribution? If there are imbalances, should the correc- tions be addressed by universities, other private sector forces, the state or the federal government, or a combination thereof? Our ultimate objective in seeking a fair, unbiased trend in man- power development must be to provide appropriate, high-quality medical care for all Americans. Legislative Perspective PAUL G. ROGERS It is true that the Congress perceived a need for an increase in health manpower in the early 1970s. When a national need exists, the Congress usually reacts, as it did in this situation, with hear- ings and investigations, trying to bring before itself and the public experts to educate the nation on the problem. Many studies, the Carnegie as well as others, developed evidence that there was a shortage of some 40,000 to 50,000 doctors, and shortages of den- tists and nurses as well. The aaMc spearheaded the effort to allevi- ate this shortage, and I recall no significant groups that opposed the health manpower legislation. It was the general consensus of the scientific and medical groups that something needed to be done. While Congress often reacts to testimony from experts, that does not necessarily get the legislation passed. What really gets legislation passed is the fact that those congressmen and senators hear from their constituents at home about what is happening in their states and districts; in this case, example after example of shortages came forth from those districts and from the states. Some communities were advertising to get a doctor to come to their community. Many offered to build an office for them, build a clinic or even a hospital, and the Congress heard this. They heard that many people couldn’t get their needs met in their areas. In some rather affluent areas doctors were taking no new patients. Without question, the need for more physicians existed. This was not just a perceived need but an actual need, and the Congress was convinced about it. At the time we were considering health manpower, medical schools often had to use the fiction of getting research grants for their staff, although the research dollars often were used to sup- 10 Legislative Perspective port the educational activities of the medical schools. There was no other source of those dollars for education. It was the capita- tion program that finally supported education for education's sake so that the professor didn’t have to go to the research application to maintain himself and his activity. We were also beginning to set up distress grants at that time, and some sixty medical colleges made application for distress grants. Without question the need was well established, the Con- gress acted, and the medical universities responded in a magnificent way. Today things are changing so rapidly that before we come to a quick conclusion that we are going to have too many doctors we should look at the situation very carefully. If, for example, the Veterans Administration undergoes the change that has been pro- posed in taking care of veterans, a vast number of veterans will come back to the private sector for medical care. The Congress is now considering what it will do about future medical education. This too needs to be followed carefully. Historical Perspective ROSEMARY A. STEVENS, PH.D. The 1910 Flexner report, to which Dr. Anlyan referred in his presentation, made three points that I would like to challenge because I think they are relevant not only to the history of medi- cine in the United States but to our discussions as well. First, the report mentioned the right of the state to deal with medical education; second, it referred to the physician as a social instrument; and third, it defined the medical school as a public service corporation. I would like to suggest, first, that the role of the physician as a social instrument is dead. When Flexner talked in 1910 about the physician as a social instrument, he was talking in a climate of enormous faith in medical research and enormous change through advancements in medical science. The advancement of medical science was seen as a social good. The physician, as an instrument of scientific advance, was a social instrument and, indeed therefore, the reformation, upgrading, and building of the medical schools was automatically seen as being in the public good. Dr. Anlyan’s presentation very nicely portrayed the fact that many of the questions of medical science existing in 1910 were dealt with very successfully. New issues that we face today raise different questions of the relationships between medicine and what used to be called ‘‘the social betterment.” The physician is only one social instrument, a segment of the system and increasingly an employee. I think that we, like Great Britain, are moving increas- ingly into a whole spectrum of closed systems as far as physician employment is concerned. The employment market is changing. My second point concerns Flexner’s allusion to the medical school as a public service corporation. History shows that the medical school has been exquisitely sensitive as a public service 12 Historical Perspective corporation over the years, but perhaps not in the way that Flexner intended. It has been enormously responsive to outside funding, to outside stimuli, to external influences over the years. Medical schools have entrepreneurial faculties, a fascinating creative blend able to adjust very quickly to outside demands. The medical school has become a public service corporation but with very special characteristics. So much has the medical school become latched into outside influences of various kinds that we also need to con- sider the growing market in medical education, the relationship between the U.S. schools and the offshore schools, and new mar- kets for medical education. My third point is that history shows that it is an illusion to assume that there is a proper public function in the role and production of physicians. First, in terms of how many doctors do we need, who is “we’’? Is it government? Is it third party? Is it the foundations? Is it the system? There is no one ‘‘we’”’ and there never has been. Iam not sure that looking at the medical schools in terms of the numbers of physicians produced is as important as looking at what the constraints are going to be in the medical system in terms of employment practices and the specialty choices of indi- vidual physicians. Are we going to see major changes in medicine as there have been in dentistry, where there is a major change in the applicants’ pool? Are we going to continue to expect every- body who goes to medical school to practice medicine? Are they going to be locked into it anyway because of the amount of debt they incur in medical school? Are we going to be willing to put up with unemployed physicians? And what are they going to be doing in the American society of the nineties? Do we have a self-correcting system here in which people decide that going into medicine is really not worth it? History teaches us a great deal about the complexities of the issue, the difficulty of prediction and, perhaps most of all, that questions of physician supply are, at the root, really questions of power and money. GMENAC Revisited ALVIN R. TARLOV, M.D. I would like to divide my presentation into three parts. First, to review the GMENAC findings and update them; second, to use internal medicine as a paradigm of the dynamics that are taking place in health manpower; and third, to focus a bit on a third compartment which has developed since the GMENAC report and which is increasing in importance. The GMENAC Findings The GMENAC charge in 1977 by the Secretary of HEwW, Joseph Califano, was to advise the secretary on the need for physician services in the United States and how those needs could best be met in terms of specialty distribution, geographic distribution, and the financing of graduate medical education. The first question—how many physicians are needed—was by far the most complex and the most difficult to deal with. In the base year of 1978 there were 375,000 full-time equivalent, actively practicing physicians in the United States which was, according to our calculations, somewhat less than the requirement for physician services. GMENAC projected that by 1990 there would be an increase in the supply of physicians to 536,000 and by the year 2000 to 643,000 physicians. The assumptions made at that time, which drove the supply model, were five: first, the class size in U.S. medical schools; second, similar size in the class of the osteopathic schools; third, the return of foreign medical graduates to the United States for training and practice; fourth, women physicians’ productivity; and fifth, retirement assumptions. Reviewing these in 1985, I think the following conclusions can I4 GMENAC Revisited be reached. First, we were probably a thousand students per class high in estimating medical school class size. Since that has an effect on the entering of physicians for about an eight-year period, the estimate of 643,000 physicians in the year 2000 is probably too high by about 8,000 to 10,000. Second, the assumptions made in 1978 in regard to the enroll- ment in osteopathic medicine in 1988 are probably correct —about 1,868 in each class. That enrollment objective was established by the osteopathic schools in order to reach a total osteopath popula- tion of about ro percent of the total practicing physicians in the United States. The third assumption, having to do with the foreign medical graduate, is still difficult to unravel, but it would appear that our estimate of the entry rate of 4,100 per year of both US. citizens and aliens is close to the mark. Fourth, in regard to women physician productivity, we calcu- lated the total lifetime productivity of female physicians at 78 percent of that of their male counterparts. We were on target in projecting the number of women physicians in practice in 1990 and 2000. No information has appeared in the intervening years to alter that estimate. Fifth, regarding retirement age, we based our estimate on the experience of the late 1970s. We did recognize that the growing supply of physicians and the increasing malpractice insurance rates and other changes in the value system of our physician population would likely lead to changes in retirement age, particularly stimu- lated by the development of rras and Keogh retirement plans. It is a little too early for us to reassess that assumption. Based on those five assumptions I believe we are headed for about 630,000 full-time equivalent, actively practicing physicians in the United States in the year 2000. Now, on the question of requirements for physicians, we had to make a judgment very early on whether to base our modeling on a need-based model or a demand-based model. We elected to use a need-based model that is based on epidemiologic information and that models each medical condition in terms of its prevalence or incidence and multiplies that by the norms of care for that particu- lar condition, whether it requires hospital visits or visits to the doctor’s office or to other facilities. We derived the norms of care by taking a careful look at what they were in 1978 and then, using a panel of experts, predicting changes that likely would occur by the year 1990 and the year 2000. After taking the prevalence and Alvin R. Tarlov 15 multiplying it by the norms of care, the model divides that by physician productivity, assigns certain segments to nonphysician health care providers, and leaves the remainder for the physician. The assumptions in that model and the judgments that had to be made were recognized at that time as being rather risky. Since our charge was specifically heavily weighted toward the require- ments for physician services, however, we felt that we had to pro- ceed with that model. The important point is that the model was based on need rather than demand. We concluded, after calculat- ing the supply and the requirements, that the supply of doctors would exceed the requirements in 1990 and even more so in the year 2000. At that time, in 1978 and 1980, there were two compartments of medicine in practice. The first compartment, which was the larger, was the fee-for-service compartment, which comprised about 95 percent of all practicing physicians and cared for about that percentage of the US. population. The second compartment — the federal compartment—had about 18,000 or 19,000 physicians in military service, the Veterans Administration, the Public Health Service, and other federal positions, serving about 3 percent of the US. population. The characteristics of those two compartments were quite different. Physicians in the second compartment were salaried, hired, and paid by the federal government to provide services to those individuals in the employ of or retired from government service. The system lacked elasticity; the precise number of doc- tors in each specialty was prescribed. In the first compartment, however, there was a great deal of elasticity. Physicians had the liberty to practice where they wished, the fee structure was respon- sive to the demands of the physicians, and there seemed to be no limit to the absorptive capacity of this compartment to handle newly trained physicians in any number. Since that time, the data sources have markedly improved. The AMA data file and annual report continue to provide extraordinar- ily good information and are being refined all the time. The Educa- tional Commission on Foreign Medical Graduates’ information is also better now than it was in 1978. The information that Paul Ellwood has been collecting for almost a decade is proving to be extraordinarily useful. 16 GMENAC Revisited Internal Medicine Some of the dynamics in the health manpower field are worth looking at, and I am going to use internal medicine as a paradigm. For the past ten years we have been collecting information on internal medicine residency training each year—R-1 through R-5 —by U.S. medical graduates, by U.S. citizens who studied abroad, and by alien foreign medical graduates. There continues to be some growth in the number of internal medicine programs: 438 today compared with 432 two years ago. There also continues to be growth in the number of residents entering internal medicine each year: roughly 35 percent of U.S. medical graduates enter their first year of training in internal medicine. There is a rather sharp drop- off of about 18 percent between the first and the second year, which reflects the fact that a great number of future specialists take their first postgraduate year in internal medicine as prepara- tion for their subsequent specialty. Another 20 percent of all the house officers on duty at any moment are individuals who are taking a one-month or three-month rotation on the internal medi- cine service. Thus, about 45 percent of the first year of postgraduate training is provided on the internal medicine services, although in the end about 25 or 26 percent of all physicians are internists. This num- ber is expected to continue to grow until 1987 or 1988 because it annually reflects a similar proportion of the graduating class size. Since entering class size did not begin to level off until 1984 or 1985, this number will continue to rise until the graduating class size begins to level. Another important point is the distribution of U.S. and alien FMGs. Ten years ago 20 percent of the physicians-in-training in internal medicine were graduates of foreign medical schools but almost all of them, 18 out of the 20 percent, were aliens. In the intervening ten years, the growth of the schools in the Caribbean and in Mexico has changed that. Today alien foreign medical graduates have a predilection to increasing subspecialization while, as yet at least, the USFMG returning to this country for house officer training has in the last five years gone out into practice as a generalist. Today the UsFMGs also seem to be increasing in subspecialty training. When you put all this information together you can see that the number of U.S. medical graduates is continuing to rise. The num- ber of residents also continues to rise, as it will continue to do for Alvin R. Tarlov 17 another three years. The number of fellows in internal medicine subspecialization also is continuing to rise roughly in proportion to the total number of residents. If we look at the number of subspecialty fellows in any year and compare it with the number of third year residents in the previous year, we get an indication of how many internists are entering one of the fifteen subspecialty fields. The results of that calculation —which we call the subspecialization rate—show that about 60 percent of all internists have elected to study and ultimately to practice in a subspecialty of internal medicine. In the mid-seventies, the ratio was even higher. It seems to have leveled off in the current academic year. We have also collected data on the sources of residents’ stipends in 1976 and again seven years later, 1983—84. Today more of it is coming from hospital revenues and less from state and federal governments, federal training grants, and research grants. When one calculates residents’ stipends and converts the 1976 figure, using the consumer price index, to 1983 dollars, one immediately detects that residents’ stipends have actually declined by about 20 percent in real dollars in that seven-year period. We have seen a similar decline of about ro or 12 percent in fellowship stipends during that same period. Using our data, we can project into any future year the expected active U.S. internist population. We can also calculate the number of entrants into the practice of medicine and the number of exitors, using such standard actuarial techniques as the death, disability, and retirement rates. The net increase in each year reflects the number of entrants minus the number of exitors. Since the average duration of medical practice is about thirty-five years, the entry rate into practice for the 1980s, 1990s, and into the year 2000 depends on present and recent rates of entry into medical school plus the entry of FMGs, while the exitor rate reflects the rates of entry into medical school during the fifties and sixties. As a result of that thirty-five-year gap, the number of entrants will exceed the number of exitors until approximately the year 2015. In other words, the system will be in a positive net balance until that time. The Third Compartment When the GMENAC study was done the third compartment of medicine in practice—prepaid capitated care—was relatively small. The growth in this compartment for the past decade has 18 GMENAC Revisited been both profound and sustained. With about fifteen million enrollees in 1985 and a 15 percent annual growth rate, that num- ber will double in about five years. There are no signs at the present time to indicate that there either has been or will be a slowdown in the enrollment in HMos. This introduces a huge change in manpower considerations because this third compart- ment, like the second compartment, is inelastic and very lean. In the year 2000 one would expect 127,000 physicians to be employed in the prepaid plans giving roughly 1.1 physician per thousand population, with the physician per thousand people ratio being three times greater in the first compartment. On a world basis, this is a relatively high physician/population ratio although it has been achieved in some of the Western European countries already. Conclusions The supply projections made by GMENAC in 1980, based on five important assumptions, receive near-term partial validation in 1985. In the year 2000, there will be 630,000 full-time equivalent, actively practicing physicians yielding a ratio of 233 physicians per 100,000 population. The rapid growth in prepaid capitated care will enlarge the third compartment to near predominance by the turn of the century. Because the third compartment is fixed relative to the entry rate of physicians per size of population, an exaggerated, disproportion- ate distribution of physicians will occur. The residual physicians left for the first compartment in the year 2000 will have a physician/population ratio three times greater than in the third compartment, or a ratio of 334 compared with 106 per 100,000 population. For some of the specialized fields, the ratio will become five, ten, or fifteen times greater. Furthermore, the elasticity or the absorptive capacity of the first compartment is coming under sharp constraint. The need-based model used by GMENAC for estimating the requirement for physicians is now obsolete because it no longer fits the factors that are governing the system. A demand or eco- nomic model is more appropriate. The Physician Surplus: Another View WILLIAM B. SCHWARTZ, M.D. Expert opinions on needs are of interest but if you want to esti- mate a surplus or deficit of physicians it is necessary to know something about the demand as well as the supply side of the equation. For this purpose, standard economic forecasting tech- niques are the best tools we have available. We know there are a number of factors that influence demand— individual per capita income, technologic change, the aging of the population, insur- ance coverage of physicians’ services, and the growth of the population. Historical data can be used to estimate the effect that changes in each of these factors have on demand, and all of that can be aggregated to get a projection of future demand. This can then be assessed in relation to the estimated growth in supply. Some years ago, soon after the GMENAC report, Frank Sloan of Vanderbilt and I used this kind of forecasting model to look at what the physician demand-supply balance would be in 1990 and our conclusion was that demand would be slightly higher than supply, a very different conclusion from the need-based estimate of GMENAC. As of the end of 1983 the data indicated very much what our estimate showed: demand had grown by 5 or 7 percent more than supply, despite the fact that during that four-year period, supply had grown considerably. At least at the early period of the decade, this kind of forecasting technique appeared to have been reason- ably reliable. As a baseline for looking toward the year 2000, we did the same sort of thing, making rather conservative estimates. We assumed, for example, that there would be no increase in coverage for physicians’ services, and we made a rather conservative estimate of the growth of individual per capita income. Our demand-based 20 The Physician Surplus model suggests that if business were going on as usual and if none of the reimbursement or regulatory changes had occurred, we would anticipate a demand surplus of about 25 or 30 percent, that is, a more rapid growth in demand than in supply. This stands in rather sharp contrast to the GMENAC projection of demand-supply balance. Now, obviously, a business-as-usual scenario does not apply. The world has changed dramatically and, in 1985, I think we have to look forward to cost containment and regulatory and competi- tive strategies that will alter the picture from the one that this business-as-usual scenario yields. So, we have done some backing off from that projection. The first backing off we did was to project that there are 30 percent fewer hospital days, that the country as a whole becomes a gigantic HMO in its behavior toward hospital- ization, and that we assume a shift in care to the ambulatory sector from the inpatient side. We also assume a 15 percent demand reduction for physicians’ services, which would decrease our excess demand to ro or 15 percent in the year 2000. The second backoff we made was in relation to technologic change. Technologic change over the past few years has been con- tributing three to four percentage points in real growth in hospital expenditures. We have backed off and said suppose a third or a half of that is eliminated by cost containment and constriction of available capital. We start with 1985 as the base year and assume that over the period of 1985 to 2000 demand will grow by 15 or 20 percent less in the aggregate than it would have in the absence of capital constraints and various cost containment strategies. If we back off from our estimate by that additional amount, we are down to either a balance point between demand and supply or, more likely, a 5 or 10 percent surplus of physicians relative to demand. Finally, we made another estimate in which we assumed supply is growing by ro or 15 percent more than in the GMENAC projec- tion for the 1990 to 2000 period. That could bring our surplus of physicians to a figure of about 15 or 20 percent. If we take our best point estimate after the backoffs that I have suggested, we come out somewhere between zero—that is, balance —and a 10 percent surplus of physicians, which is not really a very striking change. However, given the uncertainties in these numbers and the time trends we are dealing with, the possibility that the number could be more like a 15 or 20 percent surplus is certainly a real one. Given that we think there will be probably at least a slight William B. Schwartz 21 surplus, it is interesting to consider what the beneficial and per- verse effects of such a surplus will be. The market is going to equilibrate. A surplus is not permanent. Things happen, and some of these things, from society’s point of view, are very good indeed. You can expect a shorter waiting time to get an appointment with a physician; a shorter waiting time in the physician’s office until you get to see him or her; and longer visits on the part of the physician with the patient. With pros and capitation there will be a drop in expenditures for physician services per visit, so that with fees falling and these time-price changes, the patient is the beneficiary and the physician is the loser. One of the additional advantages is that you do not have to put signs up in small towns asking doctors to come in or to build offices for them; competi- tion helps in distributing physicians as well. Now, on the side of adverse effects, the main concern is quality of care. As caseload per physician falls, what happens to quality of care? Do we have to think of that as a dimension that the market will not take care of for us but that the government or others have to deal with by specific interventions on entry or residency mix? There are really no good data on caseload impact per doctor. A recently published study used regression analysis to try to isolate the effect of mortality rate, per se, from those factors that might simply have correlated with it to see what the impact of volume alone was on the excess mortality rate. Using variables such as the geographic location of the hospital, whether it had residents, and so forth, the authors found that, with the exception of open heart surgery, the volume relationship could account for only o to 5 percent of the excess mortality. The second issue has to do with the quality of applicants enter- ing medical school. The ratio of applicants to acceptances is now two to one. That is lower than a decade ago, but higher than the 1.6 to 1.7 applicants per acceptance of the late 1950s and early 1960s. We still have some distance to go in terms of falloff in applications before we reach that level, and whether we should be concerned now about some further modest drop in quality or number of applicants is an open question. No one is complaining that the middle-aged physician of today, who was drawn from the smaller pool of applicants of the early sixties and late fifties, is not doing a competent job. To summarize, the best point estimate for the year 2000 is proba- bly a slight surplus of physicians. This probably will have socie- tally beneficial effects in terms of cost, waiting time, and diffu- 22 The Physician Surplus sion of physicians. A larger surplus is certainly a possibility, but it does not look as if it is going to be huge. Even if this surplus exists, the documentation of the negative effect still remains unclear. We do not know what a 15 or 20 percent surplus will do in terms of quality of care or effect on applicants. It seems wise to watch the progress in those areas and to begin more systematic studies of individual physicians and how the overall outcomes, not just inpatient mortality rates, are affected by lower volumes. If we see evidence of a deterioration of quality, intervention at that point will certainly be warranted as it will if the quality of our applicants clearly drops very sharply. Presently there is very little evidence that we should be inter- vening, at either the federal or the university level, in constraining entry of medical students. We should, however, be alert to the possibility of intervening later. If by chance society resists the kind of draconian cost containment efforts that are being undertaken, it is possible that demand will rise more than projected in my estimates. If we have restricted entry and the demand rises more than expected, we might find ourselves in a situation where we wish we had some of those doctors we did not allow to enter medical school. Market Forces and Geographic Distribution of Physicians JOSEPH P. NEWHOUSE, PH.D. By modifying a rather simple economic model we will be able to understand both what has happened with respect to physician location and what is likely to happen. This model (figure 6) is a little too simple, but it will serve as a starting point and yield some testable implications that will prove out. The top row of figure 6 illustrates a line that can be thought of as a highway with three towns on it; the towns have populations of five thousand, ten thousand, and thirty thousand. I am going to make some simplified assumptions. First, nobody lives outside these towns. Second, the average demand per person in each of these towns is the same. Third, each of the towns is equally attractive to live in to physicians. I am also going to make behavioral assumptions: first, that physicians locate in a way that maximizes the demand for their services and, second, that people who have to travel for care demand less of it. Now under those assumptions the first doctor into this little Figure 6. Physician Diffusion as Supply Grows: Towns of 5,000, 10,000, and 30,000 Population. Number of physicians of a given specialty Total number in towns with populations of: of physicians 5,000 10,000 30,000 Nee «OO: ‘© a oS ed "Or © AW YN of} YY we 4 Vv 4 Go 24 Market Forces and Geographic Distribution world of three towns (top row) goes to the town of thirty thousand. If he went to one of the other two towns, the people in the town of thirty thousand would have to travel and they would demand less of him than if he went to the town of thirty thousand. The second doctor (second row) also goes to the town of thirty thousand. Only when we have four doctors (third row) does a first doctor appear in the town of ten thousand. Finally, when we get up to nine doctors (fourth row) there is a doctor out in the town of five thousand. Thereafter as the number of doctors grows the towns gain doctors proportionately in this simple model. There are three testable implications that I want to draw from this little model. The first is that for any given number of doctors there is a critical town size; towns above that size have a doctor and towns below it do not. For example, in figure 6 the critical town size is between five and ten thousand when we have four doctors. Critical town size falls as the number of doctors increases sO as we increase the number of doctors beyond nine the critical town size decreases below five thousand. Second, if we look at figure 6 as different specialties instead of one specialty growing over time, and if the different specialties do not compete with each other, we will find, for example, the gen- eral surgeons out in smaller towns and the neurosurgeons in larger towns. Finally, if we regard this model as concerning one specialty that grows through time, smaller towns are going to gain physicians at a proportionately faster rate. For example, if we combine the towns of five and ten thousand populations and look at the change between four doctors and nine doctors, we would go from one doctor in those two towns to three doctors, a tripling, whereas in the town of thirty thousand we would only increase from three to six, or a doubling. This simple model was somewhat discredited in the late 1960s and 1970s, I think wrongly. These are the facts on which it was discredited. First, there is the well-known disparity in metropoli- tan and nonmetropolitan physician ratios, but this is not in and of itself inconsistent with the model. This disparity disturbed many people in terms of the access of nonmetropolitan people to physicians. Particularly in the late sixties, we were producing a lot more doctors, and almost all of them were going to metropolitan areas. Even in the seventies, when this disparity evened up and then disappeared, physician numbers were still growing a little bit faster in the metropolitan areas. Joseph P Newhouse 25 Using the AMA master data file in 1970 and 1979 we looked at towns in twenty-three rural states that have a population of more than 2,500. Forty percent of towns of five to ten thousand had an internist in 1970 and that had grown to 52 percent in 1979. The first implication is that there is a critical town size: towns above it have a physician, towns below it do not. That is not exactly observed, however, and the reason it is not is because my simplify- ing assumptions are not exactly correct: there are populations outside towns; towns are not equally attractive to physicians to locate in; and the demand per person in different towns is not the same. What that means is that there is a tendency for larger towns to have a physician of a given type. Another implication is that the larger specialties will be further out in the smaller communities, but that the smaller towns are less likely to have certain subspecialists. Everything that I have said up until now has been premised on the notion that the various specialties do not compete with each other, but in fact they do and this has two kinds of consequences. First, consider the competition between an internist and a general practitioner located along a highway with the population uni- formly distributed along this highway. If the population cannot tell the difference between the two physicians or if they do not care about the difference and if physicians who are referring also do not care about the difference, then I would draw a line about halfway down the highway and people on one side of the line would go to the internist while the people on the other side of the line would go to the general practitioner. If, however, the internist faces more demand, then I would draw the line closer to the gen- eral practitioner and the internist would get a larger share of the population. The implication of this is that where the internist goes head-to-head against the general practitioner in a city and tends to win that competition, the general practitioner is going to be more willing to go to a smaller town than the internist. The same would be true if the board-certified physician tends to win against the noncertified physician, or if the usmMc tends to win against the FMG. There is a second phenomenon that exists with this interspecialty competition: specialists such as internists and pediatricians pro- duce some services that general practitioners either do not pro- duce at all or tend not to produce. The internist with unique services is like that first specialist into the town of five thousand, ten thousand, and thirty thousand: he goes into the larger town but once he is there he also produces basic services. Seeing the 26 Market Forces and Geographic Distribution internist in the larger town, the general practitioner tends to go toward the smaller town. The result of most of these phenomena is that small towns tend to have almost all general practitioners. In towns with fewer than 25,000 people, a majority of the physicians are general practitioners. Even in cities with 25,000 to 50,000 population, a near majority of the physicians are general practitioners. In the large metropolitan areas, however, only around 20 percent of the physicians are general practitioners. During the sixties and seventies there was, of course, a decline in the number of general practitioners. The family practice pro- grams came into being in the seventies but they did not produce enough graduates in the seventies to offset the decline of the general practitioner, although they started to by the 1980s. The specialties, meanwhile, were growing quite rapidly in the 1970s and growth is clearly continuing in the 198os. Indeed all specialties are going to grow in the 1980s. The drastic decrease in the number of general practitioners being produced greatly affected the small town physician supply and we saw the beginning of advertising for physicians. The critical town size for the specialist was falling, but it was not low enough to compensate for the decline in the number of general and family practitioners. Around 20 percent or so of the population does not live in towns of 25,000 or more. In fact, most of those people do not live in towns at all. The question is, How far do those people have to travel to reach the nearest physician of a given kind? In 1979, 70 to 80 percent of the rural population was within roughly half an hour of an internist. Almost half the population in 1979 was within thirty straight line miles of a neurosurgeon. One of the problems that we run into when we look at the question of medical care for metropolitan versus nonmetropolitan residents is the assumption that all the nonmetropolitan resi- dents get their care from nonmetropolitan physicians. For some of them, however, the closest physician is in a metropolitan area. When looking at physician/population ratios, those people need to be taken out of the denominator of the nonmetropolitan ratio and moved over to the metropolitan ratio. That does not much affect the metropolitan ratio but it does affect the nonmetropoli- tan ratio. The important point is that physicians do seem to choose their Joseph P Newhouse 27 location in accordance with standard economic models. As all specialties increase in numbers in the 1980s and 1990s, we can expect the critical town size to fall still further and the numbers of physicians in the rural areas to increase still further. Correcting ‘‘Surpluses” and ‘‘Shortages”” in Medical Specialties I. The Surgeons C. ROLLINS HANLON, M.D. Fifteen years ago in the cutting edge of developing physicians—in the residency area—the surgeons constituted slightly over 40 per- cent of all residencies whereas primary care physicians—not including obstetrics-gynecology —constituted 25 percent of all the residencies. In 1983—84, the figure for surgeons is down to 28 percent of all residencies, whereas for primary care it has risen to almost 43 percent. We have seen an inversion of the manpower situation of fifteen years ago. When the prophets on manpower have so obviously missed the mark thus far, why are we to assume that their projections will be any more accurate this time around? We all know that it takes five years or more before it is obvious that events have overrun fore- casts in the most predictable way; no one seems capable of foretell- ing the future in manpower. The historical landscape is littered with the burned out reports of eminent commissions and leaders of American medicine who embarked on fortune telling and suc- ceeded only in demonstrating that their crystal ball had a defect in its lattice structure. What can an international organization of some 55,000 sur- geons do about surpluses and shortages? What we should do first is to change what Paul Rogers has called the public perception. Changing that perception of the public at large—including the media, academic and other planners, the various segments of local and federal government, the insurance industry, the business com- munity whether directly involved or indirectly related to health, and the organized patients such as the 18 million individuals in the American Association of Retired Persons and other consumer groups—is a formidable and probably an impossible task, particu- The Surgeons 29 larly if we are not certain whether we can project a surplus or a shortage. Where are the possible points of influence by the American College of Surgeons? First, we can analyze and in some instances collect data on surgical services, both operative and nonoperative. We can attempt to determine whether these services are being provided by real surgeons, by surgical pretenders among physicians, or by nonphysicians. We can track the number and activities of surgical residents, the growing edge of surgical manpower resources. The College has always had a strong interest in and a substan- tial influence on the quality of surgical education. For the past several years, we have resumed the maintenance of accurate data about individuals in the surgical residency system, placing it in context with other databases of the Association of American Medical Colleges, the ama, the National Resident Matching Pro- gram (NRMP), and the surgical specialty societies themselves. Not only do we know accurately where these young men and women are but we have individual reports from some 8,000 of them about what they are doing in the way of caseload and activity. The acs has representatives on various boards and on the Resi- dency Review Committees where, with due regard for quality, fair- ness, and the brooding presence of the Federal Trade Commission, a monitoring effect may be exerted on surgical quality, with pos- sible secondary effects on the numbers of programs and then on the total number of individuals who emerge from surgical graduate medical education. This is a very sensitive area; if the residency review committees and the boards begin to deliberately exert surgi- cal birth control on those programs, I think they run a great risk of being under the gun of a Federal Trade Commission summons. Surgical program directors are concluding on their own that it is desirable to restrict the number of physicians. They may carry out this restriction either by instituting a pyramidal system—under which individuals are displaced from the pyramid at the second, third, or some higher level —or by truncating it deliberately from the outside by nonfunding of physicians at some arbitrary level. Those of us who are convinced that you cannot educate a surgeon adequately in, say, three years after medical school, find the rigid prescription of arbitrary time intervals more offensive than cut- ting down on the number of individuals allowed into the program. On the one hand, we end up with a reasonable number of inade- quately educated surgeons—I am not saying ‘‘trained’”’ surgeons although training is a component there; on the other hand, we 30 ~©Correcting ‘‘Surpluses” and “Shortages” have an inadequate number of reasonably well educated surgeons. In each instance, the decision will be based on a preconception about the proper proportion of various specialists in a presumed ideal mix that is established by something akin to a need-based formula. These formulas may not last, however. They are strongly influenced by the backgrounds and prejudices of those who are involved in their construction. I come back to the starting point of the American College of Surgeons, which is quality, and our underlying conviction that education to provide the highest quality of surgical care in every individual who comes before us will ultimately result in better care for the patient. While this does not absolutely prove that there is a direct relationship between high quality of education and delivery of better care, I believe quite strongly that that is the case. Should we, if we have the influence and the chance, cut down on the number of surgeons or press for expansion? I think we will leave that question unanswered. II. The Physicians ROBERT H. MOSER, M.D. The position of both Colleges is quite similar. We really do not have any direct impact on the manpower situation. We are, of course, vitally interested in the subject, and we do participate in such things as the residency review committee, and we have an informal relationship with our board, the American Board of Inter- nal Medicine. I think it goes without saying that we have primed the pump a little too effectively. If we were to shut down all the medical schools tomorrow, we would still apparently have an overall surplus of doctors by the year 2000. I stress the word ‘‘overall.’ In my mind, by far the larger issue is the geographical, specialty, and subspecialty distribution. There may well be too many psychiatrists in Beverly Hills. There may well be too many cardiologists in Boston. But are there enough psychiatrists in El Centro and are there enough cardi- ologists in Laramie? If not, how do we equilibrate the situation and get them to switch around? What impact will telecommunica- tions—the capability to have consultation via computer—and the expanding new technologies have on rural practices and the distribution of specialists and subspecialists across the country? The Physicians 31 What impact will the growth of other health care practitioners have on physician manpower? All of these factors are most difficult to assess. From the aspect of internal medicine, the subject has been dis- cussed at great length. It is the subject of a major position paper and has been bandied about at the Federated Council of Internal Medicine. At the moment we have no solutions, but there are some points worthy of discussion. There are still about 35 or 36 percent of young people out of the 16,000 annual medical school graduates who go into internal medicine as pc-1s. About 61 per- cent of these physicians then go into the subspecialties, leaving about 39 percent who go into primary care internal medicine. We have a feeling that this number might be inadequate for the future. There are several mechanisms whereby we can try to slow down the flow of personnel. First, what can the American Board of Internal Medicine do? As with the surgical board, they could arbi- trarily increase the cut-off score for certification or they could make exams much tougher. They could make it more difficult to get a certificate of special competence; they could simply produce fewer certificates both in internal medicine and in its subspecialties; or they could simply increase the time between graduation from medical school and the time when one can sit for the ABIM. Now one can sit for the board three years out of medical school and take the exam for the certificate of special competence within another two years. It can all be compressed to a total of four years. Lengthening this period of time would certainly slow down the pipeline. We are all pretty much opposed in principle to arbitrarily increasing the flunk rate as a means of control. How- ever, there are many of us who think that the time frame could be lengthened, both for aBim and for a certificate of special compe- tence certification. Second, what about the rrcim? A residency review committee could certainly set tougher standards for internal medicine resi- dency programs. Now that all the subspecialty programs are com- ing under our purview, we could do the same for them with the blessing of the Accreditation Council on Graduate Medical Educa- tion (ACGME). Both the rrciM and the aBIM have refused to become vehicles for the solution of this vexing social and economic problem, and I endorse that and so does our College. The rRcIM could tighten the screws. They could put perhaps 30 percent of the marginal programs that they see on probation if 32 Correcting ‘‘Surpluses” and ‘‘Shortages’’ they decided to get a great deal tougher. In addition, the ACGME can arbitrarily say that they will not accept individuals for post- graduate medical education unless they come from LCME medical schools, a very difficult political decision from many aspects. Iam inclined to agree with the stance of the aBim and the rrc that they cannot afford to become politicized, not only from the aspect of conflict with the rrc but also because they still stand as major bastions of excellence. Third, what could program directors do? Those in the Associa- tion of Professors of Medicine, those in the Association of Program Directors in Internal Medicine, and those who run subspecialty fellowships could tighten the screws and decrease the number of available slots in their individual programs. But can you expect individual program directors to act either positively or negatively in response to a social need? They have never done this before. They have mandated the number of participants in their programs on the basis of resources, patient needs, equipment, and faculty, and they have never yielded to the necessity to respond to social responsibility. I am not sure they could if they had to, unless driven by some external force. Although there has been some cutback in some programs, there is also an acceleration. In 1984, we had 1,518 individual sub- specialty programs in internal medicine; in 1985 it is up to 1,557. Rather than decreasing, it has increased by some thirty-nine programs. There are a few less cardiology programs and a few more oncology programs, but basically there has been only a little change. Predictably, however, there will be a change: there will be a cutback. Indeed, there will be a major change in many programs if the Part B Medicare reimbursement for direct resident pay actually comes to pass, and most of us think that will happen. Many programs will be obliged either to cut back or to become very creative in seeking new sources of funding for graduate medical education. In addition, the personnel flow is sensitive to other sources of federal and state funding. The flow of primary care residents in family practice, internal medicine, and pediatrics was expanded through specific federal granting programs. We were concerned about that because it possibly served to the detriment of other residency programs. It is conceivable that in the new world the control could become even more exquisite, with selective federal or even state funding of specific undersubscribed subspecialties The Physicians 33 under payback provisions that would oblige recipients of this lar- gesse to serve where needed for specified periods, somewhat analo- gous to the National Health Service Corps. This possibly could succeed because the need is greater. This mechanism also has precedent in the private sector. Some communities lacking a physi- cian have sponsored young people all the way through medical school and residency training just to ensure that they will have a physician in the community. The same mechanism will also occur in industry and other organizations seeking to employ physicians with specific specialty and subspecialty skills. There may be a time when we will accept foreign medical gradu- ates with the specific intent that they be trained here and then returned to their country. It may well be that the State Depart- ment could use this as a foreign aid program. Wouldn’t it be help- ful if we could continue to expand residencies and provide a sup- ply of well-trained individuals for countries that needed them? A final mechanism that we have heard about would be through the market place, as some equilibration occurs through reimburse- ment revision and policy. Perhaps there will be fewer ophthalmol- ogy or cardiovascular aspirants as financial incentives decline. This alone might increase the attractiveness of primary care careers. Thus, there are many factors with the potential for influencing the numbers, both in absolute terms and with respect to geographi- cal and subspecialty applicants. At the American College of Physicians, we will continue to monitor the physician supply while ensuring that the high quality of internal medicine in this coun- try will continue. Life-style Choices and Evolving Practice Patterns BRITAIN NICHOLSON, M.D. There is really no question that we are in the midst of a very dramatic change regarding physician life-styles and practice patterns. Physicians are restructuring their practices to allow greater flexibility and more predictable hours; residents are asking for more personal time, primarily in the forms of maternity and pater- nity leaves; and medical students are electing careers that allow for more time for personal relationships, balanced with sufficient secure salaries to allow them to pay back ever-mounting and stag- gering educational debts. What this means for the late eighties and the nineties is really a trend toward shift work—salaried posi- tions with predictable hours in institutional settings—or else the more procedure-oriented subspecialty careers that for now will give a much higher reimbursement and much higher income prediction than the more primary care-oriented subspecialties. At the risk of being too simplistic and sounding like a funda- mentalist minister, I think that the two forces responsible for this evolution are women and money. It would be a mistake to assume that this is just a women’s issue. This issue really affects men and women equally, but there is no question that the relatively recent entry of women into working and professional roles has influenced the current trends we perceive. Female physicians have increased dramatically in number in the past decade. In 1983 women com- prised 32 percent of all medical students nationally and 25 per- cent of all residents. Thirty-three percent of these women were married and ro percent were engaged at the time of graduation. These women are graduating in their mid to late twenties and anticipating residency and career building during the height of their childbearing years, and this necessitates a greater flexibility in life-style and in practice pattern. Britain Nicholson 35 The 1984 AMA report on maternity leave for residents found that 26 percent of female residents have children and 63 percent of practicing female physicians have children. The resident number is probably a low estimate given that the pool of respondents was generally an older group of women who were active in the AMA. Nevertheless, 45 percent of women with children had their first child during their training and 27 percent of women had at least two children during their training. At the time of this survey, four-fifths of women reportedly did not adjust their schedules before or after delivery and only took a two- to eight-week maternity leave, tending toward the two-week side, and I submit that this is just plain unhealthy. However, this survey probably does not reflect the informal ways in which women shave off hours here and there to fulfill their responsibilities. As life-style choices are more explicitly discussed, part-time roles will definitely become more explicitly defined. A second important fact is the increasing number of married male residents and medical students. In 1983, 41 percent of gradu- ating male medical students were married and 9 percent were engaged. This is in bold contradistinction to the pattern of earlier years when bachelor house officers virtually lived in hospitals. Because of the increase of women in the work place, many of these marriages are dual-career marriages. Eleven percent of these men are married to women with doctorate level education and 47 percent of the women are married to men with doctorate level education. As dual-career families, many of these graduating stu- dents are more explicitly including family concerns in their career decisions. At Harvard Medical School, a nine-week workshop is offered to students on the impact of a demanding medical career on per- sonal relationships. It is always filled with its quota of twenty students, often all couples. The major concerns of these couples are when to get married, when to have children, and what will be the institutional reaction to maternity leave or schedule adjust- ments for the family. They express concern about how they can meet the expectations at work of themselves, their peers, their mentors, and their patients and, on a very personal level, their societal expectations for themselves as spouses and parents. One of the faculty leaders of this course has the impression that stu- dents who are married or plan to be married are increasingly openly rejecting careers that do not allow flexibility, such as the subspecialty-oriented surgical practices or solo or small-group pri- mary care practices. 36 Life-style Choices Many who desire academic careers fear that they cannot com- pete as they face the traditional triple threat in teaching, patient care, and research as well as maintaining a healthy family life. Those who choose a generalist’s role choose an institutional or large-group setting with salary, controlled hours, and infrequent on-call demands. If they anticipate a twenty-five year period of children’s expenses, culminating in the backbreaker—college —they want to earn enough money while they are working to manage these expenses. Money as defined by educational debt, income, and future reimbursement policies are all having their effect. Though educational debt is rising, it currently does not emerge as a strong predictor of specialty choice or practice setting. In fact, in a recent survey carried out by the aamc for the Department of Health and Human Services (HHs), the amount of debt as a predic- tor of career choice—the career choice being high pay versus low pay—ranked seventh. The two top predictors turned out to be sex and whether you attended a public or a private medical school. Future reimbursement policies will also have some effect on career choice. The procedure-oriented subspecialties are currently very attractive, primarily for their high income. The eighties and nineties will see increasing numbers of men and women choosing, first, shift work because of flexible hours; second, large-group or institutional settings; and third, subspecialties with reimburse- ment procedures. I think there will be a decreasing number of physicians choos- ing entrepreneurial activities like solo practice, because of its financial risk and inflexibility. This will culminate in a continu- ing decrease of those individuals going into primary care or gener- alist roles. Market Influences of EMGs SAMUEL P. ASPER, M.D. Twenty-five years ago our nation addressed the problem of a physi- cian shortage by fostering both development of new schools of medicine and enlarged enrollment of students in existing schools. To date these advances have resulted in the production of about 60,000 more physicians than would have graduated under former conditions. Concomitantly, our government permitted, and to some extent encouraged, the immigration, naturalization, and licensure of qualified alien physicians, who today number about 100,000. This latter approach to solving our health manpower needs was abetted by strong market forces, as alien physicians saw opportuni- ties for job and financial security in the United States. By the mid-seventies the number of physicians from abroad achieving licensure annually nearly equaled that of graduates of U.S. schools, and suddenly the portent of a physician surplus was envisioned. Our Congress, in writing Public Law 94-484 in 1976, stated: ‘There is no longer an insufficient number of physicians and surgeons in the United States’ and discontinued the granting of immigration preference to alien doctors. A decrease in immigration followed. During the past decade, however, many more U.S. citizens began the study of medicine abroad than formerly, and a steady annual increase in the number of U.S. citizen FMGs obtaining licensure ensued. The ama Division of Survey and Data Resources! indicates that in 1983, of 519,545 physicians in the United States, 112,005 were FMGs. Of these, 80,044 were aliens or former aliens and 31,961 were U.S. citizen FMGs. Thus, FMGs, with a 22 percent repre- 1 AMA Physician Masterfile, Data Release Services, Division of Survey and Data Resources (Chicago: American Medical Association, 1985). 38 Market Influences of FMGs sentation, clearly influence the U.S. professional marketplace. Some complain the market is disturbed, others believe it is benefitted, but all agree that the role of FMGs is significant and that new issues result when one in every five physicians among us attains professional status by a route not familiar to most of us. The tracking of FMGs in the marketplace is not an easy task. The Immigration and Naturalization Service has discontinued its valuable practice of collecting data on immigrant physicians. There is no central clearinghouse on the total number of exchange visitors, as approximately 250 different organizations can and do sponsor them. Also difficult to track are usFmGs. Moreover, little is known about the hundreds of FMGs who never achieve ECFMG certification and are denied entry into graduate clinical training; as unlicensed physicians, do they participate in paramedical activities? In this presentation I shall highlight briefly some market influ- ences of foreign medical graduates. Equally important to note, however, is the fact that the market also influences foreign medi- cal graduates; one perspective requires the other. Four areas will be covered: foreign medical graduates in graduate medical education, medical practice, academe, and international medicine. In graduate education it is predicted that economic and other factors in hospitals may soon force a reduction in the number of residencies. To date, however, no decrease has occurred, but there are striking changes in the number of alien and usFMGs entering residencies. Although the number of applications from alien physi- cians is increasing sharply, the percentage of those obtaining posi- tions is decreasing; in contrast, the number of usEMGs is steadily increasing. It appears that program directors increasingly are giv- ing preference to USFMGs for appointment to residencies, despite the oft-voiced criticism of the educational programs at the off- shore medical schools. In 1983 the 13,221 FMGs in residency pro- grams comprised 18.4 percent of the total number of residents. Of the 13,221, 6,231 were alien FMGs and 6,990 were USEMGS.” Recent studies have confirmed that FmGs enter residencies that are less attractive to graduates of US. schools, chiefly physical medicine, anesthesiology, pathology, therapeutic radiology, nuclear medicine, and psychiatry. Accordingly, FMGs assist these underserved specialties. 2 Data from Physician Characteristics and Distribution in the U.S., 1983 Edition (Chicago: American Medical Association, Division of Survey and Data Resources, 1984). Samuel P Asper 39 In the practice of medicine alien FMcs also fill underserved regional and specialty areas. They practice among the urban poor, such as in some of the boroughs of New York City, and in sparsely populated rural areas, such as in West Virginia. Many anecdotal reports describe their beneficial roles. The usFmcs, however, appear to follow practice patterns of graduates of U.S. schools. Moreover, there is some evidence that alien FMGs, after working several years in underserved areas, move to the traditional patterns followed by US. physicians. The extent to which alien FMGs participate in academe may come as a surprise, for new data from the aamc show that 15.7 percent of all full-time positions in medical schools are held by alien rmcs.° The distribution by rank shows that 27.6 percent are professors, 26.2 percent are associate professors, 35.7 percent are assistant professors, 9.3 percent are instructors and below, and the remainder are in the category of ‘other.’ The explanation for significant participation of alien FMGs in academe is, I believe, that those FMcs who perform well in residency posts are subse- quently retained by departmental chairmen for teaching, patient care, and research positions. Few data are available on the extent of FMG participation in research. AAMC faculty roster data show that 69 percent of full- time FMG faculty are involved in research on at least a part-time basis. Among appointees to the staff of the National Institutes of Health, 50 percent are foreign nationals who studied medical sci- ence abroad. My unsubstantiated opinion is that UsFMGs in full- time research are few and that the number of alien physicians is far greater than is generally appreciated. An examination of the December 1984 issue of the Journal of Clinical Endocrinology and Metabolism provided informative data on the research activities of FMGs. Of thirty-four scientific articles, twenty came from medical centers in the United States and four- teen from abroad. Of ninety-one contributing authors and coau- thors to the twenty papers from U.S. medical centers, thirteen (15 percent) were alien physicians holding EcFrMc certification. Not to be overlooked is the influence of FMGs in the international, professional, and economic marketplace. Alien foreign medical graduates, following graduate study in the United States, benefit our economy through their purchase of medical journals and texts, 3. AAMC Faculty Roster System (Washington, D.C.: Association of American Medical Colleges, 1985). 40 Market Influences of FMGs medical equipment, and pharmaceuticals. Moreover, many physi- cians who have never studied in the United States hold American medicine in high regard. For example, ECFMG certification is proudly displayed as evidence of meeting a U.S. standard, and qualifying for election to membership in one of our medical spe- cialty societies is a high honor. Also, as mentioned, fourteen, or 4o percent, of the papers in the December 1984 issue of the Journal of Clinical Endocrinology and Metabolism came from abroad, showing, among other things, the interest of foreign investigators in publishing in a U.S. journal as well as the acceptability of their research for publication. Conversely, Americans studying medicine at proprietary medi- cal schools abroad have a significant effect on the economy of the nations in which these schools are located, especially the smaller nations in the Caribbean. Tuition at these offshore schools can be as high as $15,000 annually, and the estimates of the number of Americans enrolled in these schools range from 8,000 to 15,000. Fluctuations in the FrMc marketplace characterize events of the past twenty-five years. Educational opportunity, work availability, and job security are the major contributing factors that determine entry of both foreign national physicians on either temporary or permanent status in the United States and Americans studying medicine abroad who seek to enter practice upon their return. Legislation also has produced fluctuations, from former free entry of qualified alien physicians to controlled entry at present, to pro- posed regulations that would limit the number of residencies in US. medical centers to graduates of LcME-accredited schools. This is indeed an important moment to reassess our role in international medicine, to evaluate the impact on our health care system of increasing numbers of returning Americans who have studied medicine abroad, and to reconsider our responsibility to global medicine. To the greatest extent possible, we must wisely attempt to create and guide market forces so that our superb edu- cational and research system becomes an effective tool in interna- tional diplomacy and goodwill. Methodological Problems in Assessing Physician Demand, Need, and Supply: Policy Implications UWE E. REINHARDT, PH.D. The question that rises in my mind is why are we so concerned about physician manpower in general? I will focus on methodologi- cal issues in health manpower forecasting, which is a can of worms itself. What one is asked to do in a manpower forecast is begin with a population which begets the physicians and also begets the clients of the physician—the patients. Right here one has a problem. First of all, one must know about the immigration of physicians, which is a political issue; then one needs occupa- tional choice models, not only to know how many people in the potential pool choose to become physicians, but also their spe- cialty choices. There are many ways to intervene in manpower planning. One could, for example, raise tuition. Assume that every medical stu- dent would have to borrow $25,000 at the beginning of the four medical school undergraduate years. Then charge ro percent inter- est on the money with no repayment during the residency. In the ninth year out, they would begin to amortize that debt, which, by that time, would be $200,000. While that seems like a big number, in fact it is not if you amortize it over twenty years at Io percent, like a mortgage. The annual payment is then about to percent of a physician’s gross income. That is one way one might influence the supply of physicians. On the demand side, we have epidemiological models that give us morbidity patterns. The translation of a morbidity pattern into demand for services is quite complicated and may, in fact, depend on how many physicians there are. At least some of us believe that there is a potential on the part of the physicians to manipulate demand. Given that you have a demand for services and a supply of physicians and some model that translates human bodies into 42 Methodological Problems output of services, what we would really like to know is are the supply and demand of medical services in balance? The supply of physicians depends on the number of living physi- cians multiplied by the number of those who are professionally active, which depends, for example, on retirement patterns and on how often physicians drop out during childbearing years. This is multiplied by the number of physicians who give patient care. The resulting sum is the supply side of the forecasting equation (figure 7). On the demand side you need a population forecast, even though they are not always accurate. If you know the number of services needed, you divide that by the number of services delivered per physician to get the number of physicians needed. This productiv- ity parameter depends on hourly productivity times number of hours worked per week times number of weeks worked per year. All of these factors can and do vary. According to AMA data, the number of hours physicians work has decreased over time. To show you also the tenuousness of the link between man- power and services, in the seventies in New England there was a population/physician ratio of 161:1, in the East South Central states of only 95:1. Despite this disparity, visits per doctor were roughly Figure 7. A Simple Forecasting Equation for Physician Manpower. Percentage Percentage Projected Projected Projected _ Size of the of living of active number of average phy- percapita population physicians physicians physicians sician pro- demand for tobe who are rendering alive at ductivity at physicians’ served at profession- patient care time t time t services at timet ally active time Projected [ Ja surplus (X >o0) —————. K&, = a * S, -__ * N, or deficit (X