National Academies Press: OpenBook
« Previous: 1. Introduction and Summary of Numerical Recommendations, Fiscal Years 1982-1985
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 23
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 24
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 25
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 26
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 27
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 28
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 29
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 30
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 31
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 32
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 33
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 34
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 35
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 36
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 37
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 38
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 39
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 40
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 41
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 42
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 43
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 44
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 45
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 46
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 47
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 48
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 49
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 50
Suggested Citation:"2. Clinical Sciences." Institute of Medicine. 1981. Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report. Washington, DC: The National Academies Press. doi: 10.17226/9917.
×
Page 51

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

2. CLINICAL SCIENCES The Committee def ines clinical investigation as research on patients; on samples derived from patients as part of a study on the causes, mechanisms, diagnosis, treatment, prevention, and control of disease; or on laboratory animals by scientists identif table as clinical investigators on the basis of their other work. As a mode of scholarly inquiry, clinical investigation takes many forms. It may consist of the systematic observation of individual patients in a controlled environment or large-scale clinical trials involving thousands of subjects and the pooled effort of investigators in multiple institutions. It includes carefully controlled studies to elucidate the mechanisms of disease, or epidemiolog ic research to uncover leads as to the etiology of disease. An essential aim of clinical investigation is to translate new knowledge, through applied research, into new technology and modes of treatment, as well as the validation of new technology through clinical trials. The clinical investigator] generally teas an M.D. or other health professional doctorate. The Committee recognizes that basic scientists also partic mate in clinical investigation. As part of the clinical investigation team, they have a multifaceted role that may range f ram studying isolated tissue components derived from patients to close collaboration with a physician-investigator in research requiring experimental manipulation of human subjects and their environment. Moreover, they often help to ensure the application of state-of-the-art technology to clinical investigation. While such participation is to be encouraged, there will always be an overriding need for the physician investigator who is uniquely prepared to recognize research opportunities presented by human disease, to formulate researchable questions, to design experiments for answering those questions, and to perform personally the work entailed in those experimental designs. In many cases it will be necessary to study patients intensively, permitting the clinical situation to guide the nature of the questions, as well as the manner of seeking their answers. In addition, legal, ethical, and regulatory constraints establish the primacy of the physician in clinical investigation. In short, the physician investigator has a very special role both in bringing clinical insights to bear in the laboratory and in translating new knowledge into the context of medical practice. Given this irreplaceable role, the Committee reiterates its concern over the precipitous decrease in postdoctoral training of physicians. As can be seen in Table 2.1, the number of persons with a health professional doctorate receiving training under NIH training programs has declined in both absolute numbers and as a percentage of total postdoctoral trainees/fellows supported by NIH from 1911 to 1979. While some of the decline can be attributed to the elimination of predominantly 23

TABLE 2.1 Distribution of NIH Postdoctoral Trainees and Fellows by Degree Type, 1971-79a Year Total M.D.,sb Ph.D.'sb 1971 # 7,532 4,631 2,901 % 100.0 61.5 38.5 1972 # 7,AgS 4,466 2,927 % 100.0 60.4 39.6 1973 # 5,466 3,612 1,854 % 100.0 66.1 33.9 1974 # 6,313 3,522 2,791 % 100.0 55.8 44.2 1975 # 5,957 2,874 3,083 % 100.0 48.2 51.8 1976 # 4,884 1,950 2,934 ~ 100.0 39.9 60.1 1977 # S,~gS 1,906 3,389 ~0 100.0 36.0 64.0 1978 # 5,683 1,954 3,729 ~0 100.0 34.4 65.6 1979 # 5,613 1,923 3,690 % 100.0 34.3 65.7 aThese data represent individuals who actually served in NIH-supported traineeship or fellowship positions. Thus these counts may differ slightly from those shown in Chapter 1 which represent awards, not individuals on duty. Includes Fogarty Inten~abonal Center programs. b`'M.D." and `'Ph.D." Also include equivalent doctorate degrees. Persons with bow M.D. and Ph.D. degrees are shown under 'M.D.". SOURCE: National Institutes of Health (1966-81). clinical training programs early in that period, the decrease in more recent years probably reflects a diminished interest in research careers On the part of the young physician ~ see p. 28) . The Committee believes that the dec line has serious implications for the f uture of patient- centered and disease-oriented research. 24

THE TRAINING OF PHYS ICIAN INVESTIGATORS With the movement of clinical medicine "from anecdotal empiricism to an approach based on analysis and experimentation, n clinical investigation has become the bridge between basic and clinical science (Tinier, 1980~. In that process it has become enormously sophisticated and complex. It is appropriate therefore to ask how the young physician acquires the diverse scientific and clinical skills needed for a successful career in clinical investigation. Clinical research training is heterogeneous in nature, duration, sources of support, and degree of formality of the training experience. The newly trained physician investigator is not the product of a distinct course of study, such as the curriculum leading to the M.D. degree or graduate medical education programs aimed at eligibility for primary specialty board certif ~cation. Although the physician enters research training most commonly at the end of the last year of the residency program, there are many points along the continuum of medical education at which the decision for a research career may be made. Individuals enter research training at different stages of their educational and professional careers and with varying amounts of research experience. Moreover, there is a diversity of paths for transition f rom the trainee pool to the community of productive clinical researchers. Medical School Despite this diversity, the development of a clinical investigator in most cases beg ins in medical school where research opportunities can take several forms. Some medical students receive their f irst exposure to research by repeating classical experiments in basic science or by casual laboratory interactions with faculty members. Other students may develop a f irsthand acquaintance with the laboratory through summer extramural experiences, such as research fellowships. Short-term training for periods of up to 3 months, for example, is currently supported by NRSA institutional awards for 964 trainees in 65 health professional schools. In some medical schools fellowships similar to the former NIH Post-Sophomore Re search Fellowships are available to spec tally selected students. These awards provide an opportunity f or a student to inter rupt regular coursework with a f ull year of advanced study and research. Students who have developed a strong interest in research during their premedical education have an opportunity in three-quarters of United States medical schools to combine medical and graduate studies in programs leading concurrently to the M.D. and Ph.D. degrees (M.D./Ph.D.~. A selected subset of 24 medical schools have 6 year dual degree programs funded through the NRSA Medical Scientist Training Program of the National Institute of General Medical Sciences. Another source of potential clinical investigators is comprised of students who enter medical school with the Ph.D. degree (Ph.D./M.D.~. The postgrad- uation research plans of this group are compared in a later section of this chapter, with those of concurrent M.D./Ph.D. 's and medical students who do not earn a Ph.D. degree. 25

Residency The pattern for including a researab experience within a standard residency varies considerably among clinical specialties. Some residencies, including a number of surgical specialties, routinely include from 3 months to 1 year of clinical research experience. The rationale, however, is to prepare clinicians who can keep abreast of advances in the specialty, rather than to produce clinical investiga- tors. By contrast, research experience is no longer considered as part of the general training requirements in other specialties such as pediatrics and internal medicine. During the period of residency, as a general rule, the young physician will have had little or no training in the discipline of the laboratory, the design and execution of experiments, and the understanding of complex instruments and research technology. In short, to become a productive investigator, he/she needs to start a second and new career. Post-Residency ; . ~- ...... . . Research training has traditionally been intermixed with clinical specialty training for most physicians who later pursued careers in academic medicine. This ref. lects the fact that many clinical research activities must be conducted in patient care settings. In internal medicine, for example, research-oriented residents completing the general program subsequently entered subspecialty training, e.g., cardiology, gastroenterology, rheumatology, etc., with the goal of additional certif ication in that subspecialty and acquisition of research experience. Many of these subspecialty fellowship programs received support from federal training grants. Recent debate over the need for additional subspecialists has prompted questions by federal and other funding agencies as to whether it is appropriate for public funds to be used for research training provided in connection with subspecialty training. This has resulted in diminished funding of subspecialty fellowships which, in turn, has reduced the number of opportunities for training in clinical investigation. At the end of the residency, physic fans with research training in their background can work as postdoctoral research fellows under individual preceptors. NRSA fellowships are well-suited as a support mechanism for this purpose. Such support is particularly applicable to g raduates of joint M. D . /Ph. D . prog rams, most of whom take at least 3 years of residency training in ~ clinical specialty. Residency schedules, however, are relatively inflexible, permitting little or no contact with laboratory research during this period. This discontinuity makes reentry into the laboratory extremely difficult, especially in light of the rapid pace of conceptual and methodologic advance in biomedical research. To compete successfully as a physician investi- gator, a postdoctoral fellowship in a basic or clinical research laboratory is virtually a necessity. Most physicians emerging from residency programs, unlike post-Ph.D. trainees who have already been in the laboratory for several years, are not adequately prepared to write research fellowship proposals. For them 2 or 3 years of individually tailored training in either a basic science or clinical department would be needed to accomplish the coursework, seminar participation, and independent study/research for a career in clinical investigation. 26

DETERMINANTS OF CAREER CHOICE The Co~Tunittee in previous reports noted the increasing indications that young physicians may be less likely to view clinical investigation as an attractive career option. Various factors presumed to be affecting their career decisions were highlighted, including testimony at a public meeting on the effect of increasing national emphasis on the training of primary care physicians. Recognizing the need for a clearer insight into the underlying attitudes and values, the Committee assigned hi gh priority to research in this area. Since publication of its last report, three studies commissioned by the Committee for this purpose have been completed. Career Plans of Medical Students The f irst study, which was carried out by the Association of American Medical Colleges, was based on a December, 1979, survey of all medical school seniors who expected to graduate in 1980 (AAMC, 1981b). Included in this survey were questions on choice of various career paths, as well as attitudes toward research. The responses were matched with records of answers provided by each student to a questionnaire administered with the Medical College Admissions Test (MCAT), the individual ' s MCAT scores, and with records from the application for admission to medical school. Students were asked to indicate their f irst, second, and third pre- ference for alternative careers in research, clinical practice, and administration. For purposes of analysis, respondents were assigned on the basis of their f irst and second choices of career alternatives to three groups--"Research Firsts, n "Research Seconds, " and "Non-Research- ers. n Research was checked as the f irst choice of 21. 8 percent of respondents; 12.2 percent checked research as their second choice but not as f irst; 66. 0 percent checked only clinical practice as f irst and second choices. From a comparison of the three groups, some of the f indings were as follows. Test Scores The groups differ signif icantly in their mean MCAT scores. Research Firsts" and "Research Seconds" had the highest and second highest mean for all four {Verbal, Quantitative, General, and Science) tests. Gender There was no statistically signif icant difference between the sexes regarding preferences for research career. "Research Firsts" and Research Seconds" both had virtually the same distribution of males and females as in the total group of respondents. Experiences During Medical School Somewhat less than one-quarter of all respondents reported having participated in a research project as an investigator during medical school. Grouped according to career plans, 45. 3 percent of "Research Firsts" and 24.2 percent of "Research Seconds" reported such an experience, compared with 16. 3 percent of the "Non-Researchers. " similar pattern can be observed with respect to publications. 27

Approximately 15 percent of all respondents had been sole or joint authors of research papers during their undergraduate medical years. For the three groups the percentage were Research First.,. 31.4 percent; Research Seconds, " 15. 0 percent; and ~Non-Researchers, ~ 10 percent. Debt The data do not indicate that anticipated accumulated debt at graduation from medical school operates as a disincentive to choice of research career. Factors Inf luencing Research Career Choice Respondents were asked to indicate the relative inf luence each of nine factors played in their consideration of research involvement in their ensuing medical careers. The three factors of highest importance to both Research Firsts" and Research Seconds. were: opportunity to work in the academic community, challenge of search for new knowledge, and research experience while a premedical or medical student. These f actors were substantially more important, however, to "Research Firsts. than to "Research Seconds, n who scaled them all to be of "moderate, rather than "ma jor, " inf luence. A fourth factor, "availability of research training support, ~ was scaled to be of "moderates importance to "Research Firsts and HResearcb Seconds and of "minor. inf luence to "Non-Researchers. n2 Finally, all three groups rated as being of Minors or No importances four factors proposed by many observers as hypotheses to explain the declining entry of physicians into research training. These factors were: uncertain availability of research funds after completion of training, increasing f rustrations of researchers in conducting clinical research, f inancia1 disadvantages of a researob career, and obligation to pay back research training support by continued research activity. Plans of M.D./Ph.D. 's and Ph.~./M.D. 's Ninety-eight percent of graduates of the NTH-sponsored Medical Scientist Training Program (MSTP) indicated research careers as their first choice, compared to 90 percent of other M.D./Ph.D. 's. Ph.D. 's who subsequently earn the M.D. degree are about half as likely as concurrent M.D./Ph.D. 's to select a research career first (53 percent), but they are over twice as likely to choose research as are 1980 graduates earning the M.D. degree only (21 percent). Comparison with Earlier Graduates The AAMC Graduation Questionnaire has been administered annually since 1978. Comparable data on career pref erence are also available f rom a questionnaire sent to the entire graduating classes of a sample of 28 medical schools in 1960. Responses to these questionnaires, taken together, permit an assessment of chang ing student attitudes toward resea rch ca ree r s. The classes of 1960, 1978, 1979, and 1980 compare as follows: Graduation Class 1960 1978 1979 1980 28 Percent Favoring Research Careers 39.0 22.4 20.3 21.8

There is no clear trend in the last 3 years, but for all 3 the level of research interest is noticeably less than it was in 1960. The percentage of graduating seniors in the 3 recent years who will fulf ill their expressed plans is unknown, but it would have to be a high proportion to equal the level of researab participation of the Class of 1960. 3 Economic Incentives/Disincentives Scheffler ~1975), in an earlier study using 1971 income data, presented empirical estimates of the rate of return or loss to post-M.I). trainees in biomedical science who subsequently pursue careers as faculty members of a medical school, researchers at NIH, or employees in private industry. In each of the career paths the trai ning produced an economic net loss to the trainee over his/her professional lifetime relative to the alternative of a career in private practice. Moreover, the discounted value of the net loss was relatively large. 4 The loss varied substantially by specialty f ield and increased with the length of post-M.D. training. Under sponsorship of the Committee, the author has updated his earlier work and has applied a similar method of calculation to physicians, dentists, and veterinarians (Scbeffler, 1981~. In general, the results of the later study show that the net lifetime economic loss, in teens of 19 79 income data, is less than that reported earlier. For the three professions there was net loss in income assoc ~ ated with postdoctoral training. Training was assumed to last 2 years, with stipends at 1980 NIH levels and a discount rate of 12 percent. Physicians pursuing a career at NIH sustain a lifetime economic loss of about $170, 000, which is $100, 000 more than the loss incurred by medical school faculty on a base salary. Dentists employed at NIH stand to forego about $111,000 in lifetime income in comparison to dental practice careers, with little economic differences between NIH employment and dental school faculty careers. {3sing different discount rates, the same comparisons yield the following results in terms of lifetime economic loss: Physic fans at NIH at Medical School Dent i sts at NIH Lifetime Economic Loss D i scount Rate 12% 8% o% ~ _ _ $170, 000 $242, 000 $645, 000 $ 70, 000 $127, 000 $239, 000 $111, 000 $164, 000 $550, coo Compared to physicians and dentists, veterinarians have small economic losses, whether employed by the f ederal government or a veterinary school. The economic losses varied considerably by spec tatty for physic fans undertaking 2 years of postdoctoral training and selecting ~ career either at NIH or on a medical school faculty ~ see Appendix Table A9) . The largest economic loss--$241, 000 at the 12 percent discount rate--would be incurred by anesthesiologists choosing an NIH career. Psychiatrists would realize an economic gain of approximately $25, 000 as medical school faculty members (base salary plus supplemental earnings), but ~ contrast would lose $34, 000 in a career at NIH. For several specialties, such as radiology, obstetrics/gynecology, and internal 29

medicine, NIH careers produce an economic loss of over $155,000, averaging 2.8 times the losses to medical faculty. Comparing the losses in 1971 and 1979, the author found that for f ire of the seven specialties studied, the economic losses increased for employment at NIH, whereas for f ive specialties, losses declined for employment at medical schools over this period. The percent of loss accruing during the f irst ~ years following postdoctoral training is also of some interest. This derives f rom the fact that some individuals may select careers using a shorter time horizon because they do not assume they will remain in this career path. Based on the same rate of return methodology as for the earlier calculations, 42 percent of the loss in a NIH career accrues during the first 5 years after training, while 61 percent of the career loss accrues for employment at a medical school (base salary). How large would postdoctoral research training stipends need to be, it was asked, to reduce to zero the economic losses calculated for physicians, dentists, and veterinarians? In other words, what size of stipend would make the monetary rewards in each of these careers equivalent to those in private practice? Scheffler's analysis suggests that physicians who pursue a research career at NIH would need to receive a training stipend of about $63,000 per year for 2 years In order to reduce their economic losses to zero in comparison to careers in medical practice. For those employed in medical schools, it would be about $30,000 per year. Current stipend levels are about $16,000. To produce the same ef f eat for dentists employed at NIH or on a dental school faculty would require a stipend in excess of $50, 000. Veterinarians would need a stipend level of about $25, 000. The suitability of rate of return models has been criticized on grounds that earnings are only part of the determinant of rational personal investment in human capital. Few people choose a career solely on the consideration of earnings. Instead, most give greater weight to their anticipated overall career satisfaction. A significant part of the return on training may in fact be personal satisfaction of dedicated biomedical research scientists as they enhance their knowledge and skills. Also, as noted elsewhere in this chapter, many consider the challenge of the search for new knowledge to be a nonpecuniary benefit tbat more than offsets the higher remuneration obtainable from clinical practice. Thus, to the degree that nonpecuniary benefits of careers in biomedical science offset the monetary aspects, the measured rate of return to investment in post-.. training will have little consequence in influencing the decision to undertake research training. In sum, the effect of ignoring the career satisfaction aspect of training in calculating private rates of return is to underestimate the rate of return to investment in training. Research Careers in Internal Med. ic ine Under Committee sponsorship, researchers at Michigan State University have studied the factors inf luencing choice between careers in academic medic ine and private practice. Responses to a mail questionnaire were received from more than 200 physicians, representing a national sample of those who had completed their subspecialty training in internal medicine by June, 1979, who had already decided for or against going into academic medicine, and who were regarded by their training program directors as possessing the qualifications for a career in academic medicine. The sample was designed to provide an approximately equal distribution of academic and nonacademic deciders. 30

A notable f inditing was the absence of a signif ic ant difference between the two groups in either current salary or aggregate debt. Moreover, there were no apparent f inancial pressures inf luenc ing the career choices. Respondents appeared to be fully aware of potential differences in future income. Those moving into private practice expected higher income than their academic colleagues both 3 and 10 years f ram the time of the survey. Those choosing academic medic ine expected their income to be about 30 percent less in 3 years than those in private practice, and about 17 percent less af ter 10 years. Income expectations appeared to be of moderate importance to those choosing a practice career and of relatively minor importance to those selecting an academic career. The study noted a primary emphasis on intellectual stimulation by those selecting careers in academic medic ine. It is the importance of intellectual stimulation, along with considerably g reater interest in clinical research, that most clearly distinguishes those selecting academic medicine. This is in contrast to those selecting practice, who seem to be most oriented toward patient care and interaction, and who also consider other f actors important such as personal autonomy, availability of "call" coverage, ability to locate in one community, and salary. In rank-ordering the desirability of various career activities, the biggest difference between the two groups was in the rating of primary care. It was as though primary care was perceived as the reciprocal of research. According ly, the f act of having made the choice of primary care virtually excludes for that individual the possibility of performing research. D1 scusslon The study of economic rate of retur n, c ited above, has documented that the early and lifelong earnings of physicians in academic medicine are less than comparable average earning s of physic fans in private practice. This fact, plus a high level of debt accumulated before graduation f ram medical school, have been commonly perceived to serve as disincentives for following a research career. However, other studies raise some question as to the extent to which f inancial factors, such as debt burden and differential earnings, at current levels serve as deterrents to selection of a research career. Certainly, one should not forget that in the heyday of National Research Council fellowship support for M.~. investigators, the relatively enormous income differential between research and practice careers seemed not to operate as a disincentive (Cain and Bowen, 19611. Moreover, income figures may be deceptive, since averages for clinical practice can be distorted by the earnings in a relatively few subspecialties. In addition, faculty salaries in clinical departments have increased to a point where it is not uncommon for a starting salary of an assistant professor to triple the salary received in the last year of residency training. Data from the Michigan State University study and from analysis of the AAMC Graduation Questionnaire underline the key role of intellectual stimulation as an incentive for a research career. Its importance as a discriminant between academic and practice "deciders was the basis for a recommendation in the MSU study that intellectual curiosity be emphasized in the selection process for medical students and fellows. 31

Data from comparisons of future physician researchers and nonresearchers document the need for providing undergraduate medical students with the time and opportunity to acquire f irsthand knowledge of the excitement of working in a research laboratory. -This conclusion is reinforced by a f inding f rom a study of factors affecting the career decisions of more than 1,100 members of the American Society for Clinical Investigation and the American Federation for Clinical Research (Davis and Kelley, 1981) . The largest f Faction of respondents indicated that their research career decision had been made during medical school and that the most important events influencing this decision had occurred during the same period. In both the Michigan State University study and the AAMC survey of graduating seniors, payback was seen not to have been a serious deterrent to research career planning. In addition, the survey of successful clinical investigators by Davis and Kelley showed that the payback obligation was perceived to be a relatively minor factor in the decision-making process. The f indings for n recent deciders" in internal medicine may be f [awed, however, owing to the very small number of respondents who had received NRSA training. Moreover, the AA~MC f inding should be interpreted with some caution, inasmuch as the dec ision on incurring a payback obligation was for most medical school seniors some 3 or more years in the future. Similarly, the results f ram the survey of established investigators may have limited value, since they really represent answers to hypothetical questions. In sum, although it is possible that payback is less of a deterrent than is widely believed, evidence for such a view remains inconclusive. COMPARISON OF TRAINING PROGRAMS The National Institutes of Health have for many years administered both extramural and intramural research training programs for physicians. In addition to postdoctoral traineeships and fellowships (Tr~inees/Fellows), extramural funds hay'? supported since 1964 a Medical Scientist Training Program {MSTP) in which trainees concurrently pursue the M.D. and Ph.D. degrees. In addition, intramural funds support two other research training programs. One is for Research Associates who devote most of their time, under a preceptor, to laboratory research in biomedical science. The second is for Clinical Associates who- participate in both clinical . and laboratory research, as well as in clinical activities under the immediate supervision of senior investigators. The Medical Scientist Training Program averages 6 years in length, compared with just over 2 years for the other three programs. Given the decline since 1974 in numbers of physicians receiving re search training, the continuing need for physic fan investigators, and the limitations on NIH research training funds, the Committee considered it important to compare output of these princ ipal routes for research training of physicians. A recent study by the Association of American Medical Colleges undertook such a comparison, usi ng the following output measures: ~ a) percent of trainees who remain in research, ~ b) research productivity in terms of publications and type of journals in which papers are published, (c) rate of advance in academic or research positions, and (d) success in competing for NIH research grants (AAMC, 19 Bib) . The 53 persons completing the Medical Sc dentist Training Prog ram between 1968 and 1973 served as the reference group. For each of these 53 trainees, a comparable student who subsequently was trained during 32

those years in each of the other three programs was identified. Abe comparison student was of the same age and sex, entered the same (or a similar) medical school with comparable undergraduate education, and had comparable Medical Co' lege Admission Test (MCAT) scores taken at about the same time. The fourfold comparison revealed that: ( 1) All four programs were highly successful in producing physician investigators. MST P bad the highest pro- portion of graduates--74 percent--who could be conf irmed as currently involved in research. Com- parable figures for Research Associates, Clinical Associates, and Trainees/Fellows were 58 percent, 58 percent, and 47 percent, respectively. When program graduates engaged in other medical school faculty activity are included, the percentages rise to 89 percent, 75 per- cent, 68 percent, and 62 percent, respectively. As a percent of graduates with known current career status, 94 percent of MSTP and 95 percent of Research Associates were in research and teaching. Comparable percentages f or Clinical Associates and Trainees/Fellows were 86 percent and 83 percent, respectively. 2) By 19 81 a significantly larger proportion of MSTP graduates than of participants in the three other programs had attained tenured f acuity rank. About 42 percent of MSTP graduates who had joined faculties achieved tenure, compared with 17 percent of former Research Assoc iates and Trainees/Fellows and 4 percent of Clinical Associates. 3) Research grant success rates at NIH did not appear to dis- criminate among the four groups. MSTP graduates, Clinical Associates, and Trainees/Fellows had higher approval rates and better average priority scores than those for all applicants. Judged by only a small number of appli- cations, the pert onnance of former Research Assoc fates was mixed--relatively low approval rate but excellent average priority scores for approved applications. 4} Publication performance for graduates of the four training programs showed marked differences in the quantity and sci- entif ic depth of the articles they published. MST P gradu- ates were the most prolif ic authors, having published 995 articles by 1981, compared with 716 articles by Research Associates, 673 by Clinical Associates, and 408 by Trainees/ Fellows. All groups published articles in journals of all four scientific "levels, '5 but the distribution by "level" differed substantially. MSTP graduates and Research Associates, for example, published predominantly in journals characterized by a high degree of scientif ic rigor, e. g ., Journal of Biolog ical Chemistry and American Journal of Physiology. Former Clinical Assoc fates and Trainees/ Fellows published most f requently in journals, such as the New England Journal of Medicine, which contain a mixture of reports on clinical investigation and clinical observation. Of the four groups of trainees, former Clinical Associates published most f requently in journals emphasizing clinical observation, e.g., the Journal of the American Medical Asso- . . ~ c Dacron. 33

With respect to these f indings several comments are in order. First, the Harp reference group included all of the f irst 53 successful gradu- ates of this very selective program. Because of the rigorous matching criteria, however, the three comparison groups may not be entirely representative of all the graduates of each of these training programs. Second, the signif icantly larger proportion of MSTP graduates achieving tenured academic rank by 1981 may have resulted mainly f rom the advantage of their having undergone the rigorous research experience required to earn the Ph.D. degree. Third, differences in the publication history suggest strongly that the four programs had different products. MSTP graduates and Research Associates appeared to conduct much basic, as well as clinical, research. Research by former Clinical Associates and Trainees/Fellows included a relatively higher proportion of clinical observations and clinical investigations. MSTP graduates were clearly the most prolif ic authors, whereas former Trainees/Fellows made notably fewer contributions to the medical and scientif ic literature. Fourth, the f inding that Trainees/Fellows, even in the absence of a payback obligation, continued at a substantial level their teaching and research involvement runs counter to a widespread belief that a majority of these people entered practice rather than academic careers. Verif ication of this f inding was accomplished through a direct matching of all recipients of NIH extramural research training support with the Medical Faculty Roster of the Association of American Medical Colleges. Of the 26,307 physicians who received such postdoctoral training support through 1975-1976, 13, 527, or 51 percent, were on the Faculty Roster in 1981. This f igure is compatible with the 47 percent noted in item 1, above, for the comparison sample of the study. This corroboration of the sandpile data provides a basis for believing that NIH fellowship and traineeship support over several decades has been successf ul in producing researchers and teachers. RESEARCH TRAINING AND RESEARCH QREER LONGEVITY Follow-up studies such as the above suggest that partic ipation in formal research training prog rams enhances the trainees' subsequent research productivity. Proof of this proposition is hard to obtain because there are many f actors, obvious and subtle, that could also affect an individual's productivity. Since measuring their separate effects is extremely cliff icult, it is necessary to rely instead on certain measurable relationships that g ive evidence for or against the proposition. One such link is that between the duration of research grant support and previous research training experience. Recently developed data pre- sented in Table 2. 2 show that those scientists with NIH-supported postdoctoral training survived much longer in the competition for NIT research grants than those without such postdoctoral training. Ph.D. scientists, it should be noted, exhibit a similar pattern. Futhennore, the difference is consistent over the period of years shown. To the extent that survival time as a principal investigator is an acceptable measure of research productivity, these data clearly imply that postdoctoral research training is an important factor in subsequent research success. 34

TABLE 2.2 Length of Career as NIH Pnncipal Invest~gator With and Without NIH Postdoctoral Training With NIH Postdoctoral No NIH-Supported Fiscal Yesr of T an~ung Postdoctoral Training First Research M.D. s Ph.D. s M.D. s PhD. s 1957 Yrs. as NIH P.I. (median) 16.8 16.7 11.8 10.8 No. of P.L s 76 135 447 515 1958 Yrs as NIH P.L (median) 20i6 17978 120i6 '1 7 l9S9 Yrs.asNIHP.l.(median) 16.5 13.5 10.4 8.7 No.ofP.I.s 88 134 418 565 1960 YrsasNIHP.l.(median) 10.3 12.8 7.9 6.9 No. of P.I. s 120 162 467 568 1961 Yrs NIH P.l. (mea ) No.ofP.I.s 111ol6 lll;3 36i3 4597 1962 Yrs. as NIH P.l. (medi ~ 5.9 No.ofP.Ls 1959 lloi4 3651 508 1963 Yrs. as NIH P.l. (median) 10.7 11.2 5.7 5.4 No. of PI.ts 160 215 290 455 1964 YrsasNIHP.l.(medi ~ 9.2 No.ofP.Ls 203 l2lo7 34848 64i40 1965 Yrs.asNIHP.l.(median) 8.4 11.3 4.0 4.7 No. of P.I. s 210 215 301 436 1966 Yrs as NIH P.l. (median) 2is l2lo3 24698 3494 1967 Yrs as NIH P.l. (median) 28i5 10;0 233 45240 1968 YrsasNlHP.l.(median) 7.2 10.2 5.7 7.1 No.ofP.Ls 166 166 145 279 1969 Yrs.asNIHP.l.(medi ~ No ofPLs 1694 28.8 1624 2693 1970 YrsasNIHP.L(median) 8.5 8.6 8.4 6.1 No. of P.L s 155 171 89 206 SOURCE: NationalResearchCouncil(1958-80 1974-81 1979a specialtabulation5/19/81). 35

RESEARCH INVOLVEMENT Knowledge of the relative time/ef f art devoted to research by M. D. f acuity is important to an understanding of the clinical research manpower pool. Since publication of the Conm~ittee's last report, information has become available on the reported percentage of total hours devoted by M.D. faculty to research (AAMC, 1981b). In addition to current involvement, estimates of percent of time spent by faculty members on research in the 1st, 5th, 10th, and 15th year of faculty employment have been obtained. The survey f rom which these data were derived def ined research activity broadly to include all research-related teaching, patient care, and administration ~ including service on human subjects review committees). The data, it should be emphasized, describe magnitude of research ef fort only. No attempt was made In this case to measure outcomes such as research quality and productivity. Average research time varied by specialty group, ranging from 13 to 34 hours per week for all physicians on the faculty. M.D. faculty in the basic sciences spent 62 percent of their time in 1980 in research, while medical specialists (pediatrics, internal medicine, neurology, etc. ~ spent 39 percent; surgical specialists {surgery, orthopedics, urology, etc. ), 22 percent; hospital-based specialists (anesthesiology, pathology, radiology), 23 percent; and behavioral specialists (psychiatry), 30 percent of their work time in research activities. M.D. basic science faculty had the highest reported percentage of time in research activities over the span of a career, ranging f ram 75 percent at an estimated average age of 40 to 55 percent at age 55. Medical spec ialists under age 40 averaged 37 to 46 percent of their time in research, with a decline to 27 percent at average age 55. Surgical specialists, in general, have the longest work week and the lowest percent of time in research. In all of the specialty groups, including both clinical specialties and basic sciences, the older cohorts spend proportionally less time in research than do younger cohorts. NON-M. D. SCIENTISTS IN CLINICAL IN~STIGATION As seen in Table 2.1., the proportion of NIH postdoctoral traineeship and fellowship positions filled by M.D. 's has dropped from 66.1 percent In 1973 to 34.1 percent in 1978. During this period the number of Ph.D. bioscience researchers continued to expand, with an annual net growth of more than 9 percent in the pool of postdoctoral appointees in the academic sector (NRC, 1981a). In addition, reduced opportunities for tenure-track appointments in basic science departments have given rise to a growing movement of young Ph.D. scientists into clinical investigative units. Against this backs round it is not surprising that one of the questions being asked of the Committee concerns the role of the Ph. D. scientist in clinical investigation. The subject was explored at a joint meeting in April 1981 of the Co~runittee' s Panel on Basic Biomedical Sciences and Panel on Clinical Sciences. While the Committee plans to issue a separate report on this subject, the following sublunary reflects the scope of the panelists' discussions. 36

Available Data Sources At least five files contain data on numbers of non-M.D. scientists engaged in research conforming to various definitions and classifications of clinical investigation. It would also be possible from this data base to compare Pb.D. faculty in clinical and basic science departments with respect to age, NIH grant success, postdoctoral support years, f ield switching, academic rank, etc. Variation by Clinical Specialty The research role of Ph.~.'s in clinical departments varies with the specialty. In contrast to medicine, surgery, and pediatrics, the Ph.~. 's role appears to be more sharply delineated and related to the nature of the specialty in departments of psychiatry, comparative medicine, obstetrics/gynecology, nuclear medicine, and epidemiology (commonly a part of community medicine, preventive medicine and public health, etc.~. For example, research in reproduction, especially the study of mechanisms of hormone action and their control of normal and abnormal function, is now frequently centered in obstetrics/gynecology departments. For research of this type the suitably trained Ph.D. has an obvious role. Future Outlook The outlook for recognition of Ph.D. scientists in clinical departments as aspirants for senior faculty rank and tenure is not encouraging. Economic factors, such as lack of income-producing options for the Ph. D. scientist, when and if grant support is discontinued, are dominant. Where exceptions exist, they usually involve Ph.D.'s with a clinical or service potential, as in the case of a Ph. D. audiolog ist or speech therapist in a department of otolaryr~gology or a radiation physicist in a therapeutic radiology division. This situation derives in part from the fact that patient fee income teas grown since 1968 at a rate of more than 25 percent per year and has overtaken R & D funds as a source of revenue for medical schools. Need for Additional Information In earlier discussions of this topic, too much emphasis may have been given to the wrong issue. The concern should be over how to maximize the Ph.D.'s participation in clinical investigation, rather than on the departmental setting in which he/she has a primary faculty appointment. Consequently, it seems worthwhile to identify models where Ph.D. involvement in clinical investigation has been demonstrably effective and to study on a case-by-case basis the factors that have promoted productive interactions between M.D. and non-M.D. researchers. Examples of sites for such study are the National Heart, Lung, and Blood Institute's Specialized Centers of Research, the National Institute of Child Health and Human Development Mental Retardation Centers, and departments of comparative medicine. 37

THE MARKET FOR CLINICAL SCIENTISTS This Committee teas repeatedly expressed concern over both the apparent dwindling interest in research careers on the part of young physicians and the growing number of budgeted vacancies on clinical faculties in medical schools. Its projections of demand for full-time clinical faculty made in 1978 indicated continued growth over the next 5 years due to expanding enrollments, clinical R & D expenditures, and · ~ service Income. Cur rent I nd icators The latest data exhibit some conf listing trends, as can be seen in Table 2. 3. There has been an apparent increase in the number of physicians reporting research as their primary activity to the American Medical Association (line 3a) 6, yet clinical research expenditures at medical schools dropped by more than 6 percent in 1979 after adjusting for inflation (line 2a). NIH research grants awarded to M.D.'s increased in absolute numbers but declined as a percentage of total competing grants to all investigators (lines 3c and ad). However as Wyngaarden ( 1979) points out, the number of NIH grants awarded on behalf of new M. D. principal investigators has not changed appreciably since 1968. The number of medical students, residents, and clinical fellows went up sharply in 1979 (line la) but the number of M.D. 's receiving research training under NIH-supported programs remained at low levels ~ line lb) . On the demand side, growth in medical school income derived from patient care activities has slowed noticeably (in real terms) in the last few years after rising rapidly in the early 1970's.7 Clinical R and D expenditures also turned down in 1979. The net result is that since 1976 tbe total funds available to support clinical faculty {above that generated by tuition} have grown at about 3 percent per year, a slower pace than that of a few years earlier (line 2c). On the supply side, the number of M.D. ' s applying for and receiving NIH research grants has turned upward since 1976 (lines 3c and 3e). However, the number of young physicians entering research training programs is still far below the level that the Committee believes is appropriate in light of its past recommendations. As noted in Table 2.1, the number of M.D.' s and other health professionals participating in NIH research training programs is currently only 40 percent of the number that were doing so in 1971. This is a continuing cause for concern because a period of postdoctoral research training is considered essential for an M.D. to become a successful clinical researcher. Sorting out the implications of these patterns for the f uture supply and demand for clinical investigators is the goal of the remainder of this chapter. Projections Through FY 1985 It is well known that medical school faculty8 members generally distribute their activities in varying proportions among teaching, research, patient care, and administration. The CoTnmittee's model of demand for clinical faculty is therefore based on the hypothesis that the 38

E ^~ ~ t it ~ -- ~ , ° ° ° =~ ° of = ~ If; 3 §6 ~ ~ , ~ 7 ~ o -~= Ad. Ad. ~ 3 ~ t 4- ° ~ 3 ~ ~ o ~ ~ ~ ~ x ~ _ ~ ~ x ~ on ~ ~ _ ~ ~ ~ ~ x o ~ o ~ ~ 6 ~ $$ Go on o' 8 ·C~) lo Cat o' _ .5 o off 5 _ I: - U. .S Cq it: v o — Z ~ ~ ~ ~ o - ^ ° =- ~ ~ ~ o o - z ~ ~ ~ ~ o ~ o^ ~^ ~ t - ~^ ~ o - o o g x o ~ x ~ x v) ~ ~ ~ ~ ~ o Yo o ~ ', ~4 ~ ~ C'' ~ v) =~ o ~~ { a £ Z ._ _ ~: i '3~ ~igiD jl~-9~ = ~zg~it' ~ 39 C~ ~ ~ _' ~ — ~^ ·3 oo C~ ~ Z ~ ~o ~ ~ o ~ §~3] ~ 'g=3 T^ a¢~ _ i 1 ~ ~. ~ ~ a~ ~ co' ~ g

number of full-time clinical faculty in medical schools in any year t is dependent on medical student enrollment in year t (representing the teaching component of demand), funds available to support clinical research (representing the research component), and funds generated by patient care activities (representing the patient care component).9 In the case of the latter two components, it is further assumed that these funds impinge on the demand for faculty over a period of years. In other words, funds generated in year t affect the demand for faculty in years _, t + 1, and t + 2._ The projection model relies for its input on assumptions about future growth in clinical funds (clinical R and D revenue plus service income) and medical student enrollment. Figures 2.1-2.3 may assist in making these judgments about future patterns. They show the trends in funding and enrollment variables from 1961 to 1979 and projections to 1985 under several assumptions about future growth rates. Figure 2.1 shows that the growth in clinical R and D expenditures in medical schools, expressed in constant dollars, has been approximately linear from 1961 to 1979. If this linear trend is projected to 1985, the clinical R and D figure would reach $375 million, which corresponds to a growth rate of almost 7 percent per year. However, in view of the extreme budgetary pressures that exist today, the Committee and its Panel on Clinical Sciences believe it is unrealistic to assume that such growth will continue over the next few years. A more realistic appraisal is that there will be very little real growth in clinical R and D expenditures through 1985. Under the most pessimistic conditions foreseen by the Committee, there would be a decline of 1 percent per year in these funds f ram the 1979 level after ad Busting for inf ration. Income from patient care services reported by medical schools grew rapidly f rom 1968 to 1976 (Figure 2. 2~. Since 1976, the growth has slowed somewhat, although in 1979 it was still up a strong 6.3 percent, which is quite close to the most likely growth rate (6 percent) assumed by the Committee in its 1978 report. Figure 2.2 shows the projections of fee income to 1985 based on growth rates of 12 percent, 6 percent, and 2 percent per year. Combining the growth estimates of clinical R & ~ funds with those of service income, we arrive at the following real growth rate estimates for the monetary variable of the model through 1985: 8.6 percent per year (high estimate), 4.0 percent per year (most likely estimate), 0.8 percent per year (low estimate). With regard to medical school enrollments, the third component of the demand model, the Committee this year has broadened its definition to include not only medical students but also residents and clinical fellows since all three contribute varying amounts to the demand for clinical faculty. The number of medical students has grown from less than 44,000 in 1972 to over 60,000 in 1979, a growth rate of more than 5 percent per year. But several recent developments have occurred that are likely to produce slower growth in the 1980's. For one thing, a study published late in 1980 by the Graduate Medical Education National Advisory Committee (GMENAC) found evidence of a potential excess of 145, 000 physicians by the year 2000 unless steps are taken now to curtail the supply {GMENAC, 1981, Vol. I, p. 10~. The GMENAC report recommended that entering medical school classes be reduced 10 per-cent from 1978 levels by 1984, that no new schools be established, and that the influx of foreign medical graduates be restricted. 40

500 40Q Z 300 o - 200 100 Actual Projected O ~ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 FISCAL YEAR FIGURE 2.1 Clinical R and D expenditures it __ High Estimate (3%/yr.) Middle Estimate (1%1yr.) Low Estimate (-1%/yr.) 600 500 400 oh Zo J 300 200 100 Actual - Projected n medical schools {1972 A. See Appendix Table A7. / High Estimate ( t 2%/yr.) ! ~ Middle Estimate (6%/yr.) ; ~ l / / '' /~' _ _ Low Estimate (2%/yr.} I . 1 I L I I ~ . _ I 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 FISCAL YEAR FIGURE 2.2 Service income reported by medical schools (1972 A. See Appendix Table A7. 41

1 0 : OL 130 High Estimate (4%/Yr.) ~ ~ ~ Middle Estimate (2.5%/yr.) 1 10 _ ''' - Actual ~— --—— Low Estimate (0%/yr.) _ 100 ~ ~ Projected >/ ~ 90 _ / \ i3 '' Medical Students, Interns and Residents, and ~ 80 ~ / Clinical Fellows Z 70 ~ / _ High Estimate (3%/yr.) _~_ Middle Estimate (1.5%/yr.) 6 ~ '~~— —Low Estimate (0%lyr.) 0 50 _~ . ~;: u, ~ Medical Students 40 _ 30 _ _ ~,~` _~ I nterns and Residents 20 _ =' _ Clinical Fellows LIT—~ 1 1 1 1 1 1 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 FISCAL YEAR FIGURE 2.3 Medical students, interns and residents, and clinical fellows. See Appendix Table A6. 10O 90 80 70 60 50 40 30 20 10 o —Actual Projected D _ _ Volunteer Faculty / _ ~ _ / '' High Estimate (5.2°/O/yr.) '''__—Middle Estimate (2.8%/yr.) _ _ ~ ~~~~ Low Estimate (-0.05%/yr.) C ~! To; ~~ ~., _ Full-Time Faculty -_ . 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 FISCAL YEAR FIGURE 2.4 Clinical faculty in medical schools. Faculty is defined here as a full-time appointment in a clinical department regardless of tenure status. See Appendix Table A6. 42

Second, the funds provided by Congress to encourage medical schools to expand their enrollments (capitation grants), which have been declining throughout the 1970' s, were rescinded for EY 1981 and have not been reauthorized for FY 1982. On the other band, medical schools are also under heavy financial pressure and are likely to resist any attempts to reduce enrollments. The net effect, as seen by the Committee and its Panel on Clinical Sc fences, is that the current growth in medical student enrollment will not continue but will average about 1.5 percent per year through 1985 (Figure 2. 3) . The number of residents and clinical f ellows is almost as large as the number of medical students and has been growing faster in recent years. Therefore, the Committee assumes that the total number of medical students, residents, and clinical fellows will grow at an average rate of 2.5 percent per year through 1985. The upper and lower limits on this growth rate are assumed to be 4 percent and 0 percent, respectively. Given these assumptions about future growth in the monetary and en- rollment variables, the model produces the estimates of demand for full-time clinical faculty through 1985 shown in Figure 2.4 and Table 2.4. Under the most likely circumstances foreseen by the Committee (the middle panel of Table 2. 4, assumptions II and B), full-time clinical faculty in medical schools will grow to almost 40, 000 by 1985. This means an average of about 1,000 new positions will be created each year due to expansion of faculty and another 2, 000 vacancies still occur each year because of attrition. Exactly how these estimated 3, 000 positions will be f illed is crucial to determining the national needs for training clinical scientists. For this we must know something about the dynamics of f acuity accession and attrition in medical schools. The Committee is fortunate to have recently received such information from a study it commissioned the Association of American Medical Colleges (AAMC, 1981b} to do. That study provided data on several topics concerning medical school faculty, some of which are discussed elsewhere in this report. Of immediate interest here are data on hiring of clinical faculty members with research training background. Up to now it has been assumed that all demand created by expansion and attrition due to death and retirement would be filled by former postdoctoral trainees. But the new data show that this is an unrealistic assumption. From 1975 to 1978 only about 21 percent of all new appointments in clinical departments had postdoctoral research training. The AAMC data also show that a signif icant amount of attrition f rom medical school faculties (about 4 percent per year) takes place because of changes in employment and causes other than death and retirement. Attrition rates therefore have been modified in Table 2.4 to reflect this. There are still some important gaps in our knowledge of the system that need to be f illed in as this study develops. Nevertheless suf f ic lent data have accumulated to attempt a more rigorous estimate of postdoctoral training needs based on the projections than has been possible heretofore. Using the most likely projections from Table 2.4, Table 2.5 shows how these estimates translate into estimates of the number of clinical science postdoctoral trainees required to meet the expected demand under certain conditions. At each step of the calculation, the estimates are compared to the actual data where ava i table. Line 1 of Table 2. 5 shows the annual demand for full-time clinical f acuity broken down into demand c reated by expansion, death and 43

TABLE 2A Projected Growth in Medical School Clinical Faculty, 1980~85, Based on Projections of Medical School Enrollment, Conical R ~d D Expenditures, and Medical Service Income In Medical .S6hoolsa Gumptions about Real R and D Expenditures and Medical Sauce Income in Medical Schools ~ . I II III sump~don~ about Medi—1 Student Enrollment (medical students, students, "d Clinical fellows) Will expand at win expand at Will expand about 8O6~/yr. about 4.0~0/yr. Lightly, about to S964 ma- to S744 mil- Q8~/yr. to $616 lion ~ 1985 lion in 1985 million in 1985 wm grow at 4%/yr., Expected me of Minimal faculty reacting 143,000 hi medical "hoofs (CF) m students by 198S 1985 45,950 43,610 41,600 Annual grown rate in CF from 1980 to 1985 5.2% 4.3% 3.5% Average annual mcremcnt due to faculty expansion Annual replacement needs due to:b death and retirement other attrition Expected number of positions to become available annuity on clinical faculties 2,000 1,620 1,280 400 390 380 1,760 1,700 1,660 4,160 3,710 3,320 . Be Will grow at 2.S%/yr., reaching 131,000 students by 1985 Expected Size of clinical faculty m medical schools (CF) in 1985 Annual grown late in CP from 1980 to 1985 42,120 39,970 38,120 3.796 2.8% 2.0~o - Average eMnual inGrement due to faculty expansion 1,370 1,010 700 Annual replacement needs due to:b death and retirement 380 370 360 other attrition 1,670 1,620 1,590 Expected number of positions to become available annually on dinica1 faculties 3,420 3,000 2,650 C. WB1 Mow eaenlially Expected We of conceal faculty no growth from 1979 to hi medical schools (CE) no 1985, leveling off at 1985 113,000 students Answer growth ram In CF from 1980 to 1985 36~320 1.1% 34,470 Oe 3% 32~870 ~OeO54% Average annual increment due to faculty expansion 400 90 -170 Annual replacement needs due to: b death ant retirement 350 340 330 other atbidon 1,550 1,510 1,470 Expected Rumba of positions to become available annually on clinical faculties 2,300 1,940 1,630 aFaculty in this table ~ defied as a [ull-~dme appointment In a conical department regardless of tenure status. These projections are based on the following relationship: (CF/S)t ~ exp(-LS9683~103440/Dt) + 0.13, where CF = give of conical faculty in medical schools; S = medical students, residents, and clinical fellows; D = a weighted average of me last 3 years of clinical R and D expenditures b plus medical ~ ce income in medical schools, be., Dt = 1/4(Dt + 2Dt_1 ~ Dt_2). Based on an estimated replacement rate of 1.0% annually due to death and re~drement, and 4.4% annually due to other athi~don. See A - octagon of American Medical Colleges (1981a). \ 44

TABLE 2.5 Estimated Number of Clinical Research Postdoctoral Trainees Needed to Meet Expected -Demand for Clinical Faculty Through 1985 Under Vanous Conditions Actual (1975-78 average) Projected (1980~85) Other Total M.D.'s Ph.D.'s Degrees Total 1. Demand for full-time clinical faculty--annual average: a. due to expansion of faculty b. due to death and retirement c. due to other attrition 2. To reduce budgeted vacancies--annual average C 3. Demand for veterinary and dental school clinical faculty 3,673 2,982 2,113 1,783 128 97 1,432 1,102 454 280 18 156 237 50 13 174 3,000 1,010 37oa 1,620b 150 (16% of med. school demand)d 500 4. Total annual accessionse 3,673 2,982 454 237 3,650 a. % with postdoctoral research training 21.4% 18.6% 55.9% - 35.0~o 5. Total accessions with postdoctoral research Paining——annual average 6. Size of clinical saence postdoctoral pool--annual average Size needed to meet academic demand assuming a 2-yr. training period and portion of trainees seeking clinical faculty positions is: a 60~o b. 50~o 7. Annual number of clinical science postdoctoral trainees to be supported under NRSA programs: a. if 50% of pool is supported under NRSA b. if 60% of pool is supported under NRSA (35% of all accessions) 3,000 - 5,000 2,310 1,280 4,270 5,120 2,140 - 2,560 2,560 - 3,070 aAssumes an attrition rate due to death and retirement of l.O~o per year. bAssumes an attrition rate due to other causes of 4.4% per year. CIn 1978 there were 2,011 budgeted vacancies in clinical departments of medical schools. The demand for clinical faculty generated by the need to reduce this level to 1,000 by 1985 is about 150 per year. dIn 1978 there were 3,544 full-time clinical faculty members in U.S. dental schools and an estimated 1,869 full-time equivalent clinical faculty members in U.S. schools of veterinary medicine. This total (5,413) was 16% of the full-time clinical faculty in U.S. medical schools. Thus the demand for dental and veterinary school clinical faculty is estimated at 16% of medical school demand, or 500 per year. eAccessions are defied as new hires or those who rejoin faculties from nonfaculty positions. Inter-faculty transfers are not counted as accessions. Data on the percentage with postdoctoral research training were derived from newly hired faculty members only, which are 85% of total accessions. We are assuming that the same percentage applies to all accessions. SOURCES: Data from 1975-78 taken from American Dental Association (1971-79a), Association of American Medical Colleges (1980, 1981a), Association of American Veterinary Medical Colleges (1978). Projections were derived by this Committee. 45

retirement, and other attrition. The projections assume a 1.0 percent death and retirement rate and a 4.4 percent attrition rate due to other f actors. Line 2 indicates that demand for an additional 150 clinical faculty members per year is created by the need to reduce budgeted vacancies in clinical departments. These have doubled in the past decade, and the Committee views this growth as an indicator of shortages. It is appropriate, therefore, to attempt to reduce budgeted vacancies in clinical departments to 1, 000 in 1985 from their 1979 level of about 2, 000. Line 3 is the estimated demand for clinical faculty at dental and veterinary schools. This demand component is estimated to be 16 percent of medical school demand, or 500 per year. Line 4 is the sum of lines 1 to 3. This total, 3,650, is the expected annual demand through 1985 under the conditions stated. Line 4a shows the percentage of accessions having some postdoctoral research training. The current figure of 21 percent represents a severe deterioration in the research preparation of clinical faculty members. Of the present faculty in medical specialties who graduated from medical school between 1963 and 1967, more than 40 percent had postdoctoral research training (AAMC, 1981b, p. 12~. The Committee believes it is important to restore this percentage to near its former level in order to maintain the research capability of medical schools. Line 5 shows the estimated number of accessions needed with postdoctoral research training assuming that 35 percent of all accessions have this training. Note that transfers from one academic institution to another are not counted as accessions. Line 6 gives the estimated size of the current pool of clinical research trainees (no hard data are available on the size of the pool). Assuming a 2-year training period and a 60 percent yield, a pool size of about 4,270 would be needed to produce the estimated 1,280 trained clinical researchers required each year, i.e., 4,270 = (1,280 x 2~/ 0.6. Finally, line 7 provides estimates of the number of clinical science postdoctoral trainees to be supported by the NRSA program. It is encouraging to note that the Committee's projections of faculty demand are supported by the AAMC study. The findings in both cases are quite close. The Committee projects an annual demand of about 3, 000 full-time medical school clinical faculty positions based on a likely growth rate of almost 3 percent per year through 1985, while the AAMC study, using age-specif ic attrition rates, conf irms that if faculty does indeed grow by 3 percent per year, the annual demand for full-time clinical faculty would be just over 3,100 per year (AAMC 1981b, p. 253) .10 But clearly the new inf ormation the CoIIunittee has received has required readjustment in some of the assumptions made about faculty dynamics. Perhaps the biggest adjustment is in the previous assumption that all demand for clinical faculty would be satisf fed by people with research training experience. If the AAMC data accurately ref. lect the current situation, only about one out of f ive new hires in clinical departments has had research training. This fact also cautions against equating research with an academic caree r, for it implies that the majority of people choosing to become members of clinical faculties do not have a research background. While it may be true that most clinical researchers are also faculty members, these data strongly suggest that only a minority of clinical faculty members have the inclination or training required for a research career. 46

Graying of the Faculty and Research Productivity The relationship between age and research productivity tales on added signif icance with the realization that faculties in most f ields are likely to be static or show only moderate growth in the coming decade. In one of the more important aspects of its study, the AAMC (1981a) showed how research output varies with the age of the clinical investigator. Peak productivity, as measured by the annual number of publications in selective journals, is generally reached around mid-career for most clinical investigators and thereafter declines. 11 These results suggest that medical school faculties tend to become less productive as the age distribution increases. To prevent ag ing of faculties, there must be a continuous infusion of younger members, preferably with research training experience. In the most probable c ircumstances foreseen by the Committee, clinical faculty growth will average almost 3 percent per year through 1985. At this rate, the average age of clinical faculties frill increase only slightly, and research output will increase at about the same rate as faculty growth according to the AAilC report (AAMC, 1981b, p. 64~. But should faculty growth be slower than expected in the next decade, a signif icant amount of faculty aging will occur, with a resultant decline in average productivity. The AAMC estimates that if there is no f acuity growth f ram 1980 to 1990, the average research output of physicians in medical schools will drop by more than 6 percent, and in surgical specialties by almost 14 percent. The research training programs can thus be viewed as a kind of insurance policy that tends to stabilize the research capability of medical schools despite varying demographic and economic conditions. RECOMMENDATIONS In view of the outlook for continued shortages of clinical scientists to staff the faculty of medical schools and conduct clinical investigations, arm the need to encourage physicians and other health professionals to undertake research training in preparation for research careers, in the sections below the Committee makes two recommendations for training in the clinical so fences area. Tra ineeships and Fe llowsh ips Recommendation. The Committee recommends that 2, 400 traineeships and 400 fellowships be awarded annually for postdoctoral research training in the clinical sc fences f rom FY 1982 through FY 1985. The Committee also urges a greater emphasis on f illing these training positions with physic tans. Most of the f unding should be in the f arm of training grants, which are particularly suited to meet the needs of the physic fan whose doctoral training usually involves only limited research participation, and, as pre- viously recommended, approximately 15 percent for individual fellowships. 47

In recommending federal support of 2, 800 postdoctoral trainees and fellows in the clinical sciences, the Committee is well aware of the fact that not all of these awards will be made to physicians. It is appropriate that some of these trainees will hold the Ph.D. and other nonhealth professional doctorates, just as some M.D. ' s will receive training under grants awarded to basic biomedical so fence departments. Currently, physics ens hold about 70 percent of all the postdoctoral traineeships awarded by NIH in the clinical sciences. 12 While it ~ s cliff icult to def ine the optimal proportion, 85 percent may represent a reasonable proportion of M.D. 's in the pool of clinical science trainees and fellows supported under NRSA programs. It would be compatible, for example, with the observed ratio of M.D. 's to total full-time faculty in clinical departments of medical schools {AAMC, 1981a). There are cliff iculties in achieving a prompt increase in the proportion of clinical science training positions held by physicians. The relatively small number of physician applicants for NIH postdoctoral research fellowships and traineeships in recent years is one aspect of the problem. Another is NIH' s lack of leverage for inf luencing the change directly, since trainee appointments are the responsibility of the individual training program director. It is nevertheless of critical importance that NIH examine all possible means for augmenting the physic fan component of the clinical sc fences training pool. For example, consideration could be ~ iven to implementing a 1979 recommendation of the Committee to offer a postdoctoral clinical research experience for MSTP graduates, most of whom go on to a clinical residency. Since their research training up to that point has generally been in teas ~ c sc fence, such an experience would be advantageous as a bridge for those aspiring to become clinical investigators. Since inclusion of a clinical research component would requ' re support outside of the usual source of residency funding, the Committee recommended consideration of a special type of award for this purpose. To foster progress in this direction, the Committee reiterates its suggestion that Spec ialty Boards and Residency Review Committees be encouraged to develop training policies for the clinical investigator that would lead to Board certification without appreciably lengthening the training process. Another suggested approach is to change NIH guidelines regarding full-time involvement in traineeships and fellowships. A requirement of less than full-time part icipation could attract women physicians with research interests, who are temporarily out of the labor force. Also, it could facilitate the combination of research and clinical training within a given year of residency. Medical Scientist Training Program Initiated in FY 1964 with three grantee institutions and support for 17 trainees, the Medical Scientist Training Program has experienced uninterrupted growth. By PY 1976 it involved 18 institutions and 465 trainees. As a result of continued expansion thereafter, endorsed with enthusiasm by the Committee, trainee enrollment is now approaching 700. In its 1979 report, the Committee proposed a ceiling of 725 trainees in the Program through FY 1982. Further expansion, it was felt, should await the development of more analytic information regarding the Program, including comparative costs of other research training programs for physicians. As reported elsewhere in this chapter, a start has been made toward assembling the information. 48

Additional analyses are needed, however, to explore possibilities for enhancing cost-effectiveness through administrative modifications in the Program and to determine its optimal size In relation to the totality of NIH training activities. Until these further analyses can be made, and in view of the steady build-up that has occurred, the Committee believes it prudent to retain the currently recommended enrollment level through FY 1985. Recommendation. The Committee recommends that an annual · enrollment of 725 trainees In the Medical Scientist Training Program (MSTP) should be maintained through FY 1985 and high priority g iven to preserving this level should it become necessary to reduce the overall number of NRSA trainees. 49

NOTES 1. Unless otherwise Specified, the term clinical investigator refers in this chapter mainly to physicians. It should be noted, however, that clinical investigation under the Committee' s def inition is also performed by other health professionals, e.g., dentists and veterinarians, as well as by scientists holding the Ph.~. and equivalent degrees. 2. Since opportunity costs for physicians who pursue doctoral research training are relatively high, it is reasonable to infer that a loss of f ellowship and traineeship support could diminish the already small proportion of medical school seniors looking to a resea rch career. 3. A study comparing career preferences at graduation with career performance after 17 years was reported in 1979 (Sherman and Morgan, 1979~. Approximately 28 percent of 1960 medical school graduates bad contributed to the medical and scientif ic literature by 1977. The study also found that 44 percent of the 1960 graduates expressing research plans did eventually conduct and publish research, and 8.3 percent of those without research plans published research in addition to carrying a medical practice. 4. A discount rate is an interest rate used to compute the value today of a sum of money to be received in the future. For example, 1 dollar to be received a year f ram now is currently worth about 89 cents at a d i scount rate of 12 pe rcent . 5. For a full description of research n level" and its derivation, see Narin ( 1976) and She rman and Morgan ~ 19793 . 6. The AMA data themselves present some conf licts. Although the number of physicians reporting research as a primary activity has increased, the number with unknown activity has increased even faster, which creates some uncertainty about how to interpret these data. The increase in research activity reported to the AMA is puzzling, since it goes in the opposite direction of data the Committee has compiled on the number of physicians participating in the research training programs of NIH. Furthermore, since the AMA survey contains no standard def inition of research activity, the interpretation is subject to wide variation among the respondents, and in many cases it may differ from the Co~Tunittee' s def inn' ~ on. 7. Patient fee income data are taken from AMA (1960-80) but are known to be understated there because of the way medical schools account for such funds. Some schools report all patient fee income, while others receive and report only a f raction of the total. The remainder is channeled through independent corporations and hence is not accountable be the Dean's Office. The difference between reported and actual fee income is unknown, but the Committee's Panel on Clinical Sciences believes that more and more of these funds are moving out of the schools' spheres of direct accountabil ity. 8. The clinical faculty data used in this analysis are taken f rom the annual surrey of medical schools conducted by the Liaison Committee on Medical Education and published annually in the Journal of the American Medical Association (AMA, 1960-80) . The def inition of clinical faculty 50

used in that survey is an appointment in a clinical department at any rank. Residents and f ellows are excluded unless actually serving as faculty members. The definition does not mention tenure or tenure-track positions, and so presumably all appointments, whether on the tenure-track or not, are counted as f acuity. 9. The specific form of the model teas been developed from the additional hypothesis that the ratio of full-time clinical faculty to medical student enrollment (F/S) wi 11 vary in accordance with the amount of R and D funds and fee income (M). But because of tenure and other limitations on faculty hiring, the rate of growth in the _/S ratio in response to increasing funds we 11 not be linear but will decrease as the funds increase. This suggests a logistics or Gompertz-type relationship that typif ies many growth processes. Fitting this model to the data for 1961 to 1978 produces the following specif ic f unctional form: (F/S) t = e where: (-1. 6 - 103440/Mt) + 0.13, F = full-time clinical faculty in medical schools; S = medical student enrollment including residents and clinical f ellows; and M = weighted average of clinical R&D expenditures plus fee income in medical schools (R); Mt = 1/4 (Rt+25t-1 + Rt_2, _ — 10. The data shown on p. 253 pertain to all full-time faculty in medical schools, while the Committee's projections are for full-time clinical faculty in medical schools only. Averaging the AAMC data over the years 1980-1985 and allocating proportionately to clinical science faculty gives a projected annual need for 3,166 based on a 3 percent growth estimate. 11. These results should be interpreted with caution. It is easy to confuse "cohorts effects with age effects, e.g., older people may have poorer publication records because the conditions under which they were trained, recruited, and promoted may have differed substantially. 12. Spec ial tabulation prepared for this Committee by NIH, February 2, 1981, of trainees occupy) ng positions awarded in FY 1978. 51

Next: 3. Basic Biomedical Sciences »
Personnel Needs and Training for Biomedical and Behavioral Research: 1981 Report Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!