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 As of 1982, the market opportunities for clinical investigators continued to be favorable, with medical school faculties still growing and providing places for young scientists interested in research careers. The immediate problem remains one of recruiting physicians to undertake research training. In the longer term our projections indicate that the s~cuar~on Is likely to undergo appreciable change. With the possibility of a physician surplus developing, the number of medical schools will stop increasing, enrollments will stabilize or decline, and faculty growth will be slower, thereby reducing the number of positions available for new entrants into the clinical investigator pool. At the same time, greater financial uncertainty in medical schools is likely to aggravate the long-standing difficulty of attracting and retaining high-quality clinical investigators. The problem appears to be related more to the difficulty of obtaining funds for research than to the availability of training positions. Despite reduced employment opportunities, however, shortages of clinical investigators are likely to persist for the next few gears. INTRODUCTION AND OVERVIEW Clinical investigation, as defined by the Committee, includes 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 animal studies by scientists identifiable as clinical investigators on the basis of their other work. Clinical investigation is generally performed in academic health centers. In that environment collaboration with basic scientists is facilitated, appropriate resources for human studies are available, and multi- disciplinary teams are at hand to provide skills needed for comparative assessment of old and new methods. Accordingly, the Committee's assessment of demand for clinical investigators is focused on the 18
19 medical school sector. The latest data on medical school enrollment, as well as faculty professional fee income, and clinical R and D expenditures' in constant 1972 dollars, suggest that this market continued through 1982 to exhibit a relatively strong demand for clinical investigators (Table 2.1~. Highlights of the new data presented in Table 2.1 are as follows: . demand for faculty in clinical departments continued to be strong · clinical R and D expenditures rebounded sharply in 1980 after a decline in 1979, but have fallen back since then . professional fee income continues to be one of the fastest growing revenue items in medical schools, and now far exceeds the level of clinical R and D expenditures · the number of physicians applying for NIH research grants has increased in the last 5 years, but the number of grants awarded on behalf of M.D. principal investigators has not changed appreciably. A point of particular interest in the most recent (1982) data is the finding that the number of full-time faculty in clinical Clinical R and D expenditures are estimated by applying a correction factor to total R and D expenditures reported by the Association of American Medical Colleges. The correction factor for any year is the proportion of total NIH obligations that goes to support clinical research, using NIH's Central Scientific Classification System as the basis for characterizing individual research grants. Financial data are given in constant 1972 dollars unless noted otherwise. 2 Faculty in this report means academically employed, regardless of tenure status or rank. In a medical school, full-time faculty refers to faculty whose salary is paid either in full or in part by the medical school or its affiliated institutions and hospitals. Included are faculty on both strict and geographic full-time. Strict full-time medical school faculty are those who receive their entire professional income as a fixed annual amount from funds controlled by the medical school or its parent institution, who devote their full time to programs of the medical school, and whose professional activities are under the direct auspices and control of the medical school. Geographic full-time medical school faculty are those who receive a guaranteed base salary, all or most of which is paid from funds controlled by the medical school, but who may earn income from professional activities, who conduct all of their professional activities in the institution paying the base salary, and whose professional activities are under the direct auspices and control of the medical school.
20 c: ~ o ~ ~ (t C) et ~ ,r, ~ ~ ~ ~ 3 ~ a: 3 ~ ~ ~ o · ~ _ =: 30 c: ~ O en of o Go at or C ~ Cog.) Cry S . ~ Cal o or _ ,= Cal a V) A: ._ in m ~ ~ ~ o 0 ~ ~ ' o 0 0 ~ ~ 1 6g ~ O _ ~ I I ~ ~r U) U) 00 O. ~ er _ O _ ~ O _ _ ~ ~ ~ ~ O ~ O ~ _ I _ _ I I _ I 1 - } O' ~ ~ 00 °O ~ ~ ~ ~ d2 O Crx X `.S~ C] ~ d O~ d' ~ ~ _ ~ ~ ~i r~ l \ l l l l C~ ~ ~ ~ d 00 et 00 ~ C~ Ct Ce d' ~1' oo x - o ~ ~ ~ ~ ~ oo ') - - - oo o oo ~ ~ ~ ~ ~ o o - ~ ~ ~ - ~ ~ ~ ~ - O O O ~D C~ ~ O ~ ~ ~ ~ ~ _ ~ oN ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ =. ON ~<) c~ ~] °° C~ ~ ~ t ~ ~ ~ ~ ° ~ x r~ ) - ~ ~ ~ - - ~ ~ ~ ~ ~ o cr~ cr ~ ~ ~ ~ ~ ~ oo ~ oo d' ~ ~ ~ 6,~ 6~ ~ ~ ° ~ ~ ~ _ _ ~ ~ ~ _ ~ ~ ~ _ ~ 0 _ ~ ~ ~ 00 ~ 00 ~ 0 ~ oo 0 0 ~ ~ ~ 0 ~ ~ _ ~ ~ t - ) ~ ~ 00 oo ~ 0 o~ ~ ~ ~ _ ~ ~ U) o~ ~ ~ ~ ~ O ~ ~ - ^ O d' ~ ~ ~ r~ 0 ~ _ ~ ~ ~ _ _. _ ~ ~ oo 0 ~ ~ ~ ~ ~ ~ 00 0 X 0 ~ ~ ~ ~ _ ~ ~ 00 ~t ( ON ( - , ~ ~ ~ 0- kD ~ ~ \0 ~ ~ - - ~ O - ^ ~ ~ ~1 ~ O _ ~ _ ~ O ~ \0 _ ~ ~ ~ ~ ° ~ ~ ~ ~ C~ ~ ~ ~ ~ _ ~ U, _ _ ~ o~ ~ ~ ~ oo ~ ~ 0 ~ ~ ~ ~ cr Os L~ (-~4 (~) ·~ oo ~ ~ ~ ) ~ ] _ ~ _ ~ ~ 0 ~ ^ 6O ~ ~ ^ _ C~ - ~' 0 ~ 0 ~ oo ~ ~ _ ~ ~ d' O o ~ ~ ~ oo ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ oo O ~ 00 ~ 0 _ O ~ O cr~ o~ ~ d. O ~ ^ ^ ^ ~ ~ 0 - - o~ _ ~ ~ ~ _ 0 ~ ~ ~ _ ~ ~ ~ 0 d- ~~) O 00 _ ~ ~ O0 ~ _ O ~ ~ ~D 0\ ^ ^ ~ 0 0 _ ~ 0 ~ ~ _ ~ ~r _ ', _ ~ ~ ~ ~ z y ' w ~ o _ =,E ~ . E ' ' ~ ~ ~ ~ ~ 3 ~ _ ~ ,,, ~ ~ ~ ~ ~ .= ~ ~ ·s ~ ~ ~ o ~ ~ ;: a O - z ° 0 ~ ~ ~ R e ~ E O ~ E R = 0 ~ E O ~ ~ ~0 Z ~ Z ~ ~ ~0 ~ ~-a .> ~ C; m ~ O ~ ~ ~ Z ~ ~ Z ~ ~ O ~ ~ ~0 ,~- ~: .E ~ ~ Lu ~ ~ ~ ~ ~ m ~ ~ ~ ~ ~ ~ z ~ ~ ~ _ ~ ~ ~ :~ U~ ~ eD ~ c~ ,D ~ ~ ~V ~ c~s.D {.) ~ ~ Z ~ .D . . . . _ ~ ~ ~ +~ _ U, +_ V, _ _ 11 ~_ U) "4 s3 _ 3 _ c~ 3 ce ^ ._ oo ._ c~ cr~ _ ce _ ~ _ c~ _ s °° c~ .u, _ O ~ ~ 0 ° ._ ,) ~ a.~ ~ c~ ct 0 U~ q,) _ a' ^ oo ~ 0 ~ 0 c~ ~3 t o I-4 ~ o oo ~ - c) ~ ~ :- G) 00 ~ ~ oo ~ - - s ~ ~ c3,o o o cO ·4,) c: o c) ~ ee t ~ ~D · e~ c) ~ ~ - c ~ ~ ~ ~ ·c) . c · = E- '-) ^~= ~ - - · ~ ~ E `: ~ '~' R ~ *p c, oo - - z
21 departments has grown somewhat faster than expected over the past 3 years, apparently financed by higher than anticipated levels of both faculty fee income and clinical R and D expenditures. Full-time clinical faculty was up 17.9 percent in 1982 from the 1979 level--an average growth rate of better than 5 percent per year. There are indications, however, that the market outlook for the next 5 years may be changing, as described below. Slower growth in the medical education complex can be anticipated. Piefe~ionalFeeIncomein Medico Schools During the early 1960s, professional fee income generated by faculty members grew steadily at more than 10 percent per year in real terms. But starting in 1968, a dramatic upswing occurred. These funds grew at an annual rate of more than 25 percent, and overtook funds for clinical research as a source of revenue for medical schools. The latest data show a 15 percent real increase in 1982 over 1981. Although fee income in most schools is the only part of the departmental budget that has been increasing, the prospect for clinical faculties to generate additional revenue seems clouded. The impact of cost containment devices, such as the diagnosis-related group (DRG) Medicare prospective payment plan, as well as increased competition between teaching hospitals and community physicians, make it unlikely that this source of institutional revenue will continue to grow at its previous rates. ClinicaIR and D Expenditures Approximately half of total medical school revenues were derived from research funding in the late 1960s. By contrast, such funds now amount to one-fifth of total revenues. Clinical R and D currently accounts for about 38 percent of total biomedical R and D expenditures (Appendix Table All. In constant dollar terms, estimated clinical R and D expenditures dropped $7 million in 1981 and rose only $5 million in FY 1982. Its real growth rate over the past 7 years has averaged about 3 percent per year. The Committee foresees a somewhat slower growth rate through 1988. Enro ~ ents After more than 30 years of expansion, the nation's medical schools are showing some decline in applications and enrollments. The total of medical students, residents, and clinical fellows has grown at an average annual rate of only 3.2 percent since 1979. Hit simultaneously by steep borrowing rates and substantial tuition increases, student indebtedness continues to grow. Perhaps because of rising costs, as well as possibly changing perceptions of alternative career options, the number of medical school applicants for the
22 1981-82 academic year was 16 percent below that of 1972-73, and the entering class of 1~ 2-83 numbered about 100 fewer' students than in the preceding year (Rorcok, 19833. Based on survey responses from all medical schools, the expectation is for net decreases in first-year enrollment of 47 and 85 in 1983-84 and 1984-85, respectively {AAMC, 1~ 33. Moreover, the possibility of a physician surplus seems to rule out an expansion of medical education in the foreseeable future. Some growth in the number of residents may occur, however, despite evidence of recent reductions in available positions because of cost constraints. This may result from incentives for expansion of graduate medical training contained in the Medicare prospective payment plan recently passed by Congress.3 Also, with the growing complexity of medical services, prolongation of training programs and increase in number of residency years are a possibility. MAINTAINING THE FLOW OF NEW CLINICAL INVESTIGATORS ' The future vitality of clinical investigation depends upon the medical schools' ability to maintain a flow of qualified physician investigators. One measure of the flow is the number of new M.D. principal ' investigators on NIB research grants. Although the number of newly hired physicians in medical schools has more than doubled since 1965, the number of new M.D. principal investigators has remained at roughly 400 (Bryll, 1983~. The proportion of physicians among all first-time principal investigators has gone down from 29.3 percent in 1977 to 24.2 percent in 1980. The number of physicians and other professional doctorates participating in NIB research training programs hasidec~ined on average by more than 6 percent per year since 1975 {Table 2.1, line la). A no-growth situation in medical schools could have serious implications for the future demand for physician investigators. Inasmuch as NIH-supported principal investigators are in large part members of medical school faculties, the future demand is linked closely to the hiring of new faculty which, in turn, is dependent on net increase in faculty size and number needed to replace losses due to attrition. Our projections have shown that a cessation of growth in size of faculty would reduce faculty hiring by more than half (NRC, 1975-81~. New M.D. principal investigators have in recent years constituted only 10-15 percent of the number of anew hires.--a finding consistent with the fact that only 20-25 percent of physician anew 3 In addition to a medical education pass-through, Congress also doubled the current adjustment for indirect teaching costs. With the new adjustment, teaching hospitals will receive additional payment of 12-13 percent for each 0.1 increase in the residents-to-beds' ratio.
23 hires. have had postdoctoral research trainings Because medical schools include the vast majority of physicians holding academic appointments, a decline in hiring of new faculty will affect the demand for new M.D. investigators far more than other new researchers. Further, the increasing dependence of medical schools on practice income is likely to favor the recruitment of clinically-or~ented physicians over physician investigators for the limited number of faculty openings. MONITORING THE CLINICAL INVESTIGATOR POOL The:Committee has for several years been tracking the number of physicians reporting research as a primary activity to the American Medical Association (AMA). These data have been used as an indicator of shifts in interest in research careers on the part of physicians. A consistent annual decrease since 1968 has been followed since 1975 by a steady increase. The latest available data show almost a 6 per- cent rise in 1980 over 1979 (Table 2.1~. This pattern does not conform to data compiled by the Committee on the number of physicians participating in NIH research training programs, nor is it compatible with other indications of physicians' research activities available from the Association of American Medical Colleges. Efforts are underway to identify the causes of the discrepancy. The need to know much more about the population of clinical investigators has prompted the Committee to consider the establishment of a system by which such information could be readily obtained. As a first step, a roster could be established that would include physicians, dentists, veterinarians, and other health professionals with interest or training in research. The data files on NIH/ADAMHA trainees and fellows and principal investigators maintained by the Committee could be combined with the AAMC'S Medical School Faculty Roster, the Dental School Faculty Roster, and files maintained by various professional societies to form a composite data base that would include almost all clinical investigators. Samples could be drawn from the population for use in periodic surveys of research activity, sources of support, training background, and other information vital to the task of monitoring this important pool of scientists. The cooperation of the AMA, AAMC and other professional organizations would be required in developing this project. - 4 The percentage of all new-hired M.D.s with postdoctoral research training fell from 28 percent in 1970 to a low of 20 percent in 1979, with a subsequent rise to 25 percent in 1981 (Sherman and Bowden, 1932~.
24 SHORT-TERM RESEARCH TRAINING The Committee in its 1979 report welcomed NIH's resumption of support for short-term research training as an important step in ' helping'to revitalize interest in the pursuit of clinical investigation careers. Data from a Committee-sponsored study suggest that if the necessary number of clinical investigators is to be maintained, undergraduate medical students should be provided the timpani opportunity to acquire firsthand knowledge of the excitement of working in a research laboratory (AAMC, 1981b). This is consistent with the findings of another study that indicates that research career decisions had largely been made during undergraduate medical school years (Davis and Kelley, 1982~. Starting in 1979 with 5 grantee institutions and 16 trainees, the current program has increased to 60 institutions and 1,026 trainees in FY 1982. As authorized in the 1978 amendments to the NRSA Act, students could pursue training for periods up to 3 months during summer and off-quarters without incurring a payback-obligation. Training' supported by these institutional awards is not restricted to a single' discipline or department. The potential for this type of training has been enhanced by two developments since the Committee';s- last report. Under the most recent amendments to the NRSA Act, the payback obligation is now applicable only to awards in excess of 12 months, thereby increasing a program director's flexibility in planning trainee's research experience. Also, waivers may now permit the payment of short-term stipends within the regular (T-32) training grants. The Committee believes that specific recommendations on the administration and size of this program should await the results' of an evaluation. Accordingly, a two-pronged study is planned to determine the fraction of trainees who maintain their interests in research and who follow career pathways that include research activity. One approach entails an examination of records of trainees from the pre-NRSA era. In this connection, it is to be noted that NIB training grants provided support for periods of 1-3 months to 12,6'45 ' individuals during the years 1960-1974. Subsequent postdoctoral training, NIH/ADAMHA grant activity, academic appointments, and publications will be scrutinized. The other approach will compare the research plans of NRSA short-term trainees with'those of'their ' non-trainee classmates and graduates of the non-grantee institutions. This area of the study will also involve analysis of responses to' the ' AAMC annual survey of graduating seniors. MEDICAL SCIENTIST TRAM PROGRAM The Medical Scientist Training Program (MSTP) supports combined medical and scientific training leading to both the M.D. and Ph.D. degrees. Sponsored by the National Institute of General Medical
25 Sciences (NIGMS), the program has grown from 3 grantee institutions and 17 trainees in FY 1964 to 23 institutions and about 700 trainees currently. Based on an outcome study of the first 53 graduates of MSTP with respect to research retention, rate of advance in academic positions, research grant success, and publication performance, the program has been successful (NRC, 1975-81, 1~ 1 Report). Reflecting our enthusiasm for the program we have previously recommended that high priority be given to protecting MSTP training slots, should it become necessary for budget reasons to reduce the overall number of NRSA trainees. That recommendation has been implemented, even though the support of all research training has generally been under severe restraint. Moreover, trainee slots have been maintained at a time when total expenditures per ~graduate. have become significantly higher on MSTP grants than on other predoctoral training grants, reflecting both a larger annual per trainee cost and a longer period of stipend support for MSTP participants. In its 1979 report, the Committee called for a moratorium on further expansion, pending the development of more analytic information regarding the program. A study during the past year by NIGMS staff indicates that the relative cost of MSTP has been rising steadily over the last few years. Expressed as a percent of total NIGMS funds for predoctoral training, the MSTP share has increased from 16 percent in FY 1977 to 25-28 percent currently. Continuation of this trend could place in jeopardy the support of regular predoctoral programs, which, it should be emphasized, are essential to the continuing vitality of MSTP. The need to curb this growth in costs has therefore become a matter of great concern. A reasonable means for ensuring an appropriate balance would be to retain for MSTP over the near future a share of NIGMS predoctoral training funds that does not exceed 25 percent. The Committee believes that the suggested stabilization of relative cost can be achieved without detriment to quality through introducing various modifications in program administration. One such measure, limiting the period of MSTP support for an individual trainee to a total of 6 years and authorizing discontinuity in support, became effective July 1, 1983. A direct effect of that change will be to encourage greater flexibility on the part of institutions in the operation of their MSTP grants. It should be feasible, for example, for program directors to include within the training sequence periods of support from non-MSTP sources, such as research grants and institutional remission of tuition. In the Committee's view, these changes have the potential of enhancing program output, i.e., to increase the number of graduates per MSTP dollar by means of "freeing up" some trainee slots. Reference was made earlier to a limited 1% 1 study sponsored by the Committee regarding the quality of MSTP output. As the program enters its third decade of operation, the Committee believes the tome Is appropriate to undertake a study of broader scope. It would be useful, for example, to obtain a comprehensive picture of costs, training completion rates, post-training employment histories, scientific accomplishment, etc.
26 INVOLVEMENT OF PH.D. SCIENTISTS IN CLINICAL INVESTIGATION Implicit in the Committee's numerical recommendations for post- doctoral training in the clinical sciences has been the recognition that individuals with other than health professional doctorates engage in clinical investigation. Indeed, early evidence of an appreciable involvement of non-health professionals in clinical investigation emerged from the Committee's 1976 survey of recent Ph.D. recipients in biomedical and behavioral fields (NRC, 197ab). Approximately 31 percent of the respondents in that survey described themselves as engaged in research that directly involved human subjects (or animals in the case of veterinary science research) or samples derived therefrom, as part of a clinically-oriented study. That such involvement had been expanding became apparent from the Committee's analysis of staffing patterns in NIH-funded clinical projects over the 1973-78 period.5 As can be seen from Figure 2.1, Ph.D.s were the - 50 40 30 LLJ lo' 20 LLJ 10 o 73 74 75 76 FISCAL YEAR Other Professional s `_ Ph.D.7 M.~D.s & M.D./Ph.D.s I I I I J 77 78 FIGURE 2.1 Participation of Ph.D., M.D., and other scien- tists on clinical research grants sponsored by NIH, 1973-78. Data are shown as percentage of total paid full-time equiva- lent employment on the grants. See Appendix Table A15. s An Annual NIH Survey, Manpower Report, collected data from principal investigators regarding persons receiving salary from each grant during those years. The NIH Central Scientific Classification System was used in this analysis to identify clinical grants--i.e., those involving human subjects as individuals or as groups. \
27 only degree category of personnel to exhibit a relative increase in full-time equivalent employment on those clinical research grants, rising from 27 percent in 1973 to almost 3S percent in 1978. In earlier reports, the Committee has commented on the declining attractiveness of clinical investigation as a career option for the young physician. It is to be noted in this connection that the proportion of NIH postdoctoral research traineeship and fellowship positions filled by M.D.s in 1981 was about one-half that in 1973 (Table 2.2~. Dur ing that period the number of Ph.D. bioscience researchers increased substantially, with a net growth of 450 per year in the pool of postdoctoral appointees in the academic sector (see Chapter 3, Table 3.1~. AS will be noted below, a growing number of Ph.D. scientists has moved into clinical departments of medical schools (Appendix Table Ark. Under these circumstances, one of the questions examined by the Committee concerned the extent to which the steady increase in the postdoctoral training pool might be used to mitigate a possible shortage of physician investigators. In this section of the report, the Committee addresses that question. Ph.D. Faculty Members with Appointments in Clinical Departments of Medical Schools The clinical departments of medical schools provide a useful starting point for examining the role that Ph.D. scientists play in clinical research. Not only is the preponderance of clinical investigation in the United States performed at that site, but it is also possible to chart in some detail the growing frequency of Ph.D.s with faculty appointments in clinical departments. For example, full-tzme faculty in clinical departments of U.S. medical schools, including pathology departments, trebled between 1967-68 and 1981-82. Compared to other degree types, only the Ph.D.s exhibited an increase in share of the total. As a proportionate share of clinical department faculty, Ph.D.s rose from 11.5 percent in 1968 to 15.1 percent in 1982 (Appendix Table Ark. The M.D., M.D./Ph.D., and mother n groups all declined slightly in percentage terms during this period. The distribution of Ph.D. faculty in the clinical departments is uneven. The top four departments in number of Ph.D.s--psychiatry, internal medicine, pathology, and family practice--accounted for 61 percent of the total in 1981-82 (Appendix Table Arm. As a fraction of the faculty of each clinical department, Ph.D.s ranged from about 5 percent in anesthesiology to over 31 percent in otolaryngology and in psychiatry. Do Ph.D. faculty members in clinical departments differ from their counterparts in basic science departments in medical schools? In terms of the same selected characteristics, how do they differ from their physician colleagues? Relevant to these questions, three groups
28 of medical school faculty members will be compared in the following paragraphs: 1. Ph.D.s with primary appointments in clinical departments 2. Ph.D.s with primary appointments in basic science departments 3. M.D.s with primary appointments in clinical departments. TABLE 2.2 Distnbution of NIH Postdoctoral Trainees and Fellows, by Degree Type, 1971-81a Fiscal Year Total M.D.sb Ph.D.sb . 1971 # 7,540 4,634 2,906 ~ 100.0 6 1.5 38.5 1972 # 7,407 4,474 2,933 % 100.0 60.4 39.6 1973 # 5,478 3,630 1,858 5to 100.0 66.1 33.9 1974 ~ 6,364 3,551 2,813 ~0 100.0 55.8 44.2 1975 # 5,971 2,884 3,087 ~ 100.0 48.3 5 1.7 1976 # 4,910 1,970 2,940 % 100.0 40.1 59.9 1977 # 5,325 1,927 3,398 ~ 100.0 36.2 63.8 Its, 1978 # 5,758 1,984 3,774 % 100.0 34.5 65.5 1979 # 5,814 2,005 3,809 % 100.0 34.5 65.5 1980 # 5,831 2,172 3,659 ~0 100.0 37.2 62.8 1981 # 5,265 1,961 3,304 % 100.0 37.2 62.8 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 International Center programs. b"M.D." and "Ph.D." also include equivalent doctorate degrees. Persons with both M.D. and Ph.D. degrees are shown under "M.D." SOURCE: NIH (special tabulation, 10/7/82).
29 The Faculty Roster maintained by the Association of American Medical Colleges permits a comparison of the three groups in relation to such characteristics as career age, academic rank, postdoctoral fellowship training, tenure status, extent of research participation, and research grant activity (Tables 2.3 and 2.4~. For purposes of the comparison, clinical departments were divided into five departmental categories: medical, surgical, hospital- based, psychiatry, and other clinical .6 Except for psychiatry, inter-category differences were minor. In departments of psychiatry, Ph.D.s were considerably more numerous, constituting almost one-fourth of the 5,857 Ph.D.s in clinical departments in 1~ 1-82. Extent of research participation sharply differentiated psychiatry department faculty members--both Ph.D. and M.D.--from those of all other clinical departments. The percentage of Ph.D. faculty in psychiatry reporting no research participation was almost triple that for all other clinical department Ph.D.s. The percentages of M.D. faculty reporting no research participation were 48 percent and 36 percent for psychiatry departments and for all other clinical departments, respectively. For these reasons, data for faculty in psychiatry departments are shown separately from other clinical departments in Tables 2.3 and 2.4. Clinical department Ph.D.s are younger in career age than both basic science department Ph.D.s and M.D. faculty in clinical departments. Consistent with younger career age, Ph.D.s in clinical departments generally hold lower academic rank than faculty in the other two groups. Ph.D. faculty in clinical departments are less likely to have taken a postdoctoral appointment than PheDes in basic science departments, but more likely to have had such training than M.D.s in clinical departments. Ph.D.s in clinical departments are found much less frequently in tenured or tenure-track status than those in basic science departments. Compared with clinical department M.D.s, however, there is very little difference in this regard. Clinical department Ph.D.s are more likely to have no research involvement than is true for Ph.D.s in basic science departments. The extent of their participation in research, however, is considerably greater than for M.D. faculty in clinical departments, particularly with respect to the percentage for whom research is the primary professional responsibility. Finally, clinical department Ph.D.s stand midway between the other two groups in relation to research grant activity, with basic science Ph.D.s substantially higher than clinical department M.D.s in-their application, approval, and award rates for NIH/ADAMHA grants. 6 Data applicable to the individm 1 departmental categories can be found in Appendix Tables A16 and A17.
30 TABLE 2.3 Statistical Profile of Full-Time Ph.D. Faculty in Medical School Departments, 1982 Department Basic Sci. ainicala Psychiatry Other Total N TO N ~ N TO N TO N ~0 _ CAREER AGE less than 5 744 10.9 842 18.7 275 20.2 130 14.6 1,991 14.6 (Yrs. since 6-10 1,242 18.1 1,025 22.8 286 21.0 173 19.5 2,726 20.1 Ph.D.) 11 or more 4,864 71.0 2,626 58.4 803 58.9 586 65.9 8,879 65.3 _ . TOTAL 6,850 100.0 4,493 100.0 1,364 100.0 889 100.0 13,596 100.0 ACADEMIC Professor 2,328 34.0 786 17.5 323 23.7 133 15.0 3,570 26.3 RANK Assoc. Prof. 2,168 31.7 1,292 28.8 328 24.1 149 16.8 3,937 29.0 Asst. Prof. 1,995 29.1 1,986 44.2 566 41.S 170 19.1 4,717 34.7 Instructor 224 3.3 325 7.2 132 9.7 18 2.0 699 5.1 Other & Unk. 135 1.9 104 2.3 15 1.1 419 47.1 673 4.9 _ TOTAL 6,850 100.0 4,493 100.0 1,364 100.0 889 100.0 13,596 100.0 , . . YEARS OF None POSTDOCTORAL 1-2 RESEARCH 34 TRAINING 5 or more (1981 faculty) 3,160 47.8 2,1 16 32.0 976 14.8 354 5.4 TOTAL 6,606 100.0 2,581 63.4 1,081 8 1.7 637 7 1.4 7,459 57.8 934 22.9 175 13.2 175 19.6 3,400 26.4 390 9.6 41 3.1 57 6.4 1,464 1 1.4 165 4.1 26 2.0 23 2.6 568 4.4 4,070 100.0 1,323 100.0 892 100.0 12,891 100.0 _ TENURE Tenured 3,431 50.1 1,171 26.1 411 30.1 421 47.4 5,434 40.0 STATUS Tenure Track 1,473 21.5 1,045 23.3 240 17.6 148 16.6 2,906 21.4 No Tenure 863 12.6 1,422 31.6 444 32.6 149 16.8 2,878 21.2 Other & Unk. 1,083 15.8 855 19.0 269 19.7 171 19.2 403 17.4 TOTAL 6,850 100.0 4,493 100.0 1,364 100.0 889 100.0 13,596 100.0 . RESEARCH None 321 4.7 438 9.7 358 26.2 217 24.4 1,334 9.8 PARTICIPATION Some 5,381 78.6 2,755 61.3 757 55.5 521 58.6 9j414 69.3 Primary 1,003 14.6 1,060 23.6 184 13.5 94 10.6 2,341 17.2 Other & Unk. 145 2.1 240 5.3 65 4.8 57 6.4 507 3.7 TOTAL 6,850 100.0 4,493 100.0 1,364 100.0 889 100.0 13,596 100.0 NIH/ADAMHA Rate (%) Rate (DO) Rate (Jo) Rate (DO) Rate (%) RESEARCH ApplicationsC 3,979 58.1 1,835 40.8 270 19.8 151 17.0 6,235 45.9 GRANT Awarded 1,241 31.2 455 24.8 78 28.9 22 14.6 1,796 28.8 ACTIVITYb Approvalse 3,497 87.9 1,527 83.2 182 67.4 112 74.2 5,318 85.3 (Competing applications only) aExcludes departments of psychiatry. Ph.D.s in psychiatry departments appear to have different characteristics from those in over clinical departments. See Appendix Table A16 for more details. bFor these data, Basic Science includes pathology departments. CApplication rate = # applications/se faculty members. dAward rate = ~ awards/se applications. eApproval rate = # approved applications/se applications. SOURCES: AAMC (1966-83, special tabulation prepared by George Bowden, NIH, 6i9/83); NRC (1979-83, special tabulation, 6/11/83).
31 TABLE 2.4 Statistical Profile of Full-Time M.D. Faculty in Medical School Departments, 1982a Department Basic Sci. ainicalb Psychiatry Other Total N % N TO N % N TO N % . . CAREER AGE less than 5 6 0.9 1,133 4.4 81 3.4 10 2.6 1,230 4.2 (Yrs. since 6-10 33 5.2 4,154 16.0 371 15.4 10 2.6 4,568 15.4 M.D.) 11 or more 595 93.9 20,662 79.6 1,965 81.3 369 94.9 23,591 80.3 TOTAL 634 100.0 25,949 100.0 2,417 100.0 389 100.0 29,389 100.0 ACADEMIC Professor 424 66.9 7,490 28.9 618 25.6 84 21.6 8,616 29.3 RANK Assoc. Prof. 122 19.2 6,334 24.4 569 23.5 29 7.5 7,054 24.0 Asst. Prof. 71 11.2 9,704 37.4 1,016 42.0 22 5.7 10,813 36.8 Instructor 11 1.7 2,25 0 8.7 203 8.4 0 0.0 2,464 8.4 Other & Unlc. 6 0.9 171 0.7 11 0.5 254 65.3 442 1.5 . TOTAL 634 100.0 25,949 100.0 2,417 100.0 389 100.0 29,389 100.0 YEARS OF None 295 43.8 17,685 72.6 2,108 87.5 283 73.3 20,371 73.2 POSTDOCTORAL 1-2 189 28.1 4,454 18.3 198 8.2 72 18.7 4,913 17.7 RESEARCH 3~4 115 17.1 1,555 6.4 51 2.1 19 4.9 1,740 6.2 TRAINING 5 or more 74 11.0 649 2.7 51 2.1 12 3.1 786 2.8 (1981 faculty) TOTAL 673 100.0 24,343 100.0 2,408 100.0 386 100.0 27,810 100.0 TENURE Tenured 405 63.9 7,917 30.5 665 27.5 211 54.2 9,918 31.3 STATUS Tenure Track 57 9.0 5,800 22.4 526 21.8 30 7.7 6,413 21.8 No Tenure 55 8.7 7,078 27.3 763 31.6 61 15.7 7,957 27.1 Other & Unk. 117 18.4 5,154 19.9 463 19.2 87 22.4 5,821 19.9 TOTAL 634 100.0 25,949 100.0 2,417 100.0 389 100.0 29,389 100.0 RESEARCH None 60 9.5 9,274 35.7 1,158 47.9 228 58.6 10,720 36.5 PARTICIPATION Some 482 76.0 14,356 55.3 1,092 45.2 142 36.5 16,072 54.7 Primary 73 11.5 1,083 4.2 76 3.1 5 1.3 1,237 4.2 Other & IJnk. 19 3.0 1,236 4.8 91 3.8 14 3.6 1,360 4.6 TOTAL 634 100.0 25,949 100.0 2,417 100.0 389 100.0 29,389 100.0 NIH/ADAMHA Rate (%) Rate (%) Rate (%) Rate (DO) Rate (%) RESEARCH Applicationsd 741 116.9 2,962 11.4 222 9.2 39 10.0 3,964 13.5 GRANT Awardse 256 34.5 901 30.4 64 28.8 13 33.3 1,234 31.1 ACTIVITYC Approvalsf 660 89.1 2,434 82.1 119 5 3.6 36 92.3 3,249 82.0 (Competing applications only) aExcludes M.D.s who also hold a Ph.D. degree. bExcludes departments of psychiatry. Faculty members in psychiatry departments appear to have different characteristics from those in other clinical departments. See Appendix Table A17 for more details. CFor these data, Basic Science includes pathology departments. dApplication rate = # applications/se faculty members. eAward rate = # awards/se applications. fApprova1 rate = # approved applications/se applications. SOURCES: AAMC (1966-83, special tabulation prepared by George Bowden, NIH, 7/20/83); NRC (1979~3, special tabulation, 7/29/83).
32 Role ofPh.D.sin Climcaltavestigation The Ph.D. scientist in a clinical department generally engages in clinical investigation as an independent researcher, as a co-investiga- tor, or as a provider of specialized skills involved with complex laboratory instrumentation and technologies. As independent investi- gators, Ph.D. epidemiologists have been largely responsible for the rapid pace of advancement in this discipline, particularly in the development of methodologic and analytic strategies. A 1979 symposium report provides examples of the Ph.D.'s role as a catalyst in initiating collaborative research within the clinical setting (Gillis, -1979~. The report cites a collaborative study of pulmonary clearance of vasoactive hormones in patients that was a direct outgrowth of the basic scientist collaborator's earlier animal work. This role is also demonstrated in phase T drug studies. Here the Ph.D. pharmacologist is uniquely suited to gather information on how a drug introduced for the first time in man is absorbed, metabolized, and eliminated. Functioning as a source of laboratory expertise, the Ph e D ~ scientist often provides access to research tools that the clinically trained investigator might otherwise be less likely to exploit. Opportunity for the application of monoclonal antibody s, spectroscopic analysis of metabolic function, and molecular genetics to clinical problems are but a few examples. It is relevant in this connection that the National Council of the American Federation of Clinical Research decided in the past year to include faculty openings in Clinical Research's positions Available Listing. for Ph.D. scientists in . . departments of medicine and pediatrics (Zusman, 1% 3). Clinical investigation takes many forms, ranging from a study of isolated tissue components to large-scale drug trials involving thousands of subjects and the pooled effort of many investigators in multiple sites. In that connection, two questions arise. How can these diverse forms of research be categorized? How is research effort--projects and dollars--distributed in terms of those categories? Answers to those questions would indicate, for example, whether the trend was now to use the human subject primarily as a source of materials or whether participation of the intact human continued to be essential to the research. A recently formulated taxonomy by Landau, as described by Bever (1980), is useful for this purpose. It consists of six major categories and numerous subgroup- ings within each category. The first two categories, for example, include: 1} research in which the subjects or their environments are manipulated experimentally, and 2) research on the management of a disease through available diagnostic and therapeutic modalities. Legal, ethical, and regulatory constraints require that a physician play the pr imary investigative role for pro jects in those two categories. The remaining four categories of clinical investiga- tion involve materials derived from humans; animal studies or models for human experimentation; epidemiologic studies; and miscellaneous projects (communications, data systems, etc.~. For research in those categories, a Ph.D. frequently, or even predominantly, can serve as the principal investigator.
33 The distribution of NIH awards for clinical investigation in these categories is pertinent to the present discussion. Lipsett (1981) found that 1979 NIH grant-supported projects falling into Landau's first two categories represented the overwhelming majority of ~human. projects supported by most Institutes. The funds involved tended to parallel the number of projects. It was his conclusion, therefore, that a physician must take the lead role in most of the NIH-sponsored research on human subjects. This conclusion is reinforced by the findings of an earlier study on NIAMDD grants for human-related studies over a 9-year period (Bever, 1980~. For projects falling in Landau's first two categories, Bever determined that the principal investigators were almost exclunivelv M D ~ or Irwin with inch The M.D. and Ph.D. degrees. The Ph.D. scientist's role in clinical research has been enhanced not only by the increasing complexity and sophistication of biomedical technology and instrumentation, but also by a changing emphasis in clinical investigation. Analysis of research abstracts associated in_ _ __ _ `, _ _~ ^-^ e ~ ~ ~ ~ rat ~ With the annual Atlantic City Spring Meetings. of the American Federation for Clinical Research, the American Society for Clinical Investigation, and the Association of American Physicians for various years from 1953 through 1969 is informative (Feinstein and Ross, 1967, 1970, Feinstein et al, 1967~. The analysis reveal ed that the proportion of ~clinical. topics--as evidenced by papers that were either patient-centered, disease-oriented, or concerned with human material--had progressively declined, concomitant with a steady increase in the proportion of "basic. investigations, the materials for which were neither human nor diseased. More recently, Fishman and Jolly (1981) used the table of contents of one issue of the Journal of Clinical Research to demonstrate that research in clinical departments frequently "falls within the traditional purview of basic science departments." In keeping with the trend toward integration of the clinical and basic sciences, the NIH has announced a new program--the Physician Scientist Award--designed to promote opportunities for young clinicians to develop research skills and experience in a fundamental science. The Committee has repeatedly stressed the importance of the physician-scientist in clinical investigation. This emphasis derives from the M.D.s irreplaceable role in bringing clinical insights to bear in the laboratory and in translating basic observations into clinical practice. At the same time, factors such as the fall off in number of physicians seeking research training, increasing clinical activity on the part of medical school faculty, and the growing complexity of biomedical technology have opened new opportunities for the basic scientist in clinical investigation. A growing movement of young Ph.D. scientists into clinical departments of medical schools--for teaching, service, and especially for research activities--has occurred over the past 10-15 years. Indeed, research is now the primary functional activity for almost one-fourth of Ph.D. faculty in clinical departments. Of that group, some serve as independent investigators, conducting research on tissues and other substances of
34 human origin, developing biophysical and biochemical tests, pursuing animal studies, etc. Most function as collaborators in inter- disciplinary investigations. Moreover, the participation of Ph.D. scientists helps to ensure the application of state-of-the-art technology to clinical studies. Barring unforeseen change, the Committee expects that for reasons cited earlier, this increasing involvement in clinical investigation by non-medically trained scientists will continue for the near term and will contribute to the advancement of clinical science. THE MARKET OUTLOOK A substantial part of the growth in medical school enrollment since 1960 has been due to the growth in number of schools, in addition to the expansion of class size. But with the demise of capitation grants, the pressure for class expansion will lessen. Therefore, any future enrollment growth for the next few years will likely be determined largely by the number of new schools to be formed, and there will probably not be many. Almost all of the new schools formed since 1961 have been public schools {see Figure 2.2~. In 1962 there were 44 public and 42 private schools. In 1~ 1 these had increased to 75 public and 51 private schools--average annual growth rates of 2.8 percent and 0.8 percent, respectively. Therefore we would expect the public schools to have more faculty, students, funding, etc. And that is what we observe except in the case of faculty. Whereas the number of students in public medical schools began to increase faster than private school students in the mid-1960s (Figure 2.3), the number of clinical faculty in both cases has remained quite close (Figure 2.4~. Private schools thus have much higher faculty/student (F/S) ratios than do public schools (Figure 2~51. These higher ratios are apparently financed by higher levels of R and D funding and professional fee income in private schools (Figure 2.63. Public medical schools tend in general to focus more on teaching than research activities. The Demand Mode} for Clinical Faculty The above results are entirely consistent with the model we have developed to help analyze demand for clinical faculty in medical schools. According to that model, the F/S ratio is dependent on funds generated by R and D expenditures and fee income (hereafter denoted as "clinical funds. Therefore a high F/S ratio should be associated with a high level of funds, and since private schools have higher F/S ratios, we would expect to find higher levels of these funds in private schools than in public schools. As can be seen from Figure 2.6, that is definitely the case. For the period 1971-80, the average level of clinical funds per school was 30 to 60 percent higher in
c 35 130 120 110 100 80 70 60 50 40 Al l School s , - Publ ic School s J ~~ Private Schools o l ~ 62 63 64 65 66 67 68 69 70 71 72 73 74 75 FISCAL YEAR 1 ~ 76 77 78 79 80 81 82 FIGURE 2.2 Number of accredited and provisionally accredited U.S. medical schools, by control of institution, 1962-82. See Appendix Table A4. 130 120 110 100 90 80 70 En 40 30 10 - - _ - - Al 1 ~y - - - - - - Publ ic Schnol s_ - O I I , . , , , I · I , · , , I , ~ , , 1 1 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 FISCAL YEAR FIGURE 2.3 Medical students, residents, and clinical fellows at U.S. medical schools, by control of institution, 1961-82. See Appendix Table Al.
36 45 40t 35 v) 30 v 25 20 15 Al l School s_ ~ Pri Yate School s - . 10 ~ ~~==~ Public Schools' . I I I I I 1 1 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 FISCAL YEAR FIGURE 2.4 Full-time clinical faculty in U.S. medical schools, by control of institution, 1961-82. See Appendix Table A2. 0.40 0.38 0.36 0.34 0.32 0.30 0.28 0.26 Hi. 0. 24 0.22 0.201. 0.18 0.16 0.1: _ '' Private School so_ hi - 63 64 65 66 67 68 69 70 71 72 73 FISCAL YEAR 74 75 76 77 78 79 80 81 82 FIGURE 2.5 Ratio of full-time clinical faculty (F) relative to total of medical students, residents, and clinical fellows (S) at U.S. medical schools, by control of institution, 1963-82. See Appendix Table A3.
37 11,000 10,000 9,000 8,000 7,000 6,000 I: ° 5,000 4,000 _ 3,000 Private Schools I_ iAl l School so Publ ic School sat - O ' ~ ~ I ~ ~ ~ I ~ ~ · ~ I · . . . . . . 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 FISCAL YEAR FIGURE 2.6 Average funds per school for clinical R & D expenditures + fee income, by control of institution, 1962-82 (1972 $, thousands). See Appendix Table A9. private medical schools. During the same period, the private school F/S ratio was 20 to 50 percent higher than the public school ratio.7 Data for the period 1962-80 confirm that such a relationship between the F/S ratio and clinical funds does indeed exist for all U.S. medical schools and takes the form of a constrained growth model--the F/S ratio rose rapidly in the 1960s at low levels of clini- cal funds and less rapidly in the 1970s at higher levels of clinical funds {Figure 2.7~. The flattening of the curve in recent years is probably due to rapidly rising indirect costs which generally siphon funds away from faculty support. In past reports, we have constructed the F/S ratio using annual values of F and S. This year we have modified the ratio such that the denominator is a weighted average of the last four annual values of S. The rationale for this modification is the belief that faculty size does not change in response to changes in current year 7 Faculty is defined in this model as full-time faculty in clinical departments, while students are defined as the total of medical students, residents, and clinical fellows.
38 ! 0.4 - o at_ 0.2 0.1 .*~- //! 1{ 0 1 2 3 4 5 6 7 8 9 10 11 CLINICAL FUNDS PER SCHOOL (M), (1972 $, millions) /./ _ . . .. . FIGURE 2.7 Clinical faculty/student ratio (CF/WS) vs. clinical funds per school (M) for all medical schools. The ratio is defined as follows: CF = full-time faculty in clinical de- partments of U.S. medical schools; WS = 4-year weighted average of students, i.e., Beset = 1/6(St + 2St-1 + 2St-2 + St 3), where S = total of medical students, residents, and clinical fellows. Clinical funds are defined as clinical R & D expenditures (R) plus professional service income (P) in medical schools. M is defined as a Smear weighted average of R+P: Mt = 1/4~(R+P)t + 2(R+P)t 1 + (R+P)t 2] . Solid line represents a growth curve of the form Y = exp (a-b/x)+c fitted to the data for 1964-80. Parameter values are: a = ~.98517, b = 983.18, c = 0.01. Broken lines represent 95% confidence limits on the fitted curve. See Appendix Tables AS and A9.
39 enrollments, but that the effect is spread over several years. We use a growth function of the following form to model these data: CF/WS = ea-b/M +C where: OF = full-time clinical faculty in U.S. medical schools WS = 4-year weighted average of students, i.e., (WS) t=1/6 (St+2St-l+2St-2+St-3 ~ where S = total of medical students, residents, and clinical fellows M = 3-yr. weighted average of clinical R and D expenditures plus professional service income per school (1972 $, millions) c = scaling constant: CF/WS = c when M = 0 a,b = parameters to be determined empirically With this type of growth function, the percentage change in CF/WS is assumed to increase proportionately to 1/{M2~. The function has an inflection point at b/2 and is asymptotic to eat Similar curves also were derived separately for public and private medical schools. These are shown in Figure 2.8 along with the 95 percent confidence Digits in each case.9 Fitting this model to data from 1964-80, we get the following estimates for the parameters: a = -0.98517; b = 0.% 32; c = 0.01. The values were derived from 17 annual observations by a least-squares regression procedure which yields an R2 of 0.932. R2--the coefficient of determination--must lie between O and 1 and is a measure of how well the assumed function fits the data, with R2 = 1 representing a perfect fit e The dotted lines in Figure 2.7 represent the 95 percent confidence limits on the estimated curve. The curve is asymptotic to CF/WS = 0.38 and has an inflection point at M = $0.5 million per school. 9 The functional form of these curves is the same as before (although with different parameter values), i.e.: CF/WS = ea~b/M +c. Parameter estimates for each case were also derived from 17 annual observations between 1964 and 1980. These are as follows: Public schools a = -1.266 b = 0.8691 c = 0.05 R2 = 0.84 Private schools a = -0.8088 b = 1.2892 c = 0 R2 = 0.93
40 Note in Figure 2.8 that at low levels of clinical funds--below $2 million per school for example--the CF/WS ratios in both public and private schools are quite similar. At levels above 32 million per school, the ratios diverge, with private schools having substantially higher ratios for the same level of clinical funds. This implies a difference in emphasis between public and private medical schools with regard to the way clinical funds are expended. One hypothesis it that private schools tend to use the funds more for faculty support while public schools tend to use the funds more for other purposes. 0.42 0.40 0.38 0.36 3 4 5 6 CLINICAL FUNDS PER SCHOOL (M), (1972 $, millions) Private School' Publ ic School s *= Private Schools 61 = Publ i c School s 7 8 9 FIGURE 2.8 Clinical faculty/student ratio (CF/WS) vs. clinical funds per school (M), by control of institution. The ratio is defined as follows: CF = full-time faculty in clinical departments of U.S. medical schools; WS = 4-year weighted average of students, i.e., (WS)t = 1/6(St + 2St 1 + 2St 2 + St 3), where S = total of medical students, residents, and clinical fellows. Clinical funds are defined as clinical R & D expenditures (R) plus professional service income (P) in medical schools. M is defined as a 3-year weighted average of R+P: Mt = 1/4~(R+P)t + 2 (R+P)t 1 + (R+P)t 2] . Solid line represents a growth curve of the form Y = exp~a-b/x)+c fitted to the data for 1964-80. Broken lines represent 95~o confidence limits on the fitted curve. See Appendix Tables AS and A9.
41 Assumptions To use the model for projections, it is necessary to make assumptions about the future behavior of the three elements that drive the model--students, R and D funding, and professional fee income. 1. Enrollments: medical school enrollments, defined as medical students, residents, and clinical fellows, are expected to show no growth between 1982 and 1988. The upper and lower visits on the expected growth rate are +2.5 percent per year and -2.5 percent per year, respectively (see Figure 2.9~. so an AL 120 110 100 _ 90 o - ~n z cat o cat :~: z 80 70 60 50 40 30 20 10 O ' I i , , 60 62 64 66 68 70 _ Actual -- Projected - Al l School s / _: Publ ic _' School s)/ _' ~;~! --en ~ School s hit,. ;~ ~ ~~ satyr.) _e__________. r ~~J 72 74 76 78 80 82 84 86 88 FISCAL YEAR FIGURE 2.9 Medical students, residents, and clinical fellows, by control of institution, 1961-81, with projections to 1988. See Appendix Table Al.
42 2. Clinical R and D expenditures: the best-guess assumption is for a 2 percent per year growth in these expenditures between 1981 and 1988 after adjusting for inflation. The upper and lower limits are 4 percent per year and O percent per year, respectively.(see Figure 2.10~. so 4 2 1 o Actual Projected Private J f School s )~ /~ Off ~ W/~N ^/~ 62 64 ,_. High Estimate (4X/yr. ) _, ,_--' ~ ma- ME ddl e Estimate ( 2%/yr. ) - Low Estimate (0%/yr. ) Publ i c School s 66 68 70 72 74 76 78 FISCAL YEAR 80 82 84 86 88 FIGURE 2.10 Clinical R & D expenditures per school in U.S. medical schools, by control of institution, 1962-80, with projections to 1988 (1972 $, millions). See Appendix Table A9. 3. Professional service income: the best-guess assumption is for real growth (after adjusting for inflation) of 2 percent per year from 1982 to 1988. Expected upper and lower limits are 5 percent per year and -1 percent per year, respectively (see Figure 2.11~.
43 9f 8t 7 6: 3~ 2: 1 o -- Actual I,' Projected ,,_ ,~ <2~0lyr ) Private ~ ~ - _iniate (~~%/yr.) School scam ~ ~ School s Ad// Publ icy) ,~ School s 3'q I , . . . . . . . ~ . . ~ 62 64 66 68 70 72 74 76 78 80 82 84 86 88 PI SCAL YEAR FIGURE 2.11 Professional income per school reported by U.S. medical schools, by control of institution, 1962-80, with projections to 1988 (1972 $, millions). See Appendix Table A9. Projections of Demand for Clinical Faculty to 1988 Given the Panel's assumptions about enrollments, clinical R and D expenditures, and professional income in medical schools, we now use the model to make projections of demand for clinical faculty. Following our usual practice, projections are made for about 5 years ahead of the report, so this year the projections go through 1988 as shown in Figure 2.12 and Table 2.5. Under the most optimistic assumptions about clinical R and D expenditures and professional income (assumption I in Table 2.5), these aggregated clinical funds would grow by 4.7 percent per year through 1988 to about $11 million per school, driving the CF/WS ratio to 0.35 from its current value of 0.33. The 98 percent confidence fits on this estimate are 0.366 and 0.334, respectively. Since the most optimistic assumptions attempt to define an upper limit on our projections, we use the upper 95 percent confidence licit on CF/WS (0.366) as the most optimistic estimate. We project academic demand by using the most optimistic estimate of enrollment growth--2.5 percent per year (assumption A of Table 2.5--together with the estimated CF/WS ratio of 0.366. This produces an estimated upper limit for clinical faculty size of 49,640 members by 1988, for a faculty growth rate of 3.9 percent per year.
44 55 50 45 40 35 30 25 20 15 10 Actual em-- Projected Al l School s Ad/ / Private Schools ID o INTO\{ Apia (;'' -'' Middle Estimate (1.3~/yr.) -_ 1. 2X/~ Publ ic School s , , 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 FISCAL YEAR FIGURE 2.12 Clinical faculty in U.S. medical schools, by control of institution, 1961-81, with projections to 1988. Faculty is defined here as full-time appointments in clinical de- partments regardless of tenure status. See Appendix Table A2. About 1,620 positions would be created by expansion, with another 430 created by attrition due to death and retirement, and 1,900 created by other faculty attrition. The total number of clinical faculty positions that would become available each year under these high growth asssumptions is estimated at 3,950. Under the middle or best-guess assumptions {IT-B of Table 2.5), clinical funds would expand by about 2 percent per year through 1~8-- yielding ~ CF/WS ratio of 0.344--and enrollments would remain at 1981 levels. The best estimate of clinical faculty size under these assumptions is 40,700, an increase of 500 positions per year or 1.3 percent per year over the 1980 level. Attrition from all causes would add another 2,090 positions to give an estimated total annual demand for clinical faculty of 2,590. This is the Committee's most likely projection. Under the low growth assumptions (III-C of Table 2.5), clinical funds would expand very slightly to about $0.7 million per school by 1~8. Consequently, the estimated CF/WS ratio would be 0.336 with upper and lower 95 percent confidence limits of 0.350 and 0.323, respectively. Using the lower estimate of 0.323 to represent the most pessimistic conditions, together with the lowest enrollment growth
45 TABLE 2.5 Projected Growth in Medical School Clinical Faculty, 1980-88, Based on Projections of Medical School Enrollment, Clinical R and D Expenditures, and Medical Service Income in Medical Schoolsa Assumptions about Real R and D Expenditures and Professional Service Income (in constant 1972 dollarsb) in Medical Schools ($7.6 million per school in 1980) I II III Will expand at Will expand at Will decline about 4.7~O/yr. about 2.05~0tyr. slightly by about Assumptions about Medical Student Enrollment to $11.0 million to $8.9 million 0.7~O/yr. to $7.2 (medical students, residents, and clinical fellows) per school in per school in million per (118,300 in 1981) 1988 1988 school in 1988 A. Will grow at 2.5%O/yr., Expected size of clinical faculty in reaching 141,000 medicalschools (CF)in 1988 49,640 46,600 43,800 students by 1988 Annual growth rate in CF from 1980 to 1988 3.9 3.0~o 2.2~o Average annual increment due to faculty expansion 1,620 1,240 890 Annual replacement needs due to:C death end retirement 430 420 400 other attrition 1,900 1,830 1,770 Expected number of positions to become available annually on clinical faculties 3,950 3,490 3,060 B. Will show essentially Expected size of clinical faculty no growth from 1981 in medical schools (CF) in 1988 43,300 40,700 38,200 to 1988, remaining at Annual growth rate in CF from 118,300 students 1980 to 1988 2.1% 1.3% 0.5% Average annual increment due to faculty expansion 830 500 190 Annual replacement needs due tore death and retirement 400 390 370 other attrition 1,760 1,700 1,650 Expected number of positions to become available annually on clinical faculties 2,990 2,590 2,210 C. Will decline by 2.5~O/yr. Expected size of clinical faculty in to 99,100 students by medicalschools (CF) in 1988 37,700 35,400 33,200 1988 Annual growth rate in CF from 1980 to 1988 0.3% -0.4~o -1.2% Average annual increment due to faculty expansion -130 -160 -430 Annual replacement needs due to:C death and retirement 370 360 350 other attrition 1,640 1,590 1,540 Expected number of positions to become available annually on clinical faculties 2,140 1,790 1,460 aPaculty in this table is defined as a full-tune appointment in a clinical department regardless of tenure status. These projections are based on the following relationship: (CF/WS)t = exp (-1.681 3 - 1108.8/Dt) + 0.05, where CF = size of clinical faculty in medical schools; WS = weighted average of last 4 years of enrollments, i.e., (WS)t = 1/6(St + 2St 1 + 2St 2 + St 3), where S = medical students, residents, and clinical fellows; D = weighted average of last 3 years of clinical R and D expenditures plus medical service income per school, i.e., Dt = 1/4(Dt + 2Dt 1 + Dt 2). See Appendix Tables Al, A3, and A10. Deflated by the implicit GNP Price Deflator, 1972 = 100.0. See Appendix Table A7. CBased on an estimated replacement rate of 1.0~o annually due to death and retirement, and 4.4 annually due to other attrition. See AAMC (198 la).
46 assumption of -2.5 percent per year, we estimate clinical faculty size under the worst-case conditions to be about 33,200 members--a decrease of 430 positions per year from 1~ 0 levels. But attrition would create an estimated 1,890 positions per year for a minimum net demand of 1,460 positions. Estimating Postdoctoral Support Levels Under NRSA Programs The final step in our quantitative analysis of the market is to attempt to translate the projections of academic demand into recommended levels of postdoctoral training under NRSA programs. This step requires certain additional assumptions about how the system has functioned in recent years with regard to postdoctoral training and its sources of support. The features of the system that must be considered in addition to the projections of medical school faculty growth are as follows: 2. 5. contributions to academic demand generated-by: a) the need to reduce budgeted vacancies in clinical departments b} demand for clinical faculty in dental and veterinary schools the number of accessions to clinical faculty positions who have (or should have} research training 3. the appropriate length of the research training period 4. the proportion of individuals in the research training pipeline who aspire to academic careers the proportion of support of the total pool of clinical research trainees that should be provided by the federal government. In the absence of complete knowledge of the system, we must make additional assumptions about these features--first presented in the Committee's 1~ 1 report--in order to provide a quantitative basis for the recommendations. Using the projections of academic demand derived in Table 2.5, and the same set of conditions specified in the 1981 report,-we calculate in Table 2.6 the range of clinical science postdoctoral trainees that should be supported by NRSA programs under the specified conditions. Line 1 of Table 2.6 is a summary of the projections of academic demand for the extreme cases and the best-guess estimate derived in Table 2.5. Line 2 is an estimate of the demand generated by the need to reduce budgeted vacancies in clinical science departments of medical schools. Line 3 provides an estimate of the demand for clinical faculty in veterinary and dental schools--estimated at 16 percent of medical school demand. Line 4 shows the total annual demand for clinical faculty under each set of conditions. Total annual academic demand is expected to be between 1,860 and 4,750 positions with a best guess of about 3,170 · ~ positions ~
47 TABLE 2.6 Estimated Number of Clinical Research Postdoctoral Trainees Needed to Meet Expected Demand for Clinical Faculty Through 1988 Under Various Conditions Projected Through 1988 Annual High Middle Low Average Estimate Estimate Estimate 1979-81 1. Demand for full-time clinical faculty- annual average: 3,950 2,590 1,460 3,780 a. due to expansion of faculty 1,620 500 -430 1,830 b. due to death and retirements 430 390 350 360 c. due to other attrition b 1,900 1,700 1,540 1,590 2. Demand created by unfilled positionsC 170 170 170 3. Demand for veterinary and dental school clinical faculty (16% of med. school demand)d 630 410 230 4. Total annual accessions (expected demand) 4,750 3,170 1,860 Total accessions with postdoctoral research training-annual average (assuming 35~O of all accessions have postdoctoral research training)e 6. Size of clinical science postdoctoral pool- 1,660 1,1 10 6s0 annual average 3,000-5,000 Size needed to meet academic demand assuming a 2-yr. training period and portion of trainees seeking clinical faculty positions is: a. 60% 8,300 3,700 2,170 b. 50~O 9,960 4,440 2,600 7. Annual number of clinical science postdoctoral trainees to be supported under NRSA programs: a. if 50% of pool is supported under NRSA 4,150~,980 b. if 60% of pool is supported under NRSA 4,980-5,980 . 2,866 1,850-2,220 1,080-1,300 2,220-2,660 1,300-1,560 aAssumes an attrition rate due to death and retirement of 1.0% per year. See AAMC (1981a). bAssumes an attrition rate due to other causes of 4.4% per year. See AAMC (198 la). CIn 1981 there were 2,231 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 1988 is about 170 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 410 per year. eAccessions are defined as new hires or those who rejoin faculties from nonfaculty positions. Interfaculty 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. SOURCE: Table 2.5.
48 Line 5 shows the number of clinical faculty positions to be filled ,; , . .. .. . by individuals with postdoctoral research training experience assuming that 35 percent of all accessions to academic positions will be former postdoctoral trainees. In the best-guess case, this number is estimated to be 1,110. Line 6 indicates the size of the clinical science postdoctoral pool required to supply the necessary number of individuals with postdoctoral training under certain assumptions about the length of the postdoctoral training period and the proportion of the pool seeking academic employment. If the appropriate length of postdoctoral research training in the clinical sciences is 2 years, then the pool size needed to produce 1,110 trained scientists each year would be 2,220. If only 50 percent of the trainees seek academic appointments after completing their training, then the necessary pool size must be 4,440. We assume that some support for postdoctoral research training is also available from sources other than the NRSA programs. This is dealt with in line 7 of this table. Line 7 shows the estimated number of clinical science postdoctoral trainees that should be supported annually by NRSA programs under different assumptions about the proportion of total support provided by that source. The resulting range is between 1,080 under the lowest set of assumptions, and 5,~ 0 under the highest set. The best-guess assumptions yield a range of 1,8S0-2,660 postdoctoral trainees in the clinical sciences. Long-Term Considerations The foregoing analysis is an attempt to translate the Committee's assessment of enrollments and funding in the next few years into projections of academic demand for clinical faculty and ultimately into training levels needed to satisfy that demand. Along the way we are forced to make critical assumptions about how the system has worked in the past and how it will work in the future. Clearly, the end results are quite sensitive to these assumptions. In effect, what has been done is to combine expert judgment of future trends with a conceptual view of how the training system operates and how clinical investigators are absorbed into academic positions. Although we may have identified in this analysis most of the important features of the system that must be considered--attrition rates, length of training period, proportion of support provided by federal programs, for example--our knowledge of all the parameter values in the system is admittedly incomplete. In recognition of this, we have provided a fairly wide range of estimates. This is partly a reflection of incomplete knowledge, but also reflects the uncertainty inherent in any projection exercise. These caveats notwithstanding, it seems clear that even under the most optimistic set of assumptions made by the Committee, the size of the clinical faculty in U.S. medical schools will not expand as fast as it has in the recent past.
49 The long-term problem then becomes one of how to maintain the flow of quay ified new entrants into the field of clinical investigation in the face of declining opportunities in the academic sector where most clinical research is conducted. It is a basic premise of this Committee that the maintenance of such flow is vital to the research enterprise. A potential no-growth situation in medical schools will tend to decrease this flow as well as the growth of clinical research itself. This dilemma is created by both demographic and economic factors. A projected surplus of physicians has served to preclude further expansion of medical education for the next 10 years or so. Economic conditions are not favorable for large increases in R and D funding by the federal government. The ability of medical school faculties to support themselves through revenue from practice plans will be negatively affected by efforts to contain costs in medicate and medicaid programs. This is the situation facing clinical research as seen by the Committee and other observers. Victor Fuchs, a noted health economist wrote recently: "I am particularly concerned, for example, about what will happen to medical research. Without research, without advancing the state of knowledge, medicine will begin to run up against blank walls. There is only a limited amount of improvement in health that can be purchased by increasing the number of physicians or by adding hospital beds. The great advances have always come from figuring out better and newer ways of preventing or treating disease. Somehow there has to be enough funds generated in medical centers to support research and to employ faculty who are actively engaged in research.. (Fuchs, 1982) As noted earlier in this report (Chapter 1), this Committee believes that the long-term problems facing clinical research will arise more from insufficient funding of research than from a lack of training opportunities. We agree with Fuchs when he says that ways must be found in medical schools to support research activities. But we have tried in this report to make the most realistic analysis possible with the existing information. Our recommendations for training in the clinical sciences acknowledge the continuing need to attract and train physicians for research careers. Yet the overall recommended level of training in the clinical sciences has been formulated under our best judgment about the research opportunities that are expected to become available within the next few years under the most likely set of circumstances in medico education.
