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.
4. CLINICAL SCIENCES The Committee's assessment of the market for personnel in the clinical sciences has not changed since publication of the 1977 report. A strong demand is expected' to continue over the next few years. At the same time, numerous indications point to a disturbing decline in the attractiveness of a career in clinical research, which has serious implications for the supply side of this market. The clearest indications of a diminishing supply are: o A sharp drop,,off has occurred in ' the proportion of physician-investigators among all first-time principal investigators on NIN research grants. The proportion fell from 43.9 percent in FY 1966 to 22.3 percent in FY 1975 (Douglass and James 1973; Challoner, 1976~. Concomitantly, there was a decrease in absolute numbers of physicians as first-time principal investigators from 471 to 305, despite an increase over the decade in the total number of grants from which these percentages were derived. O Data from the American Medical Association (AMA) show that the number of physicians reporting research as a primary activity has decreased from 15,441 in 1968 to 7,944 in 1975 (AMA, 1963-74 and 1975-77~. This represents an average annual decrease of 9 percent e However, the most recent data show an increase of 7 percent from 1975 to 1976. The Committee <does not know how much significance to attach to this apparent reversal of a long-term trend. O The number of postprofessionals in research training programs supported by NIH has fallen from approximately 4,600 in 1971 to 1,800 in 1977 (NIH, 1966-78~. This is an average annual decrease of 14 percent. o Budgeted vacancies in clinical departments have grown about 11 percent per year since 1971, compared with 4 percent in basic science departments at medical schools (Table 4.~. A certain amount of these budgeted vacancies are undoubtedly the result of normal market conditions, but the long-term growth pattern is an indicator of an imbalance between supply and demand. 89
- - ^ - ~ - - - u, ~ ~ ~ ret ret 1 1 1 1 1 1 1 1 t_ I_ ~ ~ 1 ~ In, ~ ~ ~ a, ~ ~ o, - - - - - - - - - ~ ret 1 1 ret o, or - - a~ ED or ° ED · · · ~ ~ ~ ~ ~ a, 1 _1 O1 ~ 1 ~ - <~ Ad By dp do ~ lo do At ~ ~ Go · · ~ · ~ · · · · . a, 1 - ~ - a, , v all al u ::c ~ 0 Is; 3 1` al . a, 0 Eli U. ~ a, . O .~1 H 1` U] 00 =4, EN . V a) al . 0 Q N ' r~ O tD a) iD ~ ~ ~ ~ ~ ~ ~ ~ o' ~ ~ ~ ~ ~ ~ ~ 0 ~ ~ ~ ~ a U~ V} a~ a) kg ~r cn ~ ~ o~ ao iD O ~r 0 u, ~ ~ ~ U~ tD ~ ~n o~ - ~r a' >, ~ ~ ~ ao 0 0 · a, u~ a, ~ a, ~ ~ 0 a) 0 .E -~ 0 s: U~ ~ 0 U. ~ · ~ ~ 1 V ~ 0 ~ ~ ~ ~ ~D C} ~ 0 ~ tn ~ oo ~5 V ~ ~ ·rl O ~ ~ ~ ~ · - ~ ~ U `: ~ ~ ~ ~ ~ s:: 0 :Z ~ ~ ~ V ' ~ ~ ~ ~ 0 V 0 ~ ee a' eB 1 ·- 1 Ut O O ~ ~ ·~1 ·. C.) ~ ~ =~ 0 ~ ^ .. ~ ~ ~ ~ ~ _. O ·- ~ ~ ~ ~ ~ ~ ~ O 0 ~ ~ O ·- ~ ~ i4 ~ Q] ~ ~ ~ ~ ~ ~ ~ O ~ ~ ~ ~ ~ u E~ ·,l ::~ o. O ~ ~ ~ .- ~ ~ ~ O ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ U =- c: O m~ o~ ~ ~ O~ ~ ~ ~ ~ e u~ ~ ~ . ~ u mm~ ~ ~ - S4 ~ ~ ~ a)~^ ~ ~ · - eq ~ . - . - ~ ~ 14 <~ S ~s ~, ~ 3 ~ tn ~ [) ~ ~ ~:: ~ :> ~ U ·. ·^ ·- 0 ~ ~ O C: ~ ~ ·- ~ ~ {Q {~O ~ ,~ ~~ v~ It Sx ~ ~ ~- ~ ~ ~ ~ ~1 ~ ~1 ~ ~ V ~ t) o. U} ES ~: u~ ,1 ~ C ~ ~ ~ ~ ~ ~ ~ O ~ V ~ t) ~ ~ -1 U C: s~ ~ ·- ·- -~ ~Q 0 ~ ~ · ~ ~ · ~£~ t" O rt5 - rl ~ ~ ~ -1 a) 0 u~ ~ ~ ~: ~ .^ ~ ~ ~ ~ ·^ v ~, ~ ~ . - s" ~ ~ ~ v ~ ~ ~ ~ ~ ~ ~ ~ · c ~ ~ · - . O ~~ ,~ ~ ~ ~ u~ ~ ~ O ~ ~ u m ~ a Q~ p, 8~ ~ ~ B3 £ E~ m ~ £ :' w ~n ~ , c: ~ 90 0 ~ IB c £ 0 ~ 14 0 i4 0 c 0 ~ os u c - 0 ~ ~ v 0 ~ lU ~ 0 ~ _ co , \9 0 ~ ~ 54 _ c 0~ 0 0 0 x . 0 0 u~ 0 ·Q ~ ~ ~ ~ m. Z u · 0 o £ ~ 5 ~n ~o i4 ~ 0 v :t H o - .! ~r ~D - · U2 ~} 3 o
o A study of factors influencing physicians' career choices records a striking change in attitudes toward research. The proportion of medical students assigning high priority to research dropped from 49 percent ire 1963 to 2 percent in 1976 ~ Funkenstein, 1978 ) . O Directors of training programs in medical schools report a dwindling interest in research careers on the part of the students. In a 1977 study of 12 academic medical centers, the Association of American Medical Colleges (AAMC) conducted interviews with directors of training programs that had the largest number of research trainees. Forty percent of these directors, despite intensified efforts to select persons motivated for academic careers, reported increased difficulty in keeping such trainees in research training programs (AAM1C, 1978) DEFINITION/C LASS IF ICATION OF CLINICAL SCI ENC ES A definition of the nature and is an essential starting point in this field. Definitions of clinical mainly with respect to the boundaries of _ under the term. The Committee limits the term to 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. Similar investigations of disease in animals, conducted mainly by-veterinary scientists, parallel clinical research on disease in humans and could therefore be included under the clinical sciences rubric. The scientific techniques and instrumentation for clinical investigation may be similar to those in basic laboratory studies. Indeed, many of the ablest cat inica' investigators have had training in basic biomedical science and may also perform research on animals or wi th i solated ti ssue components . The above def inn' i-on focuses on the nature of the research rather than on the performer. Usual 1y, clinical research is done by M.~. 's, but our definition does not exclude scientists with other degrees such as D.D. S., D.V.M., D.N., Ph.I)., and other doctorate s whose re search f i t s th i s descr i pt ion . purpose of clinical investigation in examini ng the personnel market investigation di ffer the research included 91
Need for Clinical Investigation Because of increasing but understandable constraints surrounding research on human beings, especially children, the identifi- cation, relative value, and means for investigation of animal models that may be used as surrogates for man has become a larger and more important activity in almost all institutions. There is at the present, however, no animal species or in vitro laboratory test system that is completely predictive for what will occur in the human with respect to an experimental drug or new procedure. No matter how much animal or-laboratory data are gathered, it becomes necessary at some point to address the question, what do these data mean for man? Promising developments in the basic biomedical sciences must ultimately be applied and evaluated in human beings in order to establish the utility of new diagnostic and therapeutic methods. Clinical investigation represents a very difficult kind of research. Since much of clinical research is long-term in nature, its baseline characteristics are often not controllable, with little opportunity to modify experiments with rapid feedback. The numbers of variables that may affect the outcome are usually large and the analytic problems are more complex. Moreover, because such studies may require long-term follow-up of patients, there is the constant risk that segments of the sample may be lost through various forms of attrition. Clinical research is generally best performed in academic medical centers, inasmuch as collaboration with basic scientists occurs more easily in that environment, multidisciplinary teams are at hand to provide skills needed for comparative assessment of old and new methods, and appropriate facilities for human studies are available. Clinical investigation takes many forms. It ranges from systematic observation of individual patients in a controlled environment to large-scale clinical trials involving thousands of subjects and the pooled effort of investigators in multiple institutions. elucidate the research disease. patient care. Hence, an essential aim of clinical investigation is the translation of new knowledge, through applied research, into new technology, as well as the validation of new technology through clinical trials. Other purposes of clinical investi- gation are to assess the reliability, sensitivity, and speci- ficity of diagnostic procedures; to identify the possible hazards of therapeutic drugs; and to determine the efficacy of medical and surgical procedures. It may consist of carefully controlled studies to mechanisms of disease, or field epidemiologic aimed at uncovering information about the etiology of It provides the critical link between basic science and Classification of Clinical Sciences For an operational definition of clinical sciences, the Committee has relied upon the discipline/field/specialties list used by NIH 92
to classify trainees and fellows ( see Appendix D3 ) . Accordingly, numerical recommendations for the Committee's reports have been based upon that taxonomy . At the same t ime, i t i s recogn i zed that marketplace issues, such as financial disincent Ives and the payback provision of the HRSA Act, affect M.D. 's employed as basic biomedical scientists as well as clinical investiga:tors and wil ~ have to be addressed separately from the f ield category in which the M.D. is counted. This ~ eaves open the possibility that the Committee may wi sh at a later time to make degree-specific recommendat ions to NIN . In line with the above operational definition, the Committee recogni zes that ind ividuals with doctorate degrees other than the M.D. participate in research and training activities classifiable as cl inical sciences. For example, many Ph. D . cand idates rece ive NIH-supported training in clinical fields. This points up the difficulty of distinguishing clearly in some instances between the basic biomedical and clinical sciences. Notab, e examples of fields that are difficult to Cal assify as either basic biomedical or clinical are immunology ~ pathology, pharmacology, and microbiology. The resulting problem to the Committee in determining appropr late training levels in these areas is apparent and will be given further consideration as this study continues e THE 1 978 OUTLOOK FOR THE CLINICAL SCIENCES For last year's report, the Committee reviewed trends in the supply of, and demand for, clinical investigators and concluded that de~nand quite likely woul ~ expand somewhat faster than supply if these trends were to continue through 1982. This conclusion was based pr imarily on an analysis of the relationships among clinical faculty in medical schools, medical student enrollments, and R and D expenditures in medical schools. A model similar to the one used for the bi omedica~ science s, in whi ch the faculty/student ratio ~ F/S) was related to R and D expenditures, was also developed for the clinical sciences. Between 1971 and 1975, medical student enrollments grew at about 7 percent per year and R and 1) expenditures at almost 6 percent Per year. i Both of these factors were viewed as important contributors to the 9 percent annual growth of ful I-time faculty in cat inical departments in medical schools that had occurred between ~ 971 and 1975. Revi signs to Demand Closet for Cal inica~ Faculty In order to make recommendations for research training under the NRSA Act in last year ' s report, the outlook for future growth in 93
clinical faculty was considered under various assumptions about medical student enroll Laments and R and D expenditures from 1975 to ~ 982 . A h igh, middle, an] low assumption was made in each case . The resulting pro jections are reproduced in Figure 4. ~ along with the new data on enroll Iment and faculty size that have recently become availab~ e. For the most part the new data fall quite comfortably within last year's projected limits. Medical student enrollment i s sl ightly above the middle pro ject ion for 1977 (Figure 4. Ib); and clinical faculty size for 1977 almost coincides with the middle projection ~ Figure 4. Ic) . R and ~ expenditures, however, is the one variable that does not conform to the projections, falling just below the projected lower boundary for 1976 ~ Figure 4. ld ~ Thi s compar i son between the actual and pro jested data of fers encouragement that in genera' the model upon which last year's pro jections were based is realistic and usef u] for predicting faculty/student ratios over the short run . However, some ad j ustment to the monetary var iabI e ~ R and D expend) t ures ~ seems necessary, and this is the most important one relating to the demand for researchers. Last year, the Committee was concerned that total R and D expenditures in medical schools might be too broad a measure to use for assessing needs for clinical faculty . R and D expenditures in clinical departments was thought to be more appropriate, although the difficul ty in estimating this quantity was recognized. Furthermore medical school income derived from patient-care activities on the part of clinical faculty members has risen sharply in recent years. This reflects the tendency in medical schools toward increasing emphasis on health maintenance and commun ity service programs. It also sugge st s that a good port ion of the re cent expan s ion of c l i n ice faculties has been due to the demands of these service-oriented activities rather than to research. During the early 1960' s, medical service income grew steadily at about 12 percent per year in real terms. But starting in ~968, a dramatic upswing has occurred--these funds have grown at a rate of more than 25 percent per year, and have overtaken estimated R and D funds as a source of revenue for medical schools ~ Table 4. ~ ~ . Thi s movement s igni f ies an important shi ft toward patient care activities on the part of clinical faculty members, and must be recogni zed in any pro jections of demand for clinical faculty. Thi s has been done by includ ing medical service funds in the monetary variable of the model. As revised, the monetary variable now is composed of the sum of medical service income and estimated expenditures for clinical R and D in medical schools. The latter is estimated indirectly from total R and D expenditures by a procedure expiainedl in Append ix H . Med i cal serv i ce i ncome on the other hand has been measured direct' y and reported annually in the Education Number of the Journal of the American Medical Association (JAMA) at least since 1957. 94
- z - en Go to ~ ~ A \ ~ ~ ~ ~ As \ ~~ ~g \ '= ~ ~ ~ I: cC) ~ ) o o o ~ ~ orD o ~ ct . Oll~U ~ SN01111~ $ ~ o :> 3 e ~~ ~ ~ ~ 0 0 0 ~ 0 0 0 (spuesnotll) slN3Onis '~ (I) ~lin3~ ,, v JO H38WnN ' _ I" SO 95
The statistical relationship between the faculty/stuclent ratio in clinical departments and funds availab~ e from Cal inical R and D expenditures and medical income is quite strong. ~ According to the best est imates available, the relationship is not linear, but appears to take on the typical S-shape exhibited by many growth processes. The growth pattern of the clinical faculty/student ratio impl ies that it passed through a period of rapid growth during the 1960' s and has now entered a more stable period. If this pattern should continue, the ratio would respond more slows y to future changes in the level of clinical funds available in medical schools. Rev i sed Proj ections The new projections resulting from the revised model are shown in Figures 4.2 and 4. 3 and Tabs e 4. 2. Pro jections of the faculty/student ratio result from its relationship wi th clinical f units ~ i . e ., total cl i n i Cal R and D expend i t ure s pi u s med i Cal service income) and assumptions about the future ~ evels of this monetary variable. The clinical faculty/medical student projections are translated into projected demand for clinical faculty by applying the projected levels of medical student enrollments. As before, h igh, middle, and low proj ections have been made for the monetary variable and medical student enrollments, reflecting the Committee's best estimate of likely future levels based on the behavior of these variable es in the recent past ~ Figures 4. 2 and 4. 3 ~ . With regard to clinical f unds, under the high assumption, they are expected to grow at about I ] pe rcent per ye ar from the 1976 level of $422 million to $~83 million in 1983. Under a more conservative assumption, and one which the Committee considers to be its best estimate, clinical funds would grow at about 7 percent per year to a level of $686 million by 1983. Under the most con servat ive assumpt ion, whi ch the Committee cons iders to be the lower bound, clinical funds would grow by only 2 percent per year to a level of $485 million by 1983. In the case of medical student enrobe Iment, growth is expected to slow from its present 6.3 percent annual rate to about 3 percent per year by 1983 ~ at which time enrollment would reach a. level of about 69,000 students. This slowdown is expected to result primarily from a slower rate of growth in new medical school s, wh i ch have been organ i zed at the rate of alma: st 3 per year since ~ 968. ~ Offsetting this trend to some extent are the growing number of programs for continuing medical education being of fered by medical school s O These are des igned as refresher courses for physicians and increasingly are being required by state licensing boards and accrediting bodies. It is expected that these programs wi Il generate an addi tional teaching responsibil ity for 96
360 400 ' u~ Z _ O ~ - ~ 200 _ _ 300 100 Medical J| School '' Ctin~cal R and D ~ Medical Service I ncome . ~r ' I l 1965 1970 1 975 ~ ISCAL YEAfl {a} Medical SeNiCe Income and Medical School Clinical R and D {1967 $} 500 400 200 1OO o FISCAL YEAR {c} Projected Medical Service Income {1967 $} _ I I I ,, 320 l~otal l ~ Z / O 280 J240 200 160 120 ~' 1 J _ , ~ 1 1 1 970 1 975 1 980 FISCAL YEAR (b} Projected Medical School Clinical R and D {1967 $} 800 400 300 200 I I I I' 0 300 _ - J J _ ~1 1 1 / ~ }~__.~ : ~ ~ .. ,' 700 ' . ~ 600 cn z o - ~ 500 . I I J I ,/ .1 1 1 I / I I _ I I / ~ 1, 111 I' __ K_~ / 1 1 1 ' 1 1965 1970 1975 1980 1970 1975 1980 FISCAL YEAR (d) Projected Total Medical School Clinical R and D and Medical Service Incomet1967$) FIGURE 4.2 Medical service income and medical school R and D expenditures, 1961-76, with projections to 1983e Based on data from JAMit (196~77). 97
0.6 '0.5 o 0.4 lo: 0.3 0.2 o 1- High Estimate _~ Middle Estimate // Low Estimate / 'I LI I I 1 1 1 1 1 1 1 62 66 70 74 78 82 FISCAL YEAR {a} Clinical Faculty/Medical Student Ratio 3 100 80 60 Z 40 ~ 20 1 1 1 1 1 1 1 1 1 1 1 o G At 111 62 66 70 74 78 82 FISCAL YEAR (b) Medical Students 60 40 20 o =1 1 1 1 62 66 70 74 78 82 FISCAL YEAR {c} Clinical Faculty in Medical Schools High Estimate , ~ _ _ Middle Estimate Low Estimate FIGURE 4.3 Medical student enrollment and clinical faculty, 1961-76, with projections to 1983. Based on data from JAMA (1960-77). 98
TIBIA 4.2 Projected Growth in Clinical Faculty, 1976-83, Based on Projections of Medical School Enrollment, Clinical R and D Expenditures, and Medical Service Inch in Medical Schoolea Assumptions about Real R and D Expenditures Plus Medical Service Income in Medical Schools I II III 111 expand at Will expand at Will expand at "out 119~/yr to about 7~/yr to about 2%/yr to Assumptions about Medical ~ S883 million S696 million S485 million Student Enrollment in 19433 in 1983 in 1983 A. Will continue to grow Expected size of clinical at present rate (61/ faculty in medical yr), reaching 84 , 500 schools (CF) in 1983 47 , 256 46 ,606 45 365 students by 19 83 Annual growth rate in CF from 1976 to 1983 7 4` 7.2% 6.8` Average annual increment due to faculty expansion 2,665 2,572 2,395 Annual replacement needg due to death and retirement 493 489 481 Expected total annual in- crement in clinical faculty 3,158 3,061 2,876 B. . Will grow at 3%/yr Expected s ize of clinical reaching 69,200 faculty in medical medical students by schools (CF) in 1983 38,652 38,121 37,106 1983 Annual growth rate in CF from 1976 to 1983 4.4% 4.2' 3.8. Average annual increment due to faculty expansion 1,436 1,360 1,215 Annual replacement needg due to death and retirement 437 434 427 Expected total annual in- crement in clinical faculty 1,873 1,794 1,642 Wi 11 show essentially Expected size of clinical no growth from 1976 to faculty in medical 1983, leveling off at Schools (CF) in 1983 32,558 32,110 31,2S5 58, 300 medical students Annual growth rate in CF fram 1976 to 1983 l.9S l.t, 1.3% Average annual increment due to faculty expansion 565 501 379 Annual replacement needg due to death and.retisement 398 395 389 Expected total annual in- cre~nent in clinical faculty . 963 896 768 aThese projections are based on the following relationship: (CF/M)t ~ exp(-0.92982 - 75839.6/Ot) + 0.2, where CF ~ size of clinical faculty in medical schools' M - medical student enrollments Dt ~ a weighted average of the last 3 yearn of clinical R and D expenditures plus medical service income In medical schools, i.e., Dt ~ \(Dt ~ 2Dt-1 + Dt-2 ). bBased on an estimated replacement rate of 1.3 percent annually due to death and retirement (Cartter, 1976, p. 121). 99
clinical faculty and ultimately will be reflected in additional positions in clinical departments. Because of these factors, the Committee considers it possible, but not very likely, that medical school enrollment would continue to grow at the present rate of about 6 percent per year, reaching almost 85,000 students by 1983. It is also unlikely that medical school enrollment will stabilize at its current level of 5S,000. These estimates appear to bracket the likely range of future enrollment patterns. The effects of these projections on the demand for clinical faculty is shown in Table 4.2. Under the combination of highest assumptions (IA of Table 4.2) clinical faculty is expected to expand at over 7 percent per year from 1976 to 1983. This yields an estimated annual increment in clinical faculty of about 3,150 positions due to both expansion and replacement. Under the middle set of assumptions (IIB), an estimated 1,800 new clinical faculty members would be needed each year to fill the positions created by expansion and replacement needs. The combination of lowest assumptions (IIIC) result in only about 770 new positions created annually in clinical departments. Supply of Clinical Investigators On the supply side, a similar kind of shift in the sources of funds for training clinical investigators has been noted. Federal research training grants and fellowships, which for many years were the main source of research training funds in the biomedical and clinical science areas, have been declining. They are being replaced to a certain extent by funds from research grant and patient care activities. In a survey of 12 academic centers conducted for the Committee by the Association of the American Medical Colleges (AAMC) in 1977, it was found that in constant dollars, federal research training funds declined by 3 percent annually between 1972 and 1976, while funds available~to support advanced clinical training from patient care activities and other sources increased by 13 percent annually. The net result for these 12 academic medical centers is that the proportion of all advanced clinical trainees who were receiving research training on a full-time basis declined from 49 percent in 1972 to 43 percent in 1976 (AAMC, 1978~. The Committee is also concerned over the long-term decline in the number of trainees with professional doctorates participating in the research training programs sponsored by NIH. From 1971 to 1976 this number dropped by more than 7 percent annually, from approximately 4,600 in 1971 to 1,800 in 1977 (NIH, 1966-781. 100
Training for Clinical Investigators There are several possible e explanations for the declining interest in research careers on the part of young physicians. To understand them, one must first understand the process by which a cl in i cal i nve st igator i s tra i ned and the ~ f actors that af f e at: h i s or her career cho ice . Clinical research has been regarded as pr imarily but not exclusively the province of the investigator wi th a professional degree. Knowledge of di see se states and a background] in patient care dispose the clinical scientist to raise questions different from those posed by a laboratory-trained scienti st working in a clinical milieu, as well as to play a translating role in applying f indings of laboratory and f isle research to prob] ems of the patient. Moreover, the al ert physician frequently provides the f irst essential clue as to new and unanticipated etiologic factors, adverse effects from therapeutic drugs and medical or surgical procedures, or hazardous agents in the environment. These clues can lead directly to specific hypotheses to - be -tested in epidemiologic studies. Indeed, the -most productive clinical research i s general ly performed by investigators whose observations at the bedside or in the clinical laboratory have furnished the stimulus for their studies. Moreover, the investigator with a health professions background provides the role model essential for attracting students and house staff to research careers. Most clinical investigators have the M.D. degree or a comparable health professions doctorate. The curriculum of professional schools, however, provides little formal exposure to research during the predoctoral period. A notable except-ion is the combined M.~./Ph.D. program. After graduation from medical school, the new physician pursues several years of advanced clinical training. In some specialties, - particularly the surgical areas, opportunities for research experience exist within the basic residency for specialty - certification. In other specialties, however, there is little opportunity for research experience for trainees engaged in ; fulfilling residency requirements. At a relatively late point in his/her training career, the physician who is attracted to clinical investigation may finally consider research training. In most cases the individual has satisfied the requirements for pr imary board el ig ib i l i ty i n a spe ci arty. At this point there are several options: to enter practice, to continue training in a clinical specialty as a clinical fed low, or to pursue training as a research fellow. Sel ection of the third option must frequents y be made in the face of great uncertainty. The young physician selecting a postdoctoral fellowship does not know whether he/she will succeed in research. Also, there is a perception that estab~ ished faculty members are encountering difficu, ty in obtaining support for their research . Lacking medical school research experience-- largely a consequence of curricular revi signs of the past decade--young trainees doubt their ability to compete with more accompl ished investigators. 101
In addition, numerous financial and other disincentives impinge upon this career choice. What are some of these deterrents to the pursuit of clinical research training? (1) Clinical research training competes with practice opportunities in which starting salaries of $40,000- $45,000 may not be unusual. At present the differential between a physician beginning a research fellowship and one beginning clinical practice may be $30,000 or more. The starting research training stipend is often less than the house officer's salary, which the fellow earned in the preceding year. (2) The potential candidate for clinical research training often will still have debts from this lengthy educational process. Accumulated indebtedness at graduation from medical school in 1975 averaged almost $12,000, with 15 percent of the students reporting $20,000 or more in debts (HRA, 1977~. This debt burden exacerbates over a trainee's lifetime the economic net loss that is associated with post-M.D. training for a career on a medical school faculty (Scheffler, 1975~. (3) As noted earlier, resident physicians have established their clinical ability, but must "gamble" on attaining equivalent competence in research. By accepting federal support for research training, the young physician becomes subject to the payback - provision in the NRSA Act, even if he/she subsequently demonstrates a lack of talent for research. (4) Program directors interviewed in the AAMC study felt that society has given indications that those pursuing careers of patient care are valued more h, she ' than are teachers and researchers. For example, the federal government has initiated scholarship programs for those who will care for patients in medically - underserved areas or in the military, but there have been moves to end support for research training. Federal research training awards pay less than clinical traineeships, and practitioners' earnings exceed those of faculty researchers. highly teachers and research. Per "Y=mm1 ~ the (5) Testimony at public meetings held by the Committee has called attention to the probable effects of increased emphasis on the training of primary care physicians. From the time of entry into medical school, students are exposed to social pressure to pursue careers in direct patient care, particularly in 102
the pr imary care . f ields . But pr imary care physicians are less likely than clinical sub-specialists to retain an interest in research or to participate in research activities. Thus, these laudable social goals may have had an indirect effect on the career choice of medical students, reducing even further the small number with a potential interest in research and teaching. (6 ~ There is, too, a perception of growing administrative probl ems, paperwork, and restr i ct ions pecul i ar to clinical research, such as regulations for the protection of human subjects and patients' access to re cords that may make cl i n i cal tr i al s d i f f i cul t to conduct. While none would question the important obj e ct ive s of the se rul e s, the i r ex i stence undoubted ly add s to the i ncreas i ng constra int s on cl i n i cal re sear ch . RECOMMENDATIONS The projections of future demand suggest the need for an annual increment of approximately 1~800 persons to meet requirements created by expansion and attrition of clinical faculty in medical schools over the next few years (Table 4. 2) . In common with medical school faculty in the basic biomedical sciences, most of these additional clinical faculty will spend some portion of their time and effort in research activities in addition to their other .academic responsibilities ~ teaching, service, and admi ni strat ion ~ . The amount of e f for t act ual 1 y ~ evoted to re se arch, per se, obvi ou sl y wi 1 1 vary wi d ely among al 1 med i cal school faculty member s e Realization of this annual output of clinical faculty is in large part a f unction of the average length of postdoctoral training needed by clinical researchers. In determining an appropriate length of research training in the clinical discipl ines, the Committee took as its model the experience of a typical "~fellow"--a physician who has completed basic res idency training and who then elects to continue in .a postdoctoral program to acquire research skills and experience. In contrast to the Ph.D. who hats received 4 to 7 years of rigorous predoctoral training in research, and~ frequently special i zed postdoctoral training under an individual mentor, the typical M.D. 's initial research training usually occurs in the postdoctoral period. Accordingly, if M.D. ' s ~ and other professional degree holders ~ are to become ser ious and product ive research sc ienti sts, they in a sense need to be trained for a second career . The Commi ttee bel i eve s, there fore, that under optima~ circumstances, the average length of research training for M.D. 's should be from 3 to 4 years. ~ 03
The Committee recognizes, however, that it is often impossible to realize the optimal length of research training and indeed that the more typical length of postdoctoral research training for clinical investigators, past and present, is from 2 to 3 years. In order that its recommendations strike a balance both between the desired optimum and actual experience, the Committee therefore postulates that the length of research training experience in the clinical disciplines is 3 years. Accordingly, in order to achieve an annual increment of 1,800 clinical faculty, the nationwide pool of postdoctoral research trainees in these disciplines would have to be approximately 5,400 (without making any allowances for attrition). It should be emphasized that this total pool of postdoctoral trainees traditionally has been supported from multiple sources and that this practice will continue for the foreseeable future. Accordingly, it is necessary to develop a rationale for determining the proportion of this total number who should receive Federal support under the aegis of the NRSA Act. The Committee has considered the following two facts in determining this proportion: The proportion of clinical research in colleges and universities supported by NTH was 44 percent in both 1975 and 1976 (NSF, 1975-77) e The federally supported portion of the total pool of clinical research trainees was about 67 percent in 1974-75 according to NIN and NSF.2 - For purposes of its funding recommendation, the Committee has selected 50 percent of the total pool of 5,400 as the appropriate level of support under the NRSA Act. In arriving at a percentage somewhere between the two reference points, the Committee takes cognizance of the belief of many observers that clinical faculty on the whole are somewhat less likely than are basic science faculty to devote the majority of their time and effort to research pursuits. The Committee has no firm data with which to develop a precise estimate of the number of either basic science or clinical science medical school faculty whose other academic responsibilities predominate over their research endeavors. It is reasonable to assume, however, that the patient care responsibilities of most clinical faculty make it relatively more difficult for them to be as research-intensive as are most basic scientists. Accordingly, a 50 percent support level by the Federal Government for the clinical disciplines seems appropriate e Supporting 50 percent of the total pool of postdoctoral clinical research trainees would amount to an annual NRSA support level of 2,700 individuals. It should be noted that while most such trainees are M.D.'s, there are also other doctorate holders in the pool, such as persons with D.V.M., D.D.S., and Ph.D. 104
degrees. Also, to the extent that the demand estimated from the model described above does not take into account employment sites other than medical schools where federally sponsored clinical research is conducted (e.g., schools of dentistry and veterinary medicine, industry, and governmental and nonprofit laboratories), the estimated requirement for trainees is understated. Thus, the Committee believes, on the basis of the information presently--availab~e, that its previous recommendation for federal support of 2,BOO postdoctoral trainees and fellows in the clinical sciences is reasonable and should be continued. That recommendation was derived by applying, on the-basis of professional judgment, a TO percent increase to the number of trainees supported by the NIH in FY 1975. The derivation of a similar number through use of the more analytic approach outlined in this chapter is encouraging. In last year's report the Committee had expressed concern that fewer than 2~800 trainees and fellows had been funded last year. The concern pertained mainly to the possibility that too few qualified persons were seeking research training because of low stipends and other deterrents. Available evidence, however, indicates that the FY 1977 shortfall was largely the result of budgetary limitations. In reiterating its previous recommendation, the Committee is aware that the award of training positions cannot by itself counteract the slackening of interest in clinical research careers. The Committee will continue to assign high priority to a broad study of factors contributing to this problem of supply, as well as of possible actions needed to ensure an adequate pool of clinical researchers. Recommendation. The Committee recommends that 2, 800 pos tdoctor al tra i n i ng pos i t ion s be made ava i 1 abl e i n the c 1 i n i c a sciences for FY 1980 and should be maintained at this level until new in format ion i nd i cate s the need for a change . Medical Scientist Training Program The Medical Scientist Training Program (MSTP) offers to carefully selected students an integrated program of medical and graduate training leading to the combined M. D. and Ph.D degrees. The great maj ority of graduates of these programs may be expected to pursue careers on medical school faculties. This expectation will be tested in the coming year by a follow-up study of the 53 graduates who are now 5 years or more beyond completion of the program In its 1977 report ~ the Committee recommended approximately 10 percent increases for both FY 1978 and FY 1979, using as a 105
base the 600 MSTP training positions expected to be approved for FY 1977. The Committee continues to support the basic philosophy of this program and recommends that a modest increase in training positions be allocated to it. Recommendation. The Committee recommends an increase in medical~~~scientist~ trainees from 700 in 1979 to 725 in 1980, and that the program remain at that level through 1982. ~~ RESEARCH AGENDA To strengthen and improve the basis for its projections and recommendations on personnel in-the clinical sciences, the Committee has identified issues and questions for which additional information is needed. Over the next several years, the Committee will seek to obtain this information through the research plan outlined below. Supply of Clinical Researchers One of the most pressing issues for which the Committee seeks data is the number of currently active clinical investigators. The Directory of Physicians, maintained by the AMA, is a data source for estimating the full-time equivalent of M.D.ts in research in all sectors. Since this file contains updated data on hours spent by physicians during a typical week on their - various professional activities, it can provide an estimate for M.D.'s reporting research as their primary activity by type of practice, specialty, etc. Discussions are currently underway with the Association regarding data on the research involvement of all active physicians and of physicians employed as full-time faculty in medical schools. The tabulations~will reflect changes in extent of research involvement that have occurred from 1968 to 1976. Information is also being sought from the National Specialty Survey conducted by the Division of Research in Medical Education of the University of Southern California. These studies have determined by log di ary the hours devoted by physicians to research. That information might provide a check on usefulness of the ciassi f ication employed in the AMA Physician File . The data could reflect, for example, the actual extent of research i evolvement of phys i ci ens character i zi ng themselves as be i ng pr imar il y eng aged in rese arch . The National Study of Internal Medicine, sponsored by the four constituent organizations of the Federated Council for 106
Internal Medicine Manpower, merits special attention because of its scope and special relevance to clinical research training. Precise~data on the total number of clinical research fellows have been gathered from all of the subspecialty fellowship training programs in the United States. While the data are limited to internal medicine, that specialty includes a substantial majority of all post-residency clinical research trainees. In addition, that survey will provide information on the size and character of the non-NIH-supported segment of the training pipeline. The Committee therefore looks to the possibility of formulating general estimates from these data, based on the proportion of internal medical faculty to total medical school faculty. The AAMC maintains a facu~ty~roster-that incorporates information on personal characteristics, educational experience, multiple patterns of activity, employment history, and current employment of faculty members. Data from this file will be used to determine trends in age-distribution, mobility patterns, attrition, and changes in medical school faculty size. Role of Non-M.D. Investigators The discussion of clinical investigation presented earlier in this chapter focuses on the role of the researcher with a health professions doctorate in the conduct of clinical research. Nevertheless, there is evidence of an appreciable involvement of nonhealth professionals in the clinical sciences. Data from the NIH Manpower Survey3 indicate, for example, that approximately 20 percent of the personnel employed as clinical scientists and receiving salary from NIH research grants for the 1973-75 period were Ph.D. holders. In the face of declining attractiveness of research careers for the health professional, it is probably unreasonable to expect that personnel needs in the clinical sciences can be met exclusively by investigators trained in the medical school/residency/research fellowship mode or in combined M.D./Ph.D. programs. In the Committee's view, the extent to which scientists with an academic doctorate can be used to supplement and enrich the supply of clinical investigators deserves the most careful study. Dental Research Personnel An ad hoc work group on dental research manpower needs was convened in May 1978 under the sponsorship of the Committee' s Panel on Clinical Sciences. As a result of that meeting, staff of the Committee and the American Association for Dental Research wit l cooperate in mount ing a net tonal study to provide informa- 107
Lion on number and distribution of dental research personnel; types and duration of research training; areas of research interest and sources of research support; and characteristics of dentists electing research careers. In addition, the Committee has been invited to suggest modifications in the research section of the Survey of Dental Educators, which the American Association of Dental Schools uses for compil ing an annual directory. An effort will be made to define special characteristics of the dental research training system . One example i s the 1 ink generally observed between research and clinical specialties, which creates special problems in the training of dental investigators. Whereas the young physician receives a salary during hospital residency training, salary payment for the newly graduated dentist is 1 imited to hospital-based training in oral surgery. Advanced clinical training in the other specialties, wh i ch i s gener all y carr i ed out i n dental school s, rarely provide s compensation and, indeed, may require tuition payment. It seems clear therefore that NIH training programs for the D. D. S. investigator must continue to include support for a research-related clinical component. Moreover, because of this inclusion, the training period for the D. D. S. investigator wi 11 usually exceed the 3 year limit stipulated by the NRSA Act. Attitudes of Students and House Staff Data on attitudes of medical students will be examined. AAMC' - Division of Student Studies initiated in March 1978 an annual survey of career determinants and preferences of graduating seniors in all U. S. medical schools. A pilot study involving 1, 022 students in nine medical schools has been completed for 1976-77 academic year seniors. Research-related questions will cover plans for graduate medical education, such as intention to seek a research fellowship, as well as postgraduate career plans involving academic faculty appointments or employment as a Sal ar i ed rese ar ch sc i ent i st . Plans are nearing completion for conducting a mail survey on attitudes of clinical fellows toward academic research careers. The purpose of the st udy would be to determine the range and extent of factors that may be influencing the career choice of the clinical fellow. The draft protocol calls for an incremental approach, the first phase of which would concentrate on ''recent deciders" in the specialty of internal medicine. After val idation of this approach, the study would be extended to other specialties r including those with relatively few clinical research trainees, such as obstetrics and gynecology. Depending on an apprai sat of results, the survey instrument would then be modified for application to dentistry and veterinary medicine. The Comm i t te e be l i eve s the t s uch a s urvey wi 11 be u se f ul i n 108
he' ping to assess the relative we ights to be assigned to various factors affecting the decision for or against an academic career. Personnel Needs for Veterinary Scientists Preliminary data on fur I-time equivalent veterinary personnel required for every $- mill ion of animal-related research supported by grants/contracts have been obtained through a survey of ll research institutions. These data are intended for use in a model of demand for veter inary personnel being developed by the Committee' s staff. During the coming year, efforts will be devoted to improvi ng the methodology and to compar i ng the ful 1 -time equivalent D. V. M. personnel engaged in animal-related research with the number that the model suggests are needed. An ad hoc working group on veterinary research personnel is working with the Committee' s Pane, on Clinical Sciences to assess personnel needs in the industrial, governmental, and veterinary college sectors. Federal and state governments are major employers of D. V. M. scientists. Large numbers of such individuals participate In research at all levels at NIH, the Center for Disease Control ~ the Food and Drug Administration, the Environmental Protection Agency, the Department of Agriculture the Department of Defense, state and county health departments, etc. The impact of federal pod i cy in such areas as laboratory animal care, the Good Laboratory Practices regulations, water and air quality research , etc., is liked y to create special demands for D.V. M. scientists in the near and long-term future. ! 109
FOOTNOTES l. The demand model for clinical faculty in med i cal schools i s specified to be of the following form: (CF/M)t = exp ~ ~ - ~ /Dt] + K, where (CF/M)t = ratio of clinical faculty in medical schools to medical st udent enrol iment In year t; Dt = weighted sum of clinical R and D expenditures plus medical service income in medical schools ~ ~ t ~ in the last three,years, i.e. : Dt = ]/4 ~ R t + 2R tel. + R t-2 ~ ~ 1967 $, thousands. a, 6, K = constants to be determined empirical ly . This functional fonn. of the mode] specifies a growth curve which is asymptotic to en and has an intercept vat ue of K when Dt = 0. Fitting this curve to the data for 1961-76 gives the follow) ng est imate s for the parameters: a = -0. 93 ~ = 75840 K = 0.20 2. Because of a lack of central i zed record keepi ng and common nomenclature, it is extremely difficult to estimate the size of the pool of clinical research trainees in the United States for particular year. Both the AAMC and NSF have attempted to derive such estimates, but there is considerable difference in their counts. The AAMC estimate for 1974-75 is approximately 2, 700 postprofessional c' inical research trainees (AAMC, 1978 ~ . This compares with an NSF estimate of 4,792 health sciences postdoctorals for the same year (NSF, 1973-77 ~ . Recent interviews with several respondents in the NSF survey indicate that these counts include some clinical fellows because of definitional problems and, hence, overstate the number engaged in clinical research training. The actual number of clinical research trainees for that year probably ~ ies somewhere between the two estimates. If the mid-poir~t between 2, 700 and 4, 800 i s taken as the best estimate, we get 3, 750 clinical research trainees in FY 1975. - According y, the 2, 550 postdoctoral trainees and fed lows supported by NIH represented 67 percent of the total for that year. 3. The NIH Manpower Survey has been conducted annually since 1973. It is directed toward the pr incipal investigator on each research grant and is intended: to obtain data on the personnel participating in these grants. ~0
