Appendix D
Demographic Projections of the Research Workforce in the Biomedical, Clinical, and Behavioral Sciences, 2006-2016

Projecting the research workforce is particularly difficult at this time. Available survey data on the workforce predate the economic crisis and subsequent stimulus, and no previous experience is available to indicate how the size and composition of the workforce have been affected and will change as a result in the future. Projections are reported here that rely on the statistical record, and because of these circumstances they probably carry a larger than usual, although unquantifiable, margin of error.

We consider first what the record says about the research workforce in three large fields: the basic biomedical, clinical, and behavioral and social (or simply behavioral) sciences. The disciplines that these three major fields cover are specified in the preceding appendix. By the workforce we understand all those residing in the United States who are qualified to do research by reason of the appropriate degree, as long as they are not retired. Although some without the degree may make important contributions, their numbers are probably small in comparison and in any case are not documented.

We first describe the workforce itself, then the graduates and immigrants who regularly add themselves to it, then the process of leaving the workforce through retirement and death as well as other changes in status, such as from employment to unemployment. These descriptive sections give fairly clear indications of the directions in which the workforce could be headed—absent the perturbations related to the great recession. We then describe the assumptions made in the projections and present projections of the workforce up to 2016.

THE SCIENTIFIC WORKFORCE

The scientific workforce has three major segments: Ph.D.s with U.S. doctorates, immigrant Ph.D.s with foreign doctorates, and M.D.s who do not also have Ph.D.s but have research interests. Because the NIH did not release data on M.D. researchers, we focus entirely on Ph.D.s, especially on the U.S.-trained, about whom there are the most data.1

U.S.-Trained Ph.D.s

The latest survey, in 2006, put the number of research scientists with U.S. Ph.D.s at 126,000 in the biomedical field, 24,000 in the clinical field, and 120,000 in the behavioral field (Table D-1). Five years earlier, behavioral and biomedical scientists were virtually equal at 113,000-114,000 each, but since then behavioral scientists have increased at a rate of only 1.1 percent annually, in contrast to an increase of 2.1 percent annually for biomedical scientists. Clinical scientists, by far the smallest group, grew much faster, at 4.7 percent annually.

The recent increases in U.S.-trained Ph.D.s are roughly in line with long-term trends, which suggest slowing growth in the behavioral field, slightly accelerating growth in the biomedical field, and greater acceleration in the clinical field (Figure D-1). Over the past two decades, however, growth rates have moderated in each field. Rates were higher from 1995 to 2001, at 2.5 percent annually for U.S.-trained behavioral scientists, 3.7 percent for biomedical scientists, and 5.5 percent for clinical scientists. From 2001 to 2006, the annual increments in the workforce were around 1,300 behavioral scientists, twice that number of biomedical scientists, and about 1,000 clinical scientists. Given estimates of retirements and deaths (to be considered below), this implies that, in 2001-2006, close to 25 percent of biomedical and behavioral Ph.D. graduates annually and about 15 percent of clinical graduates were not being immediately absorbed into the workforce. Exactly where they were or what they were doing instead is not evident.

1

For completeness, we note other small groups that are not covered: those without Ph.D.s who still do independent scientific work and those with Ph.D.s in other fields, such as informatics, materials science, and physics, who have been recruited into the ranks of health researchers.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 125
appendix d demographic Projections of the research Workforce in the Biomedical, Clinical, and Behavioral Sciences, 2006-2016 Projecting the research workforce is particularly difficult M.D. researchers, we focus entirely on Ph.D.s, especially on at this time. Available survey data on the workforce predate the U.S.-trained, about whom there are the most data.1 the economic crisis and subsequent stimulus, and no previous experience is available to indicate how the size and composi- u.S.-trained Ph.d.s tion of the workforce have been affected and will change as a result in the future. Projections are reported here that rely The latest survey, in 2006, put the number of research on the statistical record, and because of these circumstances scientists with U.S. Ph.D.s at 126,000 in the biomedical field, they probably carry a larger than usual, although unquantifi- 24,000 in the clinical field, and 120,000 in the behavioral able, margin of error. field (Table D-1). Five years earlier, behavioral and bio- We consider first what the record says about the research medical scientists were virtually equal at 113,000-114,000 workforce in three large fields: the basic biomedical, clini - each, but since then behavioral scientists have increased at a cal, and behavioral and social (or simply behavioral) sci - rate of only 1.1 percent annually, in contrast to an increase ences. The disciplines that these three major fields cover are of 2.1 percent annually for biomedical scientists. Clinical specified in the preceding appendix. By the workforce we scientists, by far the smallest group, grew much faster, at understand all those residing in the United States who are 4.7 percent annually. qualified to do research by reason of the appropriate degree, The recent increases in U.S.-trained Ph.D.s are roughly as long as they are not retired. Although some without the in line with long-term trends, which suggest slowing growth degree may make important contributions, their numbers in the behavioral field, slightly accelerating growth in the are probably small in comparison and in any case are not biomedical field, and greater acceleration in the clinical documented. field (Figure D-1). Over the past two decades, however, We first describe the workforce itself, then the graduates growth rates have moderated in each field. Rates were higher and immigrants who regularly add themselves to it, then the from 1995 to 2001, at 2.5 percent annually for U.S.-trained process of leaving the workforce through retirement and behavioral scientists, 3.7 percent for biomedical scientists, death as well as other changes in status, such as from employ- and 5.5 percent for clinical scientists. From 2001 to 2006, ment to unemployment. These descriptive sections give fairly the annual increments in the workforce were around 1,300 clear indications of the directions in which the workforce behavioral scientists, twice that number of biomedical scien- could be headed—absent the perturbations related to the tists, and about 1,000 clinical scientists. Given estimates of great recession. We then describe the assumptions made in retirements and deaths (to be considered below), this implies the projections and present projections of the workforce up that, in 2001-2006, close to 25 percent of biomedical and to 2016. behavioral Ph.D. graduates annually and about 15 percent of clinical graduates were not being immediately absorbed into the workforce. Exactly where they were or what they were the SCieNtifiC WorkforCe doing instead is not evident. The scientific workforce has three major segments: For completeness, we note other small groups that are not covered: Ph.D.s with U.S. doctorates, immigrant Ph.D.s with foreign 1 those without Ph.D.s who still do independent scientific work and those with doctorates, and M.D.s who do not also have Ph.D.s but have Ph.D.s in other fields, such as informatics, materials science, and physics, research interests. Because the NIH did not release data on who have been recruited into the ranks of health researchers. 

OCR for page 125
 APPENDIX D TABLE D-1 Workforce of U.S.-Trained Ph.D.s in Three Major Fields, by Sex and Employment Status, 2006 Biomedical Clinical Behavioral Status Males Females Males Females Males Females Total 80,268 45,828 9,451 14,706 57,593 62,758 Employed in Science 68,236 36,340 7,817 11,924 46,399 49,261 Postdoctorate 7,442 6,526 340 549 945 1,455 Other 0, ,8 , , , ,80 Employed Out of Science 10,772 6,604 1,600 2,172 10,668 10,715 Unemployed 464 582 N/A 124 224 449 Not in the Labor Force (but not retired) 796 2,302 34 486 302 2,333 SOURCE: Data extracted from National Science Foundation Survey of Doctorate Recipients. 140 126 .1 120.4 120 10 0 Behavioral Thousands 80 Biomedical 60 40 Clinical 24.2 20 0 1970 1975 1980 1985 1990 1995 2000 2005 2010 Year FIGURE D-1 U.S.-trained Ph.D. workforce, in thousands, in three major fields, 1973-2006: quadratic trend and annual variations. SOURCE: Data extracted from National Science Foundation Survey of Doctorate Recipients, 1973-2006 and NRC analysis. D-1.eps These estimates of the workforce cover not only those fields, those on postdoctorates were 17,000 in 2006, or 7.8 per- employed in jobs related to science (which includes those cent of all those in science employment. For scientists under on postdoctorates) but also those in nonscience jobs, as well 35, postdoctorates made up 42 percent of science employment. as the unemployed and those not looking for work—often The 2006 figures represented increases from 2001 and particu- women taking a break from employment (Figure D-2). larly 2003, when absolute numbers on postdoctorates actually The estimates exclude the retired. In 2006, those actually declined. Although 17,000 was the highest figure recorded in employed in science were 80-83 percent of the total work- the biennial surveys, in percentage terms it represented no force. This is a lower proportion than the historical average more than a return to the levels of the late 1990s. up to 2001, which is around 90 percent in each field. What Another change in the U.S.-trained workforce, but one has taken the place of jobs in science is non-science employ- that has been largely gradual, is the increasing proportion of ment. In the Ph.D. surveys from 1973 to 2001, the propor- women (Figure D-3). Shortly after 2003, behavioral scien- tion of the workforce employed in non-science jobs did tists reached a turning point, with the sex ratio falling to 100, not exceed 10 percent in any field, except once, just barely, or equal numbers of men and women. Among clinical scien- among behavioral scientists in 1995. In the two surveys since tists, this turning point was reached in 1994, and the sex ratio 2001, in contrast, this proportion has been between 13 and was down by 2006 to 64 men per 100 women and still falling. 18 percent in each field. Among biomedical scientists, the sex ratio was still elevated Postdoctorates have been an important category of science at 175 in 2006, although extrapolation of current trends sug - employment, particularly for younger scientists. Across all gests that equality could be reached around 2020.

OCR for page 125
 APPENDIX D 140 Biomedical Behavioral 120 Not in the Labor Force Unemployed 10 0 Out of Science Employed, Postdoc Thousands 80 Employed, Not Postdoc 60 40 Clinical 20 0 2001 2003 2006 2001 2003 2006 2001 2003 2006 Year FIGURE D-2 U.S.-trained Ph.D.s by employment status and major field, 2001, 2003, and 2006 (thousands). SOURCE: Data extracted from National Science Foundation Survey of Doctorate Recipients 2001, 2003, and 2006. D-2.eps Biomedical 251 250 220 Clinical Behavioral 20 4 203 200 18 8 175 Males per 10 0 Females 150 13 6 124 116 10 8 10 0 10 0 92 92 91 81 75 70 64 50 0 1995 1997 1999 2001 2003 2006 Year FIGURE D-3 Sex ratio in the U.S.-trained workforce by major field and survey year, 1995-2006. SOURCE: Data extracted from National Science Foundation Survey of Doctorate Recipients, 1995-2006. D-3.eps As the sex ratio fell, female scientists apparently became for current purposes, as being in the workforce, because more likely to find scientific employment. In the biomedical one assumes, for projection purposes, that they could read - field, 85 percent of male scientists were employed in sci - ily return to employment.) The numbers are small—about ence in 2006, but only 79 percent of female scientists were 5,000 in the three fields combined—but the proportion so employed. In the other two fields, employment in science among women in 2006 was five to nine times the propor- was almost equally likely for each gender, which was not tion among men. true in the past (Figure D-4). Female scientists are more The age of the labor force has also been changing gradu- likely than males, however, to be out of the labor force. ally. The median age in 2006 was considerably lower among (Those out of the labor force but not retired are counted, biomedical scientists, at 46.9 years, than in the other two

OCR for page 125
8 APPENDIX D 0.16 Clinical 0.14 and Females Employed in Science Difference in Propor tions of Males Biomedical 0.12 2006 Biomedical 0.10 Behavioral 2006 0.08 Clinical 0.06 2006 Behavioral 0.04 0.02 0.00 -0.02 0 200 40 0 600 800 Ratio FIGURE D-4 Differences in male and female employment in science relative to the sex ratio in the U.S.-trained workforce between 1973 and 2006. D-4.eps SOURCE: Data extracted from National Science Foundation Survey of Doctorate Recipients, 1973 and 2006. foreign-trained Ph.d.s fields. Among clinical scientists the median was 51.1 years, among behavioral scientists 52.3 years. These medians Foreign-trained Ph.D.s provide a substantial addition to represent increases, since 1995, of 1.3 years in the biomedi- the research workforce. They are not the only immigrants cal field, 3.9 years in the clinical field, and 4.5 years in the in the workforce. U.S.-trained Ph.D.s include many non- behavioral field. Those 55 years and older have been a grow- citizens, both permanent residents and temporary residents, ing proportion of the workforce, but their gains have not the latter having risen to a quarter of Ph.D. graduates. These been coming at the expense of the youngest scientists. The noncitizens have been incorporated into the preceding tabu- proportion of those under 35 has also shown some growth; lations, and this section adds only those who received their though relatively slight (Figure D-5). Ph.D.s outside the United States. Clinical Behavioral Biomedical 1.0 Age 0.8 Propor tion of Work force 55 + 0.6 0.4 35 -54 0.2 25 -34 0.0 2006 2006 2006 2003 2003 2003 2001 2001 2001 1999 1999 1999 1995 1995 1995 1997 1997 1997 Year FIGURE D-5 Proportional age distribution of U.S.-trained workforce by major field, 1995-2006. SOURCE: Data extracted from National Science Foundation Survey of Doctorate Recipients, 1995-2006. D-5.eps

OCR for page 125
 APPENDIX D The foreign-trained were not all included in surveys. They in which the foreign-trained differed from the U.S.-trained were counted in 2003 and 2006 (although not in 2001), but was in age distribution. Their median ages were five to eight in both cases, the sample frame was based on the decennial years lower, the difference mainly having to do with fewer census. Any Ph.D.s who entered the country after the census foreign-trained scientists 55 years or older and more under were therefore not counted. Estimates were generated of these 35 years. As a result, whereas foreign-trained biomedical additional migrants through a complicated process involving and clinical scientists made up 21 percent of those 55 years estimating the preceding inflows and extrapolating forward. and older in 2006, of those under 35 years, they made up 45 The specific procedures are described below. They produced percent in the biomedical field and 56 percent in the clinical upward adjustments to the survey figures of 20-40 percent field (Figure D-7). (varying by field) for 2003 and 40-90 percent for 2006. Before adjustment, the foreign-trained were equivalent to WorkforCe eNtraNtS a fourth of the U.S.-trained in the biomedical field in 2006. After adjustment, they were equivalent to about a third. In The three main groups of entrants who regularly augment the clinical field, the adjustment increases the numbers of the workforce correspond to its three major segments, and as the foreign-trained from almost half to two-thirds of the with these three segments, most of the data available cover U.S.-trained. The behavioral field is a stark contrast, where Ph.D. graduates of U.S. universities. This section focuses the foreign-trained are much fewer, and the adjustment mainly on them but adds estimates of foreign-trained Ph.D. increases their numbers from 3 to 5 percent of the U.S.- migrants. trained (Figure D-6). Taking into account the foreign-trained, To summarize the main characteristics of these entrants, the workforce in biomedical sciences, instead of being only we note that, from 2000 to 2007 (the last year for which data slightly larger than the workforce in behavioral sciences, was are available), the numbers of Ph.D. graduates in the three actually 50 percent larger in 2006. major fields have mostly been rising, but the trends have been The proportions employed in science among the foreign- somewhat erratic. NIH funding appears to play an outsize role trained were generally similar to those among the U.S.- in producing these graduates. Their demographic characteris- trained, with no consistent variation across fields. For tics indicate why the workforce is changing. As a group, they instance, in 2006 in the biomedical field, the foreign-trained have become increasingly female and, in a reversal of previous were slightly more likely to be employed in science and trends, are now slightly younger on average than earlier. slightly less likely to be employed out of science, but the reverse was true in the behavioral field. Similarly, gender trends composition was largely similar, except that, in the bio- medical field, the sex ratio was slightly higher among the From 2000 to 2007, annual Ph.D. graduates in the bio- foreign-trained than the U.S.-trained. One important way medical field increased 23 percent and in the clinical field 33 Biomedical 200 63.8 Foreign-trained Behavioral 150 48.6 39.8 U.S.-trained 5.4 4.2 Thousands 3.7 10 0 126.1 120.4 119.4 117.6 113.9 113.3 Clinical 50 16.3 15.3 13.9 24.2 21.0 19.1 0 2001 2003 2006 2001 2003 2006 2001 2003 2006 Year FIGURE D-6 U.S.-trained and foreign-trained Ph.D. workforce, by major field and year (thousands). SOURCE: Data extracted from National Science Foundation Survey of Doctorate Recipients and National Survey of College Graduates, D-6.eps 2001-2006.

OCR for page 125
0 APPENDIX D 0.6 Biomedical Clinical Behavioral 0.56 0.54 0.5 Propor tion Foreign-Trained 0.45 0.40 0.4 0.39 0.34 0.36 0.3 0.29 0.2 0.21 0.21 0.1 0.10 0.04 0.07 0.03 0.03 0.0 ages ages ages 55+ 55+ 55+ 35 -44 35 -44 35 -44 25 -34 25 -34 25 -34 45 -54 45 -54 45 -54 All All All Age (years) FIGURE D-7 Proportion foreign-trained in the workforce by age group and major field, 2006. SOURCE: Data extracted from National Science Foundation National Survey of College Graduates, 2006. D-7.eps 7,000 Biomedical 6,00 0 5,000 Ph.D. Graduates 4,0 00 Behavioral 3,000 2,00 0 1,000 Clinical 0 1960 1970 1980 1990 2000 2010 Year FIGURE D-8 Ph.D. graduates from U.S. universities by major field, 1970-2007: quadratic trend and annual variations. SOURCE: Data extracted from National Science Foundation Survey8.eps Doctorates, 2007. D- of Earned percent. In the behavioral field, however, they fell 6 percent, we show five-year moving averages.) A dip in growth around although the trend since 2005 has been positive. These trends 2000 or shortly after is evident, but since then growth has have generally been in line with earlier trends (Figure D-8). rebounded, although by 2007 it had not reached the levels Roughly around 2001-2002, however, each field experienced of the early 1990s. a drop in Ph.D. graduates that was deeper, at least in the bio- Ph.D. graduates include a number on temporary U.S. medical and behavioral fields, than previous declines dating visas. In 2007, they were 30 percent of biomedical graduates, to the 1970s. Figure D-9 shows growth rates over 15 years for 23 percent of clinical graduates, and 10 percent of behavioral graduates in each field. (Since annual rates are quite erratic, graduates. Their numbers appear to rise when the numbers

OCR for page 125
 APPENDIX D 0.08 Clinical 0.06 Biomedical 0.04 Annual Grow th Rate 0.02 Behavioral 0.00 -0.02 -0.04 1985 1990 1995 2000 2005 2010 Year FIGURE D-9 Annual growth rates for Ph.D. graduates by major field (five-year moving averages). D-9.eps SOURCE: Data extracted from National Science Foundation Survey of Earned Doctorates, 2007. 2,000 2007 Biomedical e 1,500 lin Temporar y Residents ty al i u Eq 1,000 1995 Clinical 2007 500 2007 1970 Behavioral 1970 1970 0 0 1,000 2,000 3,00 0 4,0 00 5,00 0 U.S. Citizens and Permanent Residents FIGURE D-10 Ph.D. graduates who are U.S. citizens or permanent residents versus temporary residents, by major field, 1970-2007. D-10.eps SOURCE: Data extracted from National Science Foundation Survey of Earned Doctorates, 2007. of those who are U.S. citizens or permanent residents rise, graduate), although they may respond more or less strongly although not necessarily as fast (Figure D-10). The correla- than citizens and permanent residents. The rising numbers tions between the sizes of the two groups over time are 0.94 of temporary residents among graduates show no apparent among biomedical graduates, 0.99 among clinical graduates, lasting impact, at least so far, from any recent changes in and smaller although still positive at 0.32 among behavioral immigration regulations. graduates. This suggests that temporary resident students do Most U.S. citizens and permanent residents intend to stay not mainly fill in for citizens but instead respond to similar and work in the United States immediately after graduation. changing incentives and disincentives to enroll (or at least to Only 3-4 percent of those who express an intention say they

OCR for page 125
 APPENDIX D would work elsewhere. The percentages do not vary by sex highest among biomedical graduates. Overall, intentions to or major field and have not changed much over the years, stay have been rising, generally as the proportion who are although recently they may have become marginally more temporary residents has also been rising (Figure D-11). The similar recently across fields. combination of rising proportions of temporary residents and For temporary residents, intentions to stay in the United rising proportions of them intending to stay in the United States after graduation vary across fields, being generally States implies that, among those newly minted Ph.D.s enter- lowest among behavioral graduates and female clinical ing the U.S. workforce, temporary residents are increasingly graduates, intermediate among male clinical graduates, and prominent (Figure D-12). Among biomedical graduates, they 0.9 2007 Biomedical 0.8 2007 Propor tion Intending To Stay 2007 0.7 1970 Behavioral 0.6 Clinical 0.5 1970 0.4 1970 0.3 0.2 0.00 0.05 0.10 0.15 0.20 0.25 0.30 Propor tion Who Are Temporar y Residents FIGURE D-11 Temporary-resident Ph.D. graduates and their proportion intending to stay in the United States, by major field, 1970-2007. SOURCE: Data extracted from National Science Foundation Survey of Earned Doctorates, 2007. D-11.eps Biomedical 0.30 0.25 Propor tion Temporarily Resident 0.20 Clinical 0.15 0.10 Behavioral 0.05 0.00 1960 1970 1980 1990 2000 2010 Year FIGURE D-12 Temporary residents as a proportion of those Ph.D. graduates intending to stay in the United States, by major field, 1970- 2007: quadratic trend and annual. D-12.eps SOURCE: Data extracted from National Science Foundation Survey of Earned Doctorates, 2007.

OCR for page 125
 APPENDIX D were 17 percent in 1990 and 28 percent by 2007. Even among leading the sex ratio to drift further from equality. Trends behavioral graduates they showed an increase in this period, among biomedical graduates (not shown) resemble those from 4 to 7 percent. for clinical graduates. The behavioral field shows a contrast- ing pattern. In the early 1970s, male and female graduates increased in parallel, but since then, increases in female Sex and age graduates have gone with decreases in male graduates. The The proportion of graduates who are female is rising in sex ratio is dropping among behavioral graduates as females each field. Women have been the majority among clinical replace males. graduates since 1983, just over a decade before they became Since roughly 1995, Ph.D. graduates have been getting a majority in the clinical workforce. Among behavioral younger. This is true in each field even when male and graduates women have been the majority since 1986, and female graduates are examined separately (Figure D-14). it took 17 more years before they became a majority in the The change has been relatively gradual and by 2007 still fell workforce. By 2007 in these two fields, female graduates short of reversing the rising trends in the median age over the outnumbered male graduates by two to one, and the female- preceding 15 years. By 2007, median age among biomedical male gap has continued to grow. Women are still a minority, graduates was 31.3 years, down from the high of 32.2 in 1995 but not by much, in the biomedical field, in which the sex and the lowest since 1984. Among behavioral graduates, the ratio (the ratio of males per 100 females) was down to 109 by median age of 33.6 years was two years younger than the 2007, from 163 in 1990. The sex ratio has been falling faster peak in 1990 and the lowest since 1983. In both these fields, among biomedical graduates than in the other two fields. All the median age for female graduates was lower than that the graduate sex ratios are well below those in the research for males by half a year to a year, a divergence that started workforce as a whole. mainly in the 1990s. Among clinical graduates, median age The clinical and behavioral fields differ in one important has also declined, but from considerably higher levels, as respect. Figure D-13 shows the trend in numbers of female high as 36.5 among males (in 1994) and 42.6 among females clinical graduates plotted against male graduates over time, (in 1997). The 2007 medians of 35.0 and 38.1 thus represent with the trend for each decade shown separately, and the substantial declines. As these estimates indicate, female parallel trend among behavioral graduates. In the 1970s and clinical graduates have been substantially older than males, through the early 1980s, numbers of male and female clini- unlike in the other two fields. cal graduates were mostly uncorrelated. In the late 1980s, The declines in age have involved mainly reductions in however, both numbers tended to increase in parallel, a older graduates, those in their late thirties or older. Gradu- tendency clearly continuing in the 1990s and 2000s. Factors ates have become somewhat more concentrated at relatively that increase the number of female graduates also appear to younger ages. The tendency has been for younger average ages increase the number of males, although not quite as strongly, and a greater concentration of graduates at modal ages to go 3,000 2000s Behavioral 2,500 1990s 2,000 1980s Clinical Females 1,500 2000s e 1970s lin ty li ua 1,000 Eq 1990s 1980s 500 1970s 0 0 500 1,00 0 1,500 2,000 2,500 Males FIGURE D-13 Numbers of male and female clinical and behavioral graduates, 1970-2007. SOURCE: Data extracted from National Science Foundation Survey of Earned Doctorates, 2007. D-13.eps

OCR for page 125
 APPENDIX D 43 Behavioral Biomedical Clinical 41 39 Female Median Age (years) 37 Male 35 33 31 29 1970 1980 1990 2000 1970 1980 1990 2000 1970 1980 1990 2000 Year FIGURE D-14 Median age among Ph.D. graduates by major field and sex, 1970-2007. SOURCE: Data extracted from National Science Foundation Survey of Earned Doctorates, 2007. D-14.eps 0.60 1970 B i omedical Proportion in Modal Four-Year Age Group 2007 0.50 1970 Behavioral 0.40 2007 1970 0.30 2007 C linic al 0.20 0.10 28 30 32 34 36 38 40 Median age (years) FIGURE D-15 Proportion of graduates in the modal four-year age group, by median age and major field, 1970-2007. SOURCE: Data extracted from National Science Foundation Survey of Earned Doctorates, 2007. D-15 together. In the biomedical field, which has a lower median age The declines in median age do not imply that training is than the other two fields, 53 percent of 2007 graduates were in becoming more efficient or even shorter. Offsetting slightly the modal age range of 27-31 years. In the clinical field, with earlier graduation is increasing resort to postdoctoral work. the highest median age, only 27 percent were in the modal age In 1995, 30 percent of behavioral graduates planned on a range of 30-33 years. Across fields and over time, the pattern postdoctoral fellowship. By 2007, this proportion had risen is unmistakable (Figure D-15). As median ages rose in each to 46 percent. Clinical graduates showed a smaller increase, field, the modal age covered fewer graduates, but then the from 21 to 28 percent. Although biomedical graduates did trend reversed, leaving median ages and modal concentrations not show an increase in this period, they were already at a in 2007 at intermediate levels in each field. very high level: 82 percent by 2007.

OCR for page 125
 APPENDIX D Nih Support Tabulating the 2003 stock by date of entry into the United States, we reverse-survived each immigrant cohort to date of Graduates in the biomedical field passed 5,000 in 1995, entry to determine its initial size. We used retirement rates about the same time that the number of NIH predoctoral bio- and mortality rates by age for the workforce as a whole, medical awards (National Research Science Awards [NRSAs], described below. These calculations were done separately for trainees and fellows combined) also passed 5,000. Gradu- for men and women in each field. The results suggested ates passed 6,000 in 2005, and predoctoral awards passed a rising trend among foreign-trained biomedical scientists 6,000 two years later (Figure D-16). In the 1980s and early and considerable variability but no convincing trend among 1990s, the ratio of biomedical Ph.D. graduates to NRSA- foreign clinical and behavioral scientists, who were consid- supported doctoral students was close to 0.8:1, rising to erably fewer. around 1.1:1 after 2000. This suggests that the clear majority Figure D-17 shows the estimated inflows of these foreign- of Ph.D.s in these fields were NRSA awardees. If three out of trained Ph.D.s, compared with the trends in the other entrants four awardees completed the Ph.D., then slightly more than into the workforce: the U.S.-trained Ph.D. graduates (citizens that proportion of graduates were awardees after 2000. and temporary residents) who choose to stay in the United In the clinical field, NRSAs have fluctuated somewhat States. Annual immigrant numbers are relatively volatile. in number between 500 and 900. The ratio of graduates to In the biomedical field in 2000, they were 50 percent of current NRSAs was around 1.6:1 in the 1980s and 1990s the entrants into the workforce, up from only 15 percent in and settled to 2:1 in the 2000s. Assuming all the awardees 1990. (Estimates were also made for the years 1980-1984 graduate, close to half of graduates would have received an combined and 1985-1989 combined, when inflows were award. In the behavioral field, NRSAs since 1990 have fallen even smaller.) Because U.S.-trained temporary-resident short of the number in the clinical field and are proportion- graduates were an additional 10 percent of the total in 2000, ally much less important. The ratio of graduates to current immigrants made up the majority of new workforce entrants awardees is between 5:1 and 7:1, suggesting that a minority in that year, and could have been an even larger proportion, of graduates—around 15 percent—receive such awards. because permanent residents are lumped with citizens. In the clinical field, foreign-trained immigrants outnumbered foreign-trained Ph.d.s graduates in some years in the early 1990s, and in the late 1990s were 20-40 percent of the total. In the behavioral field No counts or sample estimates are available for the flow they were much less consequential, averaging only 7 percent of foreign-trained Ph.D.s into the U.S. workforce. From the of the entrants over the decade. stock information in the 2003 survey, considered above, we These immigrant flow figures are in a sense underes- estimated the probable annual inflow from the 1990s, making timates, because those who may have returned to their assumptions about retirements and deaths. countries of origin or emigrated elsewhere were not counted 7,000 Biomedical 6,000 Behavioral 5,000 4,000 Number Graduates 3,000 2,000 Trainees Clinical & fellows 1,000 0 1970 1980 1990 2000 1970 1980 1990 2000 1970 1980 1990 2000 Year FIGURE D-16 Ph.D. graduates and NRSA predoctoral trainees and fellows by major field, 1970-2007. D-16.eps SOURCE: Data extracted from National Science Foundation Survey of Earned Doctorates, 2007, and NIH IMPACII Database.

OCR for page 125
 APPENDIX D among immigrants will be essentially constant (which is The aging of the workforce will be most notable among expected at least among clinical and behavioral scientists behavioral scientists, among whom the proportion 55 years given our assumption of a constant flow). Among U.S. and older will rise from 39 percent in 2006 to 44 percent in graduates, however, sex ratios will be falling in each field. 2016. This proportion is much smaller in the other two fields They are already well below sex ratios in the workforce as a and will rise only to 24 percent among biomedical scientists whole—35 to 60 points below—and by 2016 will be 45 to 75 and 30 percent among clinical scientists (Figure D-30). points below, with the greatest gap in the biomedical field. A Younger foreign-trained researchers clearly help keep ages predominantly male biomedical professoriate, therefore, will down in the biomedical and clinical fields, although even be training a majority female student body. without them, the workforce in these fields would be younger Median ages among behavioral scientists will rise, but in than in the behavioral field. the other two fields, trends will be less clear. Male behavioral One consequence of having more older scientists is more scientists, already the oldest in 2006 with a median age of retirements and deaths. The proportion retiring or dying 54.7 years, will become older still, reaching a median of 57 annually will rise in each field, from a range of 1.2-1.6 per- years by 2016. The median age among female behavioral cent of the workforce in 2007 to 1.4-2.8 percent in 2016. The scientists will also rise, from 49.2 to 49.9 years. The median proportion retiring or dying will be particularly high among age is lowest among female biomedical scientists, at 41.6 male behavioral scientists, at 3.8 percent by 2016. Among years, and will inch upward to 41.9 years. At the same time male behavioral scientists, this number already exceeded the median age for male biomedical scientists will be falling, the number of graduates actually entering the workforce in from 47 to 46.4 years. Trends in median age among clinical 2006 (leaving out those not intending to stay in the United scientists will also provide a contrast between males and States) and will be almost twice as numerous by 2016 (Fig- females, although the trends will not be linear. Among males, ure D-31). The 3,800 behavioral scientists, both male and median age will rise and then fall; among females it will fall female, projected to retire or die in 2016 will be barely under and then rise, in neither case producing much net change. the number of new Ph.D. graduates, which is projected at Some of the complications in age trends are due to dif- 4,200. This does not account for foreign-trained Ph.D.s, but ferences between U.S.-trained and foreign-trained scientists they are few in the behavioral field. (Figure D-29). The foreign-trained have been younger than The proportion of the workforce employed in science the U.S.-trained in each field and have helped keep the should stay roughly the same, about 80-85 percent in each median age down. Median age is projected to rise among field. The proportion working in non-science jobs should also the foreign-trained in most cases, however. At the same time, stay at 12 percent in the biomedical field but could drop in the median age should fall among the U.S.-trained biomedical other two fields. It was slightly higher in the clinical field than and clinical scientists, while it rises among behavioral in the biomedical field in 2006, at 17 percent, and could fall to scientists. 11 percent. It was also higher in the behavioral field, at 18 per- Behavioral 60 U.S.-trained Male Fe male Foreign-trained Combined Clinical 55 Male Fe male Median Age (years) Biomedical Male Fe male 50 45 40 35 2006 2006 2006 2006 2006 2006 2016 2016 2016 2016 2016 2016 Year FIGURE D-29 Median age of projected workforce by major field, sex, and source of Ph.D., 2006-2016. SOURCE: NRC analysis. D-29.eps

OCR for page 125
 APPENDIX D Biomedical Clinical Behavioral 55 + 1.0 Foreign 0.9 U.S. 0.8 35 -54 0.7 Propor tion Foreign 0.6 U.S. 0.5 0.4 25 -34 0.3 Foreign 0.2 U.S. 0.1 0.0 2006 2006 2006 2008 2008 2008 2010 2016 2010 2016 2010 2016 2012 2012 2012 2014 2014 2014 Year FIGURE D-30 Projected age distribution of the workforce by source of training and major field, 2006-2016. D-30.eps SOURCE: NRC analysis. Biomedical Clinical Behavioral 0.06 Males Females Males Females Males Females 0. Propor tion of Work force 0.04 Entering graduates 0.03 Dying 0.02 Retiring 0.01 0.00 2007 2007 2007 2007 2007 2007 2013 2013 2013 2013 2013 2013 Year FIGURE D-31 Projected proportions dying and retiring, compared to entering graduates as a proportion of the workforce, by major field and sex, 2007-2016. D-31.eps SOURCE: NRC analysis. cent, and could fall to 15 percent. Since transition rates among than those employed in science. The greatest increase will be these statuses were based on rates over more than a decade and in the biomedical field (Figure D-32). The obvious explana- not allowed to change, one would expect quite recent changes tion is the growing proportion of women in the workforce. in transition rates, such as the mid-2000s, increase in propor- Although this factor should have a role, decomposition of the tion out of science, to be reversed to some degree. change suggests it is not the main explanation. In the bio- The greatest proportional change in employment status medical field, it is due instead to a projected change among could involve the proportion not in the labor force. This women themselves. Between 2003 and 2006, the proportion small segment of the workforce, between 1.5 and 3 percent out of the labor force among female biomedical scientists fell of the workforce in 2006, could grow 50-250 percent faster by a third, from an unusually high 7.5 percent to a still very

OCR for page 125
8 APPENDIX D Number Percent 12 ,000 4 2006 2006 10 ,000 2016 3 2016 8,000 Number Percent 6,000 2 4,000 1 2,000 0 0 Biomedical Clinical Behavioral Biomedical Clinical Behavioral FIGURE D-32 Projected number and percentage of the workforce not in the labor force by major field, 2006 to 2016. SOURCE: NRC analysis. D-32.eps 350,00 0 High Medium 30 0,00 0 Low 250,00 0 No migrants Biomedical 200,00 0 Scientists 150,00 0 Behavioral 10 0,00 0 Clinical 50,0 00 0 2006 2008 2010 2012 2014 2016 Year FIGURE D-33 Alternative workforce projections by major field, 2006-2016. D-33.eps SOURCE: NRC analysis. high 5.0 percent. With no evidence that this is the start of a fields (Figure D-33). The medium scenario gives almost long-term trend or even a long-term downward adjustment, identical 61 percent and 58 percent increases in the work- we have allowed rates to return to previous levels, which force between 2006 and 2016 in the biomedical and clinical accounts for the major part of the increase in those not in fields. The high scenario gives an increase in the biomedical the labor force in this field. field of 84 percent, 23 percentage points higher than the medium scenario (Figure D-34). In the clinical field, the high scenario gives an increase that is only 13 percentage points alternative Scenarios higher than the medium scenario. In the behavioral field, The alternatives to the medium scenario provide much the medium 10-year increase is only 9 percent, and the high more variation in the biomedical field than in the other two scenario gives an increase only 2 percentage points higher.

OCR for page 125
 APPENDIX D Biomedical Behavioral Clinical 20 06 -2016 medium 20 06 -2016 medium 2006 -2016 medium 23 % 25 increase = 9 % increase = 58 % increase = 61% 13 % Percentage Point Difference 2% 0 –3 % –6 % –14% –14% –25 –32% –46% –50 High Low No High Low No High Low No Migrants Migrants Migrants Alternative Scenario FIGURE D-34 Difference of alternative scenarios from the medium projection in percentage change from 2006 to 2016, by major field. SOURCE: NRC analysis. D-34.eps On the other side of the ledger, the low scenario gives a 14 percentage growth between 2006 and 2016 would be greater percentage point lower increase in both the biomedical and or smaller under different combinations of immigrant and clinical fields than the medium scenario and a 3 percentage graduate growth. Except in the behavioral field, where the point lower increase in the behavioral field. foreign-trained workforce is only a small percentage of the One could of course select arbitrarily symmetrical alter- total, the variation between high and low numbers of gradu- native scenarios, but the ones chosen are derived from some ates has a smaller effect on the projected workforce than past experience with inflows and outflows, and therefore the variation between high and low numbers of immigrants. presumably represent more realistic possibilities. The range And high numbers of graduates would clearly not make up of past growth rates has been narrower in the clinical field in for zero immigrants. Still, because zero immigration is in the past than in the biomedical field, and still narrower in the fact an ahistorical situation, one cannot entirely rule out behavioral field, which is why the scenarios for biomedical an ahistorical increase in graduates that goes beyond past researchers produce much more variation. trends. An extreme scenario has migrant inflows ending and The proportion who will be foreign-trained varies in temporary-resident graduates all deciding to work overseas. different scenarios, especially in the biomedical field (see Combined with the low projection for total graduates, this Figure D-36 and Table D-7). In the high projection for bio- would produce larger reductions relative to the medium medical scientists, the foreign-trained would reach almost projection, as much as 46 percentage points lower in the half of the total by 2016, in the low scenario only 40 percent. biomedical field. But the workforce would still grow in In the no-migrant scenario, there will still be foreign-trained each field. researchers, because we assume that immigration does not Could an increase in graduates make up for the hypoth- cease until 2010, and in addition do not assume that foreign- esized lack of immigrants in this scenario? We have not trained researchers already in the United States all emigrate. assessed the factors that might make a substantial increase By 2016, they would still be 32 percent of the total. In the in graduates possible, such as faculty, funding, and student clinical field, the proportion foreign-trained in 2016 varies in interest. We can, however, assess whether this is likely a more limited range, from 38 percent to 31 percent (in the given past trends in graduates, as reflected in the different no-migrant scenario), and in the behavioral field the variation projection scenarios. Comparing scenarios suggests that, is even more limited. absent an ahistorical boost in graduates, they would not Across the different scenarios, variations in projected sex increase enough to fill all the niches that immigrants would ratios are relatively slight. The greatest variation will be in potentially leave vacant. Figure D-35 compares additional the biomedical field in 2016, when the sex ratio could range alternative scenarios with the medium scenario, showing how from 149 to 138. Variation in age will be somewhat greater.

OCR for page 125
0 APPENDIX D Biomedical Clinical Behavioral 25 Migrants: Graduates: Percentage Point Difference High High 0 Medium Medium –25 Low Low Zero Zero –50 Medium Migrants Medium Migrants Medium Migrants Medium Graduates Medium Graduates Medium Graduates FIGURE D-35 Difference of other projections from the medium projection in percentage change from 2006 to 2016, by major field. SOURCE: NRC analysis. D-35.eps 0.5 Clinical Biomedical Scenario High Medium 0.4 Low Propor tion of Work force No Migrants 0.3 0.2 0.1 Behavioral 0.0 2006 2006 2006 2008 2008 2008 2010 2016 2010 2016 2010 2016 2012 2012 2012 2014 2014 2014 Year FIGURE D-36 Foreign-trained Ph.D.s as a proportion of the workforce in alternative scenarios, by major field, 2006-2016. SOURCE: NRC analysis. D-36.eps In general, more rapid growth implies a younger workforce. female biomedical scientists, who are clearly younger and Regardless of the scenario, however, behavioral scientists will not catch up in age to any other group in any scenario. will be older in 2016 than in 2006 (Figure D-37). Clinical scientists will also be older in almost every scenario. But Will growth Be adequate? biomedical scientists could be younger in 2016, under the high or medium scenarios, or older, under the low or no- What workforce growth would be enough to maintain the migrant scenarios. When scientists are divided by field and research infrastructure in various disciplines, to nurture suf- sex, four of the six groups are not that different in current and ficient scientific discovery, and to allow for new disciplines projected median age. The exceptions are male behavioral to develop within each field? This would require much more scientists, who are clearly older and will get older still, and information, as well as value judgments about what is suf-

OCR for page 125
 APPENDIX D TABLE D-7 Alternative Workforce Projections by Major Field and Source of Training, 2006, 2011, and 2016 Biomedical Clinical Behavioral Projection Foreign- Foreign- Foreign- and Year Total U.S.-Trained Trained Total U.S.-Trained Trained Total U.S.-Trained Trained Medium Projection 2006 189,860 126,098 63,762 40,511 24,165 16,347 125,794 120,354 5,439 2011 244,224 149,367 94,848 51,779 31,754 20,025 133,923 127,584 6,339 2016 305,571 173,588 131,986 63,808 40,454 23,351 137,221 130,100 7,120 High Projection 2006 189,860 126,098 63,762 40,511 24,165 16,347 125,794 120,354 5,439 2011 260,153 150,901 109,241 53,660 32,161 21,500 134,533 128,106 6,428 2016 348,292 181,402 166,896 68,966 42,676 26,288 139,804 132,510 7,293 Low Projection 2006 189,860 126,098 63,762 40,511 24,165 16,347 125,794 120,354 5,439 2011 234,116 147,791 86,319 49,536 31,339 18,198 132,802 127,037 5,766 2016 278,070 166,087 111,985 58,046 38,320 19,723 133,653 127,658 5,994 No-migrants Projection 2006 189,860 126,098 63,762 40,511 24,165 16,347 125,794 120,354 5,439 2011 215,341 140,073 75,261 47,036 29,936 17,099 131,381 125,945 5,437 2016 218,034 147,388 70,645 50,887 34,942 15,941 129,988 125,134 4,853 SOURCE: NRC Analysis. Biomedical Clinical Behavioral 60 Male Female Male Female Male Female 55 Median Age (years) 2006 median 50 45 40 No Migrants No Migrants No Migrants No Migrants No Migrants No Migrants High Medium High Medium High Medium High Medium High Medium High Medium Low Low Low Low Low Low Scenario FIGURE D-37 Median age in 2016 in alternative scenarios, compared with 2006, by major field and sex. SOURCE: NRC analysis. D-37.eps ficient, to determine. We can, however, compare projected fall well below the norm. For the clinical and behavioral workforce growth with past growth, particularly of the fields, all the projections—even the high scenario—are well U.S.-trained workforce. (These comparisons are limited to below previous experience, suggesting that growth will con- the U.S.-trained only, since no long series is available for the tinue to slow, as it has for some decades, although not stop. A foreign-trained.) As Figure D-38 shows, in the biomedical possible hypothesis is that these fields are maturing or have field projected growth is roughly in the range of percentage matured, and would need striking and expansive new ideas increases in the U.S.-trained workforce in previous decades, to reenter a rapid-growth phase. except for the zero migrant scenario, in which growth would Research funding levels are of course highly relevant

OCR for page 125
 APPENDIX D Behavioral Clinical Biomedical 126% 120 119% Projected, 20 06 -2016: 10 0 Percentage Increase High 83 % 87% 80 76 % 70 % Medium 67% 60 61% 58 % Low 46 % 43 % 40 44% 42 % 39 % 26 % 26 % 20 No 11% 15% migrants 9% 6% 0 3% 1973 - 1983 - 1993 - 1973 - 1983 - 1993 - 1973 - 1983 - 1993 - 1983 1993 2003 1983 1993 2003 1983 1993 2003 Year FIGURE D-38 Percentage increases in the U.S.-trained workforce in past decades and alternative projections for the entire Ph.D. workforce for 2006-2016, by major field. D-38.eps SOURCE: NRC analysis. to the question of whether the projected workforce will maintained between research funding and the workforce is be adequate. In 2003, total U.S. funding for biomedical suggested by the fact that funding, in the 1994-2003 period, research—from the NIH and other government sources, shifted to some degree from basic biomedical research to industry, and foundations and other private sources—totaled clinical research,5 at the same time that the clinical workforce $75.5 billion.2 If this were distributed equally among Ph.D.s appeared to be growing faster than the biomedical workforce in the three major fields combined, research funds would (at least where U.S.-trained Ph.D.s are concerned). amount to $232,000 per researcher. The 2006 total of $93.4 Whether projected workforce growth keeps pace with billion would imply $262,000 per researcher in that year. research funding depends on the trend in funding. Biomedical This was an apparent increase in funds per researcher of 13.3 Ph.D.s are projected to increase 4.8 percent from 2006 to percent in three years, but, if one adjusts for rising research 2016, clinical Ph.D.s, 4.5 percent. Both these estimates are costs (using the Biomedical Research and Development lower than the annual funding growth rate (adjusted for Price Index3), the increase is reduced to a trivial 0.4 percent. research costs) of 7.5 percent in 1994-2003 but higher than In this period at least, workforce growth and research funding the more recent growth rate of 3.4 percent in 2003-2007. kept pace with each other. Since the 3.4 percent rate predates the great recession, one This may not always have been the case. From 1994 to might expect funding growth up to 2016 to slow even further. 2003, total U.S. research funding grew at an annual rate of The additional $8.6 billion from the American Recovery and 7.5 percent (adjusted for changing research costs). 4 Fig- Reinvestment Act of 2009 provides a short-term boost for ure D-39 shows this earlier trend (expressed as an index, 2009 and 2010,6 but even adding $4.3 billion more for every with 2003 levels set to 100), distinguished from the trend year thereafter on top of an annual 3.4 percent increase would since 2003 because data sources differed and gave somewhat raise the funding growth rate only to 3.7 percent. higher estimates. In 1994-2003, the U.S.-trained biomedical The implication appears to be that growth in the biomedi- workforce grew at a rate of 3.5 percent, and the clinical work- cal and clinical workforces will somewhat exceed growth force at 5.5 percent. Foreign-trained researchers might have in research funding, if funding growth (in real rather than made up the gap, but we do not have adequate data before constant dollars) follows or falls short of the growth rate for 2001 to determine this. That some correspondence was 2003-2007. The reverse situation, where funding grows faster than the workforce, may be understandable, but diminishing research funding per capita would seem to be an uncomfort- Dorsey, E.R., et al. 2010. Funding of U.S. biomedical research, JAMA 2 303(2):137-143. National Institutes of Health, Office of Budget. 2010, Biomedical Ibid., pp. 1336-1337. 3 5 Research and Deelopment Price Index (BRDPI). Available at: http:// This includes $8.2 billion for the NIH for extramural scientific research 6 officeofbudget.od.nih.gov/gbiPriceIndexes.html. and $0.4 billion for AHRQ for comparative effectiveness research, accord - Moses, H. III, et al. 2005. Financial anatomy of biomedical research. ing to NIH, 2010, PowerPoint presentation on NIH Implementation of 4 JAMA 294(11):1333-1342. ARRA, available at: http://grants.nih.gov/recovery/.

OCR for page 125
 APPENDIX D Biomedical 180 Clinical 160 Funding of U.S. 140 Index (2003 = 10 0) biomedical research (adjusted for changing research costs) 120 Behavioral 10 0 80 60 40 1993 1998 2003 2008 2013 2018 Year FIGURE D-39 Index of research funding compared with indexed past and projected growth of the research workforce (2003 = 100). SOURCES: Moses, H. III, et al. 2005. Financial anatomy of biomedical research. JAMA 294(11):1333-1342; Dorsey, E.R., et al. 2010. Funding of U.S. biomedical research. JAMA 303(2):137-143; and NRC analysis. D-39.eps able situation. In the low scenario, growth in these two fields force. Their numbers could understandably decline should is reduced to 3.8 and 3.6 percent annually, which is closer U.S. funding increase at a slower pace than elsewhere. to the expected trend in funding growth if it follows 2003- 2007 rates but could still be higher than recession-affected Projection accuracy rates. Behavioral Ph.D.s are excluded from this calculation because so much of the funding—from pharmaceutical, bio- The accuracy of these projections cannot be determined technology, and medical device firms—is likely to involve prospectively. We can, however, make two types of compari- at best limited behavioral research. (These three types of sons: of these projections with projections from other sources firms accounted for almost 60 percent of research funding in and of earlier projections that used the current methodology 2007.) However, if one included behavioral Ph.D.s, combin- with subsequent survey estimates. ing the medium projections for the three fields, the projected Alternative projections have been produced by the Bureau workforce growth rate of 3.5 percent annually for 2006-2016 of Labor Statistics9 (BLS), as part of a regular program that would also be closer to the recent funding growth rate. produces 10-year projections for all occupations and indus- Should the research workforce grow more slowly than tries. Comparisons with BLS projections are not straight- projected, one possibility would be slower growth in foreign- forward because occupational classifications differ. The trained Ph.D.s. The large role of foreign-trained Ph.D.s in the closest to our categories of biomedical and clinical scientists workforce (other than in the behavioral sciences) is probably (who as earlier noted numbered 190,000 and 40,500, respec- predictable from research funding patterns. U.S. funding for tively, in 2006) are their categories of biological scientists biomedical research represents 70-80 percent of global fund- and medical scientists (95,000 and 114,200, respectively, in ing.7 We do not have data on the global workforce in these 2008). The BLS subcategories are too limited to allow sort- fields, but of science and engineering researchers worldwide, ing into biomedical and clinical groups, and the categories U.S. researchers make up only 25 percent.8 The probable mis- include some without Ph.D.s. But totals at least are close, match between providing so much of the research funding and and we make comparisons of both BLS categories against possibly a smaller share of researchers could help explain why both of those used here. large numbers of foreign researchers are entering the work- For 2008-2018, BLS projects much slower growth in numbers of both biological and medical scientists than we project for 2006-2016 for biomedical or clinical scientists. Whereas we project annual growth rates of 4.5-4.8 percent, Dorsey et al., op. cit., p. 141, citing Schweitzer, S. O. 2007. Pharmaceu­ 7 tical Economics and Policy. New York: Oxford University Press. American Association for the Advancement of Science. 2008. Guide to 8 R&D funding data—International comparisons. Available at: http://www. Bureau of Labor Statistics. 2010. National employment matrix. Avail- 9 aaas.org/spp/rd/guiintl.htm. able at: http://www.bls.gov/emp#data.

OCR for page 125
 APPENDIX D 0.10 2004-2008 0.08 Biomedical Clinical Annual Grow th Rate Ph.D.s, Ph.D.s, 2006 –2016 2006 –2016 0.06 0.048 0.045 0.04 2008 -2018 2004-2014 Bureau 0.02 of Labor Statistics 0.019 0.033 0.016 0.029 0.053 0.0 99 0.00 Biological Medical Biological Medical Biological Medical Scientists Scientists Scientists Scientists Scientists Scientists FIGURE D-40 Annual growth rates for biological and medical scientists for various periods from the Bureau of Labor Statistics and current projections for biomedical and clinical Ph.D.s. D-40.eps SOURCES: Bureau of Labor Statistics data downloaded from www.bls.gov/emp#/data in January 2010; Hecker, D.E. 2005. Occupational employment projections to 2014. Monthly Labor Reiew 128(11):70-101; and NRC analysis. BLS projects rates of 1.9-3.3 percent. Could BLS be taking in 2008, much greater than our 125,800 behavioral scientists into account the current recession, which is not reflected in in 2006. (We have excluded speech-language pathologists, the survey data from which we derive projection parameters? who would have added 119,300 to the total.) BLS projects One way to investigate this possibility is to see how their annual growth of 1.7 percent for 2004-2014 and 1.2 percent projections have changed. A paper10 published in 2005 gives for 2008-2018 for the combined group, in contrast to our 0.9 projections for 2004-2014. This paper gives growth rates for percent for 2006-2016 (Figure D-41). The BLS projections, their two categories as 1.6 and 2.9 percent, actually less opti- in this case, indicate faster growth than our projections, but mistic than the recent ones, so recent economic upsets do not its 2004-2008 data actually show annual change of –1.1 seem to have been a factor in their projecting slow growth. percent. As with biomedical and clinical scientists, our pro- Their earlier projections do give 2004 baseline data, however, jections differ from the BLS’s in the direction of relatively which can be compared with the more recent 2008 baseline more acknowledgment of past trends. (Figure D-40). This comparison suggests surprisingly high Have projections such as those made here been accurate annual growth rates for 2004-2008 of 5.3 and 9.9 percent. in the past? The 2005 report on national needs11 contained For whatever reason, BLS appears to project far slower similar projections for 2001-2011. From this series, projec - growth for these occupations than is actually reflected in the tions up to 2006 can be compared with the 2006 survey base data they use for their projections. Our projections look results, although this is not a simple matter. The 2001 much more reasonable in comparison, and the fact that they survey did not count foreign-trained Ph.D.s, who had to be are lower than the 2004-2008 rates could be justified by the estimated previously and were reestimated in the current potential impact of the great recession. exercise from later data. The 2006 survey undercounted the BLS categories for behavioral scientists are even more foreign-trained, and immigrants arriving since 2000 had to difficult to match with ours. We combine the BLS categories be estimated and added. We make the comparisons never- of psychologists, sociologists, and anthropologists, recogniz- theless in Figure D-42. The growth rates projected in the ing that the “most significant” source of personnel in some 2005 report were too low for biomedical scientists, too high of these categories is individuals with a master’s rather than for clinical scientists, and slightly too high for behavioral a doctoral degree. The total for these categories is 180,900 scientists. The range between high and low projections was Hecker, D.E. 2005. Occupational employment projections to 2014. National Research Council. 2005. Adancing the Nation’s Health 10 11 Monthly Labor Reiew 128(11):70-101. Needs. Washington, DC: The National Academies Press.

OCR for page 125
 APPENDIX D 0.020 Behavioral 0.017 0.015 Ph.D.s, 2006 -2016 0.012 0.010 0.009 Annual Grow th Rate 0.005 2008 -2018 2004-2014 projected projected 0.000 2004-2008 Bureau of Labor Statistics actual – 0.0 05 –0.011 – 0.010 – 0.015 FIGURE D-41 Annual growth rates for psychologists, sociologists, and anthropologists for various periods from the Bureau of Labor Sta - D-41.eps tistics and current projections for behavioral Ph.D.s. SOURCES: Bureau of Labor Statistics data downloaded from www.bls.gov/emp#/data in January 2010; Hecker, D.E. 2005. Occupational employment projections to 2014. Monthly Labor Reiew 128(11):70-101; and NRC analysis. 0.07 Biomedical Clinical Behavioral 0.062 0.06 0.055 0.049 0.05 Annual Grow th Rate 0.04 3 Projected 0.041 0.039 Survey Estimates 0.04 0.03 3 0.03 0 0.03 0.019 0.02 0.017 0.015 0.013 0.01 0.00 High Medium High Medium High Medium Low Low Low FIGURE D-42 Workforce annual growth rates, 2001-2006, as previously projected and as derived from surveys, by major field. D-42.eps SOURCE: National Research Council. 2005. Adancing the Nation’s Health Needs. Washington, DC: The National Academies Press; and NRC analysis. apparently not wide enough to encapsulate actual (or at least increase in immigrants in 1999 and 2000 was therefore not estimated) trends. factored into the projection. The errors regarding the clinical The 2005 biomedical projection may have been too low projection, in contrast, may involve the size of the group, mainly because foreign-trained Ph.D.s were estimated from which is relatively small and was fast growing, with consid- 1993 survey data and projected forward to 2001 to provide erable year-to-year volatility. the base for a further projection to 2011. The substantial One implication that should be drawn, clearly, is that the

OCR for page 125
 APPENDIX D alternative high and low projections here should not be taken residents would stay in the U.S., beginning with the 2008 to represent the limits of possible future variation. They are cohort. Under these conditions, the decadal increment to not based on the most extreme trends in past data and should the biomedical workforce would drop from 116,000 to not be taken to represent the extreme future possibilities, but 28,000—which would still be double the highest projected are rather alternatives that depart somewhat modestly from increment to the behavioral workforce. For the biomedical the main scenario. A second implication is probably that workforce to actually decline by 2016, a still more extreme better, up-to-date data are needed on foreign-trained Ph.D.s. situation would have to be imagined, such as, in addition, the departure of all foreign-trained Ph.D.s now in the United States. They were a third of the biomedical workforce in CoNCluSioN 2006, and their departure, together with a halt to immigra- Projecting the research workforce in three major fields— tion, would reduce the 2016 biomedical workforce by 42,000 biomedical, clinical, and behavioral—indicates that each from its 2006 level. faces different prospects. This variation in prospects is An increasing proportion of U.S.-trained Ph.D.s in the visible from a close examination of survey data on the workforce are female. In the U.S.-trained clinical workforce, workforce and on graduates. Running projections serves to they have been the majority since 1994, and in the behavioral confirm and concretize conclusions that might be drawn from workforce, they became the majority around 2004. In both such an examination. cases, their majority is projected to become larger. In the The biomedical research workforce has grown rapidly, biomedical field, they are still the minority. Although they particularly in recent years. From 2001 to 2006, it expanded will remain so until 2016, the gap will narrow, with the sex 24 percent, adding 37,000 scientists. The behavioral research ratio falling from 183 males per 100 females in 2006 to 146 workforce, in contrast, grew only 7 percent in the same in 2016. In the biomedical workforce as a whole, however, period, adding only 8,000 scientists. The clinical research the gap will be greater, because women are a smaller minor- workforce is much smaller than the other two, in total only ity among foreign-trained Ph.D.s, both in the biomedical slightly larger than the 2001-2006 increment in the bio- and clinical field, although not in the behavioral field. Even medical workforce. It grew almost as fast as the biomedical among the foreign-trained, however, the sex ratio appears workforce, at 23 percent. generally to be falling. Reflecting this recent history, the biomedical workforce is The workforce will almost certainly age in the slow- projected to grow, over a decade from 2006 to 2016, by 61 growing behavioral field, where the proportion 55 years and percent, the clinical workforce almost as fast at 58 percent, older will reach 44 percent by 2016. Whether the workforce and the behavioral workforce by an anemic 9 percent. Slow will also age in the other two fields is less clear, since an growth in the past has gone with less volatility, and alterna- increase in the smaller proportions 55 years and older is to tive projections for behavioral and clinical scientists show some extent balanced by increases in the proportions under less variation than alternatives for biomedical scientists. age 35. Among those with U.S. Ph.D.s, behavioral scientists were Will sufficient research funding be available for the pro- almost as numerous as biomedical scientists in 2006, and jected workforce? For 2006, total U.S. biomedical research actually more numerous up to 2001. However, behavioral funding, from government, industry, and foundations, was Ph.D. graduates of U.S. universities have hardly changed in $93.4 billion, or $262,000 per scientist in the three major number since 1990, a period during which biomedical Ph.D. fields combined. Real growth in funding, from 2003 to graduates have increased strongly. In addition, and just as 2007, was 3.4 percent annually. If the growth rate stays at crucially, foreign-trained Ph.D.s are far more numerous in this level (or declines because of recession), funding growth the biomedical field than in the behavioral field and are also will be slower than the projected growth of the biomedical increasing. or clinical workforces, which will be more than 1 percentage The biomedical workforce could therefore be more point faster. strongly affected than the behavioral or clinical workforces How accurate projections of this sort can be, drawing by an interruption in immigrant flow. In the most extreme on data and trends that predate the economic crisis, it is not s ituation modeled, immigration would cease in 2010, possible to say. It may take a while for data to emerge that and none of the U.S. Ph.D. graduates who are temporary would permit more confident projections.