Research Output by Medical Doctors After PhD Graduation in Radiology: 17-Year Experience From the Netherlands

(1)

University of Groningen

Research Output by Medical Doctors After PhD Graduation in Radiology

Yakar, Derya; Kwee, Thomas C

Published in:

Academic Radiology DOI:

10.1016/j.acra.2020.03.043

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.

Document Version

Publisher's PDF, also known as Version of record

Publication date: 2021

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

Yakar, D., & Kwee, T. C. (2021). Research Output by Medical Doctors After PhD Graduation in Radiology: 17-Year Experience From the Netherlands. Academic Radiology, 28(6), 827-833.

https://doi.org/10.1016/j.acra.2020.03.043

Copyright

Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

(2)

Research Output by Medical Doctors

After PhD Graduation in Radiology:

17-Year Experience From the

Netherlands

Derya Yakar, MD, PhD, Thomas C. Kwee, MD, PhD

Rationale and Objectives: To determine the frequency and predictors of lack of research output by medical doctors after PhD graduation in radiology.

Materials and Methods: This study included all 272 PhD theses by medical doctors in the Netherlands between 2000 and 2016, with radi-ology as theﬁeld of research of the primary doctoral supervisor. Post-PhD research output was considered lacking if no original research, systematic review, or meta-analysis was published asﬁrst, second, or last author in the post-PhD period.

Results: The percentage of PhD graduates without research output was 41.9% (78/186) at 5 years, 28.6% (24/84) at 10 years, and 16.1% (5/31) at 15 years in the post-PhD period. On univariate Cox regression analysis, only female gender emerged as a signi_{ficant predictor of} a lack of research output in the post-PhD period (odds ratio: 1.456, 95% con_{fidence interval: 1.023 2.073, p = 0.038). PhD student’s age,} being a radiologist or not before PhD graduation, the radiologic subspecialty topic of the PhD thesis, the H-index of the primary doctoral supervisor, the institution at which the PhD was performed, and the number of publications in the PhD period, were not significantly asso-ciated with lack of post-PhD research output.

Conclusion: A considerable proportion of medical doctors has no active research output after obtaining a PhD degree in radiology, and this should be taken into account when relying on PhD programs to replenish the physician-scientist workforce. Females appear to be more prone to lack active research output, and this should be addressed by the scientiﬁc community and society.

Key Words: Publications; Physicians; Research; Science.

INTRODUCTION

P

hysician-scientists are individuals who have been trained in both clinical medicine and research (1). Their contribution to biomedical research is consid-ered vital for the advancement of healthcare (1), and this also applies to the ﬁeld of radiology. However, the number of physician-scientists has decreased over the years (1). In the United States, the proportion of physicians engaged in research has been reported to have dwindled from a peak of

4.7% of the overall physician workforce in the 1980s to approximately 1.5% in 2019 (1). Although it is unclear if a similar trend has occurred in radiology, it is a must train and retain the next generation of physician-scientists to advance theﬁeld of radiology.

Enrollment of medical doctors (either graduated or in training) into PhD programs may be a potential method to keep them actively engaged in research in the future. PhD programs with relevance to imaging are offered in the United States and many other countries, including several in Europe (2,3). In the Netherlands, all 8 university medical centers (which are afﬁliated to the 8 universities with a medical school in the country) offer PhD programs in radiology. These PhD programs generally take 3 4 years, and should lead to a PhD thesis that has to be approved by a scientiﬁc committee and publicly defended. These PhD programs can be started either during medical school or after MD gradua-tion, and they can also be completed during or after MD graduation. A PhD degree is generally considered as a proof of being able to independently perform research by society

Acad Radiol 2021; 28:827–833

From the Department of Radiology, Medical Imaging Center, University Medi-cal Center Groningen, University of Groningen, Groningen, The Netherlands (D.Y.); Department of Medical Oncology, University Medical Center Groningen, University of Groningen, Groningen, The Netherlands (T.C.K.). Received March 5, 2020; revised March 20, 2020; accepted March 24, 2020. Address correspondence to: T.C.K. e-mail:thomaskwee@gmail.com

(http://creativecommons.org/licenses/by/4.0/) https://doi.org/10.1016/j.acra.2020.03.043

(3)

and academic community, although it should be noted that current emphasis is on team science and not on independent work.

Although there is abundant literature on the importance and continued revitalization of academic radiology (4 10), it is currently unclear which medical doctors with a PhD degree in radiology are either engaged in or discontinue active research after PhD graduation. This information may be valu-able to gain insight into the overall contribution of PhD pro-grams in radiology to the physician-scientist workforce. It is also unclear if there are any factors that can predict a PhD graduate’s scientific activity in the post-PhD period. It can be hypothesized that variables such as age and gender of the PhD student, being a radiologist or not before PhD gradua-tion, the radiologic subspecialty topic of the PhD thesis, the productivity and scientific impact of the doctoral supervisor, the institution at which the PhD was performed, and the number of publications in the PhD period, may be of influ-ence. This information may potentially be helpful for the selection of PhD students and to address obstacles that impede their continued scientific involvement after PhD graduation.

The purpose of this study was therefore to determine the frequency and predictors of lack of research output by medi-cal doctors after PhD graduation in radiology.

MATERIALS AND METHODS

Study Design

This retrospective study used publicly available data and did not involve the investigation of human subjects requiring institutional review board approval. All PhD theses in the ﬁeld of radiology by medical doctors were extracted from the publicly available database of the Historical Committee of the Radiological Society of the Netherlands (11). PhD theses were included in this study if they were performed between January 2000 and December 2016 (which provided a follow-up time of at least 3 years), if theﬁeld of research of the pri-mary doctoral supervisor was radiology, and if the PhD grad-uate was a medical doctor. PhD theses by medical students were also eligible for inclusion.

Data Extraction

The following variables were extracted from all PhD theses that fulﬁlled the inclusion criteria: age and gender of the PhD student at the time of PhD graduation, whether or not the PhD student was a radiologist before PhD graduation, the radiologic subspecialty topic of the PhD thesis (breast imag-ing, cardiac imagimag-ing, contrast media, experimental studies, gastrointestinal imaging, genitourinary imaging, health care policy and quality, medical physics and technical develop-ments, musculoskeletal imaging, neuroradiology [including head and neck radiology], nuclear medicine and molecular imaging, thoracic imaging, vascular and interventional radiol-ogy, or other), the institution at which the PhD was

performed, and follow-up duration after PhD graduation until November 2019. Clarivate Analytics’s Web of Science (12) was used to determine the H-index of the primary doc-toral supervisor of each PhD thesis. MEDLINE/PubMed was used to determine the number of publications asfirst, second, or last author of each PhD student before PhD graduation (without starting date limitation) and in each year after PhD graduation, until November 2019. Only original research (primary research) and systematic reviews or meta-analyses (secondary research) were counted to determine this research output, because they are considered as medical research (13). Case reports or case series, editorials, letters, guidelines, non-systematic reviews, and survey studies were not counted (13). The time before and until 2 years after PhD graduation was defined as the PhD period, because it may take up to 2 years before articles written during the PhD program are published. The time beyond 2 years after PhD graduation was defined as the post-PhD period. Post-PhD research output was defined as either lacking (i.e., no publication as first, second, or last author in the post-PhD period) or present (i.e., at least one publication as first, second, or last author in the post-PhD period). Onlyfirst, second, and last authorship positions were considered, because they are generally occupied by the lead investigators in the Dutch academic system.

Statistical Analysis

The percentage of PhD graduates with a lack of research out-put as a proportion of all PhD graduates was calculated at 5, 10, and 15 years in the post-PhD period, provided follow-up time of each individual care was sufficiently long to be eligi-ble for inclusion in these calculations. Cox regression analyses (without using any predefined minimum follow-up time) were performed to determine the association of the age and gender of the PhD student, being a radiologist or not before PhD graduation, the radiologic subspecialty topic of the PhD thesis, the H-index of the primary doctoral supervisor, the university medical center at which the PhD was performed, and the number of publications in the PhD period, with the lack of post-PhD research output during follow-up.p-values less than 0.05 were regarded to indicate a statistically signifi-cant difference. Kaplan-Meier curves were created for those variables that emerged as significant in the Cox regression analyses. Statistical analyses were executed using MedCalc version 17.2 Software (MedCalc, Mariakerke, Belgium).

RESULTS

PhD Students and Theses

The database of the Historical Committee of the Radiological Society of the Netherlands contained 305 PhD theses that were performed between January 2000 and December 2016, with radiology as the ﬁeld of research of the primary doctoral supervisor. Of these 305 PhD theses, 33 were excluded because the PhD student was not a medical doctor or medical student.

(4)

(5)

The 272 PhD theses that wereﬁnally included, involved 160 men and 112 women, of whom were 270 medical doctors and 2 medical students, with a mean age§ standard deviation of 34.3§ 6.0 years (range: 26 64 years), at the time of PhD graduation. Eighty PhD students were radiologists at the time of PhD graduation, 183 were not, and of 9 PhD students it was unknown if they were a radiologist at the time of PhD graduation. Top three radiologic subspecialty topics of the 272 PhD theses were neuroradiology (n= 77), vascular and inter-ventional radiology (n= 44), and gastrointestinal imaging (n=

43), as shown inFigure 1. The 272 PhD theses were overseen by 57 different primary doctoral supervisors (of whom 39 were radiologists), with a median of 2 PhD theses (range: 1 36) per primary doctoral supervisor. The median H-index of these 57 primary doctoral supervisors was 46 (range: 1 208). The distri-bution of the 272 PhD theses among the 8 university medical centers is shown inFigure 2. The median number of publica-tions of each PhD student in the PhD period was 7 (range: 2 48). Median follow-up time after PhD graduation was 3341 days (range: 1031 7253 days).

Figure 2. Distribution of 272 PhD theses among eight university medical centers in the Netherlands. (Color version of_{ﬁgure is available online.)}

(6)

Lack of Research Output in the Post-PhD Period: Frequency and Predictors

The percentage of PhD graduates with a lack of research out-put as a proportion of all PhD graduates with available fol-low-up was 41.9% (78/186) at 5 years, 28.6% (24/84) at 10 years, and 16.1% (5/31) at 15 years in the post-PhD period. On univariate Cox regression analysis, only female gender emerged as a significant predictor of a lack of research output in the post-PhD period (odds ratio: 1.456, 95% confi-dence interval: 1.023 2.073,p = 0.038). The corresponding Kaplan-Meier curve is shown inFigure 3. All other variables (age of the PhD student, being a radiologist or not before PhD graduation, the radiologic subspecialty topic of the PhD thesis (although breast imaging seems to be trending to signif-icance at p= 0.070), the H-index of the primary doctoral supervisor, the institution at which the PhD was performed, and the number of publications in the PhD period) were not significantly associated with lack of post-PhD research out-put, as shown inTable 1.

DISCUSSION

The Dutch government attaches great importance to PhD graduates in all scientific disciplines, since they are considered to contribute to the development and innovation of science, and to shape the intellectual elite in society (14). Moreover, some professional branches have a direct need for employees with a PhD graduation (14). For each successfully defended PhD thesis, the university receives a lump sum from the Dutch government, which was fixed at €93,000 between 2009 and 2017. The number of PhD theses in the field of health sciences has increased considerably in the Nether-lands over the past years; from 668 for the period 2000/ 2001 to 1679 for the period 2015/2016, corresponding to a 250% increase (15). Despite the attributed importance, financial government support, and increasing number of PhD theses, the contribution of PhD graduates to the cli-nician-scientist workforce in radiology has remained completely unclear so far.

Figure 3. Kaplan-Meier curve for lack of research output for female versus male PhD graduates in the post-PhD period. Lack of research output in the post-PhD period was signi_{ﬁcantly higher for female PhD graduates compared to male PhD graduates (log-rank} test,_{p = 0.037). (Color version of ﬁgure is available online.)}

(7)

The results of the present study show that a considerable proportion of medical doctors with a PhD degree in radiol-ogy does not actively generate research output in their post-PhD period. In addition, active post-post-PhD research output is more frequently lacking among women than among men. These results indicate that enrollment and successful comple-tion of a PhD program in radiology does not necessarily con-tribute to the future physician-scientist generation. They also show that there are apparently some intrinsic or extrinsic bar-riers that, to some degree, hinder women to continue an active scientiﬁc career after PhD graduation. This study and itsﬁndings are rather unique because of its access to PhD data and have, to the best of our knowledge, not been previously reported.

In 2014, a questionnaire among 16,000 Dutch PhD gradu-ates in various disciplines was performed, and the results were qualitatively described (15). Interestingly, it was reported that PhD graduates in the health sciences less frequently employ the skills that they learned during their PhD program in their daily post-PhD work, compared to PhD graduates in other scientiﬁc disciplines (i.e., natural, technical, agricultural, behavioral and social, and language and cultural sciences) (15). In addition, the relationship between PhD contents and post-PhD work was reported to be relatively weak for the health sciences (15). This was explained by the fact the inter-viewed medical doctors frequently indicated a PhD degree to

be a means or a condition to get enrolled into a residency program (15). In other words, the PhD degree is used to become a medical specialist, and not to perform research. This may well explain the observed numbers of medical doc-tors with a PhD degree in radiology but without any active research output in the post-PhD period.

The availability of objectively measurable predictors of sus-tained post-PhD research involvement would be useful to select medical doctors for PhD programs in radiology, or to address obstacles that impede their continued scientiﬁc involvement after PhD graduation. Although several PhD-related variables were investigated, only female gender emerged as a signiﬁcant predictor of lack of active research output in the post-PhD period. This is in line with previous research that also reported lower research productivity by female academic radiologists (16 19) and academic physi-cians in general (20), although some studies also indicate a reversing trend (16,17,19). Lower research productivity of female radiologists compared to their male counterparts can be explained by factors such as concerns by women that it will be impossible to combine a successful career with child-bearing and family life, people expecting women to assume the major responsibilities of caring for children and running the household, women feeling less comfortable promoting themselves and their work than their male counterparts, lack of encouragement to women to become physician-scientists,

TABLE 1. Univariate Cox Regression Analysis on the Association of PhD Variables and Lack of Post-PhD Research Output

Variable Odds Ratio 95% CIa _p-Value

Age PhD student 1.006 0.976 1.037 0.700

Gender PhD student (female vs. male) 1.456 1.023 2.073 0.038 Radiologist before PhD graduation (yes vs. no) 0.959 0.656 1.404 0.831 Subspecialty topic PhD thesisb,c

-Breast imaging 1.856 0.954 3.608 0.070 -Cardiac imaging 1.368 0.732 2.558 0.329 -Gastrointestinal imaging 1.049 0.606 1.817 0.864 -Genitourinary imaging 1.198 0.469 3.065 0.707 -Musculoskeletal imaging 0.913 0.383 2.176 0.838 -Thoracic imaging 1.193 0.551 2.583 0.656

-Vascular and interventional radiology 0.907 0.519 1.585 0.733

H-factor doctoral supervisor 1.003 0.995 1.011 0.503

University medical centerd

-Academic Medical Center Amsterdam 1.252 0.739 2.121 0.406 -Free University Medical Center Amsterdam 0.931 0.461 1.882 0.844 -University Medical Center Groningen 1.486 0.647 3.413 0.353 -Leiden University Medical Center 0.809 0.452 1.448 0.478 -Maastricht University Medical Center 0.643 0.318 1.300 0.221 -Radboud University Medical Center Nijmegen 0.910 0.396 2.089 0.825 -Erasmus Medical Center Rotterdam 0.659 0.332 1.306 0.234 No. of publications in the PhD period 0.976 0.931 1.023 0.315 Notes: Gender PhD student (female vs. male) was signiﬁcant, with a p-Value of 0.038.

a Conﬁdence interval. b Relative to neuroradiology.

c Because of too low numbers, the subspecialty topics“experimental studies,” “health care policy and quality,” “medical physics and techni-cal developments,” “nuclear medicine and molecular imaging,” and “other” were not entered in the Cox regression analysis.

d Relative to the University Medical Center Utrecht.

(8)

and lack of compelling role models (21,22). Having more women role models may encourage more women to remain active as physician-scientists, but this would require more women to be recruited for key leadership positions (21,22). One way of reaching this goal may be by implementing gen-der quota, which have already been implemented by several companies and universities in the Western world. The results of the present study further underline the necessity of such policies to close this gender gap in radiology.

This study had some limitations. First, research output was only counted forfirst, second, and last authorships. Although it can be argued that other authorship positions may also have required efforts (with current emphasis on team science), for a PhD thesis onlyfirst, second, and last authorship scientific papers can generally be used as proof of significant research contribution, and this definitely applies to the Dutch system. Hence the decision to extrapolate this requirement to the definition of research output in this study. Second, although the predictive value of several PhD-related variables on post-PhD research output was calculated, further prospective research is necessary to get a better grip on how to select medical graduates for a PhD program in radiology. For exam-ple, although it is plausible that several PhD students merely use the PhD degree to become a medical specialist rather than acquiring skills to perform future research, there are cur-rently no available means to objectively assess the intrinsic research motivation of an individual who applies for a PhD program.

In conclusion, a considerable proportion of medical doc-tors has no active research output after obtaining a PhD degree in radiology, and this should be taken into account when relying on PhD programs to replenish the physician-scientist workforce. Females appear to be more prone to lack active research output, and this should be addressed by the scientiﬁc community and society.

FUNDING

This research did not receive any speciﬁc grant from funding agencies in the public, commercial, or not-for-proﬁt sectors.

IRB STATEMENT

- This retrospective study used publicly available data and did not involve the investigation of human subjects requiring institutional review board approval.

-This study has been performed in accordance with the ethical standards in the 1964 Declaration of Helsinki.

-This study has been carried out in accordance with rele-vant regulations of the US Health Insurance Portability and Accountability Act (HIPAA).

ABBREVIATIONS None.

ACKNOWLEDGMENTS None.

REFERENCES

1. Jain MK, Cheung VG, Utz PJ, et al. Saving the endangered physician-scientist - a plan for accelerating medical breakthroughs. N Engl J Med 2019; 381:399–402.

2. Rahal A, Head HW, Jung AJ, et al. Combined radiology residency/PhD program for education of academic radiologists: a response to revitaliz-ing the radiology research enterprise. Radiology 2007; 245:14–20. 3. European Society of Radiology (ESR). MD PhD programmes with

rele-vance to imaging. Results from a European survey. Insights Imaging 2013; 4:409–415.

4. Hillman BJ, Fajardo LL, Witzke DB, et al. Factors inﬂuencing radiologists to choose research careers. Invest Radiol 1989; 24:842–848.

5. Bilbey JH, Fache JS, Burhenne HJ. Are there predictors for future aca-demic radiologists? A Canadian survey. Can Assoc Radiol J 1992; 43:369–373.

6. Collins J. Academic opportunities in radiology education and education research. Acad Radiol 2002; 9:753–755.

7. Feng L, Ruzal-Shapiro C. Factors that inﬂuence radiologists' career choices. Acad Radiol 2003; 10:45–51.

8. Hillman BJ, Gay SB. Research residents' choices of academic radiology careers. J Am Coll Radiol 2005; 2:1030–1032.

9. Jain SN. Pursuing a radiology career: private practice or academic? Indian J Radiol Imaging 2014; 24:207–208.

10. Grimm LJ, Shapiro LM, Singhapricha T, et al. Predictors of an academic career on radiology residency applications. Acad Radiol 2014; 21:685– 690.

11. Historical Committee of the Radiological Society of the Netherlands. Available at: https://www.radhis.nl/links.html. Accessed December 8, 2019.

12. Clarivate Analytics’s Web of Science. Available at:https://jcr.clarivate. com/. Accessed December 8, 2019.

13. R€ohrig B, du Prel JB, Wachtlin D, et al. Types of study in medical research: part 3 of a series on evaluation of scientiﬁc publications. Dtsch Arztebl Int 2009; 106:262–268.

14. Koier E, De Jonge J. De zin van promoveren Loopbanen en arbeids-marktperspectieven van gepromoveerden. Den Haag: Rathenau Instituut 2018: 1–51.

15. Statistics Netherlands. Available at: https://opendata.cbs.nl/statline/ #/CBS/nl/dataset/71247ned/table?fromstatweb. Accessed December 8, 2019.

16. Khurshid K, Shah S, Ahmadi M, et al. Gender differences in the publica-tion rate among breast imaging radiologists in the United States and Canada. AJR Am J Roentgenol 2018; 210:2–7.

17. Yun EJ, Yoon DY, Kim B, et al. Closing the gender gap: increased female authorship in AJR and Radiology. AJR Am J Roentgenol 2015; 205:237– 241.

18. McDonald JS, McDonald RJ, Davenport MS, et al. Gender and radiology publication productivity: an examination of academic faculty from Four health systems in the United States. J Am Coll Radiol 2017; 14:1100– 1108.

19. Piper CL, Scheel JR, Lee CI, et al. Gender trends in radiology authorship: A 35-year analysis. AJR Am J Roentgenol 2016; 206:3–7.

20. Eloy JA, Svider PF, Cherla DV, et al. Gender disparities in research pro-ductivity among 9952 academic physicians. Laryngoscope 2013; 123:1865–1875.

21. Andrews NC. The other physician-scientist problem: where have all the young girls gone? Nat Med 2002; 8:439–441.

22. Anzai Y, Meltzer CC, DeStigter KK, et al. Entrepreneurial women in radi-ology: role models of success. J Am Coll Radiol 2016; 13:1378–1382.