Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Scientific publications in internal medicine and family medicine: a comparative cross-sectional study in Swiss university hospitals

Scientific publications in internal medicine and family medicine: a comparative cross-sectional... Abstract Background Family medicine is a relatively new academic medical discipline. We aimed to compare the main bibliometric indices of hospital-based senior physicians practicing internal medicine versus family medicine in Switzerland. Methods We conducted this cross-sectional study in March 2020. We selected all hospital-based senior physicians practicing internal medicine or family medicine in the six Swiss university hospitals. Using Web of Science, after removing from both groups of physicians the 5% with the highest number of publications, we extracted the number of publications, the number of publications per year, the number of citations, the number of citations per year, the number of citations per publication and the h-index. We compared the data between the two groups using negative binomial regressions and the proportion of physicians having at least one publication using chi-square tests. Results We included 349 physicians in the study (internal medicine: 51%, men: 51%). The median number of publications was three [interquartile range (IQR) = 18], the median number of citations was nine (IQR = 158) and the median h-index was one (IQR = 5). All bibliometric indices were similar in both groups, as was the proportion of physicians having at least one publication (family medicine: 87% versus 82%, P = 0.15). Conclusions We found no association between the bibliometric indices and the medical specialty. Further studies are needed to explore other important indicators of academic output, such as those more specifically assessing its quality and scientific importance. Bibliometric study, citations, family medicine, h-index, publications, productivity Key Messages Family medicine is a relatively new academic medical discipline. The main bibliometric indices are the number of publications and citations and h-index. No studies compared bibliometric indices in family medicine and internal medicine. We compared these indices for all senior physicians in Swiss university hospitals. We found no association between physicians’ indices and the medical specialty. Introduction Publication of scientific output is crucial because it allows to spread scientific knowledge (1,2) and increase the recognition of researchers (3–5). The productivity of researchers, which represents the visible result of scientific activities, has become of paramount importance to assess their global performance, including in medicine (6,7). It is also used to explore various issues linked to research policy, such as the role of gender, cross-institutional collaboration or funding on publication trajectories (8–11). The productivity of researchers can be easily measured using bibliometric methods. These methods allow for a quantitative analysis of scientific publications. Despite its known limitations, they are increasingly being used in the field of library science to determine the impact of scientific publications and to explore the productivity of researchers in various scientific fields (12). In addition to the number of publications and citations, the h-index has become popular for measuring researcher productivity because it integrates two major parameters: the productivity of researchers and the impact of their publications in terms of citations (13). Indeed, the number of publications does not necessarily reflect the quality of scientific publications and the number of citations may be affected by a few influential publications (13). The h-index was first proposed by Jorge Hirsch in 2005 to determine the scientific quality of articles published by theoretical physicists and is now used in many scientific fields, including medicine. This index is defined as the number (h) of publications that were each cited at least h times (13). Bibliometric indexes can be determined using citation databases, such as Web of Science, Scopus and Google Scholar. For a given author, they may vary depending on the database chosen due to different coverage. However, Web of Science and Scopus tend to give relatively similar results (14,15). Very few bibliometric data are available concerning our country. According to a report by the State Secretariat for Education, Research and Innovation (SERI) published in 2016, for the period 2009–13, Switzerland produced 1.2% of the world’s scientific publications (all scientific fields combined), ranking it 16th (12). However, in terms of the number of publications per capita, Switzerland was the world leader with nearly 4000 publications per million inhabitants. In terms of the number of publications per researcher, Switzerland was in third place, with 857 publications per 1000 researchers. The impact of publications produced in Switzerland was also excellent, with Switzerland in third place, just behind the USA and the Netherlands. Compared to other academic medical disciplines, family medicine is a relatively new discipline (16–18). The number of publications in this discipline has increased in recent decades, largely due to the creation of university academies of family medicine in many countries (5,19–21). This development is encouraging because it is often through academic research that we see the successful development of a medical discipline (21). Building research capacity in family medicine is essential to provide better care to patients. Indeed, although family physicians benefit from hospital-based research in internal medicine, most of these studies do not apply to patients evaluated in primary care settings. Conducting studies in primary care avoids such biases and is more prone to provide useful and generalizable data to primary care physicians (21). Academic research in internal medicine is usually recognized for its high quality. The above bibliometric indicators would make it possible to compare publications in family medicine versus internal medicine and, therefore, to assess the current state of family medicine research in a country like Switzerland. The objective of our study was to compare the productivity of hospital-based senior physicians practicing internal medicine versus family medicine in Switzerland. We hypothesize higher bibliometric indices among researchers practicing internal medicine. Methods University hospitals in Switzerland There are six university hospitals in Switzerland. Half of these hospitals are located in predominantly French-speaking cantons (Western Switzerland: Geneva, Lausanne and Fribourg) and the other half in predominantly German-speaking cantons (Eastern Switzerland: Bern, Basel and Zurich). There is no university hospital in Italian-speaking Switzerland. General medicine (also known as family medicine) and internal medicine merged in 2011 into a single discipline: general internal medicine. Despite this fusion, Swiss hospitals have separate departments of family medicine (providing outpatient care) and internal medicine (providing inpatient care). Hospital-based senior physicians include those who are in long-term employment (with a few exceptions, they pursue an academic career) and those who have completed their training and generally work for a few more years as senior physicians (=senior registrars) before starting a sub-specialization or working in a private practice. After a few years spent as senior registrars, physicians may be promoted to staff physicians, mostly on the basis of their academic output. Study site and study population We conducted our bibliometric study from the first to 14 March 2020 in the six Swiss university hospitals (all university hospitals were included in the study). We included all senior physicians (heads of division, staff physicians and senior registrars) practicing internal medicine or family medicine in one of these six university hospitals. We retrieved the list of these physicians from the hospitals’ websites. As planned, we excluded for both medical specialties the 5% of physicians with the highest number of publications in order to obtain two relatively homogeneous study samples and avoid biases that could result from the inclusion of some extreme outliers. Some senior physicians working in non-university hospitals also do research. However, as in most countries, the majority of publications are authored by physicians working in university hospitals. This study was intended to cover academic output, that is, the publications by physicians working in university hospitals only. Data collection First, using the hospitals’ websites, we recorded physicians’ socio-demographic characteristics [gender, medical department, professor (Y/N), position (head of division, staff physician, senior registrar and other (consulting physician, senior research physician and senior education physician))]. These websites are very reliable in Switzerland, regularly updated and accurately list all the physicians working in the various hospital departments. We consulted them a few days before collecting the bibliometric data. Then, using the citation report function of Web of Science, a platform consisting of several literature search databases, we extracted physicians’ bibliometric data (number of publications, year of publication of the first article, number of citations and h-index). We obtained the list of publications of each selected physician for all years up to the survey (there were no limitations in the time period covered by the study) by using the following search criteria: ‘all databases’ and ‘search by author, select from index’ and by entering the surname and first name in the search box. For physicians with two surnames, we entered the two surnames separated by a space, then by a hyphen and, finally, we entered each of the two surnames. For physicians with two first names, we entered only the first one. Two investigators from the research team (PS and NV) extracted all bibliometric data. They performed the data extraction in duplicate to exclude any errors. Before extracting the bibliometric data, they removed publications authored by homonyms. They resolved doubts about certain publications through a discussion within the research team. The bibliometric data were extracted for any publication indexed in Web of Science. In this platform, different types of literature are indexed in addition to original research articles, such as non-research articles (e.g. editorials, clinical updates and non-systematic reviews) and books. Among the various bibliometric indices available, we voluntarily included both ‘productivity’ (number of publications) and ‘quality’ variables (number of citations and h-index). High-quality research articles will tend to be cited more often and, therefore, counted more often in ‘quality’ variables than other publications. However, as Web of Science coverage is not limited to research articles, we used the term ‘publications’ to summarize the scientific productivity of researchers. Statistical analyses and sample size For the purpose of the study, we built three additional bibliometric variables. We computed ‘the number of citations per year and the number of citations per year’ by dividing the number of publications, respectively, the number of citations, by the number of years since the publication of the first article. In addition, we computed ‘the number of citations per publication’ by dividing the number of citations by the number of publications. We rounded these three variables to the nearest unit. We used frequency tables to describe physicians’ socio-demographic characteristics, and median and interquartile range (IQR) to summarize the number of years since they first published and the bibliometric data. We used chi-square tests to compare the socio-demographic characteristics and the proportion of physicians having at least one publication between the two groups of physicians (internal medicine and family medicine), the Wilcoxon rank-sum test to compare their first publications and univariate negative binomial regressions to compare their other bibliometric data (= count data with overdispersion) (22,23). Finally, we used Spearman’s rank correlation coefficients to determine the correlation between physicians’ bibliometric indices and the number of years since first publication. We considered correlations lower than 0.30 as ‘small’, between 0.30 and 0.50 as ‘medium’ and higher than 0.50 as ‘large’ (24). As planned, we excluded from both groups of physicians the 5% with the highest number of publications. However, we repeated the analyses by adding these 5% of physicians to evaluate whether or not the results comparing the bibliometric indices of physicians in internal medicine versus family medicine were modified (sensitivity analysis). We used the sample size calculation for comparing two negative binomial rates as developed by Zhu and Lakkis (25). We estimated that, by comparing two groups (physicians practising internal medicine versus family medicine), a sample of 336 (168 in each group) would be sufficient to detect a difference of five publications with Type I and II error both set at 5%. We expected the average number of publications to be 10 in internal medicine and 5 in family medicine, the ratio of the number of participants to be 1 and the dispersion parameter to be 5. We carried out all statistical analyses with STATA version 15.1 (College Station, TX, USA). Results We included 349 hospital-based senior physicians in the study, of whom 178 (51%) were affiliated with a department of internal medicine and 171 (49%) with a department of family medicine (Table 1). Two-thirds of the physicians were located in French-speaking Switzerland. There were slightly more men than women (51% versus 49%). One tenth of the physicians had the title of professor and about half of the physicians were senior registrars. There was no statistically significant difference between the two groups of physicians, except for the hierarchical position (more senior registrars in internal medicine than in family medicine). Of the 349 physicians included in the study, 294 (84%) had at least one publication, with no difference between the two groups (82% in internal medicine versus 87% in family medicine; P = 0.15). There was also no difference in the number of years since they first published [median = 8 (IQR = 11) in internal medicine versus 9 (IQR = 11) in family medicine; P = 0.35]. Table 1. Physicians’ socio-demographic characteristics, overall and stratified by medical speciality (n = 349) Characteristics . Total (n = 349) n (%) . Internal medicine (n = 178) n (%) . Family medicine (n = 171) n (%) . P-valuea . Location 0.28  French-speaking Switzerland 229 (65.6) 112 (62.9) 117 (68.4)  German-speaking Switzerland 120 (34.4) 66 (37.1) 54 (31.6) Gender 0.56  Male 179 (51.3) 94 (52.8) 85 (49.7)  Female 170 (48.7) 84 (47.2) 86 (50.3) Professor 0.09  Yes 34 (9.7) 22 (12.4) 12 (7.0)  No 315 (90.3) 156 (87.6) 159 (93.0) Hierarchical position <0.001  Heads of division 11 (3.2) 5 (2.8) 6 (3.5)  Staff physicians 85 (24.4) 41 (23.0) 44 (25.7)  Senior registrars 185 (53.0) 122 (68.6) 63 (36.9)  Otherb 68 (19.5) 10 (5.6) 58 (33.9) Characteristics . Total (n = 349) n (%) . Internal medicine (n = 178) n (%) . Family medicine (n = 171) n (%) . P-valuea . Location 0.28  French-speaking Switzerland 229 (65.6) 112 (62.9) 117 (68.4)  German-speaking Switzerland 120 (34.4) 66 (37.1) 54 (31.6) Gender 0.56  Male 179 (51.3) 94 (52.8) 85 (49.7)  Female 170 (48.7) 84 (47.2) 86 (50.3) Professor 0.09  Yes 34 (9.7) 22 (12.4) 12 (7.0)  No 315 (90.3) 156 (87.6) 159 (93.0) Hierarchical position <0.001  Heads of division 11 (3.2) 5 (2.8) 6 (3.5)  Staff physicians 85 (24.4) 41 (23.0) 44 (25.7)  Senior registrars 185 (53.0) 122 (68.6) 63 (36.9)  Otherb 68 (19.5) 10 (5.6) 58 (33.9) aχ2 tests. bConsulting physicians, senior research physicians and senior education physicians. Open in new tab Table 1. Physicians’ socio-demographic characteristics, overall and stratified by medical speciality (n = 349) Characteristics . Total (n = 349) n (%) . Internal medicine (n = 178) n (%) . Family medicine (n = 171) n (%) . P-valuea . Location 0.28  French-speaking Switzerland 229 (65.6) 112 (62.9) 117 (68.4)  German-speaking Switzerland 120 (34.4) 66 (37.1) 54 (31.6) Gender 0.56  Male 179 (51.3) 94 (52.8) 85 (49.7)  Female 170 (48.7) 84 (47.2) 86 (50.3) Professor 0.09  Yes 34 (9.7) 22 (12.4) 12 (7.0)  No 315 (90.3) 156 (87.6) 159 (93.0) Hierarchical position <0.001  Heads of division 11 (3.2) 5 (2.8) 6 (3.5)  Staff physicians 85 (24.4) 41 (23.0) 44 (25.7)  Senior registrars 185 (53.0) 122 (68.6) 63 (36.9)  Otherb 68 (19.5) 10 (5.6) 58 (33.9) Characteristics . Total (n = 349) n (%) . Internal medicine (n = 178) n (%) . Family medicine (n = 171) n (%) . P-valuea . Location 0.28  French-speaking Switzerland 229 (65.6) 112 (62.9) 117 (68.4)  German-speaking Switzerland 120 (34.4) 66 (37.1) 54 (31.6) Gender 0.56  Male 179 (51.3) 94 (52.8) 85 (49.7)  Female 170 (48.7) 84 (47.2) 86 (50.3) Professor 0.09  Yes 34 (9.7) 22 (12.4) 12 (7.0)  No 315 (90.3) 156 (87.6) 159 (93.0) Hierarchical position <0.001  Heads of division 11 (3.2) 5 (2.8) 6 (3.5)  Staff physicians 85 (24.4) 41 (23.0) 44 (25.7)  Senior registrars 185 (53.0) 122 (68.6) 63 (36.9)  Otherb 68 (19.5) 10 (5.6) 58 (33.9) aχ2 tests. bConsulting physicians, senior research physicians and senior education physicians. Open in new tab Table 2 summarizes the bibliometric data and shows the association between these data and the medical specialty. For ease of data interpretation, Figures 1–3 present in graphical form (scatter plot) the number of publications, the number of citations and the h-index by the number of years since they first published and the medical specialty. The median number of publications and publications per year was 3 (IQR = 18) and 1 (IQR = 2), respectively, while the median number of citations and citations per year was 9 (IQR = 158) and 1 (IQR = 13), respectively. On average, two authors cited each publication. The median h-index was 1 (IQR = 5). There was no association between physicians’ bibliometric indices and their medical specialty. Finally, all correlations examined between physicians’ bibliometric indices and the number of years since they first published were high (number of publications: rho = 0.67, number of citations: rho = 0.78, h-index: rho = 0.75) and the relationships were statistically significant (all P values <0.005). Table 2. Associations between medical specialty and number of publications, number of citations and h-index (list of publications extracted for all years up to March 2020) Variable . Total (n = 349) Median (IQR) . Internal medicine (n = 178) Median (IQR) . Family medicine (n = 171) Median (IQR) . IRR (95% CI) . Crude P-valuea . Publications  Number of publications 3 (18) 2 (12) 5 (22) 1.0 (0.7–1.3) 0.77  Number of publications per year 1 (2) 1 (2) 1 (2) 0.8 (0.6–1.1) 0.23 Citations  Number of citations 9 (158) 7 (165) 12 (148) 1.6 (0.9–2.8) 0.11  Number of citations per year 1 (13) 1 (15) 2 (11) 1.5 (0.9–2.4) 0.12  Number of citations per publication 2 (11) 3 (12) 2 (11) 1.4 (0.9–2.1) 0.11 Publications and citations  h-index 1 (5) 1 (5) 1 (5) 1.1 (0.8–1.5) 0.59 Variable . Total (n = 349) Median (IQR) . Internal medicine (n = 178) Median (IQR) . Family medicine (n = 171) Median (IQR) . IRR (95% CI) . Crude P-valuea . Publications  Number of publications 3 (18) 2 (12) 5 (22) 1.0 (0.7–1.3) 0.77  Number of publications per year 1 (2) 1 (2) 1 (2) 0.8 (0.6–1.1) 0.23 Citations  Number of citations 9 (158) 7 (165) 12 (148) 1.6 (0.9–2.8) 0.11  Number of citations per year 1 (13) 1 (15) 2 (11) 1.5 (0.9–2.4) 0.12  Number of citations per publication 2 (11) 3 (12) 2 (11) 1.4 (0.9–2.1) 0.11 Publications and citations  h-index 1 (5) 1 (5) 1 (5) 1.1 (0.8–1.5) 0.59 aUnivariate negative binomial regression. Open in new tab Table 2. Associations between medical specialty and number of publications, number of citations and h-index (list of publications extracted for all years up to March 2020) Variable . Total (n = 349) Median (IQR) . Internal medicine (n = 178) Median (IQR) . Family medicine (n = 171) Median (IQR) . IRR (95% CI) . Crude P-valuea . Publications  Number of publications 3 (18) 2 (12) 5 (22) 1.0 (0.7–1.3) 0.77  Number of publications per year 1 (2) 1 (2) 1 (2) 0.8 (0.6–1.1) 0.23 Citations  Number of citations 9 (158) 7 (165) 12 (148) 1.6 (0.9–2.8) 0.11  Number of citations per year 1 (13) 1 (15) 2 (11) 1.5 (0.9–2.4) 0.12  Number of citations per publication 2 (11) 3 (12) 2 (11) 1.4 (0.9–2.1) 0.11 Publications and citations  h-index 1 (5) 1 (5) 1 (5) 1.1 (0.8–1.5) 0.59 Variable . Total (n = 349) Median (IQR) . Internal medicine (n = 178) Median (IQR) . Family medicine (n = 171) Median (IQR) . IRR (95% CI) . Crude P-valuea . Publications  Number of publications 3 (18) 2 (12) 5 (22) 1.0 (0.7–1.3) 0.77  Number of publications per year 1 (2) 1 (2) 1 (2) 0.8 (0.6–1.1) 0.23 Citations  Number of citations 9 (158) 7 (165) 12 (148) 1.6 (0.9–2.8) 0.11  Number of citations per year 1 (13) 1 (15) 2 (11) 1.5 (0.9–2.4) 0.12  Number of citations per publication 2 (11) 3 (12) 2 (11) 1.4 (0.9–2.1) 0.11 Publications and citations  h-index 1 (5) 1 (5) 1 (5) 1.1 (0.8–1.5) 0.59 aUnivariate negative binomial regression. Open in new tab Figure 1. Open in new tabDownload slide Physicians’ number of publications by number of years since first publication (list of publications extracted for all years up to March 2020; data are shown separately for internal medicine and family medicine). Figure 1. Open in new tabDownload slide Physicians’ number of publications by number of years since first publication (list of publications extracted for all years up to March 2020; data are shown separately for internal medicine and family medicine). Figure 2. Open in new tabDownload slide Physicians’ number of citations by number of years since first publication (list of publications extracted for all years up to March 2020; data are shown separately for internal medicine and family medicine). Figure 2. Open in new tabDownload slide Physicians’ number of citations by number of years since first publication (list of publications extracted for all years up to March 2020; data are shown separately for internal medicine and family medicine). Figure 3. Open in new tabDownload slide Physicians’ h-index by number of years since first publication (list of publications extracted for all years up to March 2020; data are shown separately for internal medicine and family medicine). Figure 3. Open in new tabDownload slide Physicians’ h-index by number of years since first publication (list of publications extracted for all years up to March 2020; data are shown separately for internal medicine and family medicine). As planned, we excluded from both groups of physicians the 5% with the highest number of publications. However, we repeated the analyses including all physicians. These analyses confirmed the absence of statistically significant differences for all bibliometric indicators listed in Table 2, except for the number of citations, which was on average twice as high in internal medicine as in family medicine [IRR = 2.1 (95% confidence interval (CI) = 1.2–3.8), P = 0.01]. The association remained statistically significant when adjusting for the number of years since first publication, location, gender and hierarchical position [IRR = 2.3 (95% CI = 1.5–3.4), P < 0.001]. This association is related to the presence of a few outliers in the internal medicine group (physicians with a very high number of citations). Discussion Main findings In summary, in this sample of hospital-based senior physicians, we found that the median number of publications was 3, the median number of citations was 9 and the median h-index was 1. There was no statistically significant difference between physicians’ bibliometric indices and the medical specialty. We also found that these bibliometric indices were highly correlated with the number of years since they first published. Comparison with existing literature To our knowledge, this study is the first to have compared the productivity of hospital-based physicians in internal medicine with those in family medicine. The absence of statistically significant differences in the various bibliometric indices considered (number of publications, number of publications per year, number of citations, number of citations per year, number of citations per publication and h-index) can probably be explained by the rapid development of academic research in family medicine these last years. As family medicine has become an academic discipline very gradually over the last few decades, the results obtained in this study could have been even more favourable to family medicine if we had restricted the study to the last years of publications. However, many questions remain unanswered in this field. Among these questions, it is unclear whether research in family medicine adequately meets the needs of community-based family physicians. In addition, our study focused mainly on the quantitative evaluation of scientific research. It is, of course, important to evaluate both the quantity and quality of research. The quality of research could in theory be measured using the number of citations and h-index, but these indicators are imperfect. Further studies are, therefore, needed to explore the quality and scientific importance of family medicine research by using or developing accurate indicators. One way of doing this could be to use the Grading of Recommendations Assessment, Development and Evaluation approach (26,27). Many of the recommendations for clinical practice are based on summaries of evidence as determined by this tool. Two caveats should be mentioned concerning the citation database and the bibliometric variables used in this study. First, the indices were calculated using the platform Web of Science. We believe that, although the results might vary slightly depending on the citation database used (mainly because there are differences in the journals and types of literature indexed), those concerning the comparison between internal medicine and family medicine should not be greatly modified. Indeed, comparisons of indices are probably less subject to variation than indices themselves. Second, the bibliometric variables used in our study could be subject to criticism. For example, they do not take into account the author’s position in the list of authors and some of them (number of citations and h-index) can be manipulated using self-citations. In addition, when comparing different disciplines with these bibliometric variables, they do not adjust for the number of researchers in the field and some other factors, such as the number of high-impact journals and access to research funding. However, although this lack of adjustment probably leads to an underestimation of the real productivity of family medicine researchers, we found similar bibliometric results in both groups of physicians. If the number of high-impact journals were similar in both disciplines, the results would probably have been even better for physicians in family medicine. Limitations Besides these two caveats, some limitations need to be pointed out in our study. First, the study sample consisted only of physicians practicing in Swiss university hospitals. The results could possibly not be generalized to other physicians in Switzerland or in other countries. The overrepresentation in our study of physicians practicing in French-speaking university hospitals may seem surprising since only about a quarter of the inhabitants live in French-speaking Switzerland. This finding is explained by the fact that the ratio between university and non-university hospitals in the different regions of Switzerland is not related to the size of the population living in each of them. Second, we cannot completely exclude certain errors in the publication lists, even though we standardized the search procedures, duplicated all searches and reviewed all publications of each physician to remove those published by homonyms. In addition, choosing to retrieve the publication lists using physicians’ full first name also reduced the risk of homonymy, but this strategy might have missed some publications that were referenced without the full first name. However, there is no reason to imagine that errors in the establishment of the publication lists could affect the two groups of physicians differently. Third, we did not know the age or date of birth of the physicians included in the study. Yet there was no significant difference between the average number of years since first publication (a variable that could be considered a surrogate for age) and the medical field. Finally, our aim was not to measure the methodological quality, scientific importance or novelty of the publications included in the study. These indicators could be used in future comparative studies. Conclusion In conclusion, in this sample of Swiss hospital-based senior physicians, we found no association between the various bibliometric indices and the medical field (internal medicine versus family medicine). Further studies are needed in the future to confirm that the quality and impact of research in family medicine and in internal medicine are similar and that family medicine has definitely become a very attractive academic field. Declarations Funding: none. Ethical approval: not required (under Swiss law, ethical approval is not required when collecting non-personal health data). Conflict of interest: none. Data availability: the data underlying this article will be shared on reasonable request to the corresponding author. References 1. Mabe MA . Scholarly communication: a long view . New Rev Acad Librarianship 2010 ; 16 : 132 – 44 . Google Scholar Crossref Search ADS WorldCat 2. Lafrenière D , Menuz V, Hurlimann T, Godard B. Knowledge dissemination interventions: a literature review . SAGE Open 2013 ; 3 : 2158244013498242 . Google Scholar Crossref Search ADS WorldCat 3. Vale RD . Accelerating scientific publication in biology . Proc Natl Acad Sci USA 2015 ; 112 ( 44): 13439 – 46 . Google Scholar Crossref Search ADS PubMed WorldCat 4. Bavdekar SB , Tullu MS. Research publications for academic career advancement: An idea whose time has come. But is this the right way? J Postgrad Med 2016 ; 62 ( 1): 1 – 3 . Google Scholar Crossref Search ADS PubMed WorldCat 5. Post RE , Weese TJ, Mainous AG III, Weiss BD. Publication productivity by family medicine faculty: 1999 to 2009 . Fam Med 2012 ; 44 ( 5 ): 312 – 7 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 6. Pfeiffer M , Fischer MR, Bauer D. Publication activities of German junior researchers in academic medicine: which factors impact impact factors? BMC Med Educ 2016 ; 16 : 190 . Google Scholar Crossref Search ADS PubMed WorldCat 7. Prpić K . Characteristics and determinants of eminent scientists’ productivity . Scientometrics 1996 ; 36 : 185 – 206 . Google Scholar Crossref Search ADS WorldCat 8. Lerchenmueller MJ , Sorenson O. Author disambiguation in PubMed: evidence on the precision and recall of author-ity among NIH-funded scientists . PLoS One 2016 ; 11 ( 7): e0158731 . Google Scholar Crossref Search ADS PubMed WorldCat 9. Akers KG , Sarkozy A, Wu W, Slyman A. ORCID author identifiers: a primer for librarians . Med Ref Serv Q 2016 ; 35 ( 2): 135 – 44 . Google Scholar Crossref Search ADS PubMed WorldCat 10. Gasparyan AY , Yessirkepov M, Gerasimov AN, Kostyukova EI, Kitas GD. Scientific author names: errors, corrections, and identity profiles . Biochem Med (Zagreb) 2016 ; 26 ( 2): 169 – 73 . Google Scholar Crossref Search ADS PubMed WorldCat 11. Gasparyan AY , Akazhanov NA, Voronov AA, Kitas GD. Systematic and open identification of researchers and authors: focus on open researcher and contributor ID . J Korean Med Sci 2014 ; 29 ( 11): 1453 – 6 . Google Scholar Crossref Search ADS PubMed WorldCat 12. SERI SS for E Research and Innovation . Bibliometric Analysis of Scientific Research in Switzerland 1981–2013 [Internet] . https://www.sbfi.admin.ch/sbfi/en/home/dienstleistungen/publikationen/publikationen-bestellen/publikationen-81-13.html (accessed on 1 December 2019 ). 13. Hirsch JE . An index to quantify an individual’s scientific research output . PNAS 2005 ; 102 : 16569 – 72 . Google Scholar Crossref Search ADS PubMed WorldCat 14. Martín-Martín A , Orduna-Malea E, Thelwall M, Delgado LCE. Google Scholar, Web of Science, and Scopus: a systematic comparison of citations in 252 subject categories . J Informetr 2018 ; 12 : 1160 – 77 . Google Scholar Crossref Search ADS WorldCat 15. Bar-Ilan J . Which h-index?—A comparison of WoS, Scopus and Google Scholar . Scientometrics 2008 ; 74 : 257 – 71 . Google Scholar Crossref Search ADS WorldCat 16. Taylor RB . The promise of family medicine: history, leadership, and the Age of Aquarius . J Am Board Fam Med 2006 ; 19 ( 2): 183 – 90 . Google Scholar Crossref Search ADS PubMed WorldCat 17. Ie K , Murata A, Tahara M et al. What determines medical students’ career preference for general practice residency training?: A multicenter survey in Japan . Asia Pac Fam Med 2018 ; 17 : 2 . Google Scholar Crossref Search ADS PubMed WorldCat 18. Mauksch LB , Fogarty CT. How do we know when to celebrate? Fam Syst Health 2014 ; 32 ( 2): 135 – 6 . Google Scholar Crossref Search ADS PubMed WorldCat 19. Pathman DE , Viera AJ, Newton WP. Research published in 2003 by U.S. family medicine authors . J Am Board Fam Med 2008 ; 21 ( 1): 6 – 16 . Google Scholar Crossref Search ADS PubMed WorldCat 20. Lin MH , Hwang SJ, Hwang IH, Chen YC. Family medicine publications in Taiwan: an analysis of the Web of Science database from 1993 to 2012 . J Chin Med Assoc 2014 ; 77 ( 11): 583 – 8 . Google Scholar Crossref Search ADS PubMed WorldCat 21. Mar CD , Askew D. Building family/general practice research capacity . Ann Fam Med 2004 ; 2 : S35 – 40 . Google Scholar Crossref Search ADS PubMed WorldCat 22. Negative Binomial Regression . Stata Data Analysis Examples [Internet] . https://stats.idre.ucla.edu/stata/dae/negative-binomial-regression/ (accessed on 2 November 2019 ). 23. Negative Binomial Regression . Stata Annotated Output [Internet] . https://stats.idre.ucla.edu/stata/output/negative-binomial-regression/ (accessed on 2 November 2019 ). 24. Cohen J. Statistical Power Analysis for the Behavioral Sciences . 2nd edn. Hillsdale, NJ : L. Erlbaum Associates ; 1988 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 25. Zhu H , Lakkis H. Sample size calculation for comparing two negative binomial rates . Stat Med 2014 ; 33 ( 3): 376 – 87 . Google Scholar Crossref Search ADS PubMed WorldCat 26. Siemieniuk R , Guyatt G. Grading quality of evidence and strength of recommendations . BMJ . 2004 ; 328 : 1490 . Google Scholar Crossref Search ADS PubMed WorldCat 27. BMJ Best Practice. What is GRADE?. [Internet] . https://bestpractice.bmj.com/info/us/toolkit/learn-ebm/what-is-grade/ (accessed on 11 November 2018 ). © The Author(s) 2020. Published by Oxford University Press. All rights reserved.For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Family Practice Oxford University Press

Scientific publications in internal medicine and family medicine: a comparative cross-sectional study in Swiss university hospitals

Family Practice , Volume Advance Article – Nov 13, 2020

Loading next page...
 
/lp/oxford-university-press/scientific-publications-in-internal-medicine-and-family-medicine-a-n0w19YJseh

References (23)

Publisher
Oxford University Press
Copyright
Copyright © 2021 Oxford University Press
ISSN
0263-2136
eISSN
1460-2229
DOI
10.1093/fampra/cmaa124
Publisher site
See Article on Publisher Site

Abstract

Abstract Background Family medicine is a relatively new academic medical discipline. We aimed to compare the main bibliometric indices of hospital-based senior physicians practicing internal medicine versus family medicine in Switzerland. Methods We conducted this cross-sectional study in March 2020. We selected all hospital-based senior physicians practicing internal medicine or family medicine in the six Swiss university hospitals. Using Web of Science, after removing from both groups of physicians the 5% with the highest number of publications, we extracted the number of publications, the number of publications per year, the number of citations, the number of citations per year, the number of citations per publication and the h-index. We compared the data between the two groups using negative binomial regressions and the proportion of physicians having at least one publication using chi-square tests. Results We included 349 physicians in the study (internal medicine: 51%, men: 51%). The median number of publications was three [interquartile range (IQR) = 18], the median number of citations was nine (IQR = 158) and the median h-index was one (IQR = 5). All bibliometric indices were similar in both groups, as was the proportion of physicians having at least one publication (family medicine: 87% versus 82%, P = 0.15). Conclusions We found no association between the bibliometric indices and the medical specialty. Further studies are needed to explore other important indicators of academic output, such as those more specifically assessing its quality and scientific importance. Bibliometric study, citations, family medicine, h-index, publications, productivity Key Messages Family medicine is a relatively new academic medical discipline. The main bibliometric indices are the number of publications and citations and h-index. No studies compared bibliometric indices in family medicine and internal medicine. We compared these indices for all senior physicians in Swiss university hospitals. We found no association between physicians’ indices and the medical specialty. Introduction Publication of scientific output is crucial because it allows to spread scientific knowledge (1,2) and increase the recognition of researchers (3–5). The productivity of researchers, which represents the visible result of scientific activities, has become of paramount importance to assess their global performance, including in medicine (6,7). It is also used to explore various issues linked to research policy, such as the role of gender, cross-institutional collaboration or funding on publication trajectories (8–11). The productivity of researchers can be easily measured using bibliometric methods. These methods allow for a quantitative analysis of scientific publications. Despite its known limitations, they are increasingly being used in the field of library science to determine the impact of scientific publications and to explore the productivity of researchers in various scientific fields (12). In addition to the number of publications and citations, the h-index has become popular for measuring researcher productivity because it integrates two major parameters: the productivity of researchers and the impact of their publications in terms of citations (13). Indeed, the number of publications does not necessarily reflect the quality of scientific publications and the number of citations may be affected by a few influential publications (13). The h-index was first proposed by Jorge Hirsch in 2005 to determine the scientific quality of articles published by theoretical physicists and is now used in many scientific fields, including medicine. This index is defined as the number (h) of publications that were each cited at least h times (13). Bibliometric indexes can be determined using citation databases, such as Web of Science, Scopus and Google Scholar. For a given author, they may vary depending on the database chosen due to different coverage. However, Web of Science and Scopus tend to give relatively similar results (14,15). Very few bibliometric data are available concerning our country. According to a report by the State Secretariat for Education, Research and Innovation (SERI) published in 2016, for the period 2009–13, Switzerland produced 1.2% of the world’s scientific publications (all scientific fields combined), ranking it 16th (12). However, in terms of the number of publications per capita, Switzerland was the world leader with nearly 4000 publications per million inhabitants. In terms of the number of publications per researcher, Switzerland was in third place, with 857 publications per 1000 researchers. The impact of publications produced in Switzerland was also excellent, with Switzerland in third place, just behind the USA and the Netherlands. Compared to other academic medical disciplines, family medicine is a relatively new discipline (16–18). The number of publications in this discipline has increased in recent decades, largely due to the creation of university academies of family medicine in many countries (5,19–21). This development is encouraging because it is often through academic research that we see the successful development of a medical discipline (21). Building research capacity in family medicine is essential to provide better care to patients. Indeed, although family physicians benefit from hospital-based research in internal medicine, most of these studies do not apply to patients evaluated in primary care settings. Conducting studies in primary care avoids such biases and is more prone to provide useful and generalizable data to primary care physicians (21). Academic research in internal medicine is usually recognized for its high quality. The above bibliometric indicators would make it possible to compare publications in family medicine versus internal medicine and, therefore, to assess the current state of family medicine research in a country like Switzerland. The objective of our study was to compare the productivity of hospital-based senior physicians practicing internal medicine versus family medicine in Switzerland. We hypothesize higher bibliometric indices among researchers practicing internal medicine. Methods University hospitals in Switzerland There are six university hospitals in Switzerland. Half of these hospitals are located in predominantly French-speaking cantons (Western Switzerland: Geneva, Lausanne and Fribourg) and the other half in predominantly German-speaking cantons (Eastern Switzerland: Bern, Basel and Zurich). There is no university hospital in Italian-speaking Switzerland. General medicine (also known as family medicine) and internal medicine merged in 2011 into a single discipline: general internal medicine. Despite this fusion, Swiss hospitals have separate departments of family medicine (providing outpatient care) and internal medicine (providing inpatient care). Hospital-based senior physicians include those who are in long-term employment (with a few exceptions, they pursue an academic career) and those who have completed their training and generally work for a few more years as senior physicians (=senior registrars) before starting a sub-specialization or working in a private practice. After a few years spent as senior registrars, physicians may be promoted to staff physicians, mostly on the basis of their academic output. Study site and study population We conducted our bibliometric study from the first to 14 March 2020 in the six Swiss university hospitals (all university hospitals were included in the study). We included all senior physicians (heads of division, staff physicians and senior registrars) practicing internal medicine or family medicine in one of these six university hospitals. We retrieved the list of these physicians from the hospitals’ websites. As planned, we excluded for both medical specialties the 5% of physicians with the highest number of publications in order to obtain two relatively homogeneous study samples and avoid biases that could result from the inclusion of some extreme outliers. Some senior physicians working in non-university hospitals also do research. However, as in most countries, the majority of publications are authored by physicians working in university hospitals. This study was intended to cover academic output, that is, the publications by physicians working in university hospitals only. Data collection First, using the hospitals’ websites, we recorded physicians’ socio-demographic characteristics [gender, medical department, professor (Y/N), position (head of division, staff physician, senior registrar and other (consulting physician, senior research physician and senior education physician))]. These websites are very reliable in Switzerland, regularly updated and accurately list all the physicians working in the various hospital departments. We consulted them a few days before collecting the bibliometric data. Then, using the citation report function of Web of Science, a platform consisting of several literature search databases, we extracted physicians’ bibliometric data (number of publications, year of publication of the first article, number of citations and h-index). We obtained the list of publications of each selected physician for all years up to the survey (there were no limitations in the time period covered by the study) by using the following search criteria: ‘all databases’ and ‘search by author, select from index’ and by entering the surname and first name in the search box. For physicians with two surnames, we entered the two surnames separated by a space, then by a hyphen and, finally, we entered each of the two surnames. For physicians with two first names, we entered only the first one. Two investigators from the research team (PS and NV) extracted all bibliometric data. They performed the data extraction in duplicate to exclude any errors. Before extracting the bibliometric data, they removed publications authored by homonyms. They resolved doubts about certain publications through a discussion within the research team. The bibliometric data were extracted for any publication indexed in Web of Science. In this platform, different types of literature are indexed in addition to original research articles, such as non-research articles (e.g. editorials, clinical updates and non-systematic reviews) and books. Among the various bibliometric indices available, we voluntarily included both ‘productivity’ (number of publications) and ‘quality’ variables (number of citations and h-index). High-quality research articles will tend to be cited more often and, therefore, counted more often in ‘quality’ variables than other publications. However, as Web of Science coverage is not limited to research articles, we used the term ‘publications’ to summarize the scientific productivity of researchers. Statistical analyses and sample size For the purpose of the study, we built three additional bibliometric variables. We computed ‘the number of citations per year and the number of citations per year’ by dividing the number of publications, respectively, the number of citations, by the number of years since the publication of the first article. In addition, we computed ‘the number of citations per publication’ by dividing the number of citations by the number of publications. We rounded these three variables to the nearest unit. We used frequency tables to describe physicians’ socio-demographic characteristics, and median and interquartile range (IQR) to summarize the number of years since they first published and the bibliometric data. We used chi-square tests to compare the socio-demographic characteristics and the proportion of physicians having at least one publication between the two groups of physicians (internal medicine and family medicine), the Wilcoxon rank-sum test to compare their first publications and univariate negative binomial regressions to compare their other bibliometric data (= count data with overdispersion) (22,23). Finally, we used Spearman’s rank correlation coefficients to determine the correlation between physicians’ bibliometric indices and the number of years since first publication. We considered correlations lower than 0.30 as ‘small’, between 0.30 and 0.50 as ‘medium’ and higher than 0.50 as ‘large’ (24). As planned, we excluded from both groups of physicians the 5% with the highest number of publications. However, we repeated the analyses by adding these 5% of physicians to evaluate whether or not the results comparing the bibliometric indices of physicians in internal medicine versus family medicine were modified (sensitivity analysis). We used the sample size calculation for comparing two negative binomial rates as developed by Zhu and Lakkis (25). We estimated that, by comparing two groups (physicians practising internal medicine versus family medicine), a sample of 336 (168 in each group) would be sufficient to detect a difference of five publications with Type I and II error both set at 5%. We expected the average number of publications to be 10 in internal medicine and 5 in family medicine, the ratio of the number of participants to be 1 and the dispersion parameter to be 5. We carried out all statistical analyses with STATA version 15.1 (College Station, TX, USA). Results We included 349 hospital-based senior physicians in the study, of whom 178 (51%) were affiliated with a department of internal medicine and 171 (49%) with a department of family medicine (Table 1). Two-thirds of the physicians were located in French-speaking Switzerland. There were slightly more men than women (51% versus 49%). One tenth of the physicians had the title of professor and about half of the physicians were senior registrars. There was no statistically significant difference between the two groups of physicians, except for the hierarchical position (more senior registrars in internal medicine than in family medicine). Of the 349 physicians included in the study, 294 (84%) had at least one publication, with no difference between the two groups (82% in internal medicine versus 87% in family medicine; P = 0.15). There was also no difference in the number of years since they first published [median = 8 (IQR = 11) in internal medicine versus 9 (IQR = 11) in family medicine; P = 0.35]. Table 1. Physicians’ socio-demographic characteristics, overall and stratified by medical speciality (n = 349) Characteristics . Total (n = 349) n (%) . Internal medicine (n = 178) n (%) . Family medicine (n = 171) n (%) . P-valuea . Location 0.28  French-speaking Switzerland 229 (65.6) 112 (62.9) 117 (68.4)  German-speaking Switzerland 120 (34.4) 66 (37.1) 54 (31.6) Gender 0.56  Male 179 (51.3) 94 (52.8) 85 (49.7)  Female 170 (48.7) 84 (47.2) 86 (50.3) Professor 0.09  Yes 34 (9.7) 22 (12.4) 12 (7.0)  No 315 (90.3) 156 (87.6) 159 (93.0) Hierarchical position <0.001  Heads of division 11 (3.2) 5 (2.8) 6 (3.5)  Staff physicians 85 (24.4) 41 (23.0) 44 (25.7)  Senior registrars 185 (53.0) 122 (68.6) 63 (36.9)  Otherb 68 (19.5) 10 (5.6) 58 (33.9) Characteristics . Total (n = 349) n (%) . Internal medicine (n = 178) n (%) . Family medicine (n = 171) n (%) . P-valuea . Location 0.28  French-speaking Switzerland 229 (65.6) 112 (62.9) 117 (68.4)  German-speaking Switzerland 120 (34.4) 66 (37.1) 54 (31.6) Gender 0.56  Male 179 (51.3) 94 (52.8) 85 (49.7)  Female 170 (48.7) 84 (47.2) 86 (50.3) Professor 0.09  Yes 34 (9.7) 22 (12.4) 12 (7.0)  No 315 (90.3) 156 (87.6) 159 (93.0) Hierarchical position <0.001  Heads of division 11 (3.2) 5 (2.8) 6 (3.5)  Staff physicians 85 (24.4) 41 (23.0) 44 (25.7)  Senior registrars 185 (53.0) 122 (68.6) 63 (36.9)  Otherb 68 (19.5) 10 (5.6) 58 (33.9) aχ2 tests. bConsulting physicians, senior research physicians and senior education physicians. Open in new tab Table 1. Physicians’ socio-demographic characteristics, overall and stratified by medical speciality (n = 349) Characteristics . Total (n = 349) n (%) . Internal medicine (n = 178) n (%) . Family medicine (n = 171) n (%) . P-valuea . Location 0.28  French-speaking Switzerland 229 (65.6) 112 (62.9) 117 (68.4)  German-speaking Switzerland 120 (34.4) 66 (37.1) 54 (31.6) Gender 0.56  Male 179 (51.3) 94 (52.8) 85 (49.7)  Female 170 (48.7) 84 (47.2) 86 (50.3) Professor 0.09  Yes 34 (9.7) 22 (12.4) 12 (7.0)  No 315 (90.3) 156 (87.6) 159 (93.0) Hierarchical position <0.001  Heads of division 11 (3.2) 5 (2.8) 6 (3.5)  Staff physicians 85 (24.4) 41 (23.0) 44 (25.7)  Senior registrars 185 (53.0) 122 (68.6) 63 (36.9)  Otherb 68 (19.5) 10 (5.6) 58 (33.9) Characteristics . Total (n = 349) n (%) . Internal medicine (n = 178) n (%) . Family medicine (n = 171) n (%) . P-valuea . Location 0.28  French-speaking Switzerland 229 (65.6) 112 (62.9) 117 (68.4)  German-speaking Switzerland 120 (34.4) 66 (37.1) 54 (31.6) Gender 0.56  Male 179 (51.3) 94 (52.8) 85 (49.7)  Female 170 (48.7) 84 (47.2) 86 (50.3) Professor 0.09  Yes 34 (9.7) 22 (12.4) 12 (7.0)  No 315 (90.3) 156 (87.6) 159 (93.0) Hierarchical position <0.001  Heads of division 11 (3.2) 5 (2.8) 6 (3.5)  Staff physicians 85 (24.4) 41 (23.0) 44 (25.7)  Senior registrars 185 (53.0) 122 (68.6) 63 (36.9)  Otherb 68 (19.5) 10 (5.6) 58 (33.9) aχ2 tests. bConsulting physicians, senior research physicians and senior education physicians. Open in new tab Table 2 summarizes the bibliometric data and shows the association between these data and the medical specialty. For ease of data interpretation, Figures 1–3 present in graphical form (scatter plot) the number of publications, the number of citations and the h-index by the number of years since they first published and the medical specialty. The median number of publications and publications per year was 3 (IQR = 18) and 1 (IQR = 2), respectively, while the median number of citations and citations per year was 9 (IQR = 158) and 1 (IQR = 13), respectively. On average, two authors cited each publication. The median h-index was 1 (IQR = 5). There was no association between physicians’ bibliometric indices and their medical specialty. Finally, all correlations examined between physicians’ bibliometric indices and the number of years since they first published were high (number of publications: rho = 0.67, number of citations: rho = 0.78, h-index: rho = 0.75) and the relationships were statistically significant (all P values <0.005). Table 2. Associations between medical specialty and number of publications, number of citations and h-index (list of publications extracted for all years up to March 2020) Variable . Total (n = 349) Median (IQR) . Internal medicine (n = 178) Median (IQR) . Family medicine (n = 171) Median (IQR) . IRR (95% CI) . Crude P-valuea . Publications  Number of publications 3 (18) 2 (12) 5 (22) 1.0 (0.7–1.3) 0.77  Number of publications per year 1 (2) 1 (2) 1 (2) 0.8 (0.6–1.1) 0.23 Citations  Number of citations 9 (158) 7 (165) 12 (148) 1.6 (0.9–2.8) 0.11  Number of citations per year 1 (13) 1 (15) 2 (11) 1.5 (0.9–2.4) 0.12  Number of citations per publication 2 (11) 3 (12) 2 (11) 1.4 (0.9–2.1) 0.11 Publications and citations  h-index 1 (5) 1 (5) 1 (5) 1.1 (0.8–1.5) 0.59 Variable . Total (n = 349) Median (IQR) . Internal medicine (n = 178) Median (IQR) . Family medicine (n = 171) Median (IQR) . IRR (95% CI) . Crude P-valuea . Publications  Number of publications 3 (18) 2 (12) 5 (22) 1.0 (0.7–1.3) 0.77  Number of publications per year 1 (2) 1 (2) 1 (2) 0.8 (0.6–1.1) 0.23 Citations  Number of citations 9 (158) 7 (165) 12 (148) 1.6 (0.9–2.8) 0.11  Number of citations per year 1 (13) 1 (15) 2 (11) 1.5 (0.9–2.4) 0.12  Number of citations per publication 2 (11) 3 (12) 2 (11) 1.4 (0.9–2.1) 0.11 Publications and citations  h-index 1 (5) 1 (5) 1 (5) 1.1 (0.8–1.5) 0.59 aUnivariate negative binomial regression. Open in new tab Table 2. Associations between medical specialty and number of publications, number of citations and h-index (list of publications extracted for all years up to March 2020) Variable . Total (n = 349) Median (IQR) . Internal medicine (n = 178) Median (IQR) . Family medicine (n = 171) Median (IQR) . IRR (95% CI) . Crude P-valuea . Publications  Number of publications 3 (18) 2 (12) 5 (22) 1.0 (0.7–1.3) 0.77  Number of publications per year 1 (2) 1 (2) 1 (2) 0.8 (0.6–1.1) 0.23 Citations  Number of citations 9 (158) 7 (165) 12 (148) 1.6 (0.9–2.8) 0.11  Number of citations per year 1 (13) 1 (15) 2 (11) 1.5 (0.9–2.4) 0.12  Number of citations per publication 2 (11) 3 (12) 2 (11) 1.4 (0.9–2.1) 0.11 Publications and citations  h-index 1 (5) 1 (5) 1 (5) 1.1 (0.8–1.5) 0.59 Variable . Total (n = 349) Median (IQR) . Internal medicine (n = 178) Median (IQR) . Family medicine (n = 171) Median (IQR) . IRR (95% CI) . Crude P-valuea . Publications  Number of publications 3 (18) 2 (12) 5 (22) 1.0 (0.7–1.3) 0.77  Number of publications per year 1 (2) 1 (2) 1 (2) 0.8 (0.6–1.1) 0.23 Citations  Number of citations 9 (158) 7 (165) 12 (148) 1.6 (0.9–2.8) 0.11  Number of citations per year 1 (13) 1 (15) 2 (11) 1.5 (0.9–2.4) 0.12  Number of citations per publication 2 (11) 3 (12) 2 (11) 1.4 (0.9–2.1) 0.11 Publications and citations  h-index 1 (5) 1 (5) 1 (5) 1.1 (0.8–1.5) 0.59 aUnivariate negative binomial regression. Open in new tab Figure 1. Open in new tabDownload slide Physicians’ number of publications by number of years since first publication (list of publications extracted for all years up to March 2020; data are shown separately for internal medicine and family medicine). Figure 1. Open in new tabDownload slide Physicians’ number of publications by number of years since first publication (list of publications extracted for all years up to March 2020; data are shown separately for internal medicine and family medicine). Figure 2. Open in new tabDownload slide Physicians’ number of citations by number of years since first publication (list of publications extracted for all years up to March 2020; data are shown separately for internal medicine and family medicine). Figure 2. Open in new tabDownload slide Physicians’ number of citations by number of years since first publication (list of publications extracted for all years up to March 2020; data are shown separately for internal medicine and family medicine). Figure 3. Open in new tabDownload slide Physicians’ h-index by number of years since first publication (list of publications extracted for all years up to March 2020; data are shown separately for internal medicine and family medicine). Figure 3. Open in new tabDownload slide Physicians’ h-index by number of years since first publication (list of publications extracted for all years up to March 2020; data are shown separately for internal medicine and family medicine). As planned, we excluded from both groups of physicians the 5% with the highest number of publications. However, we repeated the analyses including all physicians. These analyses confirmed the absence of statistically significant differences for all bibliometric indicators listed in Table 2, except for the number of citations, which was on average twice as high in internal medicine as in family medicine [IRR = 2.1 (95% confidence interval (CI) = 1.2–3.8), P = 0.01]. The association remained statistically significant when adjusting for the number of years since first publication, location, gender and hierarchical position [IRR = 2.3 (95% CI = 1.5–3.4), P < 0.001]. This association is related to the presence of a few outliers in the internal medicine group (physicians with a very high number of citations). Discussion Main findings In summary, in this sample of hospital-based senior physicians, we found that the median number of publications was 3, the median number of citations was 9 and the median h-index was 1. There was no statistically significant difference between physicians’ bibliometric indices and the medical specialty. We also found that these bibliometric indices were highly correlated with the number of years since they first published. Comparison with existing literature To our knowledge, this study is the first to have compared the productivity of hospital-based physicians in internal medicine with those in family medicine. The absence of statistically significant differences in the various bibliometric indices considered (number of publications, number of publications per year, number of citations, number of citations per year, number of citations per publication and h-index) can probably be explained by the rapid development of academic research in family medicine these last years. As family medicine has become an academic discipline very gradually over the last few decades, the results obtained in this study could have been even more favourable to family medicine if we had restricted the study to the last years of publications. However, many questions remain unanswered in this field. Among these questions, it is unclear whether research in family medicine adequately meets the needs of community-based family physicians. In addition, our study focused mainly on the quantitative evaluation of scientific research. It is, of course, important to evaluate both the quantity and quality of research. The quality of research could in theory be measured using the number of citations and h-index, but these indicators are imperfect. Further studies are, therefore, needed to explore the quality and scientific importance of family medicine research by using or developing accurate indicators. One way of doing this could be to use the Grading of Recommendations Assessment, Development and Evaluation approach (26,27). Many of the recommendations for clinical practice are based on summaries of evidence as determined by this tool. Two caveats should be mentioned concerning the citation database and the bibliometric variables used in this study. First, the indices were calculated using the platform Web of Science. We believe that, although the results might vary slightly depending on the citation database used (mainly because there are differences in the journals and types of literature indexed), those concerning the comparison between internal medicine and family medicine should not be greatly modified. Indeed, comparisons of indices are probably less subject to variation than indices themselves. Second, the bibliometric variables used in our study could be subject to criticism. For example, they do not take into account the author’s position in the list of authors and some of them (number of citations and h-index) can be manipulated using self-citations. In addition, when comparing different disciplines with these bibliometric variables, they do not adjust for the number of researchers in the field and some other factors, such as the number of high-impact journals and access to research funding. However, although this lack of adjustment probably leads to an underestimation of the real productivity of family medicine researchers, we found similar bibliometric results in both groups of physicians. If the number of high-impact journals were similar in both disciplines, the results would probably have been even better for physicians in family medicine. Limitations Besides these two caveats, some limitations need to be pointed out in our study. First, the study sample consisted only of physicians practicing in Swiss university hospitals. The results could possibly not be generalized to other physicians in Switzerland or in other countries. The overrepresentation in our study of physicians practicing in French-speaking university hospitals may seem surprising since only about a quarter of the inhabitants live in French-speaking Switzerland. This finding is explained by the fact that the ratio between university and non-university hospitals in the different regions of Switzerland is not related to the size of the population living in each of them. Second, we cannot completely exclude certain errors in the publication lists, even though we standardized the search procedures, duplicated all searches and reviewed all publications of each physician to remove those published by homonyms. In addition, choosing to retrieve the publication lists using physicians’ full first name also reduced the risk of homonymy, but this strategy might have missed some publications that were referenced without the full first name. However, there is no reason to imagine that errors in the establishment of the publication lists could affect the two groups of physicians differently. Third, we did not know the age or date of birth of the physicians included in the study. Yet there was no significant difference between the average number of years since first publication (a variable that could be considered a surrogate for age) and the medical field. Finally, our aim was not to measure the methodological quality, scientific importance or novelty of the publications included in the study. These indicators could be used in future comparative studies. Conclusion In conclusion, in this sample of Swiss hospital-based senior physicians, we found no association between the various bibliometric indices and the medical field (internal medicine versus family medicine). Further studies are needed in the future to confirm that the quality and impact of research in family medicine and in internal medicine are similar and that family medicine has definitely become a very attractive academic field. Declarations Funding: none. Ethical approval: not required (under Swiss law, ethical approval is not required when collecting non-personal health data). Conflict of interest: none. Data availability: the data underlying this article will be shared on reasonable request to the corresponding author. References 1. Mabe MA . Scholarly communication: a long view . New Rev Acad Librarianship 2010 ; 16 : 132 – 44 . Google Scholar Crossref Search ADS WorldCat 2. Lafrenière D , Menuz V, Hurlimann T, Godard B. Knowledge dissemination interventions: a literature review . SAGE Open 2013 ; 3 : 2158244013498242 . Google Scholar Crossref Search ADS WorldCat 3. Vale RD . Accelerating scientific publication in biology . Proc Natl Acad Sci USA 2015 ; 112 ( 44): 13439 – 46 . Google Scholar Crossref Search ADS PubMed WorldCat 4. Bavdekar SB , Tullu MS. Research publications for academic career advancement: An idea whose time has come. But is this the right way? J Postgrad Med 2016 ; 62 ( 1): 1 – 3 . Google Scholar Crossref Search ADS PubMed WorldCat 5. Post RE , Weese TJ, Mainous AG III, Weiss BD. Publication productivity by family medicine faculty: 1999 to 2009 . Fam Med 2012 ; 44 ( 5 ): 312 – 7 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 6. Pfeiffer M , Fischer MR, Bauer D. Publication activities of German junior researchers in academic medicine: which factors impact impact factors? BMC Med Educ 2016 ; 16 : 190 . Google Scholar Crossref Search ADS PubMed WorldCat 7. Prpić K . Characteristics and determinants of eminent scientists’ productivity . Scientometrics 1996 ; 36 : 185 – 206 . Google Scholar Crossref Search ADS WorldCat 8. Lerchenmueller MJ , Sorenson O. Author disambiguation in PubMed: evidence on the precision and recall of author-ity among NIH-funded scientists . PLoS One 2016 ; 11 ( 7): e0158731 . Google Scholar Crossref Search ADS PubMed WorldCat 9. Akers KG , Sarkozy A, Wu W, Slyman A. ORCID author identifiers: a primer for librarians . Med Ref Serv Q 2016 ; 35 ( 2): 135 – 44 . Google Scholar Crossref Search ADS PubMed WorldCat 10. Gasparyan AY , Yessirkepov M, Gerasimov AN, Kostyukova EI, Kitas GD. Scientific author names: errors, corrections, and identity profiles . Biochem Med (Zagreb) 2016 ; 26 ( 2): 169 – 73 . Google Scholar Crossref Search ADS PubMed WorldCat 11. Gasparyan AY , Akazhanov NA, Voronov AA, Kitas GD. Systematic and open identification of researchers and authors: focus on open researcher and contributor ID . J Korean Med Sci 2014 ; 29 ( 11): 1453 – 6 . Google Scholar Crossref Search ADS PubMed WorldCat 12. SERI SS for E Research and Innovation . Bibliometric Analysis of Scientific Research in Switzerland 1981–2013 [Internet] . https://www.sbfi.admin.ch/sbfi/en/home/dienstleistungen/publikationen/publikationen-bestellen/publikationen-81-13.html (accessed on 1 December 2019 ). 13. Hirsch JE . An index to quantify an individual’s scientific research output . PNAS 2005 ; 102 : 16569 – 72 . Google Scholar Crossref Search ADS PubMed WorldCat 14. Martín-Martín A , Orduna-Malea E, Thelwall M, Delgado LCE. Google Scholar, Web of Science, and Scopus: a systematic comparison of citations in 252 subject categories . J Informetr 2018 ; 12 : 1160 – 77 . Google Scholar Crossref Search ADS WorldCat 15. Bar-Ilan J . Which h-index?—A comparison of WoS, Scopus and Google Scholar . Scientometrics 2008 ; 74 : 257 – 71 . Google Scholar Crossref Search ADS WorldCat 16. Taylor RB . The promise of family medicine: history, leadership, and the Age of Aquarius . J Am Board Fam Med 2006 ; 19 ( 2): 183 – 90 . Google Scholar Crossref Search ADS PubMed WorldCat 17. Ie K , Murata A, Tahara M et al. What determines medical students’ career preference for general practice residency training?: A multicenter survey in Japan . Asia Pac Fam Med 2018 ; 17 : 2 . Google Scholar Crossref Search ADS PubMed WorldCat 18. Mauksch LB , Fogarty CT. How do we know when to celebrate? Fam Syst Health 2014 ; 32 ( 2): 135 – 6 . Google Scholar Crossref Search ADS PubMed WorldCat 19. Pathman DE , Viera AJ, Newton WP. Research published in 2003 by U.S. family medicine authors . J Am Board Fam Med 2008 ; 21 ( 1): 6 – 16 . Google Scholar Crossref Search ADS PubMed WorldCat 20. Lin MH , Hwang SJ, Hwang IH, Chen YC. Family medicine publications in Taiwan: an analysis of the Web of Science database from 1993 to 2012 . J Chin Med Assoc 2014 ; 77 ( 11): 583 – 8 . Google Scholar Crossref Search ADS PubMed WorldCat 21. Mar CD , Askew D. Building family/general practice research capacity . Ann Fam Med 2004 ; 2 : S35 – 40 . Google Scholar Crossref Search ADS PubMed WorldCat 22. Negative Binomial Regression . Stata Data Analysis Examples [Internet] . https://stats.idre.ucla.edu/stata/dae/negative-binomial-regression/ (accessed on 2 November 2019 ). 23. Negative Binomial Regression . Stata Annotated Output [Internet] . https://stats.idre.ucla.edu/stata/output/negative-binomial-regression/ (accessed on 2 November 2019 ). 24. Cohen J. Statistical Power Analysis for the Behavioral Sciences . 2nd edn. Hillsdale, NJ : L. Erlbaum Associates ; 1988 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 25. Zhu H , Lakkis H. Sample size calculation for comparing two negative binomial rates . Stat Med 2014 ; 33 ( 3): 376 – 87 . Google Scholar Crossref Search ADS PubMed WorldCat 26. Siemieniuk R , Guyatt G. Grading quality of evidence and strength of recommendations . BMJ . 2004 ; 328 : 1490 . Google Scholar Crossref Search ADS PubMed WorldCat 27. BMJ Best Practice. What is GRADE?. [Internet] . https://bestpractice.bmj.com/info/us/toolkit/learn-ebm/what-is-grade/ (accessed on 11 November 2018 ). © The Author(s) 2020. Published by Oxford University Press. All rights reserved.For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Journal

Family PracticeOxford University Press

Published: Nov 13, 2020

Keywords: family medicine; hospitals, university; internal medicine

There are no references for this article.