Author + information
- Received March 26, 2018
- Revision received May 18, 2018
- Accepted May 20, 2018
- Published online July 30, 2018.
- Mariam Asghar, MBBSa,
- Muhammad Shariq Usman, MBBSa,
- Rafi Aibani, MBBSa,
- Hamza Tahir Ansari, MBBSa,
- Tariq Jamal Siddiqi, MBBSa,
- Kaneez Fatima, MBBSa,
- Muhammad Shahzeb Khan, MDb and
- Vincent M. Figueredo, MDc,d,∗ (, )@JeffersonUniv@EinsteinHealth
- aDepartment of Internal Medicine, Dow University of Health Sciences, Karachi, Pakistan
- bDepartment of Internal Medicine, John H Stroger Jr. Hospital of Cook County, Chicago, Illinois
- cDepartment of Cardiology, Institute for Heart & Vascular Health, Einstein Medical Center Philadelphia, Pennsylvania
- dDepartment of Medicine, Sidney Kimmel College of Medicine at Thomas Jefferson University, Philadelphia, Pennsylvania
- ↵∗Address for correspondence:
Dr. Vincent M. Figueredo, Einstein Medical Center, 5501 Old York Road, Levy 3232, Philadelphia, Pennsylvania 19141.
Although nearly one-half of all medical graduates in the United States and the United Kingdom are women, academic medicine remains male dominated, with women accounting for ∼30% of clinical faculty (1). This gender gap is reflected in the medical literature, with women being underrepresented in key authorship positions. In both the United States and the United Kingdom, cardiology continues to have the lowest proportion of women among all specialties (2,3). Over the years, the proportion of female physicians and trainees in cardiology has increased, but it remains unknown whether the presence of women in the academic cardiology literature is growing at the same rate. The fraction of women holding the first authorship position is indicative of the active research involvement of women, whereas the proportion of women holding the last position is representative of female mentorship in cardiology research. In this analysis, we ascertain the existence and extent of a “gender gap” in cardiology literature, and its variation over the past 2 decades.
We selected Scopus as our database. Six prominent cardiology journals, namely Journal of the American College of Cardiology (JACC), European Heart Journal (EHJ), Circulation, American Journal of Cardiology (AJC), BMJ Heart (Heart), and Clinical Cardiology (CC), were included. Journals across a wide range of Impact Factors were selected to allow comparison.
We searched for articles published in these journals in the years 1996, 2006, and 2016, using the journal ISSN and e-ISSN numbers. The author lists, citation counts, and countries of origin of the articles were extracted from Scopus. The list of journal editorial board members in a particular year was extracted from the first issue published in that year. The sex of the first authors, senior (last listed) authors, and editorial board members was determined by inspection of their first name. In cases of uncertainty, attempts were made to find a picture of the author on his or her institution website or LinkedIn profile. If no decision could be made even after exhausting these resources, the respective author was contacted via e-mail. If an article had only 1 author, he or she was considered the senior author.
To ensure reliability, 2 independent investigators (M.A. and R.A.) collected data, and their lists were compared. This comparison revealed a low rate of discrepancy (3%) in which the sex of the first/senior author did not match between investigators. Discrepancies were resolved by consulting a third investigator (H.T.A.).
SPSS version 23 (IBM SPSS Statistics, IBM Corporation, Armonk, New York) was used to analyze the data. Percentages of women in the first and senior authorship positions were determined; subgroup analysis was conducted according to journal, article origin (United States, United Kingdom, or multinational), and type of article (original, review, letters, conference papers, and editorials). A chi-square linear-by-linear association test was used to test for trends over time. Percent change over time and corresponding 95% confidence intervals (CIs) were calculated by using the method described by Newcombe (4). The Pearson correlation test was used to assess whether the proportion of female first authors in a journal was dependent on the editorial board composition or Impact Factor of the journal. The Mann-Whitney U test was used to compare the citations of male and female first authors. The chi-square test was used to check if there was an association between the sex of the first and last authors. A p value <0.05 was considered significant in all cases.
We identified 11,529 articles on Scopus across all 6 journals (n1996 = 3,333; n2006 = 4,478; and n2016 = 3,718). Among these, 1,740 articles had only 1 author (who was considered a senior author). Sex of the first author was determined for 9,749 (99.6%) of 9,789 articles. Sex of the senior author was determined for 11,411 (99.0%) of 11,529 articles. Overall, 16.5% of first authors and 9.1% of senior authors were female. When analyzed according to year, significant increasing trends in female first authorship (p < 0.001) and senior authorship (p < 0.001) were observed. Over the 20-year period, the proportion of female first authors increased by 9.5% (95% CI: 7.7% to 11.3%), and the proportion of female senior authors increased by 6.6% (95% CI: 5.3% to 7.9%).
A detailed breakdown of female representation in each subgroup studied is given in Table 1. All 6 journals consistently showed a significant upward trend in female first authorship, but the magnitude of change in each journal differed (range: 5.1% to 14.5%). Similarly, all journals showed a significant upward trend in the proportion of female senior authorship (range: 4.2% to 8.1%), except for EHJ, which showed a decrease after 2006 (p = 0.286). The Central Illustration displays temporal trends in female first/last authorship.
Articles originating from the United States (p < 0.001) and the United Kingdom (p < 0.001) both showed upward trends in female first authorship, with increases of 9.5% (95% CI: 6.3% to 12.7%) and 18.8% (95% CI: 11.2% to 27.1%), respectively. Similarly, the proportion of female senior authors also increased in articles from the United States (p < 0.001) and the United Kingdom (p < 0.001) by 7.1% (95% CI: 4.8% to 9.5%) and 8.4% (95% CI: 2.9% to 11.5%). Articles of multinational origin showed an upward trend in female senior authorship (p = 0.01; change: 6.7%; 95% CI: 2.2% to 10.0%). Although the percentage of female senior authors also increased in the 20-year period (8.3%; 95% CI: 2.7% to 12.9%), a significant trend was not found (p = 0.19).
Upward trends in female first authorship were seen in all types of articles except for review articles (p = 0.290; change: 5.5%; 95% CI: –1.7% to 11.4%). The proportion of female senior authorship seemed to increase in all types of articles; however, the result for conference papers did not reach significance (p = 0.105), possibly due to the small sample size.
The proportion of female first authors was not associated with the Impact Factor of the journal in any of the 3 years studied (r1996 = 0.065, p1996 = 0.902; r2006 = 0.550, p2006 = 0.258; and r2016 = –0.296, p2016 = 0.569). Similar results were seen for female senior authors (r1996 = 0.032, p1996 = 0.951; r2006 = –0.138, p2006 = 0.794; and r2016 = –0.801, p2016 = 0.055). Furthermore, there was no correlation between the proportion of women on the editorial board of a journal and the percentage of female first authors (r1996 = 0.758, p1996 = 0.242; r2006 = –0.825, p2006 = 0.382; and r2016 = 0.325, p2016 = 0.530) or senior authors (r1996 = –0.080, p1996 = 0.920; r2006 = –0.190, p2006 = 0.878; and r2016 = 0.118, p2016 = 0.824) publishing in that journal.
Articles with women as first authors had significantly higher median citations in 1996 (39 vs. 31; p = 0.006) and 2006 (43 vs. 30; p < 0.001). Women also had higher median citations in 2016; however, this finding did not reach significance (3 vs. 2; p = 0.062). Female senior authors were more likely (p < 0.001) to have a female first author (25.0%) in the same article than male senior authors (15.4%), indicating a greater likelihood of a mentor–mentee relationship between authors of the same sex.
Physicians’ scholarly output is one of the major factors upon which their academic promotion is based (5). The pervasive gender gap in cardiology has been a subject of concern. As of 2016, 47% of medical graduates in the United States were women and 43% entered an internal medicine residency, but only 13.2% are board certified in cardiology (2,3). Women comprise only ∼16.5% of academic cardiologists, and this total is reflected in the cardiology literature (6). As of 2016, the proportion of female first authors in cardiology journals (21%) remained lower than that seen in the overall academic medicine (37%) (1), radiology (32%) (7), ophthalmology (34%) (8), and emergency medicine (24%) literature (9). Women are even more under-represented in the senior author position. This finding may not be surprising considering that female cardiology faculty members are significantly younger than their male counterparts, and many are still in the early stages of their career (6). In this analysis, encouraging trends toward gender parity are also seen, with the proportion of female first and senior authors being higher in 2016 than in 1996 in all instances, even upon subgroup analysis according to journal and national origin of articles. Another positive finding was the lack of association between the proportion of women on the editorial board and in the first/last authorship position, suggesting a bias-free editorial process in cardiology research.
The overall trends are positive, and the increasing involvement of women in the field of cardiology will play a key role in helping reach gender parity in authorship of cardiology literature. Concomitantly, the inherent and unconscious gender bias prevalent within the medical community must be addressed to help achieve this goal. For example, it has been shown that the same abstract is likely to receive a more critical review when submitted under a female name than under a male name (1). Such bias could offset the increased research productivity of women by hindering their submissions from being published. A paradigm shift in the way women are viewed within the scientific community is required to help eliminate this sort of unconscious bias. Implementing a double-blind review process may help bypass this problem. A point of concern is the lack of an upward trend in female first authorship in review articles and articles of multinational origin. These types of articles constituted a considerable proportion in our analysis (9% and 15%, respectively), and a failure of female representation to increase in this area could negatively affect the gender gap in cardiology literature as a whole.
The present analysis indicates that female first authors are more likely to have a female mentor compared with their male counterparts. Women in leadership positions probably influence and may even encourage women in their department to be more involved in research. Therefore, the attrition of female cardiologists at higher levels of the academic hierarchy may be a limiting factor in the increasing proportion of female first authors in the cardiology literature (6).
Another interesting finding of our analysis was that female first authors seemed to garner more citations than male first authors, indicating that their articles were more impactful. In contrast, a study published in Nature found that female first authors received lower citations in academic medicine compared with their male counterparts (10). A possible explanation for our findings can be the unconscious gender bias discussed earlier. Male first authors may benefit from a lenient review process, which would allow them to publish articles that are relatively less “impactful,” reducing their median citations. However, further investigations are needed to support our findings and help explain them.
The present analysis has several limitations. First, the sex of authors was determined by inspection of first names and Internet searches, and thus there may have been some error in the assignment of sex. However, our cross-check suggests that this outcome was the case for only an insignificant number of authors. Second, because this analysis comprised 6 cardiology journals, the findings may not accurately represent the entire cardiology literature. Third, a possible confounding factor could be the rate of submission and acceptance of papers with female first/last authors, which could not be accounted for in this analysis. Fourth, we assumed that the last listed author was the senior-most author; although this format is widely used, there are numerous other ways to specify authorship in publications.
The proportion of female first and senior authors is increasing in the cardiology literature. Our analysis found that articles published by female first authors tend to accumulate more citations. Female senior authors are more likely to have a female first author in the same article, suggesting a mentor–mentee relationship.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Received March 26, 2018.
- Revision received May 18, 2018.
- Accepted May 20, 2018.
- 2018 American College of Cardiology Foundation
- Filardo G.,
- Graca B.D.,
- Sass D.M.,
- et al.
- ↵Center for Workforce Studies. Physician specialty data book, Association of American Medical Colleges, 2016. Available at: https://www.aamc.org/data/workforce/reports/457712/2016-specialty-databook.html Accessed October 25, 2017.
- ↵Royal Colleges of Physicians of the United Kingdom. Census of consultant physicians in the UK, 2002. The Federation of the Royal Colleges of Physicians of the United Kingdom, 2002. Available at: www.rcplondon.ac.uk. Accessed October 25, 2017.
- Blumenthal D.M.,
- Olenski A.R.,
- Yeh R.W.,
- et al.
- Piper C.L.,
- Scheel J.R.,
- Lee C.I.,
- Forman H.P.
- Shah D.N.,
- Huang J.,
- Ying G.,
- Pietrobon R.,
- O’Brien J.M.