eAppendix. Web of Science Search Terms
eTable 1. Number and Mean Proportion of Female Speakers per Year of Conferences Divided by Specialty
eTable 2. Specialties of Conferences Included in the Final Analysis
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Ruzycki SM, Fletcher S, Earp M, Bharwani A, Lithgow KC. Trends in the Proportion of Female Speakers at Medical Conferences in the United States and in Canada, 2007 to 2017. JAMA Netw Open. 2019;2(4):e192103. doi:10.1001/jamanetworkopen.2019.2103
What were the proportions of female speakers at academic medical conferences held in the United States and in Canada during the last decade?
In this cross-sectional analysis of 181 medical conferences in 2007 and from 2013 through 2017, the proportions of female speakers significantly increased from 24.6% to 34.1%. These proportions were similar to the percentages of practicing physicians who were women during the same time frame.
Although the proportion of female speakers has increased during the last decade, women are underrepresented at medical conferences.
Gender equity is a prominent issue in the medical profession. Representation of female physicians at academic meetings has been identified as an important component of gender equity; however, this topic has not been systematically assessed.
To determine the trend during the last decade in the proportion of speakers who were women at major academic medical conferences held in Canada and in the United States.
Design, Setting, and Participants
A cross-sectional analysis was conducted examining the gender of speakers listed in meeting programs of medical conferences held in Canada and in the United States in 2007 and from 2013 through 2017. Eligible conferences were identified using a sensitive search strategy, and a previously validated tool was used to analyze each meeting speaker list and to assign a proportion of female speakers. Conferences held in English language, hosted in Canada or the United States, and targeted to a physician audience with 100 or more attendees were included. The comparison group was active physicians in Canada and in the United States.
Main Outcomes and Measures
The mean of the proportion of female speakers at each conference per year.
In total, 181 conferences with 701 individual meetings were analyzed, including 100 medical and 81 surgical specialty conferences. The proportion of women ranged from 0% to 82.6% of all speakers. The mean (SD) proportion of female conference speakers for all meetings analyzed significantly increased from 24.6% (14.6%) for 40 meetings in 2007 to 34.1% (15.1%) for 181 meetings in 2017 (P < .001). The mean proportion of female speakers at medical specialty conferences was 9.8% higher (SE, 1.9%; P < .001) than the mean proportion of female speakers at surgical specialty conferences for all years analyzed. The mean proportion of female speakers at conferences was similar to the mean proportion of active female physicians across all specialties in the United States and in Canada for all years analyzed.
Conclusions and Relevance
Although our findings indicate that the proportion of female speakers at medical conferences increased during the last decade, women continue to be underrepresented. Speaker invitation and selection at conferences represent important opportunities to influence gender equity within medicine.
Gender equity in medicine is a prominent topic of medical editorials, scientific literature, and mass media. Gender disparity in the medical profession has been well documented,1-3 including evidence for inequities in evaluation, hiring, compensation, harassment, promotion, and advancement.1,3-11 Contributors to gender inequity in medicine include implicit and explicit bias, cultural factors, unsupportive work environments, and representation of women in medicine.1,2,4,6,11-14
Visibility or representation of female physicians is an important facet of gender equity.15,16 Evidence from social science literature suggests that higher female representation leads to lower gender bias in the community.17 In academic settings, higher numerical representation of women correlates with improved outcomes, including improved career satisfaction and retention, for women and other minority groups.15 In medicine, a lack of role models and mentors has been identified as a barrier to advancement in surgical specialties and academic internal medicine that disproportionately affects female physicians.1,6 Underrepresentation occurs when the number of female physicians in visible positions is less than the proportion expected based on the number of female physicians overall.
Although women make up approximately 33% of the physician workforce in the United States and 42% in Canada, they are variably underrepresented in academia.18-20 Approximately one-third of academic physicians in Canada and the United States are women, and the proportion varies greatly by specialty1,5; only 8% of academic surgeons and 16.5% of academic cardiologists are women.21 Female scientists have fewer total publications and are less likely than their male colleagues to be listed as first author, even when first authorship is shared by coauthors of different sexes.21-23 Women are underrepresented as speakers at medical grand rounds and, when they do present, are less often addressed by their professional title than male presenters are.12,24,25
Although female representation at academic meetings has been identified as an important gender equity issue, the proportion of conference speakers who are women has not yet been systematically measured across different medical subspecialties, to our knowledge. Presenting at a medical conference is an opportunity for career advancement for female physicians and also represents an opportunity for trainees and colleagues to see female leaders in academia as role models.15,16
Underrepresentation of female physicians has been described at critical care conferences but has not been evaluated in other medical specialties.20,26 The phenomenon of all-male panels at scientific meetings recently gained attention on social media via the hashtag “NoMoreManels” and has prompted appeals to improve the gender balance in speaker programs.16,27-29 In this study, we aimed to measure the proportion of female speakers at medical conferences and to characterize trends in female representation between 2007 and 2017.
We consulted a librarian to develop a systematic search strategy to identify relevant medical conferences. Conferences were identified using Web of Science Conference abstract databases by searching with key words developed using a list of medical specialties (searched February 18, 2018; search terms given in the eAppendix in the Supplement). The search was filtered for English-language conferences held in Canada or the United States. Additional conferences were solicited by email from program directors and chief residents in all accredited medical specialties.30Conference refers to the host society or association, and meeting refers to the event held by the conference in a specific year. This report follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines for reporting the results and the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guidelines when describing the search methods and results. The data analyzed were publicly available on conference websites, and therefore our institutional ethics review board waived the need for a review of this project and for obtaining participant informed consent.
Conferences held outside Canada and the United States were excluded owing to potential cultural differences that may influence the gender of selected conference speakers. Non–English-language conferences were excluded. Conferences that did not have a 2017 meeting, were a chapter event of a larger society, were not held annually, were not intended for a primarily physician audience, or had fewer than 100 attendees were excluded. These criteria were intended to limit our analysis to major meetings without overrepresenting conferences with multiple events per year. To assess trends in the proportion of female speakers over time, we included 2007 data and 2013 through 2017 data when the conference program was available.
Each identified conference was screened for eligibility by 2 independent reviewers (S.M.R. and K.C.L.), who applied the inclusion and exclusion criteria to the conference title. The conference website was then assessed in detail for inclusion by both reviewers. Disagreements were adjudicated by a third reviewer (A.B.).
The meeting program or faculty lists from eligible conferences were downloaded from the conference website or requested from the host association via email. Names were extracted from the meeting programs for 2007 and for 2013 through 2017, if available. When known, speakers who presented more than once at each conference were listed for each credit in the program to fully represent the gender balance at each conference. When specified, speakers who only presented posters were excluded.
Each list of names was analyzed using the Gender Balance Assessment Tool (GBAT)31,32 to assign a proportion of female speakers at that meeting during that year. The GBAT is a validated web-based tool that can identify names in a text-based document and assign a probabilistic prediction that the first name belongs to a female.31 The GBAT uses the genderize.io algorithm to assign each name a probability of belonging to a gender on the basis of social media data.31,33 The GBAT then aggregates the probabilities to return a likely proportion of female names in the document.31 The GBAT returns a single value: the proportion of names that are likely to be female, not the absolute number of female speakers at each conference. All conferences were weighted equally regardless of the number of attendees or presenters. All analyses were performed using the proportion of female speakers from each included meeting. The GBAT does not provide the number of names analyzed or a confidence interval.
Compared with hand-coding of names, the GBAT is more cost-effective, faster, incrementally adaptable to trends in names, and multinational and assigns a probability of gender rather than make a dichotomous assignment.31,33 Validation of the GBAT has demonstrated an area under the curve of 0.93 and a net error rate of 1.42% that usually overestimates the probability of a name being female.31 The GBAT is unable to assign gender to names that begin with initials.31
The comparison group was the mean proportion of practicing physicians in the United States and in Canada who were women in each specialty in 2007, 2010, 2013, and 2015 (the last year for which such data were available in the United States). This proportion was calculated by determining the proportion of practicing female physicians in each specialty and calculating the mean of those proportions for each year. This calculation was performed so that female representation in each specialty contributed equally to the comparison group regardless of specialty size, which aligned with our methods for determining the mean proportion of female speakers for each specialty. The mean proportions of women for all specialties and for medical specialties were similar to the actual percentage of women for these groups. The mean proportion of female surgeons differed from the absolute percentage of women in surgical specialties by approximately 7% per year; this difference may be attributable to the low proportion of women in numerous, smaller surgical specialties (Table 1).
The Cohen κ coefficient was used to determine agreement of conference inclusion between the 2 reviewers (S.M.R. and K.C.L.). The mean and SD were calculated using the proportion of speakers who were women at each conference for all conferences, medical conferences, and surgical conferences. The change in the proportion of female conference speakers over time (per year) was analyzed using a linear mixed-effects model (LMM), with individual conferences (random effects, assuming random intercepts with fixed mean) and clinical area (surgical or medical; fixed effect) specified in the model as covariates. The year was considered a fixed effect. The LMM analysis was used because it accounts for multiple responses from the same conference being expected to be more similar than responses from different conferences. Furthermore, LMM analysis permits missing data. The LMM analyses were performed with R (R Foundation for Statistical Computing, version 3.3.4) using the lme4 package.34 The P values were obtained with the lmerTest package.35 All tests were 2-sided, and P < .05 defined statistical significance.
The search strategy identified 4942 conferences. After removing duplicates, 2887 conferences were screened, of which 2041 did not meet inclusion criteria. The conference website was reviewed for 846 conferences, and 371 met full inclusion criteria (κ for agreement between the 2 reviewers, 0.805). Of those, 701 meeting programs were available from 181 conferences, which included 100 medical and 81 surgical specialty conferences. (Figure 1; eTable 1 in the Supplement). Meeting data from the 181 conferences were incomplete for 141 conferences (78.0%) in 2007, 77 (42.6%) in 2013, 66 (36.5%) in 2014, 57 (31.5%) in 2015, and 45 (24.9%) in 2016. Identification of 2017 data was an inclusion requirement and therefore available for 100% of conferences (Table 2; eTable 2 in the Supplement). The LMM statistical analysis accounts for missing data in effect estimates.
Between 2007 and 2017, the mean (SD) proportion of conference speakers who were female significantly increased from 24.6% (14.6%) for 40 meetings in 2007 to 34.1% (15.1%) for 181 meetings in 2017 (P < .001) (Figure 2). There was a wide range (0%-82.6%) in the proportion of female speakers for all meetings included. There was a 0.97% increase per year in the mean proportion of female speakers in this time frame (SE, 0.13%, P < .001). This rate of increase in the proportion of female speakers was not significantly different between medical specialty and surgical specialty conferences. In addition, 82 meetings (12%) had more than 50% female speakers. Between 2007 and 2015, the mean (SD) proportion of female physicians practicing in the United States and in Canada also increased, from 26.1% (13.1%) to 32.4% (15.2%) (data not available for 2016 and 2017).18,19,36-38 The proportions of practicing female physicians were similar to the proportions of female conference speakers for all years in which data were available (Figure 2).
Medical specialty conferences had a mean (SE) of 9.8% (1.9%) more female speakers than surgical specialty conferences for each year studied (P < .001). The mean (SD) proportion of female speakers at medical specialty conferences significantly increased from 29.9% (14.1%) in 2007 for 18 meetings to 38.8% (12.1%) for 100 meetings in 2017, with a range of 0% to 79.4% (P < .04) (Figure 3A). Between 2007 and 2015, the mean proportion of practicing physicians in medical specialties in the United States and in Canada who were female increased from 30.4% to 37.6%.18,19,36-38
The proportion of female speakers at surgical specialty conferences (including obstetrics and gynecology) also significantly increased (P < .02) between 2007 and 2017, from 20.1% (SD, 13.7%; 22 meetings) to 28.4% (SD 14.1%; 80 meetings) and ranged from 0.1% to 82.6% (Figure 3B). From 2007 to 2015, the mean proportion of female surgeons in the United States and in Canada increased from 14.8% to 18.3%.18,19,36-38 The mean proportions of conference speakers who were women were greater than the proportions of surgeons who were women for all years analyzed.
Low representation of female physicians in academia is an important aspect of gender inequity in medicine. Gender disparities in medical academia have been described for compensation, the granting process, authorship, and career advancement.1,3,8,23,39 Reports of gender disparities among conference speakers have been largely anecdotal.28 Previous work to characterize the gender gap in medical conferences has been limited by the size and scope of conferences included, nonsystematic inclusion of conferences, and lack of a validated strategy to assess the gender of speakers.16,26 Because conferences represent an important opportunity for role modeling, mentorship, and career advancement, understanding gender disparities in conference speakership is important for understanding gender inequity in medicine.16
We systematically identified medical conferences in Canada and in the United States and determined the proportion of female speakers at these conferences using a validated name-gender algorithm. In addition, we examined the trend during a 5-year period in the proportion of conference speakers who were women. Our results showed that women compose a minority of conference speakers, with a mean proportion of 34.1% for conferences held in 2017. Of the 701 meetings analyzed, only 82 meetings had more than 50% female speakers and 619 had less than 50% female speakers.
When framing these results in the context of existing gender demographics of physicians in Canada and the United States, we noted that the gap between the mean percentage of practicing physicians across all specialties who were female and conference speakers who were female decreased over time. In 2007, there was a 1.5% gap between female conference speakers and practicing female physicians in Canada and in the United States (26.1% compared with 24.6%). By 2015, the last year for which data on the gender of practicing physicians is available, there was a slightly higher proportion of conference speakers who were women than of practicing physicians who were women (32.7% compared with 32.4%).18
The increase in the proportion of female speakers in most specialties during the last decade has mirrored trends within the physician workforce.18,19,36-38 Although there is a lower proportion of female speakers at surgical specialty conferences than at medical specialty conferences, the mean proportion of conference speakers at surgical meetings who were female was greater than the mean proportion of practicing surgeons who were female. This finding suggests that the perceived gender gap in speakers at physician conferences likely represents the overall gender gap in academic medicine rather than a bias specific to conferences. This disparity is most obvious in surgical disciplines because only 25.9% of surgeons in the United States and in Canada were women in 2015.18 Low female visibility at these academic conferences may therefore be a product of low representation in surgical specialties.
Alternately, low representation of female physicians in prestigious academic positions, such as conference speakership, may contribute to low proportions of female physicians in select surgical and medical specialties.18,19 Although women have comprised more than 50% of medical school graduates since 1996 in Canada40 and have approached parity with male graduates in recent years in the United States,41 female physicians make up less than 50% of surgeons and academics.3,18,19 Low representation of female physicians in these opportunities signals to female medical graduates that these career tracks are not friendly to women and also contributes to lack of female mentorship.6
The major strength of this study is that we used a validated tool, the GBAT, to assess the probability that a given first name is female rather than dichotomously assigning gender.31,33 Many first names have a degree of gender ambiguity; thus, previous studies that assigned gender dichotomously as male or female on the basis of the first name are subject to bias.16,20,26 Furthermore, our search strategy systematically identified a large number of academic conferences of different subspecialties; therefore, our results provided a more balanced representation of overall trends in academic medicine compared with previous publications that only assessed select meetings.16,20,26
Our study has some important limitations. The results did not account for conference size or importance because the proportions of female speakers at each conference were given equal weight when combining the results. We chose not to weigh conferences differentially because then smaller specialties could not be compared with larger specialties. Our result is therefore the mean proportion of speakers at each conference who were women and not the total proportion of all conference speakers who were women. In addition, the algorithm on which the GBAT is based has been shown to overestimate the proportion of women; however, this bias will affect all measured conferences equally and therefore should not influence our conclusions about the trend in female conference speakers over time.33 Owing to cultural, organizational, and structural factors that contribute to the gender gap and biases in medicine, our findings are not generalizable to conferences outside of North America or to nonmedical academic conferences.
The GBAT cannot assess representation of individuals belonging to different ethnic or racial backgrounds, does not capture how people self-identify their gender, and does not account for nonbinary genders.31 We acknowledge that women of additional visible minorities and nongender conforming physicians face more significant barriers and may be further underrepresented in medical conferences; this underrepresentation requires further assessment.
There was a wide range in the proportion of female speakers at medical conferences. On the basis of this result, we advocate for strategies to promote inclusivity in speaker invitation and selection for all conferences.16,26,29 We hypothesize that the low proportion of female speakers at medical conferences reflects broader gender inequity within the medical profession, particularly in subspecialties where the majority of physicians are men. It has been shown that the presence of female role models in male-dominated career streams can increase engagement of young women.17 Exposure to female speakers at medical conferences may be a means of encouraging female medical students and residents to choose specialties that have historically been male dominated. Strategies to promote inclusivity of female speakers at academic conferences may therefore represent an important opportunity to influence gender equity within medicine.
To our knowledge, this is the first study to systematically identify medical conferences and to characterize the perceived gender inequities in conference speakership using a validated instrument. Our findings indicated that the proportions of female speakers at medical conferences increased during the last decade and that the gap between the proportion of female conference speakers and practicing physicians who were women decreased. Conference organizers should strive for diversity when inviting speakers.
Accepted for Publication: February 22, 2019.
Published: April 12, 2019. doi:10.1001/jamanetworkopen.2019.2103
Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2019 Ruzycki SM et al. JAMA Network Open.
Corresponding Author: Shannon M. Ruzycki, MD, Division of General Internal Medicine, Department of Medicine, University of Calgary, Room 1422, 1410 Health Sciences Centre, 3330 University Dr NW, Calgary, AB T2N 4N1, Canada (firstname.lastname@example.org).
Author Contributions: Dr Ruzycki had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Ruzycki, Lithgow.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Ruzycki, Fletcher, Earp, Lithgow.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Ruzycki, Fletcher, Earp.
Obtained funding: Bharwani.
Administrative, technical, or material support: Ruzycki, Fletcher, Bharwani, Lithgow.
Supervision: Ruzycki, Bharwani, Lithgow.
Conflict of Interest Disclosures: None reported.
Funding/Support: This project was funded by the Department of Medicine and the General Internal Medicine Fellowship Program, both at the University of Calgary, Cumming School of Medicine.
Role of the Funder/Sponsor: The funding sources were not involved in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: Natasha Mueller, BEd, provided assistance with data collection; Melanee Thomas, PhD, Department of Political Science, University of Calgary, Alberta, Canada, provided guidance on the literature review and methods; and Kara Nerenberg, MD, MSc, Department of Medicine, Obstetrics and Gynecology and Community Health Sciences, University of Calgary, Alberta, Canada, provided comments on the manuscript. None of these individuals received financial compensation for the stated contribution.
Create a personal account or sign in to: