Gender Distribution of Scholarship and Measures of National Recognition in Hospital Medicine

We reviewed PHM-specific scholarly publications, presentations, editorial boards, planning committees, awards, and society leadership for the 5 calendar years from 2015 to 2019. The gender distribution of authors, presenters, leaders, conference planners, editorial board members, and awardees was extracted using methodology established in similar studies.^9,12–17  This study was determined to be nonhuman subjects research by our institutional review board.

Historical data on the gender distribution of PHM was not readily available. Thus, we estimated the gender distribution of the field over time using the membership of the largest professional society, the American Academy of Pediatrics (AAP) Section on Hospital Medicine (SOHM). The membership list provided by the SOHM included members from 1999 to present and their self-described gender (N. Alexander, e-mail communication, January 15, 2021). Trainees and nonphysician members were excluded. Data on gender were tied to the year the member joined the SOHM; thus, the roster provides a historical snapshot of the gender distribution of the field as it changed over time. We divided the gender distribution of the field into 2 groups: (1) senior members were defined as those who joined between 1999 and 2009, and (2) total membership of field was defined as members who joined between 1999 and 2019.

Outcomes included gender distribution of (1) scholarly publications, (2) national conference speaking engagements, (3) national awards, (4) national leadership roles, and (5) conference planning committees and editorial boards for the 5 calendar years from 2015 to 2019.

1. Scholarly productivity was defined as: either first or senior authorship of original science manuscripts published in journals with dedicated PHM content, such as Hospital Pediatrics and Journal of Hospital Medicine (pediatric content only). Authors could be counted more than once in any given year if they served as first or senior author on more than 1 paper.

2. National conference speaking opportunities were defined as presentations given at conferences with dedicated PHM content, including: Pediatric Hospital Medicine Annual Conference, Society of Hospital Medicine (SHM) Annual Meeting (pediatric track), Pediatric Academic Societies Meeting (hospitalist research platform content only), AAP National Conference and Exhibition Section H Program, Pediatrics Review and Education Program (PREP) Hospital Medicine, and AAP Clinical Pediatric Hospital Medicine. We included any presented content with a named presenter in the published conference programs for the pediatric hospitalist specific content and assigned the presentation to the first presenter listed, including first authorship of platform-presented research abstracts and clinical conundrums, unless the presentation was invited. Presentations were deemed invited if labeled as such in the program, presented in an unopposed time slot, or part of a conference with invitation-only content. For invited presentations, if >1 person presented, each speaker was included as an individual instance. The same speaker could be included multiple times if they had multiple speaking opportunities on different topics; presentations repeated at the same conference were included only once.

3. National awards were defined as: PHM-specific national awards given by the jointly planned PHM Annual Conference (PHM Award for Outstanding Research, PHM Award for Educational Excellence, PHM Award for Quality and Safety, PHM Award for Excellence in Clinical Care, Regional Impact Award, Lifetime Achievement Award); the Jennifer Daru Memorial Award, given by Hospital Pediatrics; and the Award for Best Abstract given by the AAP SOHM at the AAP National Conference and Exhibition. Awardees were determined from published conference content or communications with the awarding organization.

4. National leadership roles were defined as elected and appointed leadership positions in national PHM organizations. To account for variable term lengths, each year in a leadership role was assigned as 1 instance and the year was attributed to the incoming leader for overlapping terms in the same calendar year. Elected leadership roles were defined as: AAP SOHM Executive Committee, Academic Pediatric Association Hospital Medicine Special Interest Group chairs, and SHM pediatric board members and committee chairs. Appointed leadership roles were defined as: conference planning committees, as well as the American Board of Pediatrics subboard for PHM and the editorial boards for Hospital Pediatrics and PREP Hospital Medicine. Names of editorial board members were collected from published mastheads in the first publication of each calendar year. Conference planning committees and other leadership roles were determined by conference programs, published annual reports, or communications with the leadership and/or administration of the associated society.

The gender of subjects was determined based on gendered first name, an Internet search matching subject name with an accompanying profile photo, or from accompanying text using gendered pronouns based on methodology of similar studies.^9,12–17  Two authors independently determined gender for all data in a blinded fashion, and a third author adjudicated any discrepancies between the first 2 authors. Fifteen discrepancies underwent adjudication.

Descriptive statistics using 95% confidence intervals (CIs) for proportions are reported. Differences between observed proportions of women authors, presenters, and leaders and the proportion of women in the field at large were evaluated using a 2-proportion z-test with the null hypothesis that the proportions are the same and significance set at P < .05. Correlation between gender distribution of planning committees and editorial boards and gender distribution of authors and presenters was assessed by Pearson’s correlation coefficient. Trends in the proportion of women for a given outcome over time were assessed using linear regression.

We estimate that the gender distribution of senior members of the field is 57% women (95% CI 54%–60%) based on 1069 physician members of the SOHM who joined between 1999 and 2009 and now have 10 to 20 years of experience. Six members did not report gender. We estimate that the overall gender distribution of the field is 69% women (95% CI 68%–71%) based on 4104 physician members who joined between 1999 and 2019. Thirty-nine members did not report gender.

We collected data on 570 manuscripts for the period 2015 to 2019. The gender distribution of first and senior authorship for each year is presented in Fig 1. First authors were 67% women (95% CI 63%–71%) for the 5-year period, which was not different from the current gender distribution of the field (69%) (P ≥ .05). Senior authors were 49% women (95% CI 44%–53%) during this time, which was less than the proportion of women senior members of the field (57%) (P = .002). There was increasing representation of women as first authors over time (P = .002); there was no trend in senior authorship over time.

We collected data on 1093 presentations meeting inclusion criteria for the period from 2015 to 2019. Program content was unavailable for Pediatric Academic Societies 2016. The yearly gender distribution of presenters at national conferences for submitted and invited content from 2015 to 2019 is presented in Fig 2. Presenters of submitted content were 69% women (95% CI 65%–72%), which was not different from the gender distribution of the field at large (69%). Presenters of invited content were 44% women (95% CI 37%–51%), which was less than the gender distribution of senior members of the field (57%) (P ≤ .001). There were no trends in submitted or invited content presenters noted.

The gender distribution of national award recipients in PHM (number of women compared with number of awards given) was: 2 of 5 (2015), 4 of 6 (2016), 4 of 6 (2017), 4 of 7 (2018), and 7 of 8 (2019). The overall proportion of women given national awards for the 2015 to 2019 period was 66% (95% CI 47%–81%) and was not different from the gender distribution of either senior members or the field at large during the same period.

The gender distribution of elected national society and appointed leadership roles is presented in Table 1. PHM elected leadership roles were filled by women 61% (95% CI 48%–73%) of the time for the 5-year period, which was not different from the gender distribution of the field at large. The gender distribution of appointed leadership roles, which included the Hospital Pediatrics editorial board, the PREP PHM editorial board, and American Board of Pediatrics PHM subboard, were filled by women 49% (95% CI 40%–57%) of the time for the 5-year period, which was discordant with the gender distribution of the field at large (P < .001) but not with the gender distribution of senior members of the field (P = .07).

The gender distribution of the conference planning committee is plotted against the gender distribution of the content chosen for the corresponding meeting in Fig 3. Smaller meetings and those occurring infrequently are presented in aggregate, whereas the PHM meeting was presented annually. There was a strong positive correlation between the proportion of women on the planning committee and the proportion of women presenters (r = 0.94). Editorial board composition and authorship for Hospital Pediatrics showed a low correlation (r = 0.34). We did not analyze Journal of Hospital Medicine since we only evaluated that journal’s pediatric content.

With respect to first authorship of original research, submitted presentations, national awards, and some leadership roles, we found women were represented in concordance with their proportion in the field of PHM. However, we found fewer than expected women as senior authors and invited speakers, even accounting for the likelihood that the gender distribution of senior PHM members (57%) appears to be less female than the field at large (69%). We also noted a strong correlation between the gender distribution of planning committees and the gender distribution of the presenters chosen by these groups.

Our data mirror the field of academic medicine at large, with overall improvement in equity over time but a persistent lack of parity in senior authorship. In 2006, Jagsi et al examined the “gender gap” in authorship of academic medical literature from 1970 to 2004. The proportion of women first and senior authors of original research increased, yet women were still underrepresented when compared with the field as a whole, even in pediatrics.¹⁷  Fishman et al looked at original research in 3 high-impact pediatric journals from 2001 to 2016 and found that the proportion of women as first authors increased from 40% to 57% and senior authors from 29% to 38%, though women authors remained underrepresented when compared with the field at large during this time period.¹²  Similar trends have been documented by other authors in the fields of pediatrics and adult hospital medicine.^9,13,14,18 

The gender gap in senior authorship for women in medicine has been attributed to many causes, including lack of mentorship/sponsorship, lack of funding, gender bias, and a paucity of senior women in medicine.^{6,15,17,19–22}  For PHM, the argument could be made that the proportional representation of women as first authors means that the pipeline is full, and the gender gap for senior authorship will close over time. An alternative perspective is that women might be experiencing a “sticky floor” and are unable to make the leap to senior author.²³  Another possibility related to our narrow sampling strategy is that senior women may be publishing or presenting at other journals or conferences than the ones we sampled; although we do not have a hypothesis as to why they would be doing so at differential rates when compared with men.

In our examination of overall speaking opportunities, women are underrepresented as invited speakers, even accounting for the likelihood that invitations will go most frequently to senior members of the field. Invited speakerships can be seen as a key component of building a national reputation required for promotion.^24,25  Gender disparities in invited speaking opportunities have been well documented in medicine.^9,25–28  Lack of women in invited speaking opportunities was identified as 1 of several barriers to academic promotion and advancement, leading to the “invisibility” of mid-career women physicians.²⁹  Delays in promotion to associate and full professor lead to less compensation over a career,^30,31  decreased research funding,⁴  increased attrition,^19,4  fewer senior role models, and fewer leadership, authorship, and editorial opportunities.^6,29 

One potentially actionable finding in our study is a strong positive correlation between the gender distribution of the conference planners and the gender of presenters at national conferences. This association suggests the importance of representation in planning committees and is supported by other studies.^32–34  A similar relationship has been noted with award selection committees and gender disparities in award recipients.³⁵  In addition to emphasizing the importance of the selection committee composition, the selection process itself has been demonstrated to impact the gender distribution of conference speakers. After the implementation of open-call peer review for the SHM Annual Meeting, overall representation of women speakers increased.³⁶ 

Assessing proportional representation in PHM leadership roles is difficult because of the small number of positions, making statistical analysis challenging and requiring aggregation of the data over several years. Women appear to be proportionally represented in elected leadership for the 2015-2019 period when the data are aggregated (61% women). There is an apparent increase in the number of women in leadership roles over the study period, specifically in the SOHM Executive Committee; however, the numbers are too small to draw firm conclusions. In recent years, PHM societies and conference planning committees have established term limits for leadership roles and transparent selection processes for elected leadership, which may have contributed to the gender representation found in our study. Term limits have been shown to improve proportional representation,³⁷  and the National Institutes of Health recently announced term limits for its chiefs to increase gender and racial diversity.³⁸  Although our study was limited to gender, we recognize that other barriers to equity exist, especially for women with intersectional identities. Benchmarking both gender and racial distribution in leadership roles and selection committees is an important first step to understanding when inequities are evident.

Our paper has several limitations. First, given that there is no central registry for membership in our field, we had to use proxies for total membership and seniority. Our estimate of 69% women in the field corresponds with available published data on the current gender distribution of the field: 70% of the 2019 American Board of Pediatrics PHM Board examination applicants were women, and a random sample of 25% of university-based PHM programs demonstrated 72% women faculty members.⁷  We selected measures of academic career advancement because these are in the public domain and recognize that scholarly productivity is not limited to university-based hospitalists. Women and men may join and leave academic hospital medicine at different rates, though no published data exist on this phenomenon. We attempted to control for this possibility by estimating the gender distribution of senior members of the field by the year they joined the AAP SOHM, which did suggest fewer senior women. Even using this more conservative estimate of the proportion of senior women, we still found these women appeared underrepresented. We only assessed publications in 2 hospital medicine-focused journals and limited conference venues; therefore, publications in other journals were not included in our study because identifying an author or presenter as a hospitalist in nonhospital medicine-focused venues was not feasible. This choice poses two potential problems:

1. we may have missed a large amount of hospitalist output in the larger pediatric literature; and/or

2. we have likely included some nonhospitalists in our gender distribution calculations.

Although both issues are likely present in our analysis, we believe there is little evidence to suggest either issue would bias our results differentially for women versus men. Our gender assignment methods could result in inaccuracies despite relying on methods used by other authors.^{9,13,14,16–18}  Finally, we used gender as a binary term; however, we recognize that gender exists on a spectrum, which is not reflected in this work.

Our study demonstrated representative gender distributions for aspects of scholarly productivity in PHM, although a lack of gender parity persists in senior roles, including senior authorship and invited speaking opportunities. Additional study is needed to understand why these disparities exist and to determine the barriers to increasing the proportion of senior women.

We thank Niccole Alexander, AAP senior manager of hospital and surgical subspecialties, whose tireless efforts to collect leadership, award, and membership data were vital to this work.