Glass Ceiling in Hand Surgery: Publication Trends by Gender

Abstract

Background

Women are frequently underrepresented across surgical subspecialties and may face barriers to academic advancement. Abstracts presented at American Society for Surgery of the Hand annual meeting (ASSH-AM) highlight some of the top research in hand surgery. We sought to explore differences in abstract characteristics and publication rates based on senior author gender.

Though there have been increasing efforts at inclusivity in orthopedic and plastic surgery, women face several barriers to entering the field, publish less frequently, and are underrepresented in leadership positions. Understanding the stages at which discrepancies in research productivity exist may help to address these challenges.

Methods

Abstracts from the 2010-2017 ASSH-AMs were reviewed to determine basic characteristics. Author gender was determined through both a search of institutional websites for gender-specific pronouns and inference of gender based on first name. Subsequent full manuscript publications corresponding to the abstracts were identified through a systematic search of PubMed and Google Scholar.

Results

A total of 560/620 (90.3%) abstracts from 2010-2017 had an identifiable senior author gender (14.5% female). No differences were noted between male- and female-authored abstracts regarding study design including sample size or level of evidence. Female senior authors were more likely than males to author abstracts focused on pediatrics (19.8% vs 9.4%, p=0.01) and were more likely to collaborate with female first authors (41.3% vs 20.0%, p<0.01). Abstract publication rates were lower for female senior authors versus male senior authors (61.7% vs 74.5%, p=0.02).

Conclusion

The number of abstracts with female senior authors had similar representation to the membership proportion of women in the ASSH. There were few differences in abstract characteristics based on senior author gender, though senior authors tend to collaborate with investigators of the same gender. Abstracts authored by females were published 13% less frequently overall, meriting further exploration.

Level of Evidence: III

Introduction

Despite an increasing proportion of female medical school graduates,¹ surgical residents across nearly all subspecialties are predominantly male.² Orthopaedic surgery in particular lacks female representation, with only 14% of all orthopaedic surgery residents identifying as female (from 2016-17) compared to 40% representation (from 2010-16) in plastic surgery residencies.^3,4 Further, women compose just 6.5% of all American Academy of Orthopaedic Surgeons (AAOS) members and 24.0% of American Society of Plastic Surgery (ASPS) members.^4,5 This gender discrepancy is even more evident among leaders in the field, with only 1.7% and 8.7% of academic orthopaedic and plastic surgery department chairs identifying as female, respectively.^5,6 Female orthopaedic surgeons more frequently pursue hand (24%) and pediatrics (23%) subspecialty training,⁷ and therefore the American Society for Surgery of the Hand (ASSH) saw a rise in female representation among their active membership from 9% in 2010 to 14% in 2016.⁸

Several studies have attempted to explain not only the low proportion of women in surgical specialties, but also the gap in leadership roles.^9,10 Proposed theories include systemic discrimination and a lack of career role models that prevents or deters women from pursuing careers in surgery,⁹ with up to 88% of female surgical residents reporting gender-based discrimination during residency.¹¹ Furthermore, 91% of female attending surgeons experience gender bias in their careers, including discrimination, bullying, and sexual harassment, leading to negative effects on career promotion and job satisfaction that may lead them to leave the field entirely.¹¹ Some strides in hand surgery have been made in recent years, with female members of the ASSH applying for and achieving leadership positions at greater rates and earlier time points than male colleagues, respectively.¹² However, it takes time for efforts aimed at addressing these inequalities to lead to measurable changes in representation

Publication productivity is one well-known marker of success in academic medicine.¹³ Productivity in the fields of medical education and research are considered crucial for accomplished academic physicians, with the magnitude and quality of research contributing significantly to advancement recommendations. Historically, male faculty members have, on average, higher research productivity than their female colleagues, with varying reports on the impact of gender on academic promotion when adjusted for the total magnitude of research.^13–15 While past work in orthopaedic and plastic surgery has explored discrepancies in indices of research productivity such as the H-index,^13,15 no study has examined differences in abstract characteristics based on senior author gender or the resulting publication rates at a prominent academic society’s annual meeting. American Society for Surgery of the Hand annual meeting (ASSH-AM) podium presentations are known to be published at rates approaching >50-70%.^16,17 Given the higher likelihood for women to subspecialize in hand surgery, the ASSH-AM presents an ideal venue to explore this concept. The primary aims of this study therefore were to examine differences in abstract characteristics and rates of abstract publication with a female or male senior author with a focus on differences that might explain any observed discrepancies. We hypothesized that female senior-authored abstracts would have lower publication rates compared to their male colleagues.

Methods

An observational study was performed examining all abstracts presented at the podium (i.e. only podium presentations) at each ASSH-AM from 2010 to 2017. Abstracts were reviewed by two investigators to determine basic characteristics including first author gender, last author gender (the last author was assumed to be the senior author), subject age-group (i.e. adult versus pediatric), number of subjects, and study design (level of evidence, single-center vs. multicenter, randomized-controlled design [RCT], use of a large national database). Though what constitutes a high-quality study is debatable, many of these characteristics have been cited in the literature as contributing to higher quality research, or at the very least, research that is more generalizable. Author gender was determined through a search of institutional websites for gender-specific pronouns (he/him/his/himself and she/her/hers/herself) when available and inference of gender based on author first name when gender-specific pronouns were unavailable. If the gender of the senior author could not be determined with this methodology, the abstract was excluded from further analysis.

Abstracts were then classified based on broad topic categories with particular attention paid to topics that have received increased exposure in the hand surgery literature, including: basic science, anatomy or biomechanics (non-clinical abstracts); cost analysis; opioids and pain management; and medical education. Abstracts were also categorized using more specific, hand surgery-focused categories – in part based on categories used by two 2019 studies by Kuczmarski et al. and Lemme et al. – including: distal radius fracture; flexor tendon injury or repair; carpal tunnel syndrome; cubital tunnel syndrome; Dupuytren’s contracture; thumb carpometacarpal (CMC) arthritis; peripheral nerve injury or repair; and congenital hand or upper extremity anomalies.^16,18 Abstracts focusing on CMC arthritis, Dupuytren’s contracture, carpal tunnel release, and flexor tendon injury or repair were marked as adult-focused for instances when a specific age-range was not noted, given that the overwhelming proportion of these studies focused on adult patients.

Subsequent full manuscript publications corresponding to the abstracts (i.e., excluding abstracts published in supplemental form) were identified by two authors (JTB and AP) through a systematic search of PubMed and Google Scholar using the abstract title or a combination of author names. For instances where the first investigator could not identify an abstract with a corresponding journal publication, a second investigator performed the search utilizing a similar methodology. Time to publication was then calculated as the difference between the month of publication (when they first appeared online) and the month of the presenting meeting. Only abstracts presented at the 2010-2017 ASSH meetings were analyzed, despite more recent abstracts being available, in order to permit a minimum follow-up window of at least two years. Journal impact factors were determined from the Journal Citation Reports produced by Clarivate Analytics.¹⁹

Chi-squared and Fisher’s exact tests were utilized to compare categorical variables, while Mann-Whitney U tests were used to compare continuous variables. Binomial testing was used to compare the proportion of female senior authors to the reported proportion of US female membership in the ASSH.⁸ Binary logistic multivariable regression analysis was used to assess the variables predictive of publication. All analyses were conducted using a significance threshold of p<0.05.

Results

A total of 620 of 623 abstracts identified from 2010 to 2017 were evaluable, with three having been withdrawn or not available. Only 560 of these 620 (90.3%) evaluable abstracts (Table 1) had a senior author with an identifiable gender. Of these, there were 81 abstracts (14.5%) with a female senior author, which was similar to the reported proportion of US female membership in the ASSH (14.3%, 1-sided p=0.47). Abstracts with a female senior author were significantly more likely than abstracts with a male senior author to have a female first author (41.3% vs 20.0%, p<0.01). The proportion of female senior authorship was similar across all years evaluated, with a low in 2017 at 10.1%, high in 2013 at 19.0%, and a mean across all study years of 14.3 ± 3.1% (Figure 1). There was no change in female senior author proportion between the first half (2010-2013) and the second half (2014-2017, p=0.40) of the study period. No differences were noted between male and female-authored abstracts with regard to study design including sample size, level of evidence (LOE), RCT proportion, or multicenter collaboration (Table 1). Women and men also utilized large national databases as a data source in equal proportion. Women were more likely to author abstracts focused on pediatric patients (19.8% vs 9.4%, p=0.01).

Table 1.

Characteristics of ASSH Abstracts by Gender of Senior Author, 2010-2017

	Male	Female	P-value
Number of Abstracts	479 (85.5)	81 (14.5)	-
Percent Published	357 (74.5)	50 (61.7)	0.02
Time to Publication (months)	13.0 (18.0)	17.5 (21.0)	0.36
Impact Factor	2.1 (0.5)	2.1 (0.3)	0.97
Level of Evidence
1 or 2	93 (19.4)	12 (14.8)	0.33
3, 4, 5, or Not Clinical	386 (80.6)	69 (85.2)
Median Number of Patients*	51 (97)	48 (82)	0.99
Focus
Distal radius fracture or repair	57 (11.9)	11 (13.6)
Flexor tendon injury or repair	19 (4.0)	5 (6.2)
Carpal tunnel syndrome or release	21 (4.4)	3 (3.7)
Cubital tunnel syndrome or release	6 (1.3)	1 (1.2)	-
Dupuytren’s contracture	24 (5.0)	1 (1.2)
Thumb CMC Arthritis	17 (3.5)	6 (7.4)
Peripheral nerve injury or repair	64 (13.4)	12 (14.8)
Congenital upper extremity deformity	15 (3.1)	5 (6.2)
Topics
Basic science, anatomy, or biomechanics	100 (20.9)	18 (22.2)
Cost analysis	10 (2.1)	3 (3.7)	-
Opioids and pain management	10 (2.1)	2 (2.5)
Medical education or publication	5 (1.0)	0 (0)
Randomized controlled trial	30 (6.3)	5 (6.2)	0.98
Age Group
Pediatrics	45 (9.4)	16 (19.8)	0.01
Adult	252 (52.6)	42 (51.9)
Unclassified	182 (38.0)	23 (28.4)
Multicenter collaboration	27 (5.6)	7 (8.6)	0.30
Large National Database	22 (4.6)	5 (6.2)	0.54
Female First Author†	92 (20.0)	33 (41.3)	<0.01

Values reported as number (%) or median (IQR)

*Excluded large sample database studies

†Only available for 540 abstracts with both first and last author gender determined

Figure 1.

Proportion of Female Senior Authorship by Year. The mean proportion across all years was 14.3 ± 3.1%, which was lowest in 2017 at 10.1% and highest in 2013 at 19.0%.

Of 560 abstracts with known author gender, 407 (72.7%) were published as of January 1, 2020. The most common journals of publication were the Journal of Hand Surgery, American Volume (N=161), The Journal of Bone and Joint Surgery (N=33), and HAND (N=30, Table 2). Abstracts with male senior authors were more likely to have been published overall (p=0.02) compared to those authored by women, with publication rates of 74.5% and 61.7%, respectively. However, male senior authors were not necessarily faster to publication than females (13.0 ± 18.0 months vs 17.5 ± 21.0 months, p=0.36), and there was no difference in the impact factor of publishing journals between genders (2.1 ± 0.5 vs 2.1 ± 0.3, p=0.97). In multivariable analysis when accounting for presenting year, abstract focus and topic, pediatrics versus adult focus, level of evidence, multicenter collaboration, and senior author gender, only female senior author gender (OR 0.549, 95% Confidence Interval [CI] 0.327-0.925, p=0.02) and a peripheral nerve injury or repair focus (OR 0.480, 95% CI 0.271-0.848, p=0.01) were predictive of a lower likelihood of publication.

Table 2.

Top Ten Journals of Publication For ASSH-AM Podium Presentations With Identifiable Gender (2010-2017)

Journal	Number	Impact Factor
Journal of Hand Surgery, American Volume	161 (39.6)	2.09
Journal of Bone and Joint Surgery	33 (8.1)	4.72
HAND	30 (7.4)	0
Journal of Wrist Surgery	16 (3.9)	0
Plastic and Reconstructive Surgery	15 (3.7)	3.95
Journal of Pediatric Orthopaedics	14 (3.4)	2.05
Journal of Hand Surgery, European Volume	12 (2.9)	2.23
Annals of Plastic Surgery	7 (1.7)	1.45
Journal of Shoulder and Elbow Surgery	6 (1.5)	2.87
2 Journals tied with 5 Publications*	5 (1.2)	2.35, 3.62

*Journal of the American Academy of Orthopaedic Surgeons, Tissue Engineering Part A

Discussion

Despite ongoing efforts, gender disparities persist in surgery, though female representation is improving for students and residents.^10,20 For hand surgery in particular, this has translated to increased representation among leadership positions. Understandably there will be a lag in appropriate female representation at the highest levels of academic leadership as reforms are made.¹² Over our eight-year study, female senior authors produced a nearly identical proportion of abstracts (14.5%) compared to a reported 14.3% composition of US-based ASSH membership.⁸ This is nearly twice that of female membership in the AAOS, which may indicate that the field of hand surgery is making strides towards inclusivity.⁴

Prior studies on abstract publication rates in medicine are mixed, with some finding lower publication rates for female senior authors²¹ and others observing no difference.²² Our study shows that male senior-authored ASSH abstracts were published 13% more often (p=0.02), though there were no observed differences in design (e.g. RCT, multicenter, etc.) or level of evidence, here used as proxies of study quality. Noting the difficulties that exist in determining what constitutes higher research quality, we should point out that this study evaluated ASSH-AM podium presentations, which likely have a higher baseline study quality than traditional journal submissions because all abstracts undergo a thorough submission and review process. Other reasons for this difference must therefore be explored, as academic promotion relies at least in part on research productivity.^13,14

Previous investigations demonstrate that women publish less frequently,^13–15,23 receive less research funding²⁴ and are less frequently promoted²³ compared to men. While there was no difference in LOE or design by gender, female authors more often produced pediatric-focused abstracts. This finding is consistent with data showing higher rates of pediatric subspecialization for women across nearly all fields² and past studies demonstrating female orthopaedic surgeons most frequently subspecialize in hand or pediatrics.⁷ Thus, one potential reason for the lower publication rates of female-authored abstracts is this focus on pediatric patients and congenital topics. Pediatric hand conditions may be considered “too specialized” for certain journals and could face higher rejection rates,²⁵ which could explain, at least in part, the observed difference in publication rates. Interestingly, we also identified a focus on peripheral nerve injury or repair as an independent predictor of lower publication likelihood, separate from author gender, which similarly could be too specific for many journals.

Gender bias in any form presents an obstacle to both advancement and recruitment. We found that female senior-authored abstracts were twice as likely to have a female first author, which suggests that established female hand surgeons more commonly mentor young female investigators. Previous studies have noted the key role of mentorship in deciding future specialty as well as career satisfaction and success.⁷ Mentors not only recruit women into surgery, but also encourage aspiring female surgeons to pursue academic promotion.²⁰ Unfortunately, a lack of women in leadership roles results in decreased likelihood of choosing surgical specialties among female medical students² and contributes to the difficulties women face in finding mentors with similar experiences.²⁶ A common reason women may not pursue academic medicine is a lack of female mentors in a generally male-dominated field, with strong mentorship noted as a top motivator of subspecialty choice among female orthopaedists.⁷ Fewer mentors could lead to lower research productivity and the observed discrepancy in publication rates by senior author gender. While significant strides have been made to promote diversity in the field of hand surgery, it is important to recognize that, unfortunately, it will take time to see the full effect of these efforts with regard to research, mentorship, and leadership.^8,12,27

Another important barrier that women in medicine may face – one that has been posited by some as a significant cause of fewer publications and leadership roles – is career disruption relating to maternity leave and parenting. Several studies report that female physicians work fewer hours²⁸ and spend more time on familial obligations²⁹ than their male counterparts, inevitably leaves less time for research. This could be particularly true for longer manuscript writing in comparison to relatively faster abstract preparation. However, many female surgeons continue scholarly pursuits during maternity leave, working similar hours (>60/week) to male colleagues and pursuing advanced degrees or continuing full-time research.³⁰ While Schroen et al. reported no difference in the number of publications between female surgeons with and without children, this does not eliminate the possible impact of traditional gender roles on academic productivity.³¹

Another possible contributor to the observed difference in publication rates is bias at the level of the journal against female authors. Unfortunately, an increased proportion of women in orthopaedic surgery has not been matched by rates of senior authorship.³² Some have therefore called for a global review of editorial blinding to ensure journal-level decisions are not contributing to a gender publication gap. Although peer review is often “double-blinded” for authors and reviewers, editors in some cases are not blinded. While most journals have a rigorous editorial and reviewer blinding protocol, there are likely differences in protocols for some journals. Further research into whether this bias truly exists and how it may affect the surgical literature is needed.

Several limitations should be noted. We explored abstract characteristic differences presented between 2010-2017 in an effort to allow for adequate time for publication, though naturally more recent abstracts from 2016-2017 had shorter eligible “follow-up” than those from 2010-2017. Time to publication is also difficult to assess based on differences between the date an article first appears online versus its inclusion in a formal issue, which we standardized in this study to the former. Additionally, we attempted to classify author gender through the use of gender-specific pronouns on institutional websites first, followed by author first name, though this was not always possible due to a lack of website availability and because of a lack of familiarity with certain names. We also chose to evaluate the impact of senior author gender on publication rates as they more typically determine study topic/design. For this analysis, the last author was assumed to be the senior author, but we cannot entirely know the role each author played in design and publication. Last, the observed publication rates in this study were in line with or slightly higher than previously reported for ASSH podium presentations.^16,17 This could be due to a combination of factors, including (1) an increased number of predatory-type journals allowing for publication of more studies than might have been possible previously, (2) exclusion of abstracts where senior author gender could not be determined (~10%), which were anecdotally more likely to be those from international authors that may publish in non-English journals, and (3) perhaps a more thorough search for and inclusion of manuscripts published in non-indexed journals. That said, it is possible that some articles published in non-indexed journals were additionally missed.

Conclusion

Our work demonstrates that ASSH abstracts with female senior authors are published 13% less frequently compared to those with male senior authors. Although it is difficult to determine the cause of this discrepancy, study level of evidence and work resulting from RCTs was not different between female and male-authored abstracts. More work in this area is needed to identify the reasons for gender inequality in orthopaedic publications.