Discussion
During the past decades, the representation of women in ophthalmology has increased,14 as approximately 25.0%–30.0% of ophthalmologists are women.15 Prior work has shown growth in the representation of women publishing in ophthalmology journals16 and working in academic centres.17 Even so, woman ophthalmologists are under-represented in ophthalmic journal editorial and society board leadership positions,18 as well as in other award categories in the USA.19
In our study assessing ophthalmology research award recipients’ productivity by gender over the past 99 years, we found that overall, woman awardees had significantly lower research productivity than men when examining the h-index and the number of documents and number of citations separately. Men were found to have a significantly greater number of years active than women. Even when considering the number of years active, the research productivity measured using m-quotient was still significantly lower overall for women compared with men.
The number of woman physicians in the past century and their productivity levels are essential to understand why there are more woman award recipients in recent years. Just over 10% of women entered medical school in 1970.20 Therefore, not many women could be nominated for award positions, as shown by the low numbers of woman awardees in figure 1 up to 1970. After Title IX, which prohibits sex-based discrimination in schools or education programmes that received federal funding, there was an increase in the number of women pursuing medicine.20 In 1974, 22% of new medical students were women.20 The number of woman physicians has quadrupled in the last 20 years.21
Furthermore, there was a shift toward women after 1989 in first authorship in ophthalmology journals such as: Ophthalmology (1969, 4.6%; 1979, 5.4%; 1989, 12.3%; and 1999, 20.2%), Archives of Ophthalmology (1969, 6.6%; 1979, 5.1%; 1989, 15.6%; and 1999, 28.6%) and American Journal of Ophthalmology (1969, 5.6%; 1979, 4.2%; 1989, 9.2%; and 1999, 23.9%). The increase in woman first authorship during the five decades parallels the increase in American woman physicians.16 This explains the increase and frequency of woman awardees from the early 1980s to late 2000s in figure 1.
By 2004, women made up 34% of members in training in ophthalmology.21 In addition, from 2000 to 2009, when ophthalmology journals such as the American Journal of Ophthalmology, Archives of Ophthalmology and Ophthalmology were considered together in a study done by Shah et al, there was a notable increase in the percentage of woman first authors (21% in 2000 to 34.2% in 2009) and last authors (18.5% in 2000 to 20.5% in 2009).22 This correlates to the increase in woman awardees from the early 2000s, as reflected in figure 1.
In 2017–2019, the number of women entering medical school outnumbered the number of men.20 Additionally, in a study done by Gertig et al, first and senior authors in ophthalmology review journals: Progress in Retinal and Eye Research, Survey of Ophthalmology and Current Opinion in Ophthalmology were evaluated. In 2019, the frequency of women authorship rose to 44% for first authors and 29% for senior authors. Also, the proportion of reviews with women being the first or senior author did not vary by country of origin (p = 0.887 and p = 0.520, respectively).23 Therefore, this explains the majority spikes of woman award recipients seen during the later 2010s in figure 1.
Although a near equal h-index value for man and woman award recipients is expected, overall, woman awardees generally had a lower h-index than their man counterparts. However, as shown in figure 3A, most societies had no significant differences in h-index between genders. Only six societies had a significantly higher mean h-index for man awardees and three societies had a significantly higher mean h-index for woman awardees (figure 3A). In the past 40 years, the mean h-index of man award recipients has predominantly stayed higher than that of woman awardees (figure 2A). Higher mean h-index for women occurred in 1993, 1994, 2001 and 2002 due to the low sample size of women (n=2, n=5, n=5, n=8) having higher h-indexes than those of men although the number of man award recipients is higher (n=18, n=23, n=36, n=35) (figure 1). This can be attributed to the greater number of man award recipients and repeat awardees, although in recent years the proportion of woman award recipients has increased (figure 1). Evidence supports the continued existence of a gender disparity among societies throughout the years.
With m-quotient being a relatively more accurate measurement of productivity due to the consideration of years active, an equal m-quotient value is expected, especially in the past decade. Over the 40-year period, man awardees have generally maintained a higher mean m-quotient, although the number of woman award recipients has been increasing (figure 2B). Similar to the h-indices, higher mean m-quotient for women occurred in 1993, 1994 and 2002 due to the low sample size of women (n=2, n=5, n=8) having higher m-quotients due to lower number of years active and higher h-indexes than those of men, although the number of man award recipients is higher (n=18, n=23, n=35) (figure 1). Men maintained a significantly higher mean m-quotient in five societies, while women had a significantly higher mean m-quotient in two societies (figure 3B).
There are limitations of the h-index. The h-index is shifting, and values reflect values at the time of collection only.24 The h-index is based on long-term observations, which disadvantages newcomers and people who change their names (women more often than men).25 Moreover, h-index fails to consider the information contained in authorship rank (first, middle, last). However, despite this limitation, the h-index does not show decay in a scientist’s career and measures long-term performance.26
Limitations of m-quotient involve the instability of the statistic when considering the authorships of researchers with relatively short careers, as small changes in h-index lead to dramatic fluctuations in m-quotient.27 It also does not consider whether the academic has sustained productivity, again mostly for newer academics whose first papers would be minor contributions. Furthermore, m-quotient is not a fair metric when considering part-time researchers and those with career suspensions (eg, parental leaves traditionally affecting women more than men).8
Possible factors resulting in our findings include implicit gender association and bias, disparities in available funding and productivity association with academic rank and collaboration with those of the same gender, impacting author positions.28
Implicit gender bias occurs when authors of one gender exclude authors of the opposite gender, which leads to a lower woman research productivity and a lower chance of being awarded a research award, causing the gender disparity between man and woman research award recipients. Implicit biases are concealed from conscious awareness; these are the attitudes that one does not realise are impacting their understanding and actions. Implicit bias awareness and education should be implemented more to avoid gender disparity.
Woman investigators receive less funding than men in the research field and area. One study that explored gender differences in Canadian Institutes of Health Research awards found that less favourable assessments of women as principal investigators drove lower award rates, rather than the quality of their proposals.29 Such biases may contribute to barriers women face in progressing in their medical careers, including reduced opportunities.29 A Hong Kong-based study showed that women received more funding from their institutions, but men were more likely to receive public research funding, showing that man academics access a more comprehensive range of external funding.30 Less access to funding or less funding for women means that it is less likely the research will be completed as proposed and can lower the productivity levels of women.31
Studies also show that from assistant professor through professor, the h-index increases with subsequent academic rank.32 However, women are often under-represented in positions of high rank in ophthalmology societies and are not being promoted to positions of professorship or faculty at the same rate as men or women in other specialties.33
Quantifying differences between genders in award recognition and examining potential underlying factors (research productivity metrics, award year, country of affiliated institution, society itself), explaining these disparities are the first steps to achieve equitable representation between genders. Although some ophthalmic societies are more gender balanced than others, men tend to have greater representation among research awards. These men also have significantly higher mean h-index and m-quotient values than their woman counterparts. However, it is important to note that research productivity metrics can have limitations and negative biases towards women. It is therefore not recommended to choose award recipients solely on these potentially discriminatory metrics.
Selection of research awardees within societies is often a multistep process, ranging from nomination of candidates from a pool of society members, review of their accomplishments (including research productivity) and final selection by a select jury. At each step of this process, there are opportunities to increase equity and remove bias to allow for more deserving women to be recognised equally.34 Societies have a role in assessing the gender distribution of the members of under-represented groups (whether it is through mentorship or broader membership inclusion). Second, review of candidates’ research accomplishments should consider research productivity and other research impact that can sometimes be difficult to quantify, including leadership within research organisations, mentorship of research trainees and role in science communication. Lastly, the jury itself should be composed of diverse members to represent a variety of perspectives in selecting the award recipients.
In conclusion, there are multiple factors contributing to research productivity differences between man and woman research awardees. Further investigation into award selection processes and gender membership data is required to further understand why women are under-represented among ophthalmic research award recipients.15
