Introduction The proportion of women engaged in clinical research has increased over time, but it is unclear if women and men contribute to the same extent during the conduct of research. Our objective were: –to describe the prevalence of women authors of original articles published in 2000 and 2015; –to compare the research contributions and author positions according to gender. Methods We conducted a repeated cross-sectional study among all original articles published in the Annals of Internal Medicine in 2000 and 2015. Participants were all authors listed on the byline of the original articles. The primary outcomes were 10 contributions listed at the end of the paper and included in the criteria of the International Committee of Medical Journal Editors (ICJME). Secondary outcomes were the author position on the byline (first, second, next-to-last and last compared to middle rank). The main exposures were the author gender and the year of publication (2000 or 2015). Other variables were the academic degrees mentioned by the author on the byline, the home institution, the country of affiliation, and the type of funding. We assessed the association and its evolution over time of the 10 specific contributions to research paper by using mixed effect logistic regression models (one per contribution) where the contribution was the dependent variable, the article was the random factor and the gender was the main fixed factor. In each model, we included the year and an interaction term between gender and year to assess change over time. We reassessed these associations after adjusting for academic degrees. To identify if gender was associated with a specific position on the article byline, we performed four conditional logistic regression models where each article defined a cluster, with author position (e.g. first vs. middle rank) as the dependent variable and gender the main predictor. We included year and an interaction term between year and gender to assess if there was a change over time of the associations between gender and author position. Then we adjusted the models for the 10 authors’ contributions to research. We built four models, comparing the first, second, next-to-last and last position to middle position. In these models, articles with four or less authors were excluded from the analyses. Results The proportion of women authors increased from 31.5% to 41.2% between 2000 and 2015 (P = 0.004). In 2000, women authors were less frequently involved than men in the conception and design (55.1% vs. 61.2%, P = 0.026), critical revision (70.4% vs. 80.7%, P = 0.001), final approval (80.7% vs. 85.8%, P = 0.038), and obtaining of funding (16.1% vs. 21.6%, P = 0.024). Women tended to be more involved than men in administration and logistics (35.0% vs. 26.0%, P = 0.019) and data collection (49.8% vs. 45.8%, P = 0.053), but they were similarly involved in the analysis and interpretation of data, drafting of the manuscript, provision of materials/patients, and statistical expertise. Women were less often last authors than men (9.0% vs. 15.5%, P = 0.016). These gender differences persisted in 2015. Conclusions The representation of women among authors of medical articles has increased notably between 2000 and 2015 but remained below 50%. Women's roles differed from those of men with no change over time. The study has some limitations with the use of a single journal and self-reported contributions. These differences may be due to justifiable reasons such as seniority, specific training and skills in research, or role preferences of the researchers. However, the possibility also exists that the academic research milieu perpetuates sexist attitudes and unequal treatment of researchers based solely on their gender. This requires further exploration, and justifies the continuation of local initiatives that promote women's involvement in research and ensure fair career opportunities, regardless of gender.
Read full abstract