Cochrane Reviews' authorship has become more gender-diverse but remains geographically concentrated: A meta-research study.

Abstract

Objective: The aim of this study was to examine the distribution of country, region, language, and gender diversity in the authorship of Cochrane reviews and compare it to non-Cochrane systematic reviews.

Study design and setting: We retrieved all published articles from the Cochrane Library (until November 6, 2023) using a web crawling technique that extracted pre-specified data fields, including publication date, review category, and author affiliations. For comparison, non-Cochrane systematic reviews were identified through PubMed using E-utility calls. We determined the country, region of affiliations and gender of the first, corresponding, and last authors for Cochrane reviews; the same fields were determined for first authors only for non-Cochrane reviews due to data availability. Trends in geographical and gender diversity over time were evaluated using logistic regression. Fisher's exact test was used for comparisons. Diversity trends between Cochrane and non-Cochrane reviews were explored through visual presentation, Pearson's product-moment correlation, and the Granger Causality Test.

Results: This comprehensive analysis included 22,681 Cochrane reviews and 224,484 non-Cochrane reviews. Cochrane reviews showed increasing diversity in several areas: representation of first authors from non-English speaking countries rose substantially (from 16.7% in 1996 to 42.8% in 2023), and female first authorship more than tripled (from 15.0% in 1996 to 55.6% in 2023). Representation from LMICs in Cochrane reviews has declined recently (from a peak of 23.2% in 2012 to 18.4% in 2023). Among Cochrane Review Groups, diversity varied notably, with Sexually Transmitted Infections achieving the highest representation from LMICs (68.1% of first authors). In 2023, non-Cochrane reviews showed higher representation from non-English speaking countries (56.9%) and LMICs (50.8%) compared to Cochrane reviews. The patterns of gender diversity between Cochrane and non-Cochrane reviews showed strong correlations for female first authorship (r=0.829, P<0.001), suggesting parallel evolution over time.

Conclusion: Both Cochrane and non-Cochrane reviews demonstrate important progress in author diversity, particularly in gender representation and inclusion of authors from non-English speaking countries. While non-Cochrane reviews show stronger representation from LMICs, both review sources reflect the evolving landscape of global evidence synthesis.