27 Sex and gender reporting and analysis in cochrane reviews: a cross-sectional methods study. Preliminary results

Objectives Sex and gender health differences have been well established in the scientific literature. However, numerous studies present poor levels of inclusion of female participants and an insufficient sex/gender reporting and analysis. This lack of consideration of sex and gender in medical research reduces the applicability of findings and jeopardises its capacity to support informed decisions. Studies also suggest that women under-representation in science might be related to this deficient consideration of sex/gender in health research. This study aims to describe the degree of sex/gender reporting and analysis in Cochrane systematic reviews published in 2018 and investigate its association with gender of authorships. Method Cross sectional study. We screened Cochrane reviews published in 2018. We removed reviews addressing sex-specific medical conditions and those that had been withdrawn by Cochrane. We collected data on gender and country affiliation of first and last authors, and review type. We documented the frequency of sex/gender terms used in each section of the reviews (abstract, methods, results, and discussion). In the results section, we split up descriptive information from primary studies (sex rates) and analytic approaches (considering sex/gender in the assessment of risk of bias, presenting disaggregated data by sex/gender or subgroup or heterogeneity analyses). We used ‘not applicable’ to denote a situation where insufficient primary studies or data on estimates did not enable to conduct the intended analyses (e.g. meta-analysis, subgroup analysis). We performed descriptive statistics and regression analyses to assess associations between authors’ gender and sex/gender reporting. Results 556 reviews were screened, of which 91 were excluded due to withdrawal (19,8%) or sex-specific disease (80,2%). Our analysis comprised 465 studies, including 2 prognosis, 4 methodology, 5 overview, 20 diagnostic and 421 intervention reviews. Women represented 53,8% (n = 250) and 38,9% (n = 181) of first and last authorship, respectively, while in 25,3% of reviews both authors were women. 85,6% of authors belonged to high-income countries. 7.5% (n = 35) of reviews reported on sex in the abstract, 17.6% (n = 82) in the methods and 61.4% (n = 285) in the results section. Of these 285, 64.7% provided descriptive results and 16,3% had an analytic approach. In the discussion section, 13.5% (n = 63) of reviews addresses sex-related findings. Only 4 studies scored positive in all 4 sections. Studies with female first-last authorship had a non-significant increased probability of reporting on sex (RR [95% CI] 1,24 [0.68-1.92]) Conclusions Consideration of sex and gender in Cochrane reviews is scarce. This prevents from generating inclusive and unbiased health research and inhibits effective population-wide translation of results. While women are less likely to be last authors, studies with a female first or last author show an increased (although non-significant) probability of reporting on sex.