We have conducted a systematic review to understand how sex differences are considered and analysed in clinical research. We surveyed the first 20 original research articles in four top clinical journals (New England Journal of Medicine, Annals of Oncology, The Lancet and Nature Medicine) for each year between 2019 and 2025 (560 articles in total). We asked whether the article i) included both males and females, ii) considered sex in design or analysis, iii) tested for sex differences; if a sex difference was tested, whether the article iv) applied appropriate statistical methodology (i.e. performing an interaction between sex and intervention).
We found that clinical research between 2019-2025 largely (86%) included both males and females. However, the proportion considering sex in their design or analysis was 45% while the proportion testing for sex differences was 21%. When a sex difference was tested the statistical methods were applied correctly for 79%. We identify multiple studies which likely had sex differences (i.e. in response to treatment) but were overlooked, as well as positive examples in which significant sex differences in response to treatment were appropriately analysed and reported. We describe the statistical mistakes that are commonly repeated across research and how to apply appropriate statistics to test for sex differences.
Our results alarm us of the need to consider sex and sex differences, as well as apply appropriate statistics to advance clinical research and healthcare.