Editorial linked to: Astin-Chamberlain R, Pott J, Cole E, et al. Sex and gender reporting in UK emergency medicine trials from 2010 to 2023: a systematic review. Emergency Medicine Journal Published Online First: 11 September 2024. doi: 10.1136/emermed-2024-214054.

The under-representation of women in clinical trials has been well documented but even less appreciated is the lack of attention to potential differences in outcomes according to sex and gender. Due to differences in body size and composition, sex hormones and metabolism, as well as important social determinants of health, we cannot infer that research findings in men can always be generalised to women. Just like age, race, education, socioeconomic status, comorbidities and many other categories that our patients fall into, sex and gender can have an effect on the way patients present with the same illness, the way patients respond to medications and the potential for toxicity.

Studies have suggested that women with acute myocardial infarction are more likely than men to experience symptoms such as dyspnoea or palpitations in addition to chest pain1 and (once diagnosed) women receive evidence-based treatments less often than men.2 For reasons still not clear, we know that men were more likely to have severe disease and die from COVID-19.3 There are also documented differences in the way that males and females use healthcare.4 Despite these widely reported gender disparities, we still rarely see the results of clinical trials disaggregated by sex.

Reasons for imbalances in enrolment likely go beyond conscious or subconscious bias. For example, men are more likely to be involved in major trauma and thus will be the majority of subjects in trauma research. Legislation may preclude the inclusion of women without a pregnancy test when the teratogenic potential of an investigational medicinal product remains unknown. But would there be value in introducing quotas such that clinical trials end the enrolment of men after hitting the required quota, continuing to recruit only women to ensure adequate representation? Yes, this would require extra time and expense but would certainly be more efficient and accurate than running multiple studies with the same disproportionate representation and failing to look at the differences by sex. Indeed, the International Committee of Medical Journal Editors has recommended researchers include ‘representative populations’ in all study types and provide descriptive data for sex and other relevant demographic variables ‘unless there are compelling reasons not to stratify reporting, which should be explained’.5

Even with imbalances in enrolment, it should be feasible to conduct a sex-disaggregated analysis of those who are enrolled. One potential reason this does not occur is the lack of an a priori hypothesis as to why we should expect different outcomes in males and females. The p values are valuable; too many hypotheses reduce this value. If we don’t have an a priori reason to think we will see a difference, why look? This, of course, fuels a self-fulfilling prophecy; that is, if no difference is sought, future researchers have no evidence on which to base a hypothesis that there may be a male/female difference in outcomes. In addition, to adequately assess differences between sexes, the number of patients in the study may need to be increased. This is true for any other subgroup differences we wish to analyse, such as age or ethnicity.

In this issue of the Emergency Medicine Journal (EMJ), the article by Astin-Chamberlain and colleagues demonstrates poor adherence to the Sex and Gender Equity Reporting (SAGER) guidelines in recent emergency care trials, even in journals that require it in their own guidance.6 The SAGER guidelines were first introduced in 2016 by the European Association of Science Editors.7 Their goal is to ‘provide researchers and authors with a tool to standardise sex and gender reporting in scientific publications, whenever appropriate’. Importantly, the authors of the guidelines felt that even if a study was not designed to look at sex differences, they should be reported. The guidelines further distinguish the terms ‘sex’ versus ‘gender’.

The SAGER guidelines clearly indicate the need for a more sophisticated and nuanced approach not only to the reporting but also the collection and analysis of data related to sex and gender. Without better reporting, how could we ever hope to truly understand and thereby address what drives gender disparities in health and care? Furthermore, when authors report populations by ‘sex’, it is rarely clear how these data were collected or defined, which risks amplifying health disparities for transgender and non-binary people. These considerations highlight vital areas for future research, and we must begin now with better data, better approaches to analysis and better reporting.

Furthermore, the study by Astin-Chamberlain and colleagues has highlighted that the EMJ currently has no such guidance. The SAGER guidelines discuss the important role that journals can play in ‘promoting sex- and gender-specific analysis of research data as a matter of routine’.7 The paper by Astin-Chamberlain et al has thus been a wake-up call for us. Going forward, EMJ will recommend to all authors that they follow the SAGER guidance. The SAGER guidelines look like other research checklists but are simple and relatively short. They acknowledge that not all points will apply to a particular study. Moreover, it is not always necessary to perform hypothesis testing on the differences between sexes. Simply reporting outcomes by sex can provide a useful signal for further study.

We know that authors don’t always read the not-so-fine print in our guidance, so it will be on us as editors to remind authors to report sex-disaggregated results when possible. We welcome readers to hold us to our word, assuring that this happens.

Not applicable.