Note: In the context of this edition of the Tulane Digest and biomedical research, there is an operational distinction between the terms of “sex” and “gender.” “Gender” refers to socially constructed roles, behaviors, expression, and identity. “Sex-based” biology and medicine focus on sex assigned at birth through anatomical, chromosomal or other features. While the interaction between sex, gender, and other identity factors and their impact on health over the course of human life is complex topic requiring dedicated research in the pursuit of health equity, we maintain the narrow definition of biological sex for the scope of this article.
Slipping through the cracks: A continued lack of women in biomedical research
Female biology has long been underrepresented or downright ignored in both basic and clinical biomedical research, with researchers citing reasons such as increased variability in female biological and behavioral responses due to hormonal changes—since disproven in various studies including a meta-analysis of over 300 neuroscience articles—and claims that drugs in early development might harm fertility, despite estimates that the total number of pregnant individuals taking either prescribed or over-the-counter (and often necessary) medications is well over 50%. Though the FDA and NIH have in the past few decades released guidelines and mandates for inclusion of females in clinical trials and basic research, women are still routinely underrepresented in early-stage clinical trials. Even though female non-human animals are now often included in preclinical biomedical experiments, many of these are not followed-up with a proper statistical analysis, nearly negating the reason for their inclusion in the first place.
As a result of these persistent integration failures, women continue to find themselves navigating a healthcare system that does not fully understand their biology, despite observed differences in cardiovascular diseases, Alzheimer’s prevalence, schizophrenia, drug pharmacokinetics; even baseline metrics commonly taken for granted such as normal blood pressure range have recently been observed to differ in women. Furthermore, the hormonal changes of menopause have been shown to alter areas from cardiovascular health to cognition for reasons that are separate from changes purely due to age. At the same time, menopause appears to accelerate cellular aging. These findings produce further complications for ongoing clinical studies of women’s health over the course of their lifespans. Still, a greater effort to include women in clinical trials has the potential to produce more efficient studies and greater likelihood of long-term drug success. An FDA report found that eight of the ten drugs withdrawn from market between 1997 and 2000 (in some cases, within only a few years of approval) showed higher risk to women. A more rigorous analysis of adverse events and consideration of potential sex-based prescription frequencies could have ended trials earlier in the pipeline, saving considerable expenditure. Some studies find that women are more likely to complete clinical trials, leading to superior retention, greater statistical power, and the potential for shorter time to trial completion.
Even with proper inclusion and analysis, the complexity of interacting sex-related variables and their changes over time have the potential to cause difficulty when crafting accurate and effective research and healthcare strategies to address the unique needs of women. However, institutions are taking heed. The NIH’s All of Us program aims to capture health differences and disparities across the whole spectrum of diverse individuals across the US, and is committed to advancing women’s health research. Tulane University’s Center of Biomedical Research Excellence in Sex-based Precision Medicine is dedicated to conducting projects at the intersection of sex, gender and other genetic, biological, and social factors of health. At Mayo Clinic, the Women’s Health Research Center focuses on advancing sex-based studies and training early researchers in intersectional women’s health research through its BIRCWH program, one of 19 across the nation.
Artificial intelligence as a path to greater fairness in healthcare
In addition to dedicated research, addressing these complex issues requires innovative solutions, and recent advances in artificial intelligence (AI) may help to provide answers. The continued development of AI and machine learning—which have already made great strides in medicine, from evaluating chest X-rays to designing a drug for idiopathic pulmonary fibrosis that reached clinical trials—gives promise for better, less-biased detection of sex-based differences in biology. When trained on vast amounts of appropriate data, AI models are able to detect patterns and relationships in large, complex datasets that might otherwise be invisible to humans. Such models can detect genetic disorders through subtle differences in facial structure or determine how an individual’s cancer evolved, potentially allowing prediction of disease progression.
Within the realm of sex-based biology, artificial intelligence has already shown success. A study using AI found sex-based differences in posture, with potential to guide screening and treatment in physical therapy and sports medicine. Another recent study led by researchers at NYU Langone Health found sex differences in white matter of the brain at the microstructural level, potentially providing insight into why psychiatric and neurological disorders sometimes present differently between men and women. Researchers at the Karolinska Institute in Sweden used AI to find sex-specific gene expression patterns related to calcification of the aortic valve of the heart during aortic stenosis, offering a path toward precision therapies. Within the clinic, properly-trained AI-powered technologies such as ambient clinical documentation or large language model-assisted diagnostics could increase sex health equity by ensuring that symptoms are documented and properly considered, given that women’s health concerns are unfortunately more frequently dismissed by doctors compared with those of men.
Those who forget the past: Avoiding AI’s potential pitfalls
Despite great promise, it must also be stated that care must be taken when employing such models. Otherwise, they have the capacity to continue to perpetuate (and even amplify) historical biases against women and other underrepresented groups. An AI model trained on datasets that do not properly incorporate sex as a biological variable may do more harm than good. A recent study examined whether OpenAI’s GPT-4 encoded racial and sex biases in areas such as diagnostic reasoning and clinical plan generation, and found that it did not appropriately account for demographic diversity and tended to overrepresent stereotypes of diseases based on race and sex. If a model is trained on medical texts and diagnoses that were often based on decades of data ignoring sex as a biological variable, it will make many of the same mistakes of the humans who produced that data. But once this is recognized, and after applying more careful algorithm design and a commitment to undoing previous discrimination in medicine, AI systems hold great potential for parsing out sex differences (as well as a plethora of other critical factors impacting health) and ushering in greater health equity for women.
The beginnings of a new era in health equity
Because the dangers are being recognized, discussed and beginning to be addressed, we now have room for cautious optimism. Technological advancements combined with greater awareness of sex-based health inequity at both the level of individual researchers and the US government are adding velocity to a momentum that has been building over the past decades. Through greater outreach to women for inclusion in clinical trials, more careful inclusion of sex as a biological variable in basic research (and appreciation of its value even when a study does not explicit aim to research sex differences), and a commitment to better health care policies and access for women, we have more power than ever before to bring about change and ensure a more equitable future.
____________________________________________________
The Tulane Medicine team, who is also the Tulane Digest Team, is partnering at BIO 2024this week. You can send a request through the BIO partnering system, or email us directly to arrange a time to connect.
____________________________________________________
Recent Podcast Episode Drops:
Whether for travel entertainment or a quick listen between meetings, check out 6 recently released bite-size episodes of BIO from the BAYOU. Check them out on the BftB Website, Apple Podcasts, Spotify, or anywhere you podcast.
____________________________________________________
Share the Science: Please forward this issue and encourage others to sign up for Tulane School of Medicine’s Digest Mailing List.
____________________________________________________
Curated Research and Research-Related News Summaries, Analyses, and Syntheses. Published on behalf of The Tulane University School of Medicine. Content is generated by reviewing scientific papers and preprints, reputable media articles, and scientific news outlets. We aim to communicate the most current and relevant scientific, clinical, and public health information to the Tulane community – which, in keeping with Tulane’s motto, “Not for Oneself but for One’s Own”, is shared with the entire world.
Alexis L. Ducote, PhD: Editor-in-Chief
Special thanks to James Zanewicz, JD, LLM, RTTP and Elaine Hamm, PhD for copy-editing assistance.