About the Project

Sex/gender differences have historically been underrepresented in biomedical research, clinical practice, and medical technologies. For decades, the male body was often treated as the default reference in medicine, while sex/gender-specific factors affecting disease risk, diagnosis, progression, and treatment received less attention. As artificial intelligence becomes increasingly integrated into healthcare, there is growing concern that these historical biases may also be reproduced by AI systems trained on clinical and biomedical data.

Our research line studies how sex/gender influences biomedical data, clinical decision-making, and AI-driven medicine. We develop computational approaches that integrate sex/gender perspectives into AI, precision medicine, and digital health, with the goal of building more equitable, interpretable, and biologically informed models.

Cardiovascular disease is a clear example of sex/gender bias in medicine. Although it is one of the leading causes of death among women worldwide, it has historically been studied mainly as a men’s disease. As a result, diagnostic criteria, risk prediction tools, and clinical guidelines have often been based on male-centered cohorts. Women may therefore present different symptoms, experience delays in diagnosis, or receive less accurate risk prediction. AI systems trained on these historical datasets may inherit and amplify such inequities unless sex/gender-aware methodologies are explicitly considered.

Building on initiatives such as Bioinfo4Women at Barcelona Supercomputing Center, our work explores questions such as:

• How do sex/gender biases emerge in biomedical datasets and AI systems?
• Can AI models reproduce or amplify existing healthcare inequities?
• How do sex/gender differences affect cardiovascular disease risk and diagnosis?
• How can explainable AI help identify overlooked sex/gender-specific disease mechanisms?
• Can AI uncover sex/gender-specific biomarkers or therapeutic targets?