Sex and gender differences and biases in artificial intelligence for biomedicine and healthcare

Abstract

Precision Medicine implies a deep understanding of inter-individual differences in health and disease that are due to genetic and environmental factors. To acquire such understanding there is a need for the implementation of different types of technologies based on artificial intelligence (AI) that enable the identification of biomedically relevant patterns, facilitating progress towards individually tailored preventative and therapeutic interventions. Despite the significant scientific advances achieved so far, most of the currently used biomedical AI technologies do not account for bias detection. Furthermore, the design of the majority of algorithms ignore the sex and gender dimension and its contribution to health and disease differences among individuals. Failure in accounting for these differences will generate sub-optimal results and produce mistakes as well as discriminatory outcomes. In this review we examine the current sex and gender gaps in a subset of biomedical technologies used in relation to Precision Medicine. In addition, we provide recommendations to optimize their utilization to improve the global health and disease landscape and decrease inequalities.

Keywords: Biomarkers; Computational models; Medical ethics; Risk factors.

Fig. 1. The key determinants of health.

Health and wellbeing of individuals and communities are influenced by several factors, which include the person’s individual characteristics and behaviours and the socio-economic, and physical environment, according to the World Health Organization (WHO) (www.who.int/hia/evidence/doh/en/). Sex and gender differences interact with the whole spectrum of health determinants.

Fig. 2. Desirable and undesirable biases in artificial intelligence for health.

Fair data generation and explainable algorithms are fundamental requirements for the design and application of artificial intelligence to optimize for health and wellbeing across the sex and gender spectrum. This will facilitate the reduction of undesirable biases that propagate inequity and discrimination, and will promote desirable differentiations that help develop Precision Medicine.

Fig. 3. The digital divide in access to mobile technology around the globe.

The bar plot reports how less likely a woman is to own a mobile phone than a man, according to a survey analysis on mobile ownership conducted by the Global System for Mobile Communications Association (GSMA) in low- and middle-income countries (LMIC) in 2019, by geographical area (source: GSMA “The Mobile Gender Gap Report 2020”). For instance, in South Asia women are 23% less likely than men to be the owner of a mobile phone, while in Europe and Central Asia women are 1% more likely to be the owner of a mobile phone. Across LMICs (“Overall”), women are 8% less likely than men to own a mobile phone.