Abstract: Clinical decision support systems increasingly rely on machine learning (ML) models to recommend courses of action. As a result, these systems have the potential to exacerbate inequities in healthcare allocation and disadvantage historically and contemporarily marginalized groups. To address this risk, fair ML algorithms have been proposed that minimize differences in model performance among patient groups. I will discuss some of these methods and the challenges to implementing them in practice. Two major challenges are to measure and mitigate these differences when we consider grouping patients by intersections of demographic variables such as age, race, ethnicity, sex, and socio-economic status.