Fairness in predicting cancer mortality across racial subgroups

Fairness in predicting cancer mortality across racial subgroups
JAMA Open Network; Teja Ganta, MD; Arash Kia, MD; Prathamesh Parchure, MSc; Min-heng Wang, MA; Melanie Besculides, DrPH; Madhu Mazumdar, PhD; Cardinale B. Smith, MD; 7/24
In this cohort study, a machine learning [ML] model to predict cancer mortality for patients aged 21 years or older diagnosed with cancer ... was developed. ... The lack of significant variation in performance or fairness metrics indicated an absence of racial bias, suggesting that the model fairly identified cancer mortality risk across racial groups. The findings suggest that assessment for racial bias is feasible and should be a routine part of predictive ML model development and continue through the implementation process.