The Cox proportional hazards model is a popular semi-parametric model for survival analysis. In many situations, data for a cohort of patients resides at different institutions. We aimed at developing a federated algorithm for the Cox proportional hazards model over vertically partitioned data. Our algorithm, VERTICOX, obtains the global model parameters in a distributed fashion, without centralizing individual patient-level data.
Learning Objective: Learn challenges and possible solutions in distributed suruvial data analysis without sharing patient level data
Xiaoqian Jiang (Presenter)
Luca Bonomi, UTHealth
Lucila Ohno-Machado, UTHealth