WiP Seminar: Belal Hossain and Hanna Frank

Belal Hossain
Statistician, Advancing Health
PhD Candidate, School of Population and Public Health, UBC

Hanna Frank
PhD Candidate, School of Population and Public Health, UBC

Machine learning for analyzing health administrative data with survival outcomes

Health administrative data have emerged as valuable resources for public health research. These data reflect real-world healthcare interactions and are often valuable for creating tools such as risk scores and comorbidity indices. However, health administrative data come with many challenges, such as no information on important clinical predictors, missing information, mismeasurement, inconsistencies, and coding errors. Machine learning (ML) methods can potentially address some of these challenges. ML-powered survival models excel at handling high-dimensional data, capturing complex relationships, and supplementing unobserved clinical predictors by incorporating additional information from the wealth of variables in health administrative datasets. We will showcase how health administrative data and ML can be leveraged to create prediction models in two distinct populations: people with tuberculosis and people with multiple sclerosis. Using these examples, we will discuss the advantages of ML for these applications, as well as the unique challenges faced when using ML for survival outcomes and how to address them.

This is a hybrid event, you may attend in person or virtually. Please register and indicate your preference.