Tong Chen awarded Population Health Theme Seed funding grant
Congratulations to Tong Chen who was awarded a MCRI Population Health theme grant for his project: Advancing Automatic Causal Machine Learning for Epidemiological Longitudinal Studies.
"Modern cohort studies collect thousands of biological, social, and environmental measures. Many of these factors are closely related, making it difficult to appropriately adjust for high-dimensional confounding by large groups of variables"
This project will advance a novel causal machine learning method called automatic debiased machine learning (autoDML), which automates key procedures in debiased machine learning based on so called ‘deep learning’. We will evaluate how well autoDML works in realistic settings, combine it with ensemble machine learning approaches for better performance, and adapt it to incorporate survey weights so results better represent whole populations.
This joint work with Margarita Moreno-Betancur, Susie Ellul, and Prof Stijn Vansteelandt (Ghent) will utilise data from the Longitudinal Study of Australian Children (LSAC) and The Child Health CheckPoint.
Importantly, the project will also deliver user-friendly R software so these methods can be easily adopted by health researchers.
By improving the reliability of causal analyses in complex cohort studies, this work will help generate stronger evidence to guide health research, policy, and practice.
