Innovations

The virtual trial emulations move beyond the conventional observational study approaches by overcoming their inherent observation nature to enable interventional experiments to test and compare hypothetical and counterfactual treatment effects.

Our research is among the first to combine generative AI and causality learning to support full trial emulation with causal inference. Synthetic patient populations can be generated according to different inclusion criteria of the trials that are designed to explore different clinical questions for target populations.

This research work also contributes to delivering new knowledge about using synthetic data to complement the use of real world health data and RCTs. Our work advances the state-of-the-art in research by leveraging on the advance in causality learning and generative AI to support full trial emulations involving both the control and treatment arms. This opens new avenues for using synthetic data in support of health research.