Randomized controlled trials, the gold standard of comparative effectiveness research, provide safety and efficacy estimates for a small fraction of drug pairs and outcomes. Scalable computational methodologies can fill these gaps using real-world data to generate richer quantitative drug susceptibility and effectiveness maps.
Our platform takes each drug and estimates the probability of various outcomes (harmful and beneficial), relative to a comparator drug (or set of drugs), from real-world clinical data and using causal inference methodologies. We aim to generate comprehensive drug-outcome estimate maps and visualize these in a semantically meaningful manner. The developed methodology will also be applied to specific use-cases to estimate drug effects on subpopulations (or individuals) of interest.