Figure 2.
An overview of our framework’s emulation pipeline and the underlying components. The central module is the Randomized Controlled Trial (RCT) Emulator, which orchestrates the entire process. First, the Tested Drugs Extractor identifies a list of repurposing candidates, based on the user-provided Drug Criteria. For each such candidate, using the input Study Design parameters, the Treatment & Control Cohorts Extractor assigns patients to the respective cohorts. The Confounders & Outcomes Extractor computes a baseline and follow-up attributes for patients in both cohorts. The Drug Repurposing Engine then instantiates an RCT Emulator for each candidate, which estimates its effect on disease outcomes in the treatment versus control cohorts, adjusting for the extracted confounders and using methods implemented in the Causal Inference Library.