dFDA: A Decentralized Framework for Drug Assessment Using Two-Stage Real-World Evidence Validation
Key Points
The aim is to develop a cost-effective framework for drug assessment using real-world evidence.
Developed a Predictor Impact Score (PIS) for signal detection.
Utilized two-stage validation with pragmatic trials.
Aggregated data from over 108 embedded trials for outcome labeling.
Achieved validated outcome labels at approximately 44 times lower cost than traditional methods.
Enabled continuous pharmacovigilance and precision dosing recommendations through real-world data.
Abstract
We present the Predictor Impact Score (PIS), a novel composite metric operationalizing Bradford Hill causality criteria for automated signal detection from aggregated N-of-1 observational studies. Combined with pragmatic trial confirmation (based on evidence from 108+ embedded trials), this two-stage framework generates validated outcome labels at ~44x lower cost than traditional Phase III trials. This enables continuous, population-scale pharmacovigilance and precision dosing recommendations.