Skip to contents

Traverses the OMOP CDM concept hierarchy to identify candidate negative control outcomes with zero plausible mechanistic links to index exposures/conditions, and fits empirical null distributions to calibrate study p-values and confidence intervals.