Check a pair against the real world

A signal in the literature is a hypothesis. FAERS — the FDA's adverse event database, millions of reports — is where you test it. Enter a drug and an event; ATLAS pulls the reporting counts and computes whether the pair is reported disproportionately, against the standard signal criterion.

Disproportionality measures flag reporting patterns, not proven causation. FAERS is a spontaneous-reporting system with known biases (under-reporting, stimulated reporting, no exposure denominator). Treat a signal as a prompt for review, not a conclusion.