Global Patient Safety

Open tools for pharmacovigilance signal detection and clinical research acceleration.

★ Start here — how we read the data

Inside the AEMS Data: What 4.8 Million Reports Do and Don't Tell You

2026-07-14

The FDA's Adverse Event Monitoring System, the quarterly public data behind our analyses, and why the strongest raw 'signal' is usually an artifact — not a discovery.

Read →

Tools

faers.mobi
Live

Current FAERS (2018-2024): time-stratified Bayesian and frequentist disproportionality. Watch semaglutide + gallbladder events, GLP-1 class effects, and recent signals evolve over quarters.
aers.mobi
Beta

Historical AERS (2004-2012): classic safety-signal case studies. Includes the Vioxx / myocardial infarction era and other pre-FAERS reports.
pico-dag
Live

PICO question -> UMLS concept graph -> code lists + data-pull specification. Compresses the front end of clinical research from weeks to an afternoon.
VAERS vaccine safety
Live

CDC VAERS (1990-present): same Bayesian + frequentist signal detection applied to vaccines. COVID-19, influenza, and all other vaccine types across 35 years of reports.
Signal methods
Live

Reference documentation for the four disproportionality methods implemented across the apps: GPS/EBGM, PRR, ROR, BCPNN/IC. Live as Signal & Noise.
MAUDE device safety
Coming soon

FDA MAUDE (medical device adverse events) with the same time-stratified analysis used for drugs.

Recent articles

View all →
Inside the AEMS Data: What 4.8 Million Reports Do and Don't Tell You

2026-07-14

The FDA's Adverse Event Monitoring System, the quarterly public data behind our analyses, and why the strongest raw 'signal' is usually an artifact — not a discovery.
Tested Against VAERS: Christine Cotton's Safety Claims and the Post-Market Record

2026-06-13

The categories of adverse event the French biostatistician flagged in Pfizer's Comirnaty trial — cardiac, reproductive/menstrual, thrombotic, stroke — tested against the VAERS signal database with the project's 4-method consensus filter.
Independent Reanalysis: COVID/Shingles Vaccine Signals Amid FDA Publication Blocks

2026-05-12

Cardiac, neurological, and thrombotic signals in VAERS for nine COVID-19 vaccines and three shingles products — applied with the project's standard 4-method consensus filter.

About

We analyse 35 years of FDA and CDC adverse-event data using peer-reviewed Bayesian statistics to surface drug safety signals that matter to patients and clinicians.

Data sources
  • FDA FAERS (2018–present)
  • CDC VAERS (1990–present)
Methods
  • GPS — 5th-percentile credible bound, dual-Gamma posterior, linear RR scale (DuMouchel 1999 framework)
  • PRR, ROR, BCPNN/IC
  • Consensus: flagged by ≥ 2 of 4 methods
Privacy

No data is stored. No account required.