AEMS, VAERS, and the Vaccine-Safety Database Transition
A source roundup on FDA's planned adverse-event database, the VAERS replacement story, and the parallel fights over vaccine-injury reporting and compensation.
The strongest current item in the VAERS pass is a system-change story, not a single adverse-event claim. Google News indexed CIDRAP's March report, ["FDA announces AEMS, new adverse-event database to replace VAERS"](https://news.google.com/rss/articles/CBMiogFBVV95cUxPc21hM3k4UjNlZGpwVXRCUDJXN1JmNlVWYUFkOXJHM0JydWh2OTVLWWxydnd2b2ZkSlI4R0lyUXd1Und5U3VqekJiQ2ZudW1mRXpaeEY3d3QyNDNfTE1NTzdYOWdnWk50X2pQelZqSXB0d0VDZ0VsUTFNb1ZFU3ZTVTVMMXhtMkdrMXBhV0ttUDhBVE5pYWZyUmkwRmRZZmZWdEE?oc=5), describing FDA's AEMS as a new adverse-event database intended to replace VAERS.
That is the watch item. VAERS has always been a messy but important public signal system: easy to misuse, but also a rare place where raw post-vaccination reports are visible outside agency walls. A replacement database should be judged on plain questions. What fields are preserved? What public access survives? What search and download tools are available? What historical VAERS records remain online? What gets corrected, withheld, de-duplicated, or reclassified before the public sees it?
The same feed shows why the transition is politically loaded. U.S. News & World Report had a June 26 item, ["Calling the Shots: How RFK Jr. Is Influencing America's Vaccine Playbook"](https://news.google.com/rss/articles/CBMitgFBVV95cUxNNlllUFU4NlpoYThFeW9NUjN6MW1YYzlDRjJqSXVsUmFaLTZRVWo5el9ZLV9YR2hlR29yNEFKVENUX3ZiQTgtam45c25Ba2pRdDVidUlyZlBibjFUYnpVOXVsMXc3Z0tkR01XWE9tMXI3Q3ExQWpoTjViVjlwa1ctTGs1a04wdVJEM2NtYU5peFU2clBYMXdSUzdWZFBRWHVGNS1JUG1jV3dFY2d0Zll6VElUS0lmQQ?oc=5), framing the broader vaccine-policy shift under HHS leadership. KFF's Monitor item on ACIP engagement and claims of a VAERS "cover-up" shows the other side of the information fight: public distrust rises when the reporting system is both high-profile and poorly understood.
There is also a compensation track. Reuters had a March item indexed as ["Kennedy allies petition to broaden US vaccine injury list"](https://news.google.com/rss/articles/CBMixAFBVV95cUxOdDdhdFBpZ29oRkFaQjN4dDFqQ3J0QmMzcmJBSWNLSW1KWmNmdnMzVExhSlF1a3B6bDN3M0hpaWdRajM5X1NEMjhvN1RrN2I1aF95NXpKR0RzVmZnd3dMbDBFUFEtUUhZeDdaRkx6ZF92MUF3Q0I3VUx6X3ZiNk56VEhyWUxaNkR4UnVRa1FyOE40QnJuT3d0QmJMZmtxSWw0S3UtSGlVamF5U3A5YTVucnd4bldsLVQ1NEJWUEprOUEwZ1Bs?oc=5). Courthouse News separately covered ["Vaccine injury claims face ticking clock after EU ruling"](https://news.google.com/rss/articles/CBMikgFBVV95cUxNNGlMLW5nWnU5dEZ0bV90cnc2LU4yWW9CZ2tiZklQV0pxLUg2blJUajlCVVRkNzlyMk91UFI4NUJKVkJKM3hxRS0yWkU0dEgtdTJRU3NRSkx1WDhUbVotSVpNQ0pqTW5GbUdoYzZXTVMtNU85c3d6aG1BTl9ReV9mbTB3aGNPV1JCT3g5TWVDakw0UQ?oc=5), which gives the international legal angle. Put together, the records point to a larger policy question: how governments count injuries, who gets access to the underlying reports, and what legal route exists when someone says the system harmed them.
The AI layer is worth keeping in the file too. Mother Jones had a February item, ["Health Department Will Mine Unverified Vaccine Injury Claims With New AI Tool"](https://news.google.com/rss/articles/CBMivgFBVV95cUxPZjZiMEVPSEZUdWRITFJMWlJxTzB3TjgyVjlhakIySElfWFNzRFJLc0tHUUhhYjVZYzB5ekVCRFFlYXhCb2ptdmNScndYTFlZQkNoZTlkYk5mczNfYnZYYjQtQVBrdWo2U3pkUndiUk9PU21SNkk0VHU4OGFkWndHMVhIX2ZYeHA1SEY2a0NyUVVSWmphUF9BWWU3bjJnbGZ3eDNGcjZhcUdZc1g3R1NWUU5MU1BNbEFZLXM5aWxB?oc=5). OECD's AI Policy Observatory indexed the same general concern under HHS AI analysis of vaccine-injury claims. If AEMS is being built during a period when agencies are also testing automated analysis, the public should ask exactly where automation enters the workflow: intake, triage, signal detection, narrative classification, fraud review, public summarization, or enforcement.
The sober takeaway: the VAERS story should not be reduced to either "all reports are proof" or "all reports are junk." The important document trail is the migration itself. AEMS should be measured against VAERS on transparency, public download access, field definitions, auditability, correction logs, and continuity of historical records. The next records to pull are FDA's AEMS announcement, any Federal Register notices, database documentation, data dictionaries, HHS AI procurement or policy records, and the public rules for preserving old VAERS data.