Monitor signals and draft a validation for each one that fires.

The agent reads EudraVigilance, FAERS and the MHRA Yellow Card for each substance, computes PRR and ROR with the reference period and comparator recorded, and drafts a validation for anything that fires. Every figure links to its case, and the validations stay in the audit trail GVP IX requires.

Sample screening · illustrative
Teduglutide (Revestive, Takeda) · EV code 72908 · Sep to Nov 2026 against Jun to Aug 2026Sources [1] to [10], EMA DAP and SmPC 4.8

Individual cases, whole history

2,332

EudraVigilance, EEA

Cases in the period

25

104 distinct terms

Over-reported classes (SDR)

10

5 not in SmPC 4.8

Cases citing a publication

2

of 25 read

Where the review starts: Renal and urinary disorders · ROR 2.07 · 191 cases, whole history · over-reported and not in SmPC 4.8

Also over-reported and unlabelled: Product issues 2.29, Surgical and medical procedures 2.16, Investigations 1.39, Neoplasms 1.27. First reported in this window, two cases each: gastrointestinal stoma output abnormal, blood pressure increased, brain fog, hypervolaemia. Sepsis: two cases, serious, not labelled. Caveats noted on the run: fewer than 3,000 reports, so the ratios are unstable; confounding by indication; a reporting association, not causality.

  • EV case form EU-EC-10022375122, 2026-10-21
  • EMA DAP extract, month 202611 [1]
  • Revestive SmPC section 4.8 [3]
  • Endocrinol Diabetes Nutr 2026, D-lactic acidosis after discontinuation

From a run on 5 November 2026 · EudraVigilance 2026-11 snapshot, extracted 10:54 UTC

Part of the surveillance chain:

Spend the signal team's time on deciding.

The agent pulls the data, ranks what exceeds threshold and documents why.

Ranked

Start where the label is silent

Every run ranks the system organ classes and reactions reported more often than expected, and checks each one against the label. Those already in the label are marked as known. The rest are where a review starts, strongest first, with the disproportionality ratio and the case count beside each.

New

See what changed since the last run

Each run lists the reaction terms reported for the first time in the window, and flags serious cases of unlabelled reactions on their own. New terms are where an emerging signal shows first, so they are never buried in the totals.

Traced

Trace every figure, state every limit

Each number links to the case form, data extract or label section behind it, so your team checks a claim by opening its source. The limits are written in too: when reports are few the ratios are unstable, the disease being treated can explain the reaction, and a reporting association is not causality.

Any cadence

As often as each product's risk requires, monthly by default

Run it more often for a new product or an open safety issue, on demand before a signal meeting or a PSUR, or once on a product you take over. Each run is logged with its date, data snapshot and version, and that log is the monitoring record.

Cases that cite a publication are checked against literature monitoring. Each run's findings feed sections 15 and 16.2 of the PSUR.

Keep your detection tool.

Empirica or EVDAS stays. At each run the agent takes the export, adds the public record and the literature, and writes the monitoring record and the validations.

  1. 01

    Screen EudraVigilance, FAERS and the MHRA Yellow Card per substance, or your EVDAS export

    It runs on your registration and your schedule, and lists the duplicates it removes.

  2. 02

    Compute PRR and ROR per substance-event pair

    Records the reference period and comparator, so any run can be repeated and compared.

  3. 03

    Count the literature beside the cases

    Adds the published cases for the same events, so a case series is not missed.

  4. 04

    Draft a validation for each signal above threshold

    Gathers the cases, timing, stop-and-restart outcomes and literature, then drafts the assessment, with every number linked to its source.

  5. 05

    Build the signal log and the monitoring record

    Built-in traceability from the start.

On record

Every run, per substance

  • Monitoring recordSources, dates and comparators for every run
  • Signal logKept in the GVP IX audit trail
  • Validation and assessment draftsEvery figure traced to its case
  • Run and version logAny past run repeated and compared
  • System validation packageSupplier documentation for your QA

Questions

It includes detection on the public record and takes the export from the platform you already have. What it adds is the record of each run and the validation write-up with every figure traced. Teams that rely on Empirica or EVDAS for the statistics keep using them.

Only with your authorisation. By default it works from the public record and from the EVDAS export you choose to give it, on your registration. It can also be connected to your internal safety data, and then it sees your cases too. Either way, your database, narratives and submissions stay where they are.

With you, per substance, and recorded on every run: the reference period, the search period and the comparator. A holder whose database changed with a deal can set the periods around the deal, and the record shows that it did.

Start with a tailored analysis of one substance.

Tell us the substance and the period. We run the monitoring on it and walk your signal team through what it finds.

By submitting this form, you agree with our Terms of Service. We process your data to respond to your contact request in accordance with our Privacy Policy.