The PSUR database search, step by step
Every PSUR must survey publicly available vigilance data for your device and similar devices. Here is the search protocol that satisfies that requirement — and the tool that runs it in seconds.
What the MDR actually requires
Under EU MDR Articles 85–86, manufacturers must produce a post-market surveillance report (Class I) or a Periodic Safety Update Report (Class IIa and above, updated annually or biennially). The PSUR summarizes your post-market surveillance data — and per the MDCG 2022-21 guidance, that includes a survey of publicly available vigilance databases covering both your own device and similar devices on the market.
In practice, the workhorse databases are the FDA's — MAUDE for adverse events and the FDA recall database — because EUDAMED's vigilance module is not publicly accessible. Notified Body auditors expect to see a documented, reproducible search protocol: which databases, which search terms, which date range, retrieved when.
The manual protocol (2–4 hours per PSUR)
- Identify the FDA product code(s) covering your device type — the product code directory groups all ~7,000 codes by specialty.
- Query MAUDE for the product code across your reporting period; export and deduplicate the results; count by event type (death / injury / malfunction) and by month.
- Split "your device" from similar devices by brand and manufacturer name — remembering that name matching is indicative, not authoritative.
- Query the recall database for the same codes and period; note Class I–III enforcement classifications.
- Trend against the prior period and summarize the most-reported coded device and patient problems.
- Write the methodology block: databases, filters, dates, coverage, and exclusions — precise enough that an auditor could repeat the search.
Or: run the whole protocol in seconds
BEUDAMED's PSUR reports execute exactly this search over the complete openFDA datasets — trends, splits, recalls, and the methodology block included. Free.
Public example reports available — no account needed to generate one
What auditors look for in the data section
Three things distinguish a PSUR data section that passes review: reproducibility (the search protocol is documented, not just the results), honesty about matching (name-based device attribution is labeled as indicative), and disclosed exclusions (records dropped for bad data are counted, not silently ignored). Our generated reports do all three by construction — every report ends with a methodology block you can paste into your PSUR appendix.