How can duplicate safety reports distort signal detection?

Why

The same patient may be reported by a doctor, a manufacturer and a published case series. If treated as three independent cases, the safety database can overstate how much evidence exists for a product-event combination. Compare stable case identifiers where available, patient characteristics that can lawfully be used, onset and reporting dates, suspect products, narratives and reporter context.

Use both automated candidate matching and human review, particularly when privacy masking reduces identifying fields. Follow the established case-master and follow-up process, preserve original source records and document why records were linked or retained separately. Never delete reports simply because a name and event match.

Removing genuinely distinct patients has the opposite problem: it suppresses a possible signal. Re-run relevant analyses after reconciliation and record the effect of duplicate management.