What CRAs Actually Check When They Review a Site’s Records

On the closet walk, CRAs are not only counting boxes. They are testing whether disposition language matches physical reality: which lots are usable, quarantined, used, or destroyed — and whether the site can show that without inventing the story mid-aisle.

What looks weak: a destruction log dated the morning of the visit, lot numbers that do not match shelf labels, missing quantities, or quarantine stock still living in the “available” mental model. What looks strong: lot-level lines with dates, disposition status, and an export that could have been generated before the CRA opened their laptop.

Accession- or lot-level traceability matters more than a pretty spreadsheet. If the site cannot answer “which earliest-expiration lot was destroyed and when?” without email archaeology, the finding writes itself — even when the physical closet is orderly.

Central-lab missing-sample queries are where memory fails sites. A CRA (or the site preparing for one) needs a boring chain: visit label, sample type, ship date, who handed it off, and a tracking / AWB number — shipments, not patients.

Reconstructed AWB lists from inboxes read differently than a contemporaneous shipment log. Timestamps and tracking numbers that pre-date the query are the difference between “we think it shipped” and “here is the courier record.”

Sites that keep that log current spend less monitoring time arguing with the past. Sites that do not often burn the visit on scavenger hunts that look like process gaps even when the specimen actually left the building.

Deviation and CAPA / NTF review is less about literary quality and more about whether the site noticed, contained, and documented on a believable timeline. A zero-deviation binder can raise more questions than a binder with several clear, timely notes.

Monitors look for completeness of the operational story — what happened, immediate action, prevention thought — and for the absence of identifiers that do not belong in an administrative draft. Vague notes written weeks later are harder to defend than plain notes written the same day.

When sites hand over a structured CAPA or NTF draft that was scrubbed of PHI and exported from a working session, the conversation stays on process. When the only artifact is a paragraph rebuilt from memory the night before, the conversation stays on credibility.

Across kit, shipment, and deviation records, the same CRA instinct applies: contemporaneous systems beat heroic reconstruction. Exports with dates, lot/accession codes, and consistent status language are easier to trust than binder inserts that only appear when monitoring is on the calendar.

If you are a CRA, ask for the current export and the story of how it is maintained — not only the paper that was printed for your visit. If you are a site, keep the operational view current so monitoring week is confirmation, not invention.