.jpg){kind=link}
- Published
- Updated
Choosing records only from readily available files can exclude problematic workflows. A high number of completed reviews may therefore cover only a narrow range of risks. Teams should explain which question their selection is intended to answer rather than treating review volume as sufficient evidence.
Scope of the decision
A risk-focused sample is different from a statistical sample intended to represent every record. State its purpose and the limits of inference. Quietly substituting another file for a missing or delayed record can remove the very process problem that monitoring needs to examine.
Worked review
For example, a monitor reviews the same ten available participant files at each visit. Five other files remain in another department and are never selected. Tracking the inaccessible group separately reveals a handover responsibility gap. Few errors in the available files have not resolved that gap.
Practical checks
- Explain which risk or activity category the selected records cover.
- Record why files are inaccessible and their follow-up status instead of silently replacing them.
- Check whether repeated visits overrepresent the same convenient group.
Acceptance and limits
Accept the selection when it provides suitable coverage of the stated risk and a traceable rationale. Manage access barriers as separate actions. Do not report findings as an error rate for every site record when the selection method cannot support that inference.
Frequently asked questions
- Must every review use random sampling?
- No. Targeted selection can be useful if its rationale and supported conclusions are clear.
- Is an inaccessible record automatically incorrect?
- No. Access is an issue to investigate; a content error cannot be assumed before the record is examined.