- Published
- Updated
One error among two records produces a fifty percent rate. That percentage alone does not carry the same information as the same rate at a large site. Monitoring teams should establish what the denominator represents before deciding how urgently participant safety or reliable data may be affected.
Scope of the decision
A denominator may count participants, visits, forms or actual opportunities to perform a procedure. These units answer different questions. Define the counting unit and observation period in the review rule so that volatility remains visible. Assess an important individual event separately rather than allowing an aggregate statistic to conceal it.
Worked review
Consider a newly opened site where samples were late at two of four eligible visits. Instead of reporting only fifty percent, the team examines the four visit dates, reasons for both delays and other deliveries using the same courier. It can then investigate a specific logistics problem without prematurely judging overall site performance.
Practical checks
- Display both errors and eligible opportunities beside the calculated rate.
- Avoid directly ranking first visits against long periods from established sites.
- Review individual effects on participants and the primary endpoint separately.
Acceptance and limits
Triage can be accepted when the review note explains uncertainty and the next point for examining new evidence. This does not establish resolution. In the example, subsequent eligible deliveries are reviewed after the courier intervention; a small sample cannot demonstrate lasting success.
Frequently asked questions
- Is one minimum denominator suitable for every trial?
- No. The monitored risk, seriousness and rate of data accrual determine an appropriate study-specific rationale.
- Should small sites be excluded from monitoring signals?
- No. Explain the limits of rate interpretation and continue examining significant events and process evidence.