- Published
- Updated
PMCF and complaint systems may contain different details about the same event. Few records in one source do not automatically invalidate the clinical question raised by another. Compare scope and capture methods before interpreting a shared pattern.
Define the assessment
Link records through the question, device identity, event timing and user context. Use authorised matching without unnecessarily duplicating personal information. Receipt of a complaint, technical confirmation and clinical assessment may be different states; avoid reducing them to one open-or-closed field.
Worked example
For example, PMCF interviews describe a recurring use difficulty absent under the same wording in complaints. Review finds another coding category. The team examines whether the events truly share characteristics and compares reporting routes, rather than merely matching terms. The same event is not counted twice.
Practical checks
- Compare users, periods and reporting routes covered by each source.
- Keep potential event-matching decisions and uncertainty traceable.
- Assess effects of coding changes on historical patterns and denominator interpretation.
Decision and limitations
The combined assessment should show how each source contributes. Absence of complaints does not establish absence of a problem observed actively. Route shared patterns into risk and clinical assessment; do not force uncertain matches into apparently definitive deduplication, and explain count differences.
Frequently asked questions
- Is a complaint count an event frequency?
- Usually not directly. Reporting behaviour and the known exposure scope matter.
- Does identical wording establish the same event?
- No. Matching needs support from device, timing and clinical context.