- Published
- Updated
Treating people lost to PMCF follow-up only as missing forms may understate uncertainty in the results. Examine why follow-up was lost and what is known about that group. Do not automatically classify an unknown outcome as success or failure.
Define the assessment
Distinguish withdrawal, lost contact, missed assessment and care elsewhere. Investigate reasons within permitted information and available records. Contact attempts and valid clinical assessments are different counts. Assess missingness before assuming it is unrelated to processes associated with outcomes.
Worked example
For example, some users with early device difficulties do not answer a follow-up survey. The team does not report favourable respondent outcomes as applying to everyone. It compares previous findings, follow-up duration and contact conditions. Authorised analytical review explains which conclusions the available evidence supports and whether further assessment is needed.
Practical checks
- Classify follow-up states separately with reasons and the last valid assessment date.
- Appropriately compare important known characteristics of lost and completed groups.
- Assess missing-data assumptions and how sensitive conclusions are to those assumptions.
Decision and limitations
The report should show findings obtained and the size of the unobserved portion. Loss to follow-up does not always create bias, but its lack of effect should not be assumed without assessment. Connect limitations to clinical evaluation and subsequent PMCF decisions rather than silently removing missing people from denominators.
Frequently asked questions
- Is an unreachable user a failure?
- Not automatically. The outcome may be unknown; analytical handling needs a reasoned approach.
- Do many calls resolve missingness?
- They demonstrate contact efforts. Uncertainty remains if a valid assessment was not obtained.