- Published
- Updated
High satisfaction in a PMCF survey may reflect only people who chose to respond. Without understanding nonrespondents and access conditions, findings cannot readily be generalised to all users. Keep the invitation-to-response pathway visible.
Define the assessment
Distinguish recipients, undeliverable invitations and partial responses. Use appropriate permitted information for assessing representation rather than collecting unnecessary personal details. Reminder counts alone do not establish representativeness; the distribution channel also matters.
Worked example
For example, only frequent app users see an in-app survey. Results appear favourable, but people who stopped using the device may never have received the invitation. The team discloses this limitation, appropriately compares invitation coverage with usage information and considers another feedback route. Missing responses are not treated as satisfaction.
Practical checks
- Separate numbers invited, reached, starting and completing the survey.
- Examine survey access for groups who discontinued use or requested support.
- Explain response rates and unanswered questions within the clinical question’s context.
Decision and limitations
The report should show which group the survey may represent and where uncertainty remains. Nonresponse does not always establish bias, but it creates a limitation worth examining. Assess findings alongside other PMCF and surveillance sources rather than selecting favourable responses to produce a safety conclusion.
Frequently asked questions
- Does a large response count establish representativeness?
- No. Invitation coverage and characteristics of nonrespondents can also influence the result.
- Are nonrespondents necessarily dissatisfied?
- That cannot be assumed either. Their reasons may be unknown, and the uncertainty should remain explicit.