1. What is the difference between design verification and design validation?
Verification confirms the device was built correctly, testing design outputs against design inputs through bench testing and analysis. Validation confirms the right device was built, testing it against user needs in real or simulated use. The sequence is clear: the design is frozen, verification confirms it was built correctly, and validation follows.
2. What does medical device design verification include?
It typically includes verification planning, functional and performance testing, mechanical verification, electrical safety and EMC verification, software verification under IEC 62304, formative usability evaluation under IEC 62366, and risk control verification under ISO 14971. Each activity is documented in a verification protocol and report and traced back to a specific design input.
3. What is a design verification protocol?
A design verification protocol is a document that defines what will be tested, the method, and the acceptance criteria, traced to a specific design input or risk control. It is reviewed and approved before testing begins, and the results are recorded in a verification report that provides objective evidence the input was met.
4. What happens if a verification test fails?
A failed result is a finding, not a dead end. The result is recorded as it stands, the cause is investigated, and the outcome usually points to one of three things: the design output needs revising, the design input or its acceptance criterion was wrong, or the test method itself was unsuitable. Whichever it is, the affected records and the risk management file are updated and the activity is re-run against the agreed criterion. Verification is not complete until every input has a passing, documented result.
5. Does a design change mean repeating verification?
Not all of it. The traceability matrix shows which design inputs the change touches and which verification activities depend on them, so re-verification can be scoped to those activities rather than the whole programme. This is why the matrix is built as the work happens: without it, a late change to one component can be impossible to bound, and the safe answer becomes retesting everything.
6. How many units are needed for verification testing?
It depends on the input being verified and the risk attached to it. Some activities are satisfied by a single unit or by inspection, while others require a sample size with a documented statistical rationale, particularly where a risk control is being confirmed. Sample sizes and their justification are set out in the verification plan before testing begins, because a reviewer will ask why a given number was sufficient.