How to Run a Clinical Investigation of a Medical Device
Apply device good clinical practice, which is a different standard from the drug one.
What a how-to is not
A how-to is SPEQ’s practitioner method, not a procedure. It does not replace your own SOP, it is not a validated approach, and the judgement calls in it belong to your quality unit.
Device clinical investigations have their own good clinical practice standard, and reaching for the drug one is a scope error that survives a spot-check because both are real and both say GCP. The differences are structural: blinding is often impossible, the operator is part of the intervention, the device may change between versions during the investigation, and the evidence feeds a clinical evaluation rather than a marketing application alone.
- 1
Establish what the investigation must demonstrate
Device evidence usually addresses safety and performance against an intended purpose, and the investigation exists to close a specific gap identified in the clinical evaluation. Starting from the gap keeps the design proportionate; starting from a template produces evidence nobody needed.
- 2
Perform and document the risk analysis before the design
Device investigations are built on a documented risk analysis that identifies the hazards the investigation itself introduces, including those arising from the procedure and the operator. The mitigations feed directly into the investigational plan.
- 3
Address the operator as part of the intervention
Results depend on the person using the device, so training, qualification, and the learning curve are design variables rather than logistics. An investigation that does not define operator qualification cannot separate device performance from operator skill.
- 4
Plan for what cannot be blinded
Where blinding participants or operators is impossible, mitigate rather than ignore: blinded outcome assessment, objective endpoints, independent adjudication. Declaring an open-label design without stating the mitigation leaves an obvious source of bias unanswered.
- 5
Control device version and traceability through the investigation
Record which device and which version each participant received, and treat a modification as a change requiring assessment of its effect on data already collected. Pooling results across an undocumented design change is the device-specific analogue of an unreported protocol amendment.
- 6
Report device deficiencies and adverse events under the device definitions
The reportable categories include device deficiencies that could have led to a serious adverse device effect had circumstances differed or intervention not occurred. That near-miss category has no clean drug equivalent, and it is the one most often missed by teams arriving from drug trials.
- !The drug good clinical practice standard applied to a device investigation because both are called GCP.
- !Operator training and learning curve treated as logistics, so device performance cannot be separated from skill.
- !A device modification mid-investigation, with results pooled across versions and no documented assessment.
- !Device deficiencies not reported because no harm occurred, missing the category defined by what could have happened.
How to Run a Clinical Investigation of a Medical Device: frequently asked questions
Common questions on run a clinical investigation of a medical device.
Why not simply follow the drug GCP guideline?
Because it does not address what makes device investigations different: the operator as part of the intervention, the frequent impossibility of blinding, version control of the investigational device, and the device deficiency category. Both standards are real and both say GCP, which is exactly why substituting one is a scope error that passes a casual check.
What is a device deficiency and why does it get reported?
An inadequacy in identity, quality, durability, reliability, usability, safety or performance. It is reportable where it could have led to a serious adverse device effect had suitable action not been taken, had intervention not occurred, or had circumstances been less fortunate. The trigger is the potential, not the outcome.
How does the investigation relate to the clinical evaluation?
The clinical evaluation identifies what evidence is missing; the investigation is designed to close that gap and its results feed back into the evaluation. Designing an investigation without reference to the evaluation tends to generate data that does not answer the question the evaluation left open.