Validating Sensor-Derived Endpoints in Clinical Trials
Why a sensor reading is only as good as its verification, analytical, clinical and usability validation — and the interpretation threshold that makes it an endpoint.
REGULATORY OPERATIONSA sensor reading — from a wearable, a home instrument or a phone's own sensors — is not an endpoint until it is validated for the trial's own conditions.
Error enters at capture, pre-processing and the algorithm, and reliability alone is not enough. Hold V3+ evidence for the enrolled population, the real conditions of use and the exact device, firmware and OS fleet — plus a pre-specified, patient-anchored meaningful-change threshold — before enrollment, not after lock.
Each link of the chain can distort the result; no one claim covers all
Model, unit and OS variation; calibration drift. Equivalence is demonstrated, not presumed
Sampling-rate differences, clock drift, dropped frames. Part of the verification claim.
Parameters that suit one population, not another. Changing them is a change-control event.
Training-population bias; unseen edge cases. Validate in the enrolled population.
A valid metric that misses the concept. Clinical validation and a threshold are required.
Four positions that decide whether the endpoint survives submission
Consumer sensors are capable, but capability in general is not performance for a specific measurement in a specific population on a specific fleet; accuracy degrades outside intended use.
The measure must detect real change and carry a defined, patient-anchored meaningful within-patient change threshold — set before analysis and distinct from the treatment-effect test.
Validated in supervised healthy adults is not valid for unsupervised, elderly, motor-impaired participants at home. The trial's own conditions are the test; the bar rises with endpoint role.
Once models, firmware and OS range are specified, anything outside is unverified — live in BYOD, quieter in provisioned fleets. Updates are a change-control discipline (QPP-07).
Five duties, one per party in the chain — none is delegable
- 1Sponsor: own the fit-for-purpose decision, validation plan and interpretation threshold — accountability is not delegable.
- 2Vendor: generate and hold verification and analytical evidence for capture, pre-processing and algorithm; control every update.
- 3Manufacturer: specify performance characteristics and intended-use conditions; state plainly where use falls outside them.
- 4Biostatistics: treat device/OS heterogeneity and the threshold as explicit analysis assumptions, not footnotes, and defend them.
- 5Engage the qualification route early — FDA DDT or EMA Qualification of Novel Methodologies — as advice before full qualification.
FDA DHT guidance for Remote Data Acquisition (2023) · FDA PFDD guidance series · FDA–NIH BEST · FDA DDT and ISTAND · EU MDR Annex II · ICH E6(R3) Annex 1 and 2 · EMA Qualification of Novel Methodologies · DiMe V3+
© qointa 2026 – Public – Uncontrolled when printed · Not legal advice; this summary does not classify any device.
sales@qointa.com · qointa.com
More from the library
Digital Health Technologies in Clinical Trials — A Regulatory Position-Paper Series
One device, several perimeters: a framework for assessing the regulatory impact of the technologies a trial relies on.
Read more →Who Is the Manufacturer? Economic-Operator Roles in DHT Supply Chains
How provisioning, importing, kitting and modifying a device assign manufacturer, importer and distributor duties — often by operation of law.
Read more →You Can Delegate the Work, Not the Accountability: Vendor Qualification and Oversight under ICH E6(R3)
The sponsor’s duty to qualify and oversee its DHT vendors — distinct from who holds the economic-operator role.
Read more →Talk to a specialist
Bring one device and one protocol — a wearable, a sensor, an app, anything. We will tell you which regulatory perimeters it opens and what it takes to close them.
Book a 15-minute call