Selection driver
Protocol
Not vendor preference
Key factors
5
Endpoint · Population · Duration · Data · Support
Framework
FDA V3
Verification, validation, usefulness
Protocol-First Device Selection
Endpoint → Population → Device → Support
Wearable device selection in clinical trials is still more often shaped by what the team used last time, what the technology vendor recommended, or what was easiest to procure — than by a systematic evaluation of protocol fit.
The consequences are predictable: devices that generate the right raw data but don't wear comfortably enough for the population to use consistently; devices with impressive sensor suites that are overkill for the endpoint and add unnecessary patient burden; devices with poor battery life that require daily charging from elderly patients who won't do it reliably.
Device selection failures often don't surface until mid-study, when wear-time data reveals that patients are not using the device in the way the protocol assumed. By then, changing devices is nearly impossible and recovering the compromised data is worse.
See also: Wearable Devices · Device Integration
Answer these questions before looking at specific devices. The answers determine which device categories are eligible and which are not.
Activity level endpoints can be served by consumer-grade accelerometers. Cardiovascular endpoints — HRV, ECG-grade arrhythmia detection, SpO2 — require medical-grade sensors with validation data. Mixing endpoint precision requirements with device category is the most common selection error.
An elderly COPD patient wearing a wrist device 22 hours a day for 52 weeks has completely different tolerance and compliance requirements than a 35-year-old healthy volunteer wearing a chest patch for two weeks. The population determines the form factor, comfort requirements, and setup complexity ceiling.
Battery life, durability, waterproofing, and replacement logistics all change significantly between a 4-week and a 24-month study. Devices that work well for short-duration studies fail in longitudinal programs because they were not evaluated on long-term durability.
Some endpoints require continuous streaming; others need epoch-level summaries. The data format your analysis plan assumes needs to match what the device actually outputs. Discovering a format mismatch after data lock is a serious problem.
A technically superior device that requires complex setup, frequent charging, and app management is only viable if there is a support infrastructure in place. Matching device complexity to available support capacity is as important as matching it to the endpoint.
If the wearable data is intended to support a regulatory submission — as a primary or key secondary endpoint — FDA's V3 framework (Verification, Validation, Usefulness) should structure your device selection and validation approach from the start.
The fit-for-purpose device is not always the best device. It is the right device for this study.
Yes, for appropriate endpoints. Consumer devices are often acceptable for activity-based endpoints or exploratory digital biomarker collection. They are generally not appropriate for primary endpoints requiring medical-grade accuracy or for studies supporting regulatory submissions without specific validation.
Requirements depend on the endpoint and regulatory context. Exploratory use requires less formal validation than use as a primary endpoint in a submission. FDA's V3 framework provides structured guidance for device verification, analytical validation, and clinical validation.
Device replacement mid-study is inevitable at scale. A replacement logistics plan — including spare inventory, rapid dispatch capability, patient instructions for swap, and data continuity procedures — should be part of the study operations plan from day one.
Delve provides fit-for-purpose device strategy across cardiovascular, respiratory, metabolic, neurological, and activity endpoints — matched to your protocol requirements, population, and regulatory context.
Talk Device Strategy