- Published
- Updated
Absence of data from a wearable does not establish that the participant did not wear it. Measurement, signal quality, power and transfer may fail at different stages. Distinguishing causes can improve support decisions and interpretation.
Define the assessment
Map the chain from wearing the device through signal generation, valid measurement, storage and transfer. Explain what each available record actually establishes. Participant accounts, technical logs and central record counts are not interchangeable, and discrepancies deserve review.
Worked example
For example, three days are blank in the central system. The device log shows measurements were taken but the phone connection failed. Some hours remain absent after delayed upload. The team separates transfer delay from actual measurement loss instead of assigning the whole period one nonadherence code.
Practical checks
- Define missing intervals by reconciling device time with central system time.
- Record which technical conditions can be verified before contacting participants for support.
- Track changes to initially missing periods when delayed records arrive.
Decision and limitations
The analysis should explain uncertainty about the causes of missingness. Do not replace every unobserved period with zero activity or zero symptoms. Clinical and statistical reviewers should consider whether device use relates to clinical condition; a support ticket cannot substitute for generating a valid outcome.
Frequently asked questions
- Does no data mean the device was not worn?
- Not necessarily. Measurement, storage and transfer records may support different explanations.
- Are late-arriving data automatically accepted?
- Process them through defined checks of origin, timing and validity.