Identifying chest-worn light logger adherence: a validation study

Area of Science:

• Chronobiology
• Wearable technology
• Health informatics

Background:

• Light exposure is crucial for regulating circadian rhythms and overall health.
• Wearable light loggers offer advanced capabilities for measuring daily light exposure.
• Accurate data from these devices is vital for clinical studies.

Purpose of the Study:

• Develop methods to differentiate between correct and incorrect use of chest-worn light loggers.
• Define distinct lighting intensity thresholds for indoor versus outdoor environments.
• Enhance the reliability of data collected in future clinical research using light logging technology.

Main Methods:

• A logistic regression model was trained using photopic illuminance, device orientation, accelerometer data, and time of day.
• The model was validated to differentiate wear, non-wear, and nighttime placement of the light logger.
• The model was applied to participants with migraine, and indoor/outdoor lighting environments were analyzed for optimal distinction points.

Main Results:

• The logistic regression model demonstrated high accuracy (AUC 0.93-0.94) in differentiating device use.
• The model showed good-to-excellent performance (AUC 0.84-0.91) on the validation dataset.
• Optimal cut-points for differentiating indoor/outdoor light were identified as 442 lux (photopic illuminance) and 412 lux (melanopic equivalent daytime illuminance).

Conclusions:

• Chest-worn light loggers can reliably differentiate between wear and non-wear using internal sensor data.
• Established cut-offs for distinguishing indoor and outdoor light are lower than traditional thresholds.
• These findings improve the reliability of everyday light exposure data in clinical populations using wearable light loggers.