Digital Biomarkers for Depression Screening With Wearable Devices: Cross-sectional Study With Machine Learning...

Area of Science:

• Digital health
• Mental health technology
• Wearable sensor technology

Background:

• Depression is a widespread mental health condition, often undiagnosed and untreated.
• Wearable activity trackers collect detailed user data (digital biomarkers) for potential depression screening.
• Digital biomarkers offer a scalable and unobtrusive method for monitoring mental health.

Purpose of the Study:

• To assess the predictive power of digital biomarkers from wearables for detecting depression risk in a working population.
• To explore the association between digital biomarkers and depressive symptom severity.

Main Methods:

• A cross-sectional study involved 290 healthy working adults wearing Fitbit Charge 2 devices for 14 days.
• Collected data included physical activity, sleep patterns, and heart rate variability.
• Depressive symptoms were assessed using the Patient Health Questionnaire (PHQ-9).
• Spearman correlation, multiple regression, and supervised machine learning were employed.

Main Results:

• Greater depressive symptom severity correlated with increased nighttime heart rate variability and reduced circadian rhythm regularity.
• Digital biomarkers showed limited ability to detect depression in the overall sample.
• Machine learning models achieved 80% accuracy, 82% sensitivity, and 78% specificity in detecting high-risk individuals within balanced subsamples.

Conclusions:

• Digital biomarkers from wearables show potential for depression risk detection and screening assistance.
• Current predictive ability in broad populations is limited, but machine learning models can discriminate high-risk individuals.
• Further research is needed to refine the use of digital biomarkers for scalable mental health monitoring.