Developing prediction algorithms for late-life depression using wearable devices: a cohort study protocol.

Area of Science:

• Gerontology
• Digital Health
• Psychiatry

Background:

• Major Depressive Disorder (MDD) is prevalent in the elderly, yet undertreated due to stigma and access barriers.
• Wearable devices offer a novel approach for early MDD symptom screening in naturalistic settings.
• Existing research predominantly focuses on younger populations, leaving a gap in elderly-specific digital health solutions.

Purpose of the Study:

• To develop prediction algorithms for late-life depression using longitudinal data from wearable devices.
• To devise strategies for enhancing medical access to depression care within community settings for the elderly.
• To address the underdiagnosis and undertreatment of depression in older adults.

Main Methods:

• A 3-year longitudinal cohort study involving 685 elderly participants from the Korean Genome and Epidemiology Study.
• Collection of self-report, observational, app-based survey, and passive sensing data over three annual interviews and two years of app usage.
• Primary data analysis utilizing machine learning techniques to identify patterns and predict depression.

Main Results:

• The study is ongoing, with data collection concluding at the second follow-up interview.
• Analysis will focus on identifying digital biomarkers for early detection of late-life depression.
• Expected outcomes include validated prediction algorithms and actionable strategies for community care.

Conclusions:

• Wearable technology holds significant potential for early detection and intervention of late-life depression.
• This research aims to bridge the gap in digital mental health for the elderly population.
• Findings will inform the development of accessible and effective community-based mental healthcare strategies.