Aims:
Emerging evidence suggests that continuous monitoring of digital biomarkers, such as sleep architecture and activity patterns, can detect early signs of functional decline and acute health deterioration1. Privacy-preserving thermal sensors installed in homes offer a continuous, non-intrusive method to monitor older adults without "wearable fatigue"2. This study evaluates the feasibility of using thermal sensors to capture longitudinal digital biomarkers in older adults.
Methods:
Community-dwelling older adults, aged 65 years and above, with high falls risk, were recruited from a geriatric medicine clinic in Singapore. Baseline clinical assessments were performed, and wall-mounted thermal sensors were installed in the participants' primary living areas for continuous monitoring over six months. Feasibility was measured via study retention, data fidelity, and the successful identification of relevant digital biomarkers.
Results:
11 participants were recruited, with a mean age of 83.1 years. They were predominantly female (90.1%), and the majority lived alone (72.7%). The median FRAIL score was 2, median SARC-F score was 4, and mean TUG was 23.2 seconds. In terms of feasibility, there was a 100% participant retention rate. Data fidelity was maintained throughout the study period with 48312 hours of data collected and no device-related adverse events. The thermal sensors successfully identified activity patterns, sleep behaviour, bed motion, toilet use patterns, and lie-to-sit-to-stand-to-walk time.
Conclusion:
Privacy-preserving thermal sensing is a feasible method for the longitudinal monitoring of older adults at home. Our findings suggest that thermal-based digital biomarkers represent a promising way for moving geriatric care from reactive interventions toward proactive, data-driven management.