Video-based gait analysis using pose estimation can quantify gait differences among non-frail, pre-frail, and frail older adults

Frailty is a geriatric syndrome that increases vulnerability to adverse outcomes, and early identification of pre-frail individuals could enable timely interventions. The researchers applied pose estimation algorithms to video recordings of older adults walking, extracting gait parameters such as stride length, cadence, and gait speed. They found significant differences in these metrics across non-frail, pre-frail, and frail groups, suggesting that computer vision can provide an objective frailty assessment without specialized equipment or clinician time. This proof-of-concept could lead to low-cost, automated frailty screening in clinics or even home settings, but validation in larger, more diverse populations and comparison with established frailty scales are needed before clinical adoption.