Abstract Background and Objectives Falls and fall-related injuries are significant public health issues for older adults, especially those with mild cognitive impairment (MCI). This study examined how an Artificial Intelligence (AI)-enabled assessment using in-home sensor-derived changes in functional mobility and gait compares to performance-based clinical assessment in older adults with and without MCI. Research Design and Methods We conducted a longitudinal, single-cohort observational study, recruiting 75 community-dwelling older adults residing in retirement and senior living communities (eligible if aged 65+ and able to walk household distances). Participants were assessed for MCI. A depth sensor, installed in each participant’s residence, recorded in-home movement as silhouette depth images over a one-year period. AI/machine learning algorithms automatically identified, segmented, and analyzed walking segments to extract gait speed, stride time, stride length, and height of the individual walking. Timed Up and Go (TUG) score, gait speed, stride length, and stride time were captured continuously by the AI system and by a clinical expert at baseline, 3, 6, 9, and 12 months. Results The correlation between AI-based and performance-based TUG was strong and positive (r = 0.611, 95% CI: 0.354, 0.782). A moderate positive correlation between stride time and clinically assessed TUG was observed (r = 0.424, 95% CI: 0.193, 0.639). Discussion and Implications This study demonstrates that AI-enabled in-home monitoring can capture clinically meaningful indicators of functional mobility among older adults with and without MCI. As slowed gait speed and altered stride variability precede cognitive decline and heightened fall risk, a continuous in-home assessment can facilitate proactive interventions.
Demiris et al. (Tue,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: