Technology & Care Innovation

Machine learning-based fall risk prediction using dual-task spatiotemporal gait parameters

Publication Type: Peer-reviewed case-control study (BMC Geriatrics)

Publication Date: July 31, 2026

Author(s): Dong, G., Hu, H., Guo, Y., et al.

Source Link: https://doi.org/10.1186/s12877-026-07998-3

Summary: Researchers used motion-capture gait data collected while 209 community-dwelling older adults walked and simultaneously performed a cognitive task, then trained eight machine learning models to distinguish those with a fall history from those without.

Abstract Summary: The best-performing models achieved strong discriminatory accuracy (up to 92.7% AUC), with gait variability under dual-task conditions — walking while thinking — proving far more informative than standard clinical balance scales, which tended to miss subtle risk in higher-functioning older adults.

Why It Matters: Standard falls-screening tools already used in Kansas facilities may be missing risk in residents who look steady on a simple walk test; this suggests dual-task assessment (walking while doing a cognitive task) is worth piloting as a low-cost screening add-on. The prediction models themselves rely on specialized motion-capture equipment not readily available in most Kansas facilities, so the near-term takeaway is the assessment principle, not the specific technology.

Family caregivers’ experiences with mHealth and related digital health support for people living with dementia

Publication Type: Peer-reviewed qualitative evidence synthesis (BMC Geriatrics)

Publication Date: July 28, 2026

Author(s): Gao, H., Sang, S., Chen, L., et al.

Source Link: https://doi.org/10.1186/s12877-026-08011-7

Summary: This synthesis pooled findings from 24 qualitative studies on family caregivers’ experiences using mobile health apps and digital tools to support dementia caregiving at home.

Abstract Summary: Caregivers generally found these tools convenient, informative, and emotionally reassuring, but consistently reported device and technical barriers, and asked for more integrated functions, personalization, professional support, and guidance specific to the stage of dementia their family member was in. The synthesis stops short of claiming these tools improve outcomes compared to usual care.

Why It Matters: For Kansas caregiver-support programming, the recurring ask for stage-specific guidance and integrated (not scattered) functionality is a concrete design brief for any app or resource list a provider organization builds or recommends — generic dementia apps are likely to underdeliver against what caregivers actually say they need.