# A Smartphone Platform for Remote Motor Fitness Assessment and AI-Generated Personalized Exercise Programs for Older Adults: Randomized Controlled Trial

## TL;DR

A smartphone app that creates personalized exercise programs for older adults improves balance, arm flexibility, and strength more effectively than standard guidelines.

## Contribution

A novel smartphone-based system using AI to deliver personalized multicomponent exercises for older adults, improving neglected fitness components without lab supervision.

## Key findings

- Personalized exercise group showed significant improvements in balance, arm flexion, extension, and strength compared to controls.
- Improvements occurred with as few as 1.5 weekly sessions and within 4 weeks.
- No improvement was observed in torso rotation or sit-to-stand measurements.

## Abstract

Exercise guidelines for older adults are predominantly “one-size-fits-all,” primarily focusing on aerobic activity with limited emphasis on motor components.

We examined the hypothesis that remotely delivered, personalized multicomponent exercise—based on a simple yet highly reliable and accurate smartphone motor fitness assessment and individually tailored using machine learning—can improve balance, flexibility, and strength among older adults, obviating the need for a laboratory or professional supervision.

This randomized controlled study recruited community-dwelling healthy older adults aged ≥65 years, with normal cognition, low fall risk, and no hospitalization within the last year for cardiac/neurological illness. Participants were randomly assigned to an experimental 8-week personalized exercise group (5×/wk, multicomponent exercises), an 8-week active-control group (exercise counseling according to World Health Organization guidelines), or a control group (no intervention). Participants were assessed at baseline, 4, 8, and 12 weeks. Measurements were remotely recorded using smartphone sensors and analyzed using machine learning to create each participant’s unique fitness profile. Primary outcomes were fitness profile changes at 8 weeks.

We assessed 317 volunteers; 239 of them consented and met inclusion criteria (155 women, mean age 72.63, SD 5.38 y). Compared to both controls, the personalized exercise group significantly improved in dynamic balance (F6,404=3.232, P=.004, η2=0.046), total balance (sum of all balance measurements; F6,432=3.03, P=.006, η2=0.040), arm flexion (F6,448=2.527, P=.02, η2=0.033), arm extension (F6,450=2.753, P=.01, η2=0.035), and arm strength (F6,424=2.394, P=.03, η2=0.033). Significant improvement was observed with adherence as low as 1.5 exercise sessions/week over 8 weeks and often within just 4 weeks. No improvement was observed on torso rotation and on sit-to-stand.

A smartphone platform, with remote assessment and delivery of home-based individually tailored exercises, effectively targets the often-neglected key fitness components—balance, arm flexibility, and arm strength—in older adults. This approach has the potential to generate varied movement profiles and personalized exercise programs for both healthy individuals and those with mobility or cognitive impairments.