An exploratory analysis of the associations between accelerometer-derived measures and future health-related changes in older adults: preliminary results on prospective cohort study

Accelerometer-derived measures have been shown to be predictors for multiple health-related changes in older adults, including cognitive, mental, and physical health. However, previous studies focused on only analyzing a few health-related domains and comprehensive investigation covering physical, cognitive, and mental health are not fully studied. Herein, we collected accelerometer-derived measures for physical activity and sleep in free living settings along with self-reported, qualitative data of daily physical activities, and investigated their associations with health-related changes and events (car accidents and dementia severity) in physical, cognitive, and mental health after following up 1 year. We found that accelerometer-derived measures were associated with multiple health-related changes (P < 0.05): (i) lower moderate/vigorous physical activity with decline of cognitive score; (ii) lower vigorous physical activity with car accidents; (iii) lower moderate/vigorous physical activity with worsening dementia severity; and (iv) lower sleep efficiency with decline of cognitive score. On the other hand, lower metabolic equivalent (METs) calculated from qualitative data of daily physical activities were uniquely associated with worsening mental health and decline of cognitive score (P < 0.05), although there was no association with any accelerometer-derived measures we investigated. Our findings suggest that accelerometer-derived measures could be useful for predicting future cognitive decline and risk of car accidents. In addition, enabling to quantify METs from accelerometer data would be helpful for predicting changes in mental health.