Unobtrusive technological approach for continuous behavior change detection toward better adaptation of clinical assessments and interventions for elderly people

Abstract

Behavior change indicates continuous decline in physical, cognitive and emotional status of elderly people. Early detection of behavior change is major enabler for service providers to adapt their services and improve the quality of life of elderly people. Nowadays, existing psychogeriatric scales and questionnaires are insufficient to observe all possible changes at a daily basis. Therefore, we propose a technological approach for behavior change detection, that employs unobtrusive ambient technologies to follow up elderly people over long periods. In fact, we study significant behavior change indicators (e.g., sleep impairments, visits and go out) and investigate statistical techniques that distinguish transient and continuous changes in monitored behavior. Furthermore, we present a preliminary validation of our approach through results based on correlations between our technological observations and medical observations of two-year nursing home deployment.