P2‐388: Dementia Care Coordination: Impact on Patient and Caregiver Outcomes

Abstract

shower, etc.). A set of 11 sensors (motion, contact, and electric) were installed, for 5 to 12 months in each participant's home (see Caroux, Consel, Dupuy, Sauz eon, 2014 for details). Then, we calculated the most probabilistic routine for each selected ADLs based solely on the data collected, with minimal a priori on time frames or durations for each ADL. Reported routine was compared to the observed routine to validate formulas. Results: For all ADLs mixed together, reported and observed routines strongly correlated, r(113) ¼ .978, p < .001. Moreover, significant correlations were observed for each respective ADLs, despite moderate inter-subject variability (waking up: r(19)¼.652, breakfast: r(18)¼.652, dressing/showering: r(19)¼.819, lunch: r(20)¼.534, dinner: r(20) ¼ .550, going to sleep: r(20)¼.485). Conclusions: Results showed that simple, non-intrusive low-cost assistive technology shows strong concurrent validity when it comes to identifying older adults' daily routine and compares it to the self-reported routine. Results also suggest that older adults metacognition on their daily routine is relatively accurate. Following analyses will aim to demonstrate how assistive technologies can detect abnormal behavior in older adults with cognitive or physical declines. Poster Presentations: Monday, July 25, 2016 P794