P3‐215: Patterns of Home Computer Use Differ over Time Between Older Adults Destined to Develop Mild Cognitive Impairment and Those Who Remain Cognitively Intact

Identification of pre-symptomatic changes in healthy older adults that signal future cognitive decline and transition to MCI and AD is an important goal for early detection and intervention. Subtle changes in behavior associated with less efficient or effective performance in routine daily activities may signal incidental MCI. These changes have been correlated with cognitive tests and medial temporal lobe brain structures often affected in early AD. In this study, we examined whether patterns of responding to an in-home weekly online health questionnaire differed between older adults who remained cognitively intact and those who were destined to transition to MCI over a study period of four years. Participants were 110 community dwelling older adults enrolled in a longitudinal study of in-home monitoring technologies. At baseline in 2011, all participants were non-depressed (GDS<5) and cognitively intact (mean age = 84.8 years; 77% female). Of these individuals, 29 (26%) were classified as MCI based on neuropsychological criteria (mean age=86.0; 79% female) at an annual clinical follow-up visit (mean follow-up = 2.1 years). During the study period, a short online health questionnaire was distributed weekly through the Internet on Monday mornings at 5 am on participants’ home computers. Longitudinal generalized linear mixed effects models were generated with three online questionnaire response variables: completion time, in seconds; start time in the day, in seconds from 5am; and whether staff assistance was needed (yes/no), as unique outcomes. Models were adjusted for age, education and number of items endorsed. On average, individuals destined to develop MCI completed their questionnaires 1.4 hours later in the day than cognitively stable participants (p=.029), after adjusting for covariates. Over time, more participants needed assistance to complete the online questionnaires and this increase was significantly steeper in those destined to convert to MCI (p=0.032). The longitudinal trajectories of ambiently assessed measures of weekly home computer use differed between those who remained cognitively stable and those who were destined to transition to MCI. Remote, high frequency assessment of routine activities is a promising approach for detecting the earliest changes associated with future cognitive decline in older adults.