Context-Aware Medication Prompting

Abstract

This is a pilot project to study the relationship between medication adherence and patterns of activity in the home among normal elders and those with Mild Cognitive Impairment. Our primary hypothesis is that there is a predictive relationship between sleep, patterns of morning activity, and the likelihood that subjects will follow a morning medications regimen. A secondary hypothesis is that subjects will improve in medication adherence as a result of contextual prompting. We will recruit twenty subjects - 10 healthy elders and 10 elders with MCI. We will instrument their homes with bed sensors to detect sleep patterns; motion sensors and door switches to detect morning routines; and a MedTracker device, which is a medication dispenser that detects whenever the lid has been opened. Subjects will also wear a watch that can track their movement inside and outside the house, and that can display real-time reminders. Subjects will be given vitamin pills and instructed to use the MedTracker to store the pills, and to take a vitamin each morning and evening. During the baseline phase of the study, we will collect data on sleep patterns, adherence to the medication regimen, and morning routines such as patterns of movement from room to room. During the intervention phase of the study, we will enable the watch to display reminders and prompts to take their pills. The data analysis will determine whether particular patterns of sleep or other patterns (e.g., a lower than normal level of activity) can predict when subjects are likely to diverge from their medication regimen. We will also determine the effectiveness of the prompting in increasing the level of adherence. Additionally, we will examine the differences between healthy elders and those with MCI on these variables.