Affiliation: OHSU Biomedical Engineering
Funding Period: 2005 - 2009
Funding Source: Intel BAIC
View related topics
Abstract
This project will extend the work on context aware medication prompting done by PRI. We will specifically address the question of whether the set of inferences and the resulting prompting rules can be made more effective and less uncertain by using additional sensors and longer data collection, which may enable more accurate inference and prediction of activities. There are many parameters and choices available to select prompting strategies. These include choice of devices, modality, time, intensity, etc. Given this relatively unconstrained space, it is relatively easy to fail to prompt effectively. We will attempt to approach this problem in a principled way so as to maximize the likelihood of successful choices. We will also conduct a more extensive outcome study to determine whether the CAMP set of sensors and rules improve medication adherence in substantially longer trial periods.