The objective of this research is to make publicly available a unique and extensive dataset of fully annotated unobtrusive and continuous behavioral monitoring data. Unobtrusive and continuous monitoring is the use of environmental and physiological sensors to continuously gather health-related data about patients during their daily activities in their homes and communities. This methodology allows the collection of objective data capturing an individual's behavior in an ecologically valid way. At the Oregon Center for Aging and Technology (ORCATECH), we have developed a system for collecting these objective behavioral measures continuously in a person's home as they go about their daily activities. Over the past seven years, we have deployed this system or parts of this system into more than 1000 homes. Our most comprehensive dataset is comprised of 250 individuals with more than three years of continuous data (the """"""""ISAAC study""""""""), representing literally millions of measures for each person. We are continuing to collect data in more than 100 of these homes, and so this dataset continues to grow. Importantly, this dataset is fully annotated with weekly self-report of health changes, falls, ER visits, visitors, outings, medications, mood, and pain. Furthermore, these data are provided in the context of full clinical and neuropsychological assessments performed semiannually for each person. Although ORCATECH is not the only center collecting in-home monitoring data, the extensive annotation of our dataset makes it a unique and valuable resource. Thus, our Specific Aims are: 1) To create a cleaned dataset of sensor-based behavioral data, fully annotated with relevant and timely clinical and neuropsychological data, which can be hosted at National Archive of Computerized Data on Aging (NACDA). This will entail working closely with NACDA investigators and staff to determine the optimal way to store and represent the annotations that are a key part of this unique data resource. It will also entail reviewing the data from each home in detail to make determinations of how to handle missing data;and 2) To develop a process and software (application programming interfaces, or APIs) for continued automated upload of new data from our database to NACDA, as data are collected in our ongoing longitudinal studies. For the proposed research, we will restrict the scope to data from the ISAAC study, although we will design the interfaces to allow later addition of other APIs to upload data from the full range of our in-home sensors and clinical assessments.
The growing numbers of seniors with attendant chronic illness has given rise to a well-recognized crisis of health care in America. Continuous in-home assessment, in which behavioral and health measures are taken on an ongoing basis as a person goes about their daily activities, can improve our ability to identify health changes at the earliest possible time-points, averting crises and improving care management. The proposed project will archive the largest and best-annotated dataset of in-home monitoring data in the country, allowing researchers from many disciplines and perspectives to use this dataset to develop new models of healthy aging and care management.