Integrating & Visualizing Clinical, Environmental, and Sensor Data

点击翻译按钮获取中文摘要

PROJECT 2 ABSTRACT (PROJECT DESCRIPTION) Context and usability are important drivers behind the design of the Biomedical REAl-Time Health Evaluation for Pediatric Asthma (BREATHE) platform. First, without context, it becomes difficult to understand the data cap- tured by sensors, diminishing the utility of observations. This observation spurs BREATHE to incorporate as much information about a patient and his/her situation and surroundings in order to fully contextualize interpre- tations and make informed recommendations. Second, usability is an imperative consideration in mobile health (mHealth): if the system is perceived as unusable – for any reason – patients will fail to be compliant and the mHealth application will ultimately fail. These two ideas motivate Project 2, which develops the informatics infra- structure to enable context around sensed data, to analyze the data, and to optimize the range of interactions and activities different users will have with BREATHE. Project 2 starts with development of a unified, extensible data model for pediatric asthma, establishing a common foundation for standardizing and organizing information. The data model creates a comprehensive view across different types of sensed information (physiologic, environmental), clinical information, and self-reported infor- mation (e.g., symptoms), all accrued over time. Given this data model, Project 2 establishes an information architecture that enables new data sources (e.g., real-time air quality reports, electronic health records) to be combined with sensor data, providing additional context. Open source projects (UIMA, Taverna) are adapted to provide a flexible framework for specifying data collection protocols and associated data processing workflows on these integrated datasets. As a use case, the linkage of processing workflows to statistical and machine learning methods to aid in predictive risk modeling of asthma exacerbations is considered, along with support for real-time analytics in order to facilitate timely responses by the BREATHE system. Finally, the unified data model also enables the creation of a range of user interfaces (UI) and visualizations: ohmage is leveraged to bootstrap mobile device client and UI development, including the deployment of scheduled surveys, reminders, and data collection; and web-based UIs are proposed to support the review of data collection protocols and patient-specific timeline views that enable clinicians and researchers to “drill down” into collected information. Usability concerns are reflected in the development of these various UIs through careful requirements analysis, and iterative design and testing. In this regard, Project 2 works closely with Project 1 to develop BREATHE’s mobile device interfaces, and with Project 3 to garner feedback on system design and usability in order to refine the platform over the course of the Center’s efforts.