BIC

The ability to determine the medical state of an individual, based on information extracted from physiology data collected by biosensors, allows a prognosis and/or diagnosis to be formulated and applied to estimate the likely medical outcome of the individual in the absence of caregiver intervention. This information is vitally important when triaging multiple casualties, and while monitoring casualties during transport from the field to more advanced hospital facilities. However, such physiology data are difficult to collect, manage, and analyze.

Ongoing efforts to address the data management and analysis issues have resulted in the development of the PAS, which is a system that greatly reduces the data management burden currently faced by investigators, while at the same time providing data analysis capabilities in a flexible environment. The PAS is a server-based system that supports the storage and analysis of diverse types of physiology data. The PAS has been expressly designed for the efficient management and analysis of time-series data, and incorporates easy-to-use time-series data visualization tools and algorithms to analyze such data. In specific, besides standard 'spreadsheet' data analysis capabilities as can be found in relational databases, the PAS incorporates a time-series data handling structure and mathematical algorithms designed expressly to analyze time-series data to permit sophisticated mining of this type of data. These features reside on a server, thereby dispensing with the need for each investigator to possess and maintain copies of analysis tools in their workstation and avoiding the time-consuming process of downloading large amounts of data from the server to the workstation. In the PAS implementation, only the results of the analysis are transmitted to the user. Indeed, all an investigator needs to use all features of the PAS is a Web browser.

The PAS is ultimately a tool for the extraction of knowledge from the data stored in it. The system incorporates a logic engine in which discrete or time-series data can be analyzed by passing it through a 'chain' of sequentially-acting functions. The modular data processing capability brought about by the chained function model allows for the assembly of powerful and flexible analytic routines that are particularly useful for the analysis of time-series data. Because the quality of the knowledge extracted from the physiology data is only as good as the data it is based on, the system also incorporates an extensive set of features to rank the reliability of the data (e.g., heart rate or respiratory rate quality) on a point-by-point basis, so that only the most reliable data is used for development of the diagnostic/prognostic algorithms.