Bioinformatics-enabled wearable vital signs monitor for combat triage

Acute hemorrhage with subsequent shock remains a leading cause of battlefield mortality. It is the primary cause of death in approximately 30% of injured soldiers who die from wounds. Early diagnosis of the onset and severity of hemorrhagic shock is critical for providing effective combat triage and optimizing survival. This project will demonstrate the feasibility of applying a nonparametric, kernel-based technology called Similarity Based Modeling (SBM) to a set of readily field-measured vital signs from a wearable, mobile sensing device for improved visibility into health status in humans. Kernel-based modeling technology has been shown in commercial applications to map otherwise unknown interrelationships between instrumented variables using empirical data. Such mapping can be used to predict/infer non-instrumented signals, such as predicting an unmeasured pressure from other, measured parameters of a system. Furthermore, the empirical mapping of these interrelationships can provide ensembles of expected values that can be compared to measured values to ascertain incipient deviations from normal system behavior. These capabilities mean (a) values can be predicted for physiological parameters for which no direct measurement can practically be made, and (b) differences between measured values and “expected” values can indicate physiological instability that is about to cascade into a physiological emergency.