Trauma Data Collection and Mining to Enhance Combat Triage Decisions

The fundamental objective of the research effort proposed herein is to provide guidance to combat medics in making field treatment and triage decisions. Combat medics must prioritize patients according to the need for: (1) immediate medical intervention, and/or (2) immediate evacuation and surgical intervention. Many life-threatening conditions -- most prominently acute hemorrhagic shock, followed by circulatory collapse -- are difficult to diagnose in the field, particularly in a timely fashion. This effort seeks to find predictive markers to enhance combat medic triage decisions and entails fulfillment of several objectives: (1) collection of a high-quality database of continuous physiologic data on trauma patients; (2) extraction of predictive features of relevant clinical outcomes from these data; and (3) development of a predictive methodology to aid in trauma triage decision making. Herein, Barron Associates, Inc. (BAI) and the University of Virginia (UVA) propose to collect continuous data on trauma patients via the Pegasus aeromedical transport service and the UVA Level 1 Trauma Center. BAI will apply sophisticated signal analysis and nonlinear dynamic pre-processing and feature extraction techniques to increase the likelihood of identifying useful predictive markers as inputs to estimation and classification neural networks, the latter of which predict relevant clinical outcomes.