Neural Network Medical Decision Algorithms for Pre-Hospital Injury Severity and Risk Assessment

The central focus of the work proposed herein is the application of neural network estimation and classification methodologies to emergency medical care. On the pre-hospital scene, first-aid workers must, often in a matter of seconds with spare information at hand, make critical decisions concerning the yrgency of an injured patient's condition and the type of trauma care required. This decisional process, known as triage, is an integral component of emergency-care medicine. Avoiding undertriage is a matter of saving lives; avoiding overtriage is a matter of preventing precious emergency care resources from being overtaxed and misallocated. Although no existing triage technique has singularly proved entirely effective, there is strong evidence suggesting that use of multiple tests can collectively result in superior triaging decisions. In this vein, neural networks stand to play a valuable role in producing fast decisions based upon complex, multifaceted information. The work proposed will investigate the application of neural netwotk methods to the key decisional tasks in triage.