Identifying and Understanding Patterns and Processes in Human Shock and Trauma

The aim of this paper is to provide an overview of an on-going collaborative effort of research physicians, computer scientists, and statisticians to develop a quantitative way of understanding the clinical course of a critically ill and injured patient. The method is based on a multivariable analysis of time series of individual physiologic measurements. The ultimate goal is to reduce a large body of complex physiologic data to an information base that is relatively small and simple so that abnormal patterns may be exposed in a manner that can be directly interpreted and utilized at the bedside by the attending clinician to improve patient care. In this paper we describe some contributions made toward reaching this goal.