Computational Modeling of Coagulopathy for Decision Support

Nearly everyone, throughout their life, is at risk of being involved in a serious traumatic event, such as motor vehicle accidents, sports and occupational injuries, or natural disaster related injuries. Twenty-eight percent of trauma patients precipitously develop abnormalities in their blood coagulation system. These coagulopathies increase their mortality rate by 5fold. The current coagulopathy diagnosis protocol collects basic patient information, vital signs, and performs traditional lab and point-of-care (POC) blood testing. A high-stakes decision must then be made by the trauma surgeon, using their intuition, training, and the results from the blood drawn at least 15 minutes prior, to determine the requirement for a resuscitation treatment through coagulation inhibitors or activators. Computational modeling and system analysis of the human blood coagulation are integral to developing superior decision support tools for trauma surgeons. In short, the coagulation system consists of the following functional subsystems: 1) blood flow, 2) platelet function, 3) diffusion, 4) advection, and 5) biochemical kinetics. We utilize a combined approach of both 0-D and 3-D model development with the overarching goal of developing a validated, near real-time decision support system. The biochemical kinetics of the coagulation system is implemented in the 0-D model with a set of 113 nonlinear, coupled ordinary differential equations (ODEs), describing the time rate of change of the numerous chemical concentrations and their interaction with one another. 0-D models provide a fast, efficient means of simulating the coagulation biochemical kinetics, but these ODEs lack the ability to describe the global effects of fluid flow, advection, and diffusion. Hence, the set of 113 ODEs are modeled as source terms and combined with the Navier-Stokes and chemical advection/diffusion equations in a three-dimensional finite volume computational domain, providing a global coagulation model. Model validation studies employ parallel experimental POC blood testing and 3-D computational modeling. Results from the 0-D model are consistent with testimonials from expert trauma surgeons, whom verify the model provides appropriate reasoning for their difficulties in predicting patient outcome. Thus, validated computational models have potential as a hypothesis generator used for developing new approaches for providing trauma surgeons with sufficient information to make better informed clinical decisions, “the decision support tool,” leading to decreased mortality.