Abstract Uncontrolled hemorrhage accounts for up to 40% of trauma related mortality, and approximately 30% of trauma related deaths occur in the hospital within the first 1-4 hours of injury. The Advanced Trauma Life Support (ATLS) guidelines still recommend initial resuscitation with crystalloid fluid as an integral part of maintaining or restoring hemodynamic stability, but are starting to recognize the role of early resuscitation with blood products in patients with evidence of Class III and IV hemorrhage, in line with recent clinical trial and observational data. Although there is a growing body of evidence that early transfusion with blood products may be associated with improved survival compared with crystalloid infusion, the optimal transfusion strategy still remains unclear and is being actively investigated in clinical trials. In silico models of resuscitation offer a means to evaluate the effectiveness and safety of resuscitation protocols, including effects on body fluid compartment volumes, tissue oxygenation, and hemostatic factors. Here we describe our work extending earlier deterministic models of trauma resuscitation to incorporate hemodynamic, hemostatic, resuscitation, biomarker, and tissue-oxygenation domains into a single comprehensive model. Additionally, this model expands upon previous deterministic models to create a stochastic computational model of hemostatic resuscitation that accounts for variations in blood component composition and temporal features of transfusion therapy. Using this stochastic model, an in silico trial was piloted to compare conventional component therapy, i.e. red cell, platelet and plasma components, to cold-stored whole blood using the following starting assumptions: patient weight of 81.3±21.8 kg, hematocrit of 41.8±4.3%, initial bleeding rate between 50 and 150 mL/min, systolic blood pressure of 125±18 mmHg, fibrinogen of 293±88 mg/dL, international normalized ratio of 1.0±0.1 and platelet count of 239±62x109/L, with blood component and whole blood constituent distributions based on laboratory quality control data. This in silico trial demonstrated that resuscitation with whole blood was associated with a 28-minute reduction in the time spent in the critical window defined by platelet count <50x109/L, INR≥2, hemoglobin <9 g/dL and fibrinogen <150 mg/dL compared with conventional component resuscitation over the course of a 4 hour simulation (p<0.001 based on median regression). We demonstrate how stochastic models of hemostatic resuscitation can be used to perform in silico trials of resuscitation strategies. Our work can guide inclusion and exclusion criteria for future clinical trials and facilitate identification of patient subgroups most likely to benefit from certain resuscitation strategies. This work was partially funded by the AABB Foundation (formerly the National Blood Foundation).