A Causal Model for Prehospital Blood Transfusion and Mortality in Traumatic Hemorrhagic Shock

OBJECTIVES: Utilization of prehospital blood transfusion (PBT) as a treatment for traumatic hemorrhagic shock (THS) has been expanding nationally, yet evidence supporting the practice in some operational environments remains limited. Development of strong scientific evidence to support this practice requires a clear causal model linking prehospital transfusion to mortality and other clinically relevant outcomes. The aim of this project is to develop such a causal model, identifying pathways among the intervention, mediators, and contextual factors through which PBT may impact mortality. METHODS: The causal model was designed to link PBT with all-cause mortality, compared with standard resuscitation using crystalloid fluid, in patients with suspected THS. An initial causal model was developed based on existing literature and emergency medical services (EMS) operational workflows. The model was subsequently refined through an iterative expert-informed review process involving multidisciplinary stakeholders, including EMS clinicians, EMS physicians, trauma surgeons, transfusion medicine specialists, and investigators with expertise in EMS system and causal inference. RESULTS: The causal model includes exposure to PBT, confounders, mediators, and colliders, all of which may impact the primary outcome of interest: patient survival. The model includes multiple exposures, or types of PBT, including packed red blood cells (PRBCs), plasma, platelets, and whole blood. Potential confounders include EMS system-level factors, such as EMS system times (response time, on-scene time, and transport time), EMS clinician training and protocols, units of blood products available, and quality improvement programs associated with program operation. Patient-level confounders include type of injury, injury severity, age, and patient comorbidities. The model also includes two categories of mediators through which PBT is thought to impact patient outcomes: physiologic mediators, such as improved blood pressure, heart rate, acidosis and coagulopathy, and system mediators, including reduced crystalloid exposure and decreased time from injury to transfusion. CONCLUSIONS: A causal model was developed to provide a transparent framework and methodological foundation for pragmatic research and comparative effectiveness studies. The model identifies key confounding processes that should be considered when developing study-specific adjustment strategies for observational analyses. This model is intended to guide implementation, evaluation, and future research on the impact of PBT in patients with THS.