A Clinical Prediction Model for the Early Identification of the Need for Major Intervention in Patients with Traumatic Hemorrhage

Abstract

Background: There is a lack of well-validated clinical decision tools to assist clinicians with risk stratification of bleeding trauma patients. Objective: This thesis derives and validates a clinical prediction score in order to identify patients requiring major interventions for traumatic hemorrhage. Methods: We created a model based on the pre-specification of predictors. We conducted a systematic review of prediction models and a survey of traumatologists to identify candidate predictors. We conducted a derivation study of 748 trauma patients from 2014 to 2017. Results: The final model included systolic BP, clinical exam, lactate, FAST and CT. The c-statistic was 0.953 (naïve) and 0.952 following optimism-correction with bootstrap validation. Discussion: This thesis utilizes pre-specification to minimize reliance on small datasets and potential for over-optimism. Pre-specification is based on the best available knowledge within the literature and clinical expert community. Conclusion: A simple score is proposed for risk stratification of bleeding trauma patients.