RESUMEN
INTRODUCTION: Current decision tools to guide trauma computed tomography (CT) imaging were not validated for use in older patients. We hypothesized that specific clinical variables would be predictive of injury and could be used to guide imaging in this population to minimize risk of missed injury. METHODS: Blunt trauma patients aged 65 y and more admitted to a Level 1 trauma center intensive care unit from January 2018 to November 2020 were reviewed for histories, physical examination findings, and demographic information known at the time of presentation. Injuries were defined using the patient's final abbreviated injury score codes, obtained from the trauma registry. Abbreviated injury score codes were categorized by corresponding CT body region: Head, Face, Chest, C-Spine, Abdomen/Pelvis, or T/L-Spine. Variable groupings strongly predictive of injury were tested to identify models with high sensitivity and a negative predictive value. RESULTS: We included 608 patients. Median age was 77 y (interquartile range, 70-84.5) and 55% were male. Ground-level fall was the most common injury mechanism. The most commonly injured CT body regions were Head (52%) and Chest (42%). Variable groupings predictive of injury were identified in all body regions. We identified models with 97.8% sensitivity for Head and 98.8% for Face injuries. Sensitivities more than 90% were reached for all except C-Spine and Abdomen/Pelvis. CONCLUSIONS: Decision aids to guide imaging for older trauma patients are needed to improve consistency and quality of care. We have identified groupings of clinical variables that are predictive of injury to guide CT imaging after geriatric blunt trauma. Further study is needed to refine and validate these models.