Enhancing Utility of Interfacility Triage Guidelines Using Machine Learning: Development of the Geriatric Interfacility Trauma Triage Score (GITTS)

This study addresses the problem of under-triage of injured older adults to tertiary trauma centers, particularly in rural areas where patients are often first taken to non-tertiary centers. The current triage guidelines do not adequately prioritize risk factors nor allow individual risk prediction. The goal was to develop a risk score to help simplify the secondary triage process, determining which injured older adults should be transferred to a tertiary trauma center. Using machine learning techniques on data from the Oklahoma State Trauma Registry, the study identified key predictors of mortality or serious injury, creating a Geriatric Interfacility Trauma Triage Score (GITTS).

• 1
  The Geriatric Interfacility Trauma Triage Score (GITTS) model, with an AUC of 75.4%, includes airway intervention, traffic-related femur fracture, spinal cord injury, ED GCS <=13, and hemodynamic support as the top predictors.
• 2
  A GITTS risk score of 7 yields a sensitivity of 78% and specificity of 56% for predicting mortality or serious injury.
• 3
  The GITT model allows clinicians to calculate an absolute probability of outcome, for example, a patient with a traffic related femur fracture, ED GCS <=13, who required an airway intervention and had pre-existing cardiac disease with a score 25 would have a predicted probability of serious injury or mortality of 88%.