The penetrating injury risk of non-home discharge (PIRD) score: Development and validation of a predictive tool using the National Trauma Data Bank.

Abstract

Penetrating trauma disproportionately affects young patients, yet many survivors cannot be discharged directly home and instead require post-acute care. Identifying these patients early could expedite rehabilitation referral and discharge planning, but no validated tool predicts non-home discharge in penetrating trauma. We developed and internally validated the Penetrating Injury Risk of Non-home Discharge (PIRD) score. Using the National Trauma Data Bank (NTDB) Trauma Quality Improvement Program Participant Use Files (2019-2024), we selected adult patients (≥18 years) with penetrating trauma who survived to discharge. The outcome was non-home discharge to post-acute care (e.g., skilled nursing or inpatient rehabilitation) versus home. The cohort was split 2:1 into derivation and validation sets, stratified by outcome. Candidate predictors available early in the hospital course (demographics, insurance, mechanism, emergency-department physiology, the Injury Severity Score [ISS], body-region injury, transport, and comorbidities) underwent univariate screening, collinearity assessment, and backward-stepwise multivariable logistic regression. Unlike the orthopedic FORD score, ISS was retained because penetrating trauma is a high-severity, mortality-driven population. Coefficients were converted to integer points, scaled 0-10. Among 302,399 patients, 21,186 (7.0%) had non-home discharge. The 40-predictor model was led by spinal-cord injury, critical injury severity (ISS ≥25), advanced age, and severe traumatic brain injury, whereas stab mechanism and private-vehicle arrival lowered the odds. Validation discrimination was strong (area under the receiver operating characteristic curve [AUROC] 0.866, 95% confidence interval 0.861-0.871) with excellent calibration (Brier 0.046, calibration slope 1.00), exceeding the GTOS-II, TRIAGES, FORD, and DEPARTS scores (all p < 0.001). Non-home discharge rose from 0.7% (low risk) to 37.5% (high risk); at a score ≥4, sensitivity was 86.0% and negative predictive value 98.5%. Derived from variables available early in care, the PIRD score accurately predicts post-acute care discharge in adult penetrating trauma and outperforms existing severity scores, supporting timely discharge planning and resource allocation.