Validation of the Modified Sequential Organ Failure Assessment Score for Early Clinical Deterioration in Prehospital Patients with Seizures: A Multicenter Cohort Study.
Journal of emergency nursing · 2026-01-03 · Method & validation
Abstract
INTRODUCTION: Seizures are a common emergency condition associated with significant clinical and logistical burden, particularly in the prehospital setting. Early identification of patients at risk of serious complications remains a challenge. The modified sequential organ failure assessment score, originally developed to predict mortality, may offer a rapid and feasible tool for risk stratification in this context. Its objective is to evaluate the predictive performance of the modified sequential organ failure assessment score for identifying early clinical deterioration in prehospital patients with seizures. METHODS: A multicenter, retrospective study was conducted including 477 adult patients with seizures treated by advanced life support units in 3 Spanish provinces. Early clinical deterioration was defined as a composite outcome including: use of ≥3 anticonvulsants in the prehospital setting, mechanical ventilation (prehospital or in-hospital), intensive care unit admission, or early in-hospital mortality (within 2 days). Clinical, analytical, and outcome data were collected prospectively and analyzed using receiver operating characteristic curves and calibration models. The Youden index was used to determine the optimal modified sequential organ failure assessment threshold. RESULTS: Early clinical deterioration occurred in 13.2% of patients. The modified sequential organ failure assessment score showed good discriminative ability (area under the curve, 0.790; 95% CI, 0.726-0.853), particularly for ruling out low-risk patients (negative predictive value, 92.5%). At the optimal threshold of 5.5, the score yielded a sensitivity of 71.4% and a specificity of 73.7%. DISCUSSION: The modified sequential organ failure assessment score demonstrated moderate-to-good performance in predicting early clinical deterioration in patients with seizures in the prehospital setting. Its simplicity, real-time applicability, and high negative predictive value make it a promising tool for early triage and decision making, especially in resource-limited environments.