Article’s

Sepsis is a life-threatening condition caused by the body’s dysregulated response to infection, often progressing to multi-organ failure and death if not identified promptly. Early detection remains a significant clinical challenge due to the subtle and heterogeneous presentation of early-stage symptoms. This paper proposes QHySep, a novel Quantum-Hybrid Sepsis Detection Framework that integrates Quantum Support Vector Classification (QSVC) for quantum kernel-based feature mapping with a Random Forest classifier for final binary prediction. Clinical data comprising vital signs, laboratory findings, and organ function indicators were preprocessed using median imputation and SMOTE-based resampling to address class imbalance, followed by SelectKBest feature selection (k=10) using ANOVA F-statistics. The QHySep model achieved 97.33% classification accuracy and an AUC of 0.950, outperforming standalone Random Forest (97.15%) and RF trained on QSVC-selected features (96.67%). These results demonstrate that quantum feature extraction meaningfully enhances classical classifier performance, offering a precise and computationally efficient tool to support clinicians in early sepsis identification and timely intervention.