Assessment of Abbreviated Burn Severity Index Scoring System for Predicting Mortality Outcome in Burns at Kirtipur Hospital

Abstract

Introduction: Global burn incidence is disproportionately high in low- and middle-income countries, with Southeast Asia serving as an “epicenter” for fire-related deaths. Clinicians frequently face challenges in predicting patient outcomes upon initial presentation, necessitating the use of validated scoring systems.  This study evaluates the validity of the Abbreviated Burn Severity Index (ABSI) for predicting mortality at Kirtipur Hospital, Nepal. 
Method: This observational study reviewed 800 burn in-patients admitted between July 2024 and June 2025. A pre-formatted ABSI proforma was utilized for data collection. Statistical analyses, including univariate and multivariate logistic regression and ROC analysis were performed to identify independent mortality predictors and determine the Lethal Area 50 (LA50).
Result: The overall mortality rate was 11.1% (89 deaths). Inhalation injury (OR=14.74) and full-thickness burns (OR=9.06) were the most significant independent predictors of death. While the ABSI score demonstrated excellent univariate discrimination (AUC=0.904), its multivariate contribution was attenuated (OR=0.88, p=0.74), suggesting the score largely captures severity already reflected by TBSA and burn depth. The LA50 point estimate was 47.43% TBSA. Mortality risk stratified by ABSI scores ranged from 9.03% (score ≤10) to 100% (score ≥14).
Conclusion: The ABSI is a reliable tool for risk stratification in the local Nepalese context but provides limited independent prognostic information beyond core clinical variables like inhalation injury. Developing locally calibrated models or adopting modified scoring systems is recommended to improve bedside accuracy for family counseling and resource allocation