Do standard burn mortality formulae work on a population of severely burned children and adults?

Area of Science:

• Medical research
• Clinical epidemiology

Background:

• Accurate burn mortality prediction is crucial for audits, treatment planning, and resource allocation.
• Existing burn scoring systems (Ryan Score, ABSI, Baux) require validation in severely burned populations (≥20% TBSA) and comparison with APACHEII.
• The revised Baux (R-Baux) needs further external validation, and the pediatric Baux (P-Baux) lacks validation.

Purpose of the Study:

• To compare the performance of common burn mortality prediction formulae against APACHEII in severely burned patients.
• To re-validate existing formulae and validate the pediatric Baux score in a contemporary, severely burned cohort.
• To assess the calibration and discrimination of various burn scoring systems.

Main Methods:

• Analysis of data from 522 severely burned patients.
• Comparison of burn formulae (Ryan Score, ABSI, Baux, R-Baux, P-Baux) with APACHEII using AUROC.
• Assessment of calibration through mortality overestimation/underestimation.

Main Results:

• Burn formulae (ABSI, Baux, R-Baux) significantly outperformed APACHEII in adults (p<0.001).
• The revised Baux demonstrated excellent discrimination (AUROC 0.908) and calibration.
• ABSI and P-Baux showed good discrimination in children (AUROC >0.82) but poor calibration due to significant mortality overestimation.

Conclusions:

• Burn-specific formulae are superior to APACHEII for mortality prediction in severely burned adults.
• The revised Baux score is a reliable tool for predicting mortality in severely burned patients.
• Challenges remain in developing universally applicable and well-calibrated burn scoring systems, particularly for pediatric populations.