Burn Mortality in an Appalachian Referral Center: An Examination of Mortality Prediction Scores in a 13-Year Retrospective Study

TL;DR

This study evaluated how well different burn mortality scores predict outcomes at a hospital in West Virginia, finding that Baux, revised Baux, and BOBI scores are effective tools for identifying high-risk patients.

Contribution

The study provides region-specific validation of burn mortality scores in an Appalachian referral center, highlighting their clinical utility for resource allocation and patient care.

Findings

Baux and revised Baux scores had the highest accuracy in predicting mortality with AUCs of 0.926 and 0.927, respectively.

Patients with Baux scores above 79 had 42.6 times higher mortality odds compared to those with scores of 79 or lower.

Age and total body surface area burned were identified as key factors influencing mortality risk in burn patients.

Abstract

Introduction

Burn injuries have profound implications, prompting the use of various mortality scoring systems. This study aimed to evaluate their effectiveness within our Appalachian burn referral center, which serves as the sole burn center in the state of West Virginia. Given this unique status, understanding the efficacy of mortality scoring systems within our center is crucial for resource allocation and optimizing patient outcomes in our region.

Methods

A retrospective analysis of patients admitted to Cabell Huntington Hospital Burn Intensive Care Unit (BICU) from January 2010 to June 2023 was conducted, assessing Baux (B), revised Baux (rB), Belgian Outcome in Burn Injury (BOBI), and Abbreviated Burn Severity Index (ABSI) scores. Logistic regression and receiver operating characteristic analysis were employed to examine survival status and determine optimal cut points.

Results

Among 1,104 patients, 57 died (5% mortality rate). Deceased patients had significantly higher B/rB/BOBI scores (mean: 98/98/92) than survivors (45/46/4.19) (p < 0.001), with ABSI showing no significance (p = 0.079). Each one-point increase in B/rB/BOBI scores correlated with a 1.09/1.09/2.34 times higher mortality risk (p < 0.001). The AUC for B score in predicting mortality was 0.926 (95% CI: 0.890, 0.962), with sensitivity and specificity values of 0.789 and 0.92, respectively, and an optimal cutoff point of 79. The AUC for the rB score was 0.927 (95% CI: 0.892, 0.962), with sensitivity and specificity values of 0.789 and 0.926, respectively, and an optimal cutoff point of 80. The AUC for the BOBI score was 0.901 (95% CI: 0.865, 0.937), with sensitivity and specificity values of 0.895 and 0.775, respectively, and an optimal cutoff point of 2. For patients with B scores above 79, their odds of mortality were 42.6 times higher than those with B scores of 79 or lower (95% CI: 22.6, 85.6, p < 0.001). Similarly, for patients with rB scores exceeding 80, their odds of mortality were 42.9 times higher than those with rB scores of 80 or lower (95% CI: 22.9, 84.8, p < 0.001). Finally, for patients with BOBI scores greater than 2, their odds of mortality were 17.8 times higher than those with BOBI scores of 2 or lower (95% CI: 9.88, 33.4, p < 0.001).

Conclusion

Our study underscores the vital role of mortality scoring systems in guiding clinical decision-making and resource allocation for burn patients, particularly within the Appalachian region served by the Cabell Huntington Hospital BICU. By leveraging tools such as the Baux, revised Baux, and BOBI scores, healthcare providers can identify high-risk patients early in their treatment course, facilitating personalized interventions and improving overall patient outcomes. Moreover, our findings highlight the significance of age and total body surface area burned as key determinants of mortality risk, emphasizing the need for tailored approaches to care for elderly patients and those with extensive burns. Continued research and refinement of mortality scoring systems are essential to further enhance their effectiveness and ensure optimal patient care in the challenging field of burn management.

Figures