101 Predicting and Estimating Burn Outcomes: A Single Institution Analysis of over 4000 Cases

Abstract

Abstract Introduction Advances in burn care have improved patient outcomes, and independently validated indices, scores, and predictors of burn outcomes warrant re-evaluation. The purpose of this study is to consolidate predictors of burn outcomes and determine the factors that significantly contribute to length-of-stay (LOS) and mortality. Methods A retrospective review of all burn patients (n = 5778) admitted to a quaternary provincial burn unit from 1973 to 2017, was conducted. Removal of blank and pediatric entries yielded 4622 independent cases. Goodness-of-fit models and multivariate logistic regression was performed. Burn predictors included %TBSA, Baux (classic, revised) index, Abbreviated Burn Severity Index, and Ryan score. Primary outcomes were mortality and LOS. Variables considered in the multivariate logistic regression included: diabetes, hypertension, smoking, obesity, alcohol use, drug abuse, full-thickness burn, ventilator support, and ICU referral. Results Multivariate logistic regression for mortality showed the classic Baux index to be a significant predictor for mortality (OR = 1.118, p < 0.001). Other predictors included male sex, ICU referral, diabetes, smoking, and alcoholism (OR = 1.96, 4.97, 2.38, 1.63, 1.98, all p < 0.05). Interestingly, hypertension had a protective effect (OR = 0.24, p < 0.013). Linear regression for LOS found %TBSA, ICU referral, alcoholism, age, male sex, significant. The area under the ROC curve for Baux index was 0.945. Conclusions The regressions show that burn mortality and LOS are best predicted with the Baux index. Hypertension may have a protective effect on burn outcomes and may be attributed to increased perfusion to the periphery. Goodness-of-fit models, although variable, tended to show tighter grouping in patients with TBSA >20%. LOS ratios prove to be useful benchmarks for burn units with TBSA >20%. Similar findings are preliminary found in the NBR, national burn repository, database. Applicability of Research to Practice LOS ratios and burn index scores prove to be valuable markers to predict burn outcomes. The results of this study will directly help the clinician make decisions and communicate clinical severity to patient and family members.