Meu SciELO
Serviços Personalizados
Journal
Artigo
Indicadores
- Citado por SciELO
- Acessos
Links relacionados
- Citado por Google
- Similares em SciELO
- Similares em Google
Compartilhar
Cirugía Plástica Ibero-Latinoamericana
versão On-line ISSN 1989-2055versão impressa ISSN 0376-7892
Resumo
BRITO, Íris et al. Mortality prediction models in the burn unit. Comparative study. Cir. plást. iberolatinoam. [online]. 2020, vol.46, suppl.1, pp.91-96. Epub 22-Jun-2020. ISSN 1989-2055. https://dx.doi.org/10.4321/s0376-78922020000200015.
Background and objective
Prognostic scoring models in burn patients predict mortality risk in burn-injured populations. Mortality remains a main outcome measure for burn care.
The study aims to compare different mortality prediction models in the burn unit (BU) and identify their potential applications.
Methods
Retrospective study of 409 burn patients admitted in the BU between 2016 and 2018, including mortality assessment and comparative analysis of survivors and deceased patients’characteristics based on admission demographic data and clinical criteria. Evaluation of four models (Abbreviated Burn Severity Index (ABSI), Ryan model, Belgian Outcome in Burn Injury (BOBI) and revised-Baux) on assessment of predicted mortality versus observed, and analysis of models’ discriminative power and goodness-of-fit (ROC curves and area under the curve - AUC; and Hosmer-Lemeshow tests).
Results
Patients mean age was 59.1 years and mean total burned surface area (TBSA) was 12%; 56.5% were male; third degree burns were present in 67% and inhalation injury in 11.5%. Overall observed mortality rate was 6.4% (n=26). Deceased patients were significantly older (71.7 vs. 58.3 years; p<0.001), presented a larger TBSA (23.2 vs. 11.2%; p<0.001) and higher frequency of third degree burns (92.3 vs. 65.3%; p=0.005), with no significant difference on patient gender and inhalation injury.
Mortality prediction models presented adequate goodness-of-fit and discrimination. All models showed p-values >0.05 in Hosmer-Lemeshow test assessment. Revised-Baux and ABSI showed good discriminative power (AUC 0.84 ± 0.04 and 0.81 ± 0.04, respectively), BOBI discrimination was moderate / good (0.79 ± 0.04) and moderate with Ryan (0.73 ± 0.05).
Conclusions
The four mortality indices revealed appropriate predictive performance, with revised-Baux presenting as the most accurate model for prediction of mortality. Their use in the BU represents a valuable tool for risk stratification, quality control of burn care, appraisal of new therapeutic strategies and scientific research purposes.
Palavras-chave : Burn Unit; Burns; Mortality; Prognosis.