Competing risks and burn outcomes: more questions than answers.

Abstract

Age, total body surface area of burn injury, and inhalation injury have long been documented as independent predictors of mortality in burn injury1; all these factors are identifiable on admission. In this issue, Taylor et al2 propose the use of competing risk analysis to examinedynamic factors that may affect lengthof stay andmortality in patientswith burns. Although thismethod isnew inpatientswithburns, itwas initially applied in a large cohort of patients with trauma more than adecade ago.3Modeling outcomes in patientswith complexconditions, suchasburnsandtrauma, is inherently fraught with challenges. The key assumptionunderlying theuse of competing risk analysis is that the causes of early and late mortality are different in patients with burns as they are in patients with trauma.4Theauthors includea singleparagraphwith a simple univariate description of patients who died within 1 day of admission. Does this description provide a complete picture of patients who have early deaths following burn injury, or is thepicturemore complicated than the total body surface area of larger burn injuries and greater incidence of inhalation injury? Could depth of burns also play a role? Presumably, the 24-hour cutoff is designed to capturepatientsplacedontopalliative care early; while National Burn Repository data do not include this information, do patients for whom the decision to start palliative care is later lookmore similar topatientswho die early, or more similar to patients who die later? A trend analysishighlightingpractice changes through theyears could be developedusing the included 9 years of data andwould allowus toanswer theexceptionally importantquestion—arewe getting better at the care of patients with burns? Theuseofcompeting riskanalysismarksan importantnew way for us to consider outcomes in patients with burns. Ultimately, however, this analysis results in as many—or more— questions than answers.