Hospital length of stay--does 1% TBSA really equal 1 day?

Abstract

Length of stay (LOS) continues to be a standard variable when evaluating progress and outcomes in burn care. Common wisdom would dictate that this measure is linearly related to TBSA. Is this truly the case? A retrospective review of the National Burn Repository was conducted to evaluate factors that affect hospital LOS. The National Burn Repository data set was obtained from the American Burn Association. Data from the years 2002-2007 were extracted. Unique patients were identified by removing readmissions, outpatients, and patients not admitted. Patients whose "HOSPLOS" and/or "AREATOT" field was blank or 0 were excluded, as were nonthermally injured patients. Patients without an entry for age and dead patients were also excluded. This left a final data set of 52,712 patients for analysis. The data were then analyzed, with %TBSA burned as the independent variable. In patients who survived their entire LOS, the mean LOS increased linearly by decile. Females with a TBSA <40% have a trend toward increased LOS relative to their male counterparts of the same TBSA; however, this trend reverses for TBSA ≥40%. Age alone is not a significant predictor of increasing LOS. The cause of burns was predominantly flame related across all deciles of TBSA, and most etiologies for burn demonstrate the predicted increase in LOS per %TBSA. LOS was not significantly affected by insurance type or whether the injury was work related or not. This analysis was confounded by the small numbers of patients with burns >60% and age older than 70 years. Anticipating hospital LOS is not a simple task. Using complex statistical analysis, a linear trend associated with %TBSA can be seen; however, other variables do contribute. Until the precise role of these variables can be elucidated, anticipating patient LOS to be 1 day for every %TBSA is still a useful exercise.