OBJECTIVES: Published information on the relationship between patient characteristics such as patient age or total body surface area (TBSA) of burn on length of stay (LOS) is not available in the United States (US). Clinical expertise assumes a rule-of-thumb of 1 day per percentage TBSA, but true accuracy is unknown. The American Burn Association National Burn Repository (NBR) version 8.0 (2002-2011) was analyzed to understand the relationship between LOS and key characteristics for surviving, severe (TBSA10-40%) burn patients.

METHODS: A number of models were explored, including ordinary least squares, mixed-effects model and generalized linear models with Gaussian, Poisson, negative binomial and gamma family with log link. Independent variables considered were TBSA, TBSA of partial-thickness and full-thickness (FT) burn and age (linear, squared and cubed to account for non-linearity), hospital-acquired infection (HAI), other infection, inhalation injury, female gender and diabetes status. A backward selection stepwise procedure with p-value<0.2 was used to select variables. Several tests of model specification heteroscedasticity and goodness-of-fit were performed to select the final model.

RESULTS: A total of 26,258 patients were included in analyses (excluding 2 outliers with LOS>300). Based on model performance tests, the mixed-effects model was selected. All independent variables, except FT, were retained in the model. After adjusting for gender, age and comorbidities, predicted LOS for adults (18+) was 11.2, 19.4, 30.6 and 43.6 days for 10%, 20%, 30% and 40% TBSA respectively. Similarly for pediatrics (age<18), predicted LOS was 7.3, 16.2, 20.4 and 43.8 days for each TBSA group, respectively.

CONCLUSIONS: For adults, 1 day per percentage TBSA is approximately accurate for estimating LOS. However, this overestimates LOS for pediatrics at smaller TBSA percentages. This is the first published analysis of the relationship between LOS for severe burn patients and patient profiles in the US.