Estimated additional hospital length of stay caused by 40 individual complications in injured patients: an observational study of 204,388 patients.

Abstract

Complications are common in the care of trauma patients and increase hospital length of stay (LoS). Because many factors influence LoS and because patients may experience more than a single complication, it is difficult to estimate the effect of individual complications on LoS. We describe here a mathematically principled approach to estimating the additional LoS caused by complications and provide estimates for additional LoS caused by 40 common complications. The Pennsylvania Trauma Systems Foundation provided data on trauma patients hospitalized in one of 25 hospitals between 2002 and 2010. We estimated the attributable additional LoS in patients surviving to hospital discharge for 40 individual complications using a generalized linear model that controlled for anatomic injury and physiologic derangement at the time of admission, as well as patient factors (age, comorbidities, transfer status, type of insurance), discharge characteristics (day of week, destination), and year of admission. We also compared average risk-adjusted LoS among hospitals. Of 204,388 trauma patients surviving to discharge, 9.1% had one complication and 2.2% had multiple complications. Additional days LoS caused by individual complications ranged from less than 1 (bronchial intubation) to 16 days (wound dehiscence). Most complications added less than 1 week to LoS, but infectious complications added from 1 to 2 weeks; surgical complications added 2 to 3 weeks. If all complications could be eliminated, 24% of hospitalization days would be avoided. Individual hospitals' mean LoS differed from that predicted by our model by less than 1 day. Complications are common in the care of trauma patients and add moderately to LoS. Among all complications, surgical complications are associated with the greatest increases. Prognostic study, level III.