Costs and predictors of 30-day readmissions after craniotomy for traumatic brain injury: a nationwide analysis

OBJECTIVE There is increasing interest in the use of 30-day readmission (30dRA) as a quality metric to represent hospital and provider performance. Data regarding the incidence and risk factors for 30dRA after traumatic brain injury (TBI) are sparse. The authors sought to characterize these variables using a national database. METHODS The Nationwide Readmissions Database was used to identify patients with a primary diagnosis of TBI who underwent craniotomy or craniectomy between 2010 and 2014. Our primary outcome of interest was 30dRA. Binary logistic regression was used to identify variables related to patient demographics, comorbidities, and index hospital admission that were associated with 30dRA. RESULTS A total of 25,354 patients met the inclusion criteria. The 30dRA rate during the entire study period was 15.5%. In 2010 the 30dRA rate was 16.8% and in 2014 it decreased to 15.1% (pooled OR 0.90, 95% CI 0.87-0.94). The mean cost associated with a 30dRA increased slightly but significantly, from $9999 in 2010 to $10,114 in 2014 (p = 0.021). Factors associated with increased odds of 30dRA in the binary logistic regression included increased age, greater comorbidity burden, more severe injury, tracheostomy, gastrostomy, sodium abnormality, and venous thromboembolism. In order of decreasing frequency, the most common causes for 30dRA were neurological, injury/iatrogenic, cardiovascular/cerebrovascular, infectious, and respiratory. CONCLUSIONS The incidence of 30dRA after craniotomy for TBI decreased slightly from 2010 to 2014. This study identified several variables associated with 30dRA that require confirmation in a prospective study, which could direct attempts to prevent readmissions.