Analysis of cost variation in craniotomy for tumor using 2 national databases

BACKGROUND: There is a significant increase and large variation in craniotomy costs. However, the causes of cost differences in craniotomies remain poorly understood. OBJECTIVE: To examine the patient and hospital factors that underlie the cost variation in tumor craniotomies using 2 national databases: the National Inpatient Sample (NIS) and Vizient, Inc. (Irving, Texas). METHODS: For 41 483 patients who underwent primary surgery for supratentorial brain tumors from 2001 to 2013 in the NIS, we created univariate and multivariate models to evaluate the effect of several patient factors and hospital factors on total hospital cost. Similarly, we performed multivariate analysis with 15 087 cases in the Vizient 2012 to 2015 database. RESULTS: In the NIS, the mean inflation-adjusted cost per tumor craniotomy increased 30%, from $23 021 in 2001 to $29 971 in 2013. In 2001, the highest cost region was the Northeast ($24 486 ± $1184), and by 2013 the western United States was the highest cost region ($36 058 ± $1684). Multivariate analyses with NIS data showed that male gender, white race, private insurance, higher mortality risk, higher severity of illness, longer length of stay, elective admissions, higher wage index, urban teaching hospitals, and hospitals in the western United States were associated with higher tumor craniotomy costs (all P < .05). Multivariate analyses with Vizient data confirmed that longer length of stay and the western United States were significantly associated with higher costs (P < .001). CONCLUSION: After controlling for patient/clinical factors, hospital type, bed size, and wage index, hospitals in the western United States had higher costs than those in other parts of the country, based on analyses from 2 separate national databases.