The Impact of Provider Volume on Mortality after Intracranial Tumor Resection

Abstract

OBJECTIVE Policies of regionalization and selective referral for a number of “high-risk” surgical procedures are being explored and implemented as a result of significant variation in postoperative mortality between high- and low-volume providers. The effect of provider volume on outcomes after intracranial tumor resection is unknown and warrants investigation. METHODS By use of the Nationwide Inpatient Sample for 1996 and 1997, patients (older than 19 yr) who had a diagnosis of a malignant central nervous system neoplasm and underwent craniotomy or craniectomy were included. Hospital volume and surgeon volume were categorized by quartiles (very low, low, high, or very high volume). Unadjusted and case mix-adjusted analyses were performed with regard to postoperative in-hospital mortality. RESULTS The crude in-hospital mortality was 2.8% for a total of 7547 patients. The mean patient age was 55.8 years (66.5% <65; 33.5% ≥65). Mortality for very low- to very high-volume hospitals was as follows: 3.8, 3.2, 2.4, and 1.8% (P < 0.001). Mortality for very low- to very high-volume surgeons was as follows: 4.1, 3.9, 3.1, and 1.4% (P = 0.003). Predictors of mortality in a logistic regression model were emergent admission (odds ratio [OR], 2.97; 95% confidence interval [CI], 2.02–4.38;P < 0.001), and age 65 years or greater (OR, 1.63; 95% CI, 1.16–2.30;P = 0.005). The risk of mortality was reduced for very high-volume hospitals (OR, 0.58; 95% CI, 0.35–0.97;P = 0.038) and very high-volume surgeons (OR, 0.42; 95% CI, 0.22–0.84;P = 0.012). CONCLUSION Higher-volume providers have superior outcomes after surgical resection of malignant intracranial tumors. This reduction was maintained despite adjustment for case mix. As the regionalization of high-risk surgery moves forward, it is important for neurosurgeons to maintain leadership roles in the development of specialty-specific data collection and health policy initiatives that improve and reduce variation in outcomes.