Hospital teaching status associated with reduced inpatient mortality and perioperative complications in surgical neuro-oncology.

Abstract

Studies have demonstrated that higher surgical volumes correlate with improved neurosurgical outcomes yet none exist evaluating the effects of hospital teaching status on the surgical neuro-oncology patient. We present the first analysis comparing brain tumor surgery perioperative outcomes at academic and non-teaching centers. Brain tumor surgeries in the Nationwide Inpatient Sample (NIS) from 1998 to 2014 were identified. A teaching hospital, defined by the NIS, must have ≥ 1 Accreditation Council of Graduate Medical Education (ACGME) approved residency programs, Council of Teaching Hospitals membership, or have a ratio ≥ 0.25 of full-time residents to hospital beds. Annual treatment trends were stratified by hospital teaching status, assessing yearly caseload with linear regression. Multivariable logistic regression determined predictors of inpatient mortality/complications. Hospitals were further divided into quartiles by case volume and teaching status was compared in each. Teaching hospitals (THs) exhibited an average annual increase in brain tumor surgeries (+ 1057/year, p < 0.0001). In multivariable analysis, teaching status was associated with decreased risk of mortality (OR 0.82, p = 0.0003) and increased likelihood of discharge home (OR 1.21, p < 0.0001). In subgroup analysis, within the highest hospital quartile by caseload, higher mortality rates and lower routine discharges were again seen at non-teaching hospitals (NTHs) (p = 0.0002 and p = 0.0016, respectively). THs are performing more brain tumor surgeries over time with lower rates of inpatient mortality and perioperative complications even after controlling for hospital case volume. These results suggest a shift in neuro-oncology practice patterns favoring THs to optimize patient outcomes especially at the highest volume centers.