Understanding variability in hospital-specific costs of coronary artery bypass grafting represents an opportunity for standardizing care and improving resource use

Abstract

This study was undertaken to examine interhospital variability in inpatient costs of coronary artery bypass grafting (CABG). The Nationwide Inpatient Sample was used to identify isolated CABGs performed between 2005 and 2008 in the United States. Charges for inpatient care were supplied by the data set, and hospital charge-to-cost ratios were used to derive inpatient costs for each patient and aggregated at the hospital level. Mixed-effect linear regression models were created to evaluate variability in costs between hospitals adjusting for 34 patient, operative, complication, and hospital-related variables. A total of 633 hospitals performed isolated CABG in 183,973 patients. In unadjusted analysis, there was significant baseline variability in average inpatient costs of CABG between hospitals (SD, $12,130; P<.001). This variability represented 30% of the overall unadjusted average cost of performing CABG per hospital ($40,424). After risk adjustment, significant variability in average costs between hospitals persisted (P<.001). Of the 34 additional variables included in the model, only hospital region, postoperative sepsis, in-hospital mortality, and need for ventricular assist device, extracorporeal membrane oxygenation, permanent pacemaker, or implantable cardioverter-defibrillator were stronger predictors of increased costs compared with the hospital effect. There is a wide variation in the cost of performing CABG in the United States. We determined that individual hospital centers, independent of multiple patient- and outcome-specific factors, are drivers of these differences. Comparison of hospital-specific behavior with identification of the causes of cost discrepancies represents an opportunity for standardization of care and improvement in resource use.