Variation in hospital charges for total joint arthroplasty: an investigation of physician efficiency

Total joint arthroplasty is a common procedure for which consistent, clinically satisfactory outcomes are expected. Data from 796 total joint procedures performed by 20 surgeons at one center were evaluated to identify sources of variability in costs (as measured by hospital charges) where clinical outcome is expected to remain constant. Stepwise multivariate regression characterized the contribution of six variables to hospital charges listed in order of explanatory power: postoperative length of stay, surgical time, patient preoperative morbidity, units of blood transfused, perioperative complications, and procedure type (hip or knee) accounted for 46% of variability in hospital charges (multiple R2). In a subsequent analysis, after statistical adjustment for preoperative comorbid diagnoses, the sampling distribution of mean values for surgical time, total units of blood transfused, and total hospital charges were summarized and compared among surgeons. Despite adjustment for comorbid diagnoses, substantial variation and significant differences remained between surgeons in markers of resource utilization and "surgical efficiency." These findings suggest there is substantial variability in hospital charges not attributable to patient characteristics or category of procedure--a distinct and economically significant portion of this variability is practitioner specific.