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

Total joint arthroplasty is a common procedure for which consistent, clinic ally satisfactory outcomes are expected. Data from 796 total joint procedur es performed by 20 surgeons at one center were evaluated to identify source s of variability in costs (as measured by hospital charges) where clinical outcome is expected to remain constant. Stepwise multivariate regression ch aracterized the contribution of six variables to hospital charges listed in order of explanatory power: postoperative length of stay, surgical time, p atient preoperative morbidity, units of blood transfused, perioperative com plications, and procedure type (hip or knee) accounted for 46% of variabili ty in hospital charges (multiple R-2). In a subsequent analysis, after statistical adjustment for preoperative com orbid diagnoses, the sampling distribution of mean values for surgical time , total units of blood transfused, and total hospital charges were summariz ed 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 findin gs suggest there is substantial variability in hospital charges not attribu table to patient characteristics or category of procedure-a distinct and ec onomically significant portion of this variability is practitioner specific .