Evaluating individual surgeons based on total hospital costs: Evidence for variation in both total costs and volatility of costs

BACKGROUND: There is increasing interest in evaluating quality in health care, extending to the assessment of outcomes, including costs, for individual surgeons. STUDY DESIGN: Surgical patients entered in the private sector National Surgical Quality Improvement Program at the University of Michigan Medical Center between September 2003 and September 2004 were included. Patient level characteristics and outcomes measures were combined with internal hospital cost data. Analysis was performed at the individual surgeon level using hospital costs as the outcomes variable, controlling for patient case-mix variables and procedural complexity. We used an econometric statistical model combining ordinary least squares and quantile regression methods, which allowed us to examine the effect of individual surgeons on costs. RESULTS: Considerable variation in costs across surgeons is demonstrated, holding patient case mix and procedural complexity constant. This is shown for mean estimates (p < 0.001) and estimates of 10^th (p = 0.001), 50^th (p < 0.001), and 90^th (p = 0.013) percentiles. Examining the 10^th to 90^th interquantile range also demonstrates substantial variation in the ranges of costs for surgeons (p = 0.005), implying volatility in costs across providers, again holding patient case mix and procedural complexity constant. In dollar terms, 6 of 28 surgeons differ from a reference surgeon by 39% or more. CONCLUSIONS: Individual surgeons appear to have statistically and clinically significant differences in their costs and volatility of costs when holding patient factors and procedural complexity constant. Implications for quality improvement and incentive programs are discussed.