Background: Studies of volume-outcome relationships typically subdivide volume via non-evidence-based methods, producing categories that vary widely among studies, preclude the comparison of results, and possibly obscure the true volume-outcome relationships. The goal of the current study was to use quantitative methods to derive meaningful, risk-based categories for hospital and surgeon total hip arthroplasty (THA) volume based on relationships with mortality, complications, and revision.
Methods: Using New York statewide patient data (1997 to 2014; n = 187,557), we derived risk-based hospital and surgeon-volume categories for primary THA based on relationships with 90-day complications and mortality and 2-year revision.
Results: The following categories, based on relationships with complications, mortality, and revision, were derived for surgeon volume: 0 to 12, 13 to 25, 26 to 72, 73 to 165, 166 to 279, and ≥280 THA/year. For hospital volume, the categories derived were 0 to 11, 12 to 54, 55 to 157, 158 to 526, and ≥527 THA/year. More than 35% of THA cases in New York State were conducted by surgeons performing ≤1 THA/month (0 to 12 THA/year), and these were associated with a 2 to 2.5-fold increase in the risk for complications, mortality, and revision relative to higher-volume surgeons. Similarly, 15% of THA cases in New York State were conducted in hospitals performing ≤1 THA/week (0 to 11 or 12 to 54 THA/year), and these were associated with a nearly 1.5-fold increase in complications and between a 4 and 6-fold increase in mortality. Traditional non-evidence-based quartile categories were concentrated at lower volumes, did not capture the full magnitude of the volume-related differences, and were a poorer representation of the outcome data, as assessed by several model metrics. Thus, quartiles showed only a <2-fold increase in complications, mortality, and revision for the lowest versus the highest surgeon-volume quartile and failed to show the increased risk for lower versus higher hospital volumes.
Conclusions: The volume-outcome relationships in THA are more pronounced than previously apparent through standard statistical techniques. Volume-based strategies for improving outcomes in THA should use benchmarks that are evidence-based to achieve optimal results.
Level of evidence: Therapeutic Level III. See Instructions for Authors for a complete description of levels of evidence.