Geographic diversity of low-volume hospitals in total knee replacement: implications for regionalization policies.

Background/Objectives: Centers performing low volumes of total knee replacements (TKR) have worse outcomes of TKR than higher volume centers. Regionalization policies that shift patients to higher volume centers are being considered as a means of improving TKR outcomes. We sought to describe geographic diversity in the distribution of low-volume centers and examine state level characteristics associated with states that have a higher proportion of low-volume centers and/or a higher proportion of TKRs performed in low-volume centers. Methods: We used U.S. Census data and geocoded Medicare claims to ascertain state-level demographic factors, procedure volume, and TKR rates and to conduct our state level analysis. We defined 2 outcomes: 1) proportion of all hospitals with a low annual TKR volume (<26 per year in the Medicare population); and 2) proportion of all TKRs in the Medicare population performed in low-volume centers. We examined linear associations among the 2 outcomes and state factors, and used multivariate regression to identify factors associated independently with these outcomes. Results: Half of hospitals performing TKR in the Medicare population were low-volume centers, accounting for 13% of TKRs. Multivariate analysis revealed lower TKR rates, higher proportion of rural areas and larger state area were associated with a higher proportion of low-volume hospitals in a state. Lower proportion of elderly residents, higher population density and higher proportion of rural areas predicted a higher proportion of TKRs performed in low-volume centers. Conclusions: The distribution of low-volume hospitals among U.S. states varies substantially. Regionalization of TKR may require different strategies in states with small and large numbers of low-volume centers.