Choice of Hospital for Revision Total Hip Replacement.

Background: Little is known about how often patients have revision total hip replacement in the same hospital in which they had the primary procedure. Methods: We examined Medicare claims data to identify patients who had primary total hip replacement from July 1995 to June 1996 and subsequently had revision through December 31, 2006. We examined whether the revision was performed in the same or different hospital from the primary procedure, with different hospitals being categorized as being in a lower, a higher, or the same hospital volume stratum. Hospital strata included twenty-five or fewer cases of total hip replacement annually in the Medicare population, twenty-six to fifty cases, fifty-one to 100 cases, and >100 cases. We calculated the number of revisions generated (primary procedures eventuating in revision) by hospitals in each volume stratum and the number of revisions performed in these hospitals. Results: Of 4448 revision procedures, 3306 (74%) were performed in hospitals in the same volume stratum as the hospital where the primary procedure was performed. Four hundred twenty-nine revisions (9.6%) were performed in a lower-volume hospital, and 713(16%) were performed in a higher-volume hospital. Thirty-one (3%) of 960 patients who had revision within one year after the primary total hip replacement had the revision in a lower-volume center, compared with 204 (15%) of 1393 who had revision more than six years after the primary procedure (odds ratio = 4.6 ; 95% confidence interval, 3.0 to 6.8). The ratio of revisions performed to revisions generated was 1.21 for the highest-volume centers and 0.86 for the lowest-volume centers. Conclusions: Of 4448 revisions examined in this study, 429 (<10%) were performed in centers with a lower-volume of total hip replacement than the center at which the initial hip replacement was performed, whereas 713 (16%) were performed in higher-volume centers. Higher-volume centers performed 21% more revisions than they generated (531 revisions performed, compared with 438 generated). These data will help to inform health-care policy with regard to the utilization of resources for revision total hip replacement. [ABSTRACT FROM AUTHOR]