Purpose: Total knee replacement (TKR) is a frequently used procedure to relieve pain and improve quality of life in patients with end stage knee arthritis. Centers performing low volumes of TKR have worse outcomes than higher volume centers. Regionalization policies that shift patients to higher volume centers are being considered as a means of improving TKR outcomes. The cost-effectiveness of having TKR in high volume centers, as compared with low volume centers has not been established. Methods: We built a decision tree to estimate the incremental cost per quality-adjusted year of life gained for two TKR strategies over a two-year period following TKR: (1) having TKR in high volume center (>200 TKR annually in Medicare population); and (2) having TKR in a low volume center (<26 TKR annually). Population characteristics, rates of complications and mortality as well as quality of life after TKR stratified by hospital volume were derived from Medicare claims data. Cost data were derived from published literature. To further examine the sensitivity of our results to variation in imperfect model parameters, we performed a probabilistic sensitivity analysis, drawing values from distributions for difference in costs between high and low volume centers and the magnitude of volume-outcomes relationship. Results: Having TKR in a low volume center was a dominated strategy as long as the cost of TKR in a low volume center was at least as high as cost of TKR in high volume center. Even when the cost of TKR in low volume center was 20% lower than the cost in a high volume center, TKR in high volume center exhibited a C-E ratio of $53,000 /QALY compared to the low volume center. If decision makers are willing to pay $20,000 for each QALY gained, delivery of TKR in high volume center will be more cost-effective than in low volume center greater than 90% of the time. Conclusions: While a substantial number of TKRs are performed in low volume centers, delivery of TKR in high volume centers is not only more effective but also cost-effective. Debate surrounding regionalization polices for improving the quality of total joint replacement should include data on cost-effectiveness in addition to the volume-outcome relationship.