In an environment where decision makers are willing to reimburse ESA, our base case analysis suggests that treatment to a target Hb of 110 g/L is most likely to be cost-effective. This strategy, however, will lead to higher costs (mainly due to ESA acquisition) compared to the low Hb target strategy, and it is based on the assumption that the intermediate target will improve QoL compared with the low target, which is unproven. Given the generally modest clinical benefit of ESA and the direct relationship between dose and cost, it may be prudent to consider a maximum ESA dose (above which the dose would not be increased further, even if the Hb target were not reached). Future research should focus on this comparison. In the interim, decision makers might reasonably choose to reimburse only the low Hb strategy because of the uncertainty in the quality-of-life gains associated with the intermediate strategy. Because of the higher cost of IV epoetin, the merits of reimbursing only SC epoetin (or darbopoetin by either route) should be explored. Lastly, because even small differences in potency per unit cost of ESA can translate into large differences in total costs, head-tohead comparisons of epoetin and darbepoetin should be considered.