In 2007, the Centers for Medicare & Medicaid Services issued a national coverage determination (NCD) restricting the use of erythropoiesis-stimulating agents (ESAs) in patients with chemotherapy-induced anemia (CIA). We conducted a study to determine the impact of this policy change on the clinical outcomes of Medicare patients with CIA receiving ESA therapy. In this retrospective observational study of an electronic medical record database within a large oncology/hematology practice, cohorts were defined based upon the initiation of ESAs (pre-NCD cohort: January 1, 2007-April 30, 2007; post-NCD cohort: August 1, 2007-March 10, 2008); follow-up extended until the last ESA dose plus 35 days or July 31, 2007, or April 15, 2008, for the cohorts, respectively. A total of 401 Medicare patients received ESAs for 8 weeks (pre-NCD, 243; post-NCD, 158). Hemoglobin (Hb) levels were significantly higher in the pre-NCD cohort than in the post-NCD cohort at each measured time point (baseline: 10.7 g/dL vs 9.7 g/dL; P < 0.0001; week 4: 11.0 g/dL vs 10.2 g/dL; P < 0.0001; week 8: 11.2 vs 10.3; P < 0.0001). A significantly greater proportion of patients required transfusion in the post-NCD cohort than in the pre-NCD cohort (18.4% vs 9.5%; P = 0.0096). These findings warrant further assessment of the clinical implications relative to the policy change.