Abstract
PURPOSE: There are no population-based data on hospitalization rate for toxicity from breast cancer chemotherapy, and even large clinical trials often do not report this information. Medicare data, linked to the Surveillance, Epidemiology, and End-Results (SEER) tumor registries, are now used to assess rates of hospitalization for chemotherapy-related toxicity in a population-based setting. PATIENTS AND METHODS: A total of 35,060 women diagnosed with stages I through IV breast cancer aged >or= 65 from 1991 through 1996 were identified from the SEER-Medicare linked program and studied. Patients were defined as being hospitalized for adverse effects of chemotherapy if there was a Medicare inpatient claim for neutropenia, fever, thrombocytopenia, or adverse effect of systemic therapy less than 7 months after diagnosis of breast cancer. RESULTS: More than 9% of women with breast cancer who received chemotherapy were admitted with the diagnosis of neutropenia, fever, thrombocytopenia, or adverse effect of systemic therapy, compared with 0.5% of women with breast cancer who did not receive chemotherapy. The rates for stage I to IV were 6.3%, 8.1%, 12.3%, and 13.2% in those treated with chemotherapy, and 0.4%, 0.6%, 0.7%, and 1.5% in women not treated with chemotherapy. The hospitalization rates for adverse effects increased significantly with comorbidity score and varied more than two-fold across the nine SEER areas but did not vary by age. Use of anthracycline-containing chemotherapy agents was associated with greater odds of these toxicities (eg, odds ratio, 2.53 for neutropenia; 95% confidence interval, 1.97 to 3.26). CONCLUSION: This study demonstrated the feasibility of using Medicare data to assess rates of hospitalization for serious toxicity associated with cancer chemotherapy. Rates in actual practice were higher than those reported in clinical trials and did not vary by age.