Mortality among women diagnosed with stage II or III breast cancer based on SEER-Medicare data.

Abstract

111 Background: Increasing emphasis on assessing the quality of treatment (tx) for cancer patients highlights the need for real-world data sets that can address critical policy questions about cancer care in the US. This study examined treatment and mortality in women with incident early-stage breast cancer (BC) 2007-2009 using the Surveillance, Epidemiology, and End Results (SEER)-Medicare linked data. Methods: We identified women age ≥66 years at the diagnosis of AJCC stage II/III BC between 2007-2009, who received surgery and were enrolled with Medicare Parts A and B. Women were followed from definitive surgery until the end of 2010, death, or change in enrollment status. Risk (cumulative incidence proportion [CIP]) of all-cause death was calculated with 95% confidence intervals (CI) using the Kaplan-Meier method stratified by tumor characteristics and tx type. Results: The cohort included 10,048 women with stage II (78%) or III (22%); 58% were ≥75 years and 78% had hormone-receptor (HR) positive tumors (77% ER positive, 64% PR positive). After the definitive surgery, 1,271 deaths reflected a CIP of 5.7% (95% CI 5.3-6.2) at 1 year, 10.9% (10.2-11.5) at 2 years, and 16.9% (15.9-17.9) at 3 years. Unadjusted mortality (Table) was higher among those without chemotherapy (chemo), women with neo-adjuvant tx, and those identified as “triple negative” phenotype (HR negative, no trastuzumab). Conclusions: Linking SEER and Medicare allowed us to assess mortality according to tumor characteristics and cancer treatments, respectively, which are both related to prognosis. The use of administrative data from Medicare should be further enhanced with inclusion of test results from standard gene expression panels to ensure a better match of therapy for the patient and adequate assessment of resource utilization. [Table: see text]