BACKGROUND: The purpose of this study was to compare the rates of all-cause and breast cancer-specific mortality after breast-conserving surgery (BCS) only, BCS plus radiation therapy (RT), mastectomy, and the receipt of adjuvant tamoxifen in a large population-based cohort of older women with early-stage disease. STUDY DESIGN: This cohort study was conducted within six US integrated health-care delivery systems. Automated administrative databases, medical records, and tumor registries were used to identify women aged 65 years or older who received BCS or mastectomy to treat stage I or II breast cancer diagnosed from January 1, 1990, through December 31, 1994. We compared cause-specific 10-year mortality rates across treatment categories by fitting Cox proportional hazards models adjusted for demographics and tumor characteristics. RESULTS: We identified 1,837 women having operations for stage I or II breast cancer. Compared with women receiving mastectomy, those receiving BCS without RT were twice as likely to die of breast cancer (adjusted hazards ratio [HR]=2.19, 95% confidence interval [CI], 1.51 to 3.18). Breast cancer mortality rates were similar between women receiving BCS plus RT and women receiving mastectomy (adjusted HR=1.08, 95% CI, 0.79 to 1.48). In the subset of 886 chemotherapy-naive women treated with tamoxifen, those treated with tamoxifen for less than 1 year had a substantially higher breast cancer mortality rate than those exposed 5 years or more (adjusted HR=6.26, 95% CI, 3.10 to 12.64). CONCLUSIONS: Our findings indicate that older women receiving BCS alone have higher rates of breast cancer death than those receiving BCS + RT or mastectomy and that the survival benefit from tamoxifen increases with increasing duration of treatment.