PURPOSE: To identify variables that might be predictive of sexual health (interest, dysfunction, and satisfaction) in a large sample of breast cancer survivors, with a validation conducted in a second, independent sample. PATIENTS AND METHODS: On the basis of a conceptual framework of sexual health in breast cancer survivors, we performed multivariable regression analyses to estimate sexual interest, dysfunction, and satisfaction in both samples. Additional analyses were performed using stepwise regression and recursive partitioning to explore in each sample the relative contributions of the independent variables toward predicting the outcome measures. RESULTS: The models for sexual interest accounted for at least 33% of the variance, and the significant predictors common to the two samples were having a new partner since the diagnosis of breast cancer, mental health score, and body image score. For sexual dysfunction, the models in the two samples explained at least 33% of the variance, and the common si...