Optimal therapy for patients with metastatic breast cancer is unknown. For those with rapidly progressive disease or liver metastases, combination chemotherapy is believed to be the treatment of choice. However, most patients do not fall into that category. For them, two treatment approaches can be considered: (1) ignore estrogen receptor/progesterone receptor (ER/PR) data and treat all patients with hormonal therapy or (2) use ER/PR data and treat patients with ER-/PR- tumors with combination chemotherapy and treat all other patients with hormonal therapy. This report presents a decision analysis for assessing the usefulness of ER/PR data in this situation. The probabilities of response to hormonal therapy were obtained from literature reviews. Patients' utilities or preferences for the various outcomes were estimated by using physicians as patient surrogates. The baseline analysis favored ignoring ER/PR data and treating all patients with hormonal therapy. Extensive sensitivity analyses showed that the decision was most affected by patients' utilities for the various outcomes. The conclusion is that patients' utilities are most important in deciding the usefulness of ER/PR data in metastatic breast cancer.