Since 1986, the recommended therapy for patients with ovarian cancer has included surgery and chemotherapy with a platinum compound (cisplatin or carboplatin). The purpose of this study is to assess the use of chemotherapy in elderly patients with advanced ovarian cancer.The Surveillance, Epidemiology, and End Results-Medicare database represents approximately 14% of the United States population and provides clinical and demographic information on cancer patients covered by Medicare, along with health care-utilization data from Medicare claims files. We analyzed the association of demographic and clinical factors with treatment among patients diagnosed from 1992 to 1996 with stage III or IV ovarian cancer, who survived > or = 120 days beyond diagnosis, and were > or = 65 years of age (N = 1,775).Approximately 83% of elderly patients received some form of chemotherapy within 4 months of diagnosis. In a multiple logistic regression model with patients aged 65 to 69 years as the reference, the odds ratios of receiving chemotherapy were 0.96 (95% confidence interval [CI], 0.63 to 1.46) for ages 70 to 74, 0.65 (95% CI, 0.43 to 1.00) for 75 to 79, 0.24 (95% CI, 0.15 to 0.37) for 80 to 84, and 0.12 (95% CI, 0.07 to 0.19) for 85+. Hispanic patients were less likely to receive chemotherapy than non-Hispanic white patients. Since 1992, paclitaxel has gradually replaced cyclophosphamide as the drug most commonly used with platinum.Despite its proven efficacy in treating ovarian cancer, chemotherapy seems to be used less among patients over age 65, especially those who are nonwhite and/or in the oldest age groups. Further research is needed to elucidate to what degree this represents appropriate clinical judgment and to what degree other factors, such as patient choice, play a role.