Due to the higher risk of morbidity and perioperative mortality compared to younger patients, elderly patients with advanced ovarian cancer are challenging to treat. A population-based analysis was performed to predict treatment outcomes and establish risk factors for early death of elderly patients with advanced ovarian cancer using a cohort of 3994 women diagnosed with stage III or IV ovarian cancer between 1992 and 1999, registered with the Surveillance, Epidemiology and End Results Cancer Registries. A multivariate accelerated failure time model allowed estimation of a risk factor model for overall survival. Patient's age, stage at presentation, presence of comorbidities, and oncology treatment facility were independently associated with overall survival at 12 months from diagnosis. Patients were assigned to low (0-7 points), moderate (8-14 points) or high (>/=15 points) risk groups according to accumulation of risk factors, which showed good ability to predict 12-month mortality (receiver-operator characteristics curve [ROC] derivation cohort = 0.763; ROC validation cohort = 0.756). Across all three risk groups, patients who received both surgery and chemotherapy showed significantly improved survival as compared to patients who received only surgery or chemotherapy. For patients 80 years and over who had upfront surgery, perioperative mortality was significantly greater in the high-risk group (21%; 95% CI = 16-26%) compared to patients within the moderate (8%; 95% CI = 5-12%) and low-risk groups (0%; 95% CI = 0-11%). The risk factor profile established could be helpful to plan future clinical trials to establish optimal treatment for elderly patients with advanced stage ovarian cancer.