Background: Surveillance of trends in cancer management can help in service evaluation and quality improvement. Patients and methods: We used East Anglia cancer registration data from 1995-2006 on 3063 patients with epithelial ovarian cancer. We described treatment patterns and examined their predictors using multivariate logistic regression. Results: Use of chemotherapy (with/without surgery) and of surgery (with/without chemotherapy) remained stable over time (59% and 69% of patients, respectively). Use of surgery was less frequent in most deprived compared with most affluent patients (64% vs. 73% respectively; adjusted per increasing deprivation group quintile OR = 0.88, p = 0.001). Among patients treated by surgery, use of omentectomy and of exenteration increased over time (37% in 1995-8 vs. 56% in 2003-6 for omentectomy, and 2% vs. 8% for exenteration respectively, p < 0.001 for both). More deprived patients were less likely to be treated with omentectomy [42% vs. 53% for most deprived and more affluent patients respectively (adjusted OR = 0.91, p = 0.017)]. Conclusion: Socioeconomic differences in use of ovarian cancer surgery require further elucidation. Use of some types of surgery increased over time, potentially reflecting changes in service organisation. The findings provide a 'benchmark' for future treatment pattern surveillance, and regional or international comparisons.