Introduction: Gastroesophageal reflux disease (GERD) is a common disorder that is associated with significant patient morbidity and health care utilization. There is a paucity of population-based data on the determinants of surgical therapy in patients with GERD. The role of surgery in the treatment of reflux disease remains controversial. Our objectives were to determine the incidence of surgery in a cohort of patients with GERD and to identify factors associated with its use. Methods: A retrospective inception cohort of patients with GERD was created using administrative data in Ontario , Canada . An index event was defined by the first combination of an upper gastrointestinal endoscopy and an associated GERD-related diagnosis code between the fiscal years 1991 and 1994. Exclusion criteria included age less than 18 years, regions without physician-billing data, and a history of cancer. Follow-up data were collected to the fiscal-year ending 2004. The incidence rate of surgery was defined as the number of overall procedures in the numerator and patient years of follow up in the denominator. Comorbidity was measured using the Charlson score. Univariate and multiple logistic regression models were developed to examine the determinants of surgery. Odds ratios and 95% confidence intervals are presented. Results: We identified a cohort of 43,992 patients who underwent an upper GI endoscopy and were assigned a diagnosis of GERD during the fiscal years 1991-1994, of whom 1,878 (4.3%) subsequently underwent an antireflux procedure. The incidence of surgery was 43 procedures per 10,000 patient years of follow up. Patient age (p < 0.0001), Charlson co-morbidity score (p < 0.0001), esophageal ulcer (p=0.05), esophageal stricture (p=0.04), and income quintile in neighborhood of residence (p = 0.003) were found to be univariate predictors of surgery. In multivariable models, patient age (OR 0.97, 0.97-0.98), Charlson scores of 0 (OR 3.19, 1.80-5.66) or 1 (OR 2.26, 1.24-4.11), and the presence of an esophageal ulcer (OR 1.91, 1.17-3.12) were significantly associated with the use of surgery. Patient gender and income quintile were not associated with surgical intervention. Conclusions: We estimated the population-based incidence of antireflux surgery for a cohort of patients with GERD in Ontario , Canada . Patient age, co-morbidity and complicated esophageal disease appear to be important factors in the decision to proceed with surgery. Further investigation of healthcare delivery factors, such as the availability of specialist care, in determining the use of surgery for GERD may further explain variation in the use of antireflux surgery.