To determine the incidence of Guillain-Barré syndrome (GBS) in the Canadian provinces of Ontario and Quebec during 1983-1989 and to demonstrate the feasibility of measuring the incidence of GBS through internal record linkage of Canadian hospital service data, we conducted a record linkage study. This study used the databases of the Hospital Medical Records Institute and the Ministère de la santé et des services sociaux du Québec. We extracted records containing the International Classification of Diseases, 9th revision, code for GBS or a diagnosis likely to harbor misclassified GBS cases from each database and linked them internally using computerized algorithms. We identified a total of 1,302 and 1,031 incident cases of GBS admitted to Ontario and Quebec hospitals, respectively. The calculated mean annual GBS incidence rate in each province, after age and sex standardization to the 1986 Canadian census population, was 2.02 per 100,000 person-years in Ontario and 2.30 per 100,000 person-years in Quebec. Chart reviews revealed that the false-positive diagnosis rate might be as high as 0.26 per 100,000 person-years in Ontario and 0.21 per 100,000 person-years in Quebec. With adjustment for these false-positive rates, the incidence rate of GBS becomes 1.51 per 100,000 person-years and 1.78 per 100,000 person-years in Ontario and Quebec, respectively. In both provinces, the incidence rate was higher in older age strata (70-80 years) and in males. We saw no seasonal or geographic pattern.