Background: One of the objectives of the National Mesothelioma Surveillance Program is to estimate the national incidence of pleural mesothelioma and track its trends. In this context, it was important to assess the feasibility of using data from the Program for the Medicalization of Information Systems (PMSI), which is available at the national level. Methods: By deduplicating hospital stays from the national database with a diagnosis of mesothelioma or pleural cancer, we identified 506 patients in 1998 and 474 patients in 1999 residing in one of the 17 departments covered by the Mesothelioma Surveillance Program in 1998. These cases were matched using indirect criteria with cases recorded in the National Mesothelioma Surveillance Program during the same period. A review of patient records was conducted in one department to validate the matching procedure and document cases unknown to the Mesothelioma Surveillance Program. Results: Only two-thirds of the cases recorded in the Mesothelioma Surveillance Program had a corresponding hospital stay in the PMSI in the same year. The review of medical records showed that 1) confirmed cases from the Mesothelioma Surveillance Program were most often (83% for the two years studied) found in the PMSI with a mesothelioma diagnosis code, but 10 to 15% of patients with a mesothelioma diagnosis code in the PMSI had a different diagnosis in their medical records; 2) 65% of patients with a mesothelioma diagnosis code in the PMSI and unknown to the Mesothelioma Surveillance Program in 1998 and 55% in 1999 were prevalent cases; 3) 3 suspected cases of mesothelioma were unknown to the Mesothelioma Surveillance Program. Conclusion: The lack of diagnostic certification, coding errors, and above all the inability to distinguish incident cases from prevalent cases in the PMSI prevent reliance on this single data source to estimate the incidence of mesothelioma. It nevertheless constitutes a source of supplementary information that should be systematically consulted and cross-referenced with data from mesothelioma registries at the departmental level. (R.A.)