OBJECTIVE: We assessed the usefulness of the French hospital database (PMSI) implemented in view of a DRG-like prospective payment system for an economic evaluation of the management of haemophiliac patients. METHODS: We extracted and analysed data of the French public hospital database on hospitalisations with a principal or secondary diagnosis of haemophilia (type A) for the years 1996-1997. RESULTS: A total number of 7 000 admissions were found per year. The inpatient admissions concerned haematology and orthopaedic surgery but were scattered over a total number of 328 DRGs, reflecting both the variability of practice patterns and more likely of coding procedures. Almost all of DRGs have less than 1 % of the files registered. We identified 3 main ambulatory DRGs but none was specific to Haemophilia. The ambulatory DRGs contain more than 50% of the files related to Haemophilia. The analysis of the national database does permit to identify neither the type of therapeutic strategy for haemophilia (prophylactic treatment or on demand treatment) nor the presence of a factor VIII inhibitor. We found differences on the quality of diagnostic coding between the regional university hospitals (UH) and the local hospitals. The mean of diagnostics registered for patient in 1997 is 2.9 diagnostics in UH and 4 in other hospitals. We compared the observed costs of haemophiliac patients with the DRG reimbursement schedule and found variations from plus to minus 40%. This gap is mainly explained by the length of hospitalisation. Thus it appeared that the national DRG database is not currently appropriate for assessing the management and costs of treatments for haemophiliac patients, but could well supplement the existing prospective cohort studies.