Since 1963, information on each hospital admission has been collected from all general and university hospitals in the Netherlands (the 'Landelijke Medische Registratie'; LMR). This information includes, among others, one main discharge diagnosis and up to nine secondary diagnoses made during admission, the duration of the hospital stay, death during admission, times of admission and discharge, type of specialist who treated the patient, operations and related therapeutic procedures. This information system is not biased by financial incentives and is therefore a good resource for epidemiological studies and drug-safety research. As of 2005, a new type of administration has been introduced that records diagnosis-therapy combinations for reimbursement purposes. Because the reimbursement of closely related diagnosis-therapy combinations may differ, the system is not only sensitive to fraud but will probably also jeopardise the validity of epidemiological research with such data. Therefore, the LMR system cannot be missed.