How to approach a hospital information system as sampling frame. Selection of patients with a percutaneous renal biopsy.

The paper describes the four steps that have to be taken in the process of using a hospital information system as sampling frame: step 1, description of the medical routine and registrations related to the study population; step 2, defining the design and criteria for selection; step 3, creating the datafile and step 4, validating the datafile. To illustrate these four steps a detailed description of the selection of patients with a percutaneous renal biopsy is given, using the central database of Leiden University Hospital. All registrations relating to patients undergoing a percutaneous renal biopsy in 1985 and 1986 were taken into account and combined in defining the design and criteria for selection. The selection resulted in 182 patient-records, of which 177 were compared with the medical file. Overall 150 percutaneous biopsies were confirmed. Hospital Information Systems are in principle useful as sampling frames for clinical research. However, to use these systems successfully in selecting a study population intensive mutual consultation between clinicians, clinical epidemiologists, database experts and administrative personnel is required.