Assessment of the representativity of in-patient hospital diagnoses in the German Pharmacoepidemiological Research Database

Background: Claims data are an important source for health services research. Additionally to plausibility checks to identify errors in data processing and studies examining the validity of the data, it is important to assess the representativity of the data. Objectives: To assess the representativity of in patient hospital diagnoses in the German Pharmacoepidemiological Research Database (GePaRD). Methods: GePaRD consists of claims data from four German statutory health insurance (SHI) providers covering about 14 million insurants (approximately 17% of the German population). The analysis was based on data for the year 2006. We compared all primary hospital discharge diagnoses in the database to those published by the Federal Statistical Office in Germany (Statistisches Bundesamt). These statistics cover all in-patient hospital stays reimbursed by SHIs in Germany. Diagnoses are coded by ICD-10-GM in both data sources. Comparisons were performed within all 22 ICD chapters and for the items of the European Short List, a Europe-wide standardized list of 38 selected diagnoses developed to enable between-country comparisons. Compared was the number of diagnoses in the respective ICD chapter/European Short List Item divided by the sum of all hospital diagnoses. Results: The proportion of the total number of diagnoses per chapter with respect to all hospital diagnoses and the resulting order of chapters in the database population were in line with the data for the German population. Stratification for age and gender improved concordance especially in men. Concordance on the level of European Short List items was also good, both for the whole database population and stratified by age and gender. Conclusions: The in-patient hospital diagnoses 2006 contained in GePaRD were sufficiently representative for Germany. Some differences concerning women and selected age groups will be examined in more detail.