A MiniReview of the Use of Hospital‐based Databases in Observational Inpatient Studies of Drugs

The majority of pharmacoepidemiological data resources are based on data generated in primary health care. Although inpatient data resources have existed since the 1960s, inpatient pharmacoepidemiological studies are relatively scarce. The objectives of this M ini R eview were to describe pharmacoepidemiological studies in hospital settings and the underlying databases to provide an overview of research questions addressed by such databases. The studies were retrieved by chain searching. We included pharmacoepidemiological studies in hospital settings containing data on inpatient drug use. Twelve inpatient databases in A sia, the U nited S tates and E urope were found. Most databases were automatically collected from claims data or generated from electronic medical records. The contents of the databases varied as well as the potential for linkage with other data sources such as laboratory and outpatient data. Twenty studies were selected and discussed to illustrate the diversity of inpatient pharmacoepidemiological studies. Hospital‐based databases had mainly been used for drug utilization studies and research in adverse drug reactions. Five studies within comparative effectiveness were found. The number of pharmacoepidemiological studies in inpatient settings was low compared with studies from primary healthcare settings. These resources may be under‐utilized.