In an era of health care cost containment it is of particular interest to identify measures that reduce costs and at the same time improve health care quality. One of these cost cutting measures might be the reduction of the frequency of Adverse Drug Reactions (ADR). The objective of this paper is to summarize all original work on ADR frequencies at hospital admission and to come up with a valid estimate for the actual frequency of ADR-related hospital admissions. Additionally, we compared established concepts of ADR monitoring with respect to their utility for drug safety monitoring and pharmacoepidemiologic research. We reviewed 25 studies from the past 25 years. Analysing the effect of methodological characteristics showed that variation of reported ADR frequency mainly depends on differing study bases and the concepts of ADR monitoring. Investigations that thoroughly screened all members of the study population for the presence of adverse drug reactions (comprehensive ADR monitoring) generally yielded highest ADR proportions. Studies that concentrated screening on selected high-risk patients (preselective ADR monitoring) and those applying spontaneous or intensified spontaneous reporting detected lower ADR proportions (2.9% and 2.5%). The ADR proportion among admissions to departments of internal medicine was higher than among mixed hospital populations including surgical patients. In conclusion 4.2-6.0% (lower and upper quartile) and in median 5.8% of all admissions to medical departments are caused by adverse drug reactions. A two-step preselective ADR monitoring appears to be appropriate and efficient for both signal generation and signal validation as compared to spontaneous reporting and comprehensive monitoring. In conclusion, adverse drug reactions are a common cause of hospital admissions. As hospital care is expensive, attempts to prevent ADR and thus hospital admission need active encouragement.
Copyright 1997 John Wiley & Sons, Ltd.