Identification of adverse drug reactions by evaluation of a prescription database, demonstrated for "risk of bleeding".

OBJECTIVE: Information about adverse drug reactions plays an important role when assessing the benefit/risk profile of a drug. Identifying rare adverse drug reactions, however, is a difficult task. This paper illustrates the advantages of using a prescription database for this purpose.
METHODS: The mediplus database used in our analysis covered data from 320,644 outpatients observed between July 1999 and June 2002. The example of bleeding complications during intake of antidementia drugs is used to illustrate this approach. The comparison of cohorts and subgroups is nearly always a problem in surveys. For our analyses we considered a set of patients who had taken a selected medication for a certain period of time and compared the frequency of adverse events with those occurring when the same patients did not take this medication. Hence, the comparison with versus without a certain medication is based on the same set of patients as in a cross-over study.
RESULTS: Our evaluations indicate that the rate of bleeding complications is low when taking any of the widely used antidementia drugs, glutamate modulators, cholinesterase inhibitors, calcium antagonists or the phytomedicine Ginkgo biloba.
CONCLUSION: Basing the comparison of the rates of complications during periods with and without intake of a certain drug on the same set of patients may be a useful tool for assessing adverse drug reactions from data reported in prescription databases.