Adverse drug reactions: Comparison of two report methods

Abstract

Increasing numbers of effective therapeutic agents changed the practice of medical care. One of the costs of the widespread use of new therapies is the occurrence of adverse drug reactions (ADR). We compared a routine procedure of spontaneous reporting to a systematic registration of medical informations including adverse drug reactions in a cluster-randomized controlled trial. This study was carried out during the first semester of 1993 at the Centre Hospitalier et Universitaire de Nancy, France receiving about 100,000 patients a year. The 10 participating departments were paired-matched. For the first eight weeks of the study, reporting methods were randomly assigned within each pair. For the second eight weeks of the study, the attribution of methods was reversed. All patients, who left a participating department within one of these two periods were included. Subsequent discharges after further readmissions for the same patient were not considered. Thus, the study was based on 5967 patients. We received 46 ADR reports. After adjustment for age and sex, using a logistic regression model, the probability for reporting ADR for a patient is twice as likely by systematic registration of medical information than by the routine procedure of spontaneous reporting. Generalization of this supplementary method for reporting adverse drug reactions appears interesting.