Abstract
An adverse drug-reaction monitoring system based on spontaneous reporting to a central register of adverse reactions has been in operation for eight years. As a test of the validity of the reports and of the probability of causal relationship between drug and reaction a random sample of 82 cases were followed up in detail. The sample included 17 deaths, 26 serious reactions, and 39 reactions of moderate or only minor severity. Altogether 78% of the reactions were considered to be "probably" drug related and 13% "possibly" drug related. It is concluded that the reports are of value in the detection and evaluation of drug safety.