Postmarketing surveillance: strengths and limitations. The flucloxacillin-dicloxacillin story.

Spontaneous reporting of adverse drug reactions continues to be the principal method used for monitoring the safety of marketed drugs. Despite the many successes attributed to these schemes, they can reliably detect only a small fraction of the range of possible drug-related events and provide virtually no useful quantitative data. Some of the limitations of spontaneous reporting were demonstrated recently in relation to flucloxacillin. Reports in Australia suggested the likelihood of an unacceptable risk of flucloxacillin-associated jaundice, but the data from spontaneous reporting in countries with apparently similar use of the drug, such as New Zealand and the UK, were insufficient to confirm or refute this proposition. Spontaneous monitoring should be supplemented by the systematic monitoring of cohorts of users of new drugs, using record-linkage to track their subsequent health. Although several impediments exist to the introduction of such a scheme in Australia, consideration should be given to addressing how such a system might be implemented.