Time for Action on Drug Safety

To the Editor: While I generally agree with the advocacy position taken by Mr Moore and colleagues¹ that more resources be committed to the postmarketing surveillance of adverse drug events (ADEs) occurring with the use of marketed medications and other treatments, most of their suggestions are based on hyperbole and not on an appreciation of the principles or difficulties of surveillance. Consider their suggestion that the Food and Drug Administration (FDA) enumerate or estimate drug-induced disease and deaths much as the National Highway Transportation Safety Administration and Federal Aviation Administration (FAA) track motor vehicle and airplane crashes and deaths. Although this sounds rational and easy, it is unclear exactly how a drug-induced clinical event should be defined, especially since most events of interest are not particularly specific to a single drug or population. Whereas judging a death in an automobile crash as being caused by the crash is correct almost all the time, clinical judgment about the cause of clinical events will not have sufficient predictive value to provide meaningful estimates. The inescapable fact is that an accurate estimation of the extent of event occurrence that is attributable to drugs and treatments will require control groups and a massive expenditure of resources. Such effort is not necessary with attributing deaths to automobile or airplane crashes because few deaths that occur during crashes are attributable to causes other than the crash.