sessment of drug safety after approval as the generation of hypotheses based on early safety signals, the strengthening of safety signals, the conduct of confirmatory studies to identify and quantify new or hypothesized risks and benefits, the evaluation of risk management programs to minimize known safety risks, and the continuing evaluation of risks and benefits in light of new risk or new benefit information to ensure that the known benefits of a drug continue to outweigh the known risks. The committee concludes that although CDER is involved in a variety of activities to generate and assess postmarket safety information, the current approach is not as comprehensive and systematic as is needed to serve drug safety and public health objectives optimally. The committee offers specific recommendations to CDER and other federal departments and agencies for improving postapproval assessment of drug-related risks and benefits.
Although some safety signals are generated in laboratory tests and clinical trials conducted in the preapproval setting or from known or suspected biologic actions of a drug, the primary method by which FDA documents new adverse events in the postmarket setting is monitoring of suspected adverse drug reaction reports entered into the Adverse Event Reporting System (AERS). AERS combines the voluntary adverse drug reaction reports from MedWatch, such as direct reports from healthcare practitioners and consumers, and the required reports from manufacturers—15-day expedited reports of serious3 and unexpected adverse events and manufacturer periodic reports. The information provided by this part-voluntary, part-mandatory system of reporting forms the basis of detection of many safety signals and has been useful in identifying rare adverse events.
Spontaneous AERSs have many limitations, but they offer the possibility of identifying rare serious adverse events in a timely manner among all persons across the entire region to which the system applies. For example, if there is a one-in-a-million serious adverse event applicable to those exposed to a drug used in 10 million people per year in the United States, it might never be observed in a database of several hundred thousand, or even several million people in which the number exposed to the drug might be only a few thousand per year. But in the entire United States it is not so unlikely that at least one such event would get reported. Even a small number of reports of events that are commonly caused by drug exposure, such as liver or kidney failure, aplastic anemia, anaphylaxis, Stevens-Johnson syndrome,