With the use of any medication comes the possibility of unintended consequences that may help or harm the patient. Harmful results are referred to as adverse drug reactions (ADRs). The discovery of an ADR requires that someone notice the event, recognize it as a possible ADR, identify the suspected drug, and report it. All this is difficult for several reasons: 1) The ADR may occur long after the original use of the drug. 2) The same adverse event may occur with measurable frequency in the unexposed population. 3) There may be no predictable relationship between the ADR and the intended effect of the drug.

Several different methods have been used to identify both new and known ADRs. Case reports have been the most successful for serious, new ADRs. One needs only the suspicion that an adverse reaction may be related to the prior use of a drug to publish a case report. In 1993 the United States Food and Drug Administration (FDA) introduced the MEDWATCH program to encourage health care professionals to report events that lead to death, hospitalization, significant or permanent disability, congenital anomalies, or to ones that require medical or surgical interventions to prevent any one of these results.

Computerized surveillance systems have enhanced the ability of health care providers to screen for known ADRs, but their value for identifying new ADRs remains unclear. These systems work by looking for changes in medication orders or drug levels, orders for antidotes, and other laboratory information. Additional computer-based approaches include the use of large databases that have been in existence for years and can link events to either current or previous medication usage.

The use of postmarketing cohort studies to detect unknown ADRs has been disappointing. Their only value may lie in helping to identify relatively common ADRs that occur with increased frequency among patients exposed to a particular drug.

Meta-analysis, the quantitative analysis of 2 or more independent studies to identify an overall effect, has been suggested for the establishment of associations between drugs and ADRs, estimation of frequency of ADRs, and identification of subgroups at increased risk for ADRs.

Case-control studies have been the most effective method for assigning causality of ADRs that are otherwise unpredictable based on known toxicology studies, the structure or function of the medication, or use history of similar agents.