A new national active surveillance system designed to detect adverse drug events is good at picking up true cases, but not particularly sensitive—especially in detecting hypoglycemia caused by diabetes medications and bleeding associated with anticoagulants, the Centers for Disease Control and Prevention reported.
In 2003, the CDC worked with the Consumer Product Safety Commission and the Food and Drug Administration to develop the National Electronic Injury Surveillance System-Cooperative Adverse Drug Event Surveillance (NEISS-CADES) project.
Adverse drug events (ADEs) are often more difficult to identify than other injuries, so the CDC conducted an independent chart review in a sample of six NEISS-CADES hospitals (0.2%–1.7% of emergency department visits).
Of 4,561 charts reviewed, 68 ADE cases were identified. Of the 29 ADE cases that had been reported to NEISS-CADES prior to the chart review, 25 were among the 68 cases detected by the reviewers. The remaining four were false positives in which an injury attributed to a drug in the chief complaint section of the chart was not confirmed elsewhere in the chart (MMWR 2005;54:380–3).
The estimated sensitivity of the NEISS-CADES for ascertaining ADEs was 0.33, while the estimated positive predictive value of a reported ADE was 0.92. The relatively low sensitivity was attributed to the difficulty in detecting hypoglycemia associated with diabetes agents (just 3 of 16 were detected), and of bleeding associated with anticoagulants such as warfarin and heparin (1 of 9 were detected). When those two types of cases were excluded, sensitivity of the NEISS-CADES increased to 0.45. That figure compares favorably with the FDA's Adverse Event Reporting System, a passive surveillance system.