Correspondence: Emergency Hospitalizations for Adverse Drug Events — NEJM

To the Editor:

"We commend Budnitz and colleagues (Nov. 24 issue)1 for their contribution to our understanding of emergency hospitalizations caused by adverse drug events. However, as a result of detection bias, it is likely that the relative contribution of bleeding and hypoglycemic manifestations of adverse drug events have been overestimated and that the overall burden of disease has been underestimated. This is because the case-finding methods at institutions participating in the National Electronic Injury Surveillance System–Cooperative Adverse Drug Event Surveillance (NEISS-CADES) project rely on coders who use a short list of symptoms that include “bleeding” and “hypoglycemia” to identify adverse drug events.2 Other commonly found manifestations of adverse drug events, such as neuropsychiatric, gastrointestinal, and cardiovascular symptoms, are not listed as triggers and therefore are less likely to be identified.3
In addition, emergency physicians frequently do not attribute emergency department presentations to adverse drug events, which leads to a lack of documentation in hospital records.4 Validation of the triggers in the NEISS-CADES project against a prospective criterion standard would enable a better estimate of the sensitivity of such measures. Prospective case-finding methods may yield more accurate data on the frequency and causes of adverse drug events and on the relative contribution of various adverse drug events to the overall disease burden."