Improving self-reporting of adverse drug events in a West Virginia hospital.

Adverse drug events significantly increase length of stay and costs of hospitalization but are underreported in health care institutions. We hypothesized that hospitals could improve the accuracy of adverse drug event self-reporting by comparing adverse drug events recorded in an occurrence reporting tool with those detected by surveillance of "rescue" drugs administered to treat adverse drug events. We conducted a prospective cohort study of all adult inpatient discharges from a 200-bed rural acute care hospital in West Virginia during a 6-month period. We performed 3572 chart audits, of which 1011 included rescue drug administration. Our outcome measure was the proportion of adverse drug events in the rescue drug surveillance that were found in the occurrence reporting tool. We found that less than 4% of all adverse drug events involving use of rescue drugs were reported. We concluded that underreporting of preventable adverse drug events in this hospital is comparable to published rates and that surveillance of adverse drug events to detect underreporting is feasible.