Methodologies for the detection of adverse drug reactions: comparison of hospital databases, chart review and spontaneous reporting

Abstract

Purpose: To evaluate a methodology for adverse drug reactions (ADRs) detection through hospital databases.

Methods: A retrospective analysis was conducted to identify ADRs using diagnostic codes from databases, later validated by chart review. An independent chart review was performed for comparison, as well as assessment of spontaneous reports.

Results: 325 ADRs were identified (prevalence of 2.41%, positive predictive value of 87.6%). Independent chart review identified 9% of ADRs at a cost of 35 person-hours (versus two person-hours in databases). There were seven spontaneous reports of ADRs.

Conclusions: Although not frequently used, the detection of ADRs through databases is a relatively less expensive, fast and effective methodology that can improve current pharmacovigilance systems.