A computerized method for identifying incidents associated with adverse drug events in outpatients

Introduction: In inpatients, computer monitors have been used to improve th e detection of adverse drug events (ADEs). However, similar programs have n ot been available in outpatients. Objective: To describe an approach for de tecting incidents suggesting that an ADE may have occurred in outpatients b y adapting methods from inpatient computer monitoring and developing termin ology searches of electronic medical records. Method: One year of informati on from the outpatient electronic medical record (EMR) at one hospital and its clinics was reviewed. Altogether, 23 064 patients and 88 514 visits wer e identified. Patient demographics, medical problem lists, ICD-9 claims, pa tient allergies, medication history and all clinic visit notes were extract ed and merged. We then searched for incidents suggesting that an ADE might be present using four methods: ICD-9 claims, new allergies, computer rules linking laboratory data to known medication exposures, and a medical termin ology lexicon ((MD2)-D-2(TM)). In this report, we describe how these search methods were developed to allow for ADE identification. Conclusion: The ab ility to carry out such quality-related work is an example of the benefits of the outpatient EMR that may not be apparent to those institutions consid ering adopting it. (C) 2001 Elsevier Science Ireland Ltd. All rights reserv ed.