Evaluation of an indirect method of detecting adverse reactions to anticoagulants

An indirect system of adverse-drug-reaction (ADR) surveillance was evaluated by determining the number of "alerting orders" for suspected ADRs in patients on anticoagulant therapy that were reported by staff pharmacists compared with those detected by an on-floor pharmacist observer. Physicians' orders for patients receiving warfarin or continuous-infusion heparin therapy were monitored during a five-week period to detect "alerting orders" that appeared to indicate the presence of bleeding or excessive hypoprothrombinemia during warfarin therapy or bleeding or thrombocytopenia during continuous-infusion heparin therapy. Alerting orders were classified as antidote, laboratory-test, and dosage-reduction orders, including "hold warfarin" orders; staff pharmacists were taught to detect these orders through inservice-education programs and the use of a training manual. The on-floor observer detected alerting orders through daily monitoring of all study patients' charts and evaluated suspected ADRs using predefined criteria. A total of 1622 physicians' orders were written for 79 patients. Staff pharmacists detected significantly fewer alerting orders than did the on-floor observer (76 versus 273). Also, staff pharmacists reported as alerting orders 21 physicians' orders that did not meet alerting-order criteria. Staff pharmacists reported at least one alerting order for 52 of the 75 patients who had alerting orders identified by the on-floor observer. The use by staff pharmacists of "alerting orders" was not an effective method of detecting anticoagulant-drug adverse reactions when compared with the performance of an on-floor concurrent monitor.