Identification of patient-controlled analgesia overdoses in hospitalized patients: a computerized method of monitoring adverse events

OBJECTIVE: To describe and validate a computer-based quality assurance method that detects narcotic overdoses associated with patient-controlled analgesia (PCA) use. SETTING: Two acute care teaching hospitals. PATIENTS: 4669 patients who received PCA. INTERVENTIONS: The following patient lists were obtained during a two-year period from both hospital information systems: those who received PCA and (1) received naloxone, a narcotic antagonist, (2) were transferred to an intensive care unit, (3) had a cardiac or respiratory arrest, or (4) died. Possible overdoses were defined as patients who appeared on the PCA list and one of the other lists. Charts were reviewed if the patient's name appeared on the PCA and one of the other lists. Patients were judged to have experienced a narcotic overdose if there was an immediate improvement in blood pressure, respiratory rate, or mental status after the administration of naloxone. RESULTS: The search strategy identified 294 possible overdoses in 1499 patients who received PCA. Ten charts were unavailable for review. An actual overdose occurred in 11 patients. The accuracy of the new method was compared with that of the hospitals' present reporting methods. Eleven overdoses were identified by the computer search, but only 6 overdoses were identified in incident and adverse drug reaction reports. CONCLUSIONS: The systematic computer search identified almost twice as many adverse incidents than were reported by the traditional hospital methods.