Using an Anesthesia Information Management System to Prove a Deficit in Voluntary Reporting of Adverse Events in a Quality Assurance Program

Abstract

Objective.A deficit is suspected in the manual documentation ofadverse events in quality assurance programs in anesthesiology. In order toverify and quantify this, we retrospectively compared the incidence ofmanually recorded perioperative adverse events with automatically detectedevents. Methods.In 1998, data of all anesthetic procedures, includingthe data set for quality assurance of the German Society of Anaesthesiologyand Intensive Care Medicine (DGAI), was recorded online with the AnesthesiaInformation Management System (AIMS) NarkoData4® (Imeso GmbH). SQL(Structured Query Language) queries based on medical data were defined for theautomatic detection of common adverse events. The definition of the SQLstatements had to be in accordance with the definition of the DGAI forperioperative adverse events: A potentially harmful change of parameters ledto therapeutic interventions by an anesthesiologist. Results.During16,019 surgical procedures, anesthesiologists recorded 911 (5.7%) adverseevents manually, whereas 2966 (18.7%) events from the same database weredetected automatically. With the exception of hypoxemia, the incidence ofautomatically detected events was considerably higher than that of manuallyrecorded events. Fourteen and a half percent (435) of all automaticallydetected events were recorded manually. Conclusion.Using automaticdetection, we were able to prove a considerable deficit in the documentationof adverse events according to the guidelines of the German quality assuranceprogram in anesthesiology. Based on the data from manual recording, theresults of the quality assurance of our department match those of othercomparable German departments. Thus, we are of the opinion that manualincident reporting seriously underestimates the true occurrence rate ofincidents. This brings into question the validity of quality assurancecomparisons based on manually recorded data.