Objective: In 1994 the Department of Anaesthesiology and Intensive Care Medicine of the Justus Liebig University of Giessen decided to implement an Anaesthesia Information Management System (AIMS) to replace the previous hand-written documentation on paper. From 1997 until the end of 1998 the data sets of 41,393 anaesthesia procedures were recorded with the help of computers and imported into a data bank. Individual aspects and results of this data pool are presented under the aspect of how the system in its present form is able to guarantee documentation of quality according to the requirements of the German Society of Anaesthesiology and Intensive Care Medicine (DGAI).
Methods: Since 1997 information on all anaesthesia procedures has been documented "online" with the anaesthesia documentation software NarkoData 4 (ProLogic GmbH, Erkrath). The data sets have been stored in a relational data bank (Oracle Corporation) and statistically processed with the help of the SQL-based program Voyant (Brossco Systems, Espoo, Finland). As an example of two adverse perioperative events (AVB) we compared incidences of "hypotension" and "nausea/vomiting", recorded by staff members into the AIMS, with the incidence of comparable events that were recorded with the help of online data during anaesthesia procedures, such as blood pressure and drug application. Since 1998 data recording has been revised constantly in department meetings; advanced training has been given. The results have been analysed critically.
Results: In 1997 the incidence of adverse perioperative events entered manually into the system was 3.6% (grade III and higher 0.9%) and increased during 1998 to 22.2% (grade III and higher 1.9%). The frequency of anaesthesia procedures with manually documented AVBs was significantly below the incidence (determined with the help of online data) of comparable events: "hypotension" (1.8% vs. 8.5%) and "nausea/vomiting" (4.9% vs. 8.3%).
Conclusion: The current documentation of AVBs in almost any hospital is incomplete. In contrast to the hand-written procedure, the AIMS provides recorded data for evaluation and guarantees more detailed and complete quality documentation. In addition, the effort needed for documentation is reduced. Whether these data sets really describe and measure quality or not has to be evaluated. In addition it has to be considered whether different requirements (such as automatic AVB recognition for an AIMS) are advantageous for quality documentation regarding the data raster and the AVB recognition, with respect to different documentation procedures.