gms | German Medical Science

19th Annual Meeting of the German Drug Utilisation Research Group (GAA)

Gesellschaft für Arzneimittelforschung und Arzneimittelepidemiologie

22.11. - 23.11.2012, Jena

Computerized surveillance of adverse drug reactions in universal hospital

IT-gestütztes Screening von Laborwerten zur UAW-Erfassung

Meeting Abstract

  • author Astrid Scheuerlein - Institut für Pharmakologie/Toxikologie Arbeitsbereich Klinische Pharmakologie, Jena, Germany
  • author Dorothea Gruca - Institut für Pharmakologie/Toxikologie Arbeitsbereich Klinische Pharmakologie, Jena, Germany
  • author Stefan Döhring - Ludwig-Maximilian-Universität IBE, München, Germany
  • corresponding author Marion Hippius - Institut für Pharmakologie/Toxikologie Arbeitsbereich Klinische Pharmakologie, Jena, Germany

Gesellschaft für Arzneimittelanwendungsforschung und Arzneimittelepidemiologie e.V. (GAA). 19. Jahrestagung der Gesellschaft für Arzneimittelanwendungsforschung und Arzneimittelepidemiologie. Jena, 22.-23.11.2012. Düsseldorf: German Medical Science GMS Publishing House; 2012. Doc12gaa06

doi: 10.3205/12gaa06, urn:nbn:de:0183-12gaa061

Published: November 14, 2012

© 2012 Scheuerlein et al.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License ( You are free: to Share – to copy, distribute and transmit the work, provided the original author and source are credited.



Background: In Germany adverse drug reactions (ADR) are currently either detected by conventional spontaneous surveillance method or intensive reporting survey. Both these approaches are either ineffective or time-consuming, and as a result, too expensive. Therefore, a more efficient detection method with improved cost-performance is highly desired. Typically, pathological ADR are reflected in the changes of certain laboratory values. As a consequence, they should be detectable by computer-aided screening procedures. Not only positive results and experiences from other studies are known, furthermore, extensive IT systems such as an electronically patient’s record including the entire laboratory values are available. Therefore, this proof-of-concept study of a computer-aided laboratory values screening has been conducted with the aim to proof technical feasibility as well as practicability in terms of man-power.

Materials and Methods: The data presented here were collected within a multi-center pilot study at the Universitätsklinikum Jena. The study covers all patients that were admitted to departments of internal medicine hospital during the period of examination. For detection of ADR signals a set of relevant laboratory values were defined. For example, to detect signs of a possible liver damage ALAT, ASAT, AP and bilirubin were analyzed. As a marker for kidney damage we used elevated serum creatinine whereas deviating potassium- and sodium level suggest water-electrolyte imbalance. As a laboratory signal we define a laboratory value that is either exceeding or goes lower than a certain limit value. For all patients where a laboratory signal was observed the electronically patient’s record was cross-checked to confirm the presence of a real ADR. At the same time on the very same [Station] an intensive assessment of ADR was conducted using both periodical surveys of the medical staff in charge as well as the systematic evaluation of patient’s records or curves. For both groups the time required for investigation was monitored. Subsequently, a final evaluation of all the cases was conducted based on the electronically patient’s records. Data analysis was descriptive.

Results: Within the period of examination in total 58 ADR were detected. A percentage of 55% (32) of those ADR cases were detected by the laboratory value screening, whereas the intensive assessment displayed 64% (37) ADR. Only 19% (11) of the ADR could not be found in either one of the two groups. This might be due to the fact that i) many ADR do not necessarily result in a change of laboratory values, and ii) within the intensive reporting survey many ADR with change of laboratory values are not detected.

The time required for the laboratory value screening] was with 70.3 min/day significantly higher than for the intensive reporting survey with 36.7 min/day. This finding is furthermore reflected in the efficiency which was 0.99 UEA/hour in the control group whereas the laboratory value group showed a significantly reduced value of 0.45 UAE/hour. Different ADR-range is observed for both detection methods.

Conclusion: For both methods the screening of electronically available laboratory values as well as the intensive reporting survey approximately the same numbers of ADR could be detected. However, only very little overlap of the detected cases within both groups was observed. The screening of laboratory values was all in all more time consuming than the intensive reporting survey. Furthermore, the nature of the detected ADR differs significantly depending on the method. Notably, the ADR with changing in laboratory values were detected very well by the screening of laboratory values whereas other ADR were not detected at all. As a consequence, the laboratory value checks itself applied as a single detection method is not considered to be suitable for accurate ADR detection. Only upon combination of both methods the highest detection rate with an extensive coverage of the complete ADR spectrum will be obtained.