gms | German Medical Science

Kongress Medizin und Gesellschaft 2007

17. bis 21.09.2007, Augsburg

Establishment of a pharmacoepidemiological database in Germany - methodological potential, scientific value and practical limitations

Meeting Abstract

Search Medline for

  • Iris Pigeot - Universität Bremen, BIPS, Bremen
  • Wolfgang Ahrens - Universität Bremen, BIPS, Bremen

Kongress Medizin und Gesellschaft 2007. Augsburg, 17.-21.09.2007. Düsseldorf: German Medical Science GMS Publishing House; 2007. Doc07gmds307

The electronic version of this article is the complete one and can be found online at:

Published: September 6, 2007

© 2007 Pigeot 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.



Purpose: The withdrawals of Vioxx and Lipobay illustrate the need for a routine system to monitor drug safety of newly launched drugs. Since spontaneous reporting systems do not relate to a defined population of drug users they are unable to identify and quantify disease-specific risks. To overcome this limitation, we established a population-based pharmacoepidemiological database in Germany derived from claims data of statutory health insurances (SHIs). The potential use of this database, its scientific value and its limitations will be discussed.

Methods: Data files of 3 SHIs were linked with drug dispensation data from a pharmacies’ electronic data processing centre on an individual basis. In accordance with German data protection regulations, the linkage was done using the unique subject identification number at a trusted third party centre that transferred pseudonymized data to the research centre. Plausibility checks and first descriptive analyses were carried out.

Results: We were able to establish a database covering 3.6 million SHI-members for specific research questions. The database provides drug utilization data and data on medical events leading to hospitalizations and enables investigation of drug-related serious adverse events. Confounder information on medical conditions, concomitant medications and socio-demographic variables can be obtained from the database, while the assessment of confounders related to lifestyle requires supplementary data collection. The data provide reliable information on prescribed drugs, their formulation, dosage, package size, and (approximate) date of their dispensation.

Conclusion: The database guarantees a complete coverage of all age groups and 100% participation proportion since no informed consent is necessary. It allows for a population-based approach and reflects daily practice including off-label use of drugs. There is no risk that a systematic reporting bias on medication or outcome occurs because exposure and outcome data are collected independently. Legal conditions that allow continuous updating of the database and regular quality management need to be settled.