gms | German Medical Science

Background: Medication anamnesis are routinely conducted in epidemiological cohort studies. Almost every second data access request in SHIP (Study of Health in Pomerania) [1] contains variables on medication intake. There are currently more than 100,000 approved drugs in Germany [2]. The assessment of drugs is prone to errors and requires appropriate electronics data capture support. In addition, recall bias of drug exposure is known to be high [3]. In SHIP, participants are requested to bring all prescribed medications to the examination center, including their original packaging. During the computer assisted personal interview (CAPI) they provide information about the intake of drugs during the last 7 days, intake patterns, daily dosage and the duration of the current intake period.

Methods: We expanded an electronic data capturing system (SHIPPIE) [4] by a web application for a standardized recording of medication (SHIPPIE drug module). The underlying drug database is provided by the WIdO (Wissenschaftliches Institut der AOK). It contains over 161,000 entries such as drugs, aids, and other pharmacy products and is regularly updated. We implemented an assisted documentation of drug exposure where the interviewer is guided by the web application. Drug specific routes of administration, doses and possible packaging sizes should be accessible. Therefore an acquisition together with the participants and the requirement of an up-to-date drug database is essential for high data quality. Further requirements were:

• direct recording and clear identification via the PZN (Pharmazentralnummer)
• several search filters (copy commonalities for manual drug entry)
• default values to avoid different notations
• item dependencies (hidden or visible input fields)
• additional entries without accessing the database such as intake pattern, daily dose and duration of intake
• a summary list of the participant's medication that have already been recorded

Results: The WIdO drug database is imported into our PostgreSQL database in a JSON structure and can be accessed via the SHIPPIE drug module. Over 60% of the drugs are found via the PZN which is provided as a 2D barcode on every pharmaceutical package. This allows for an assignment of the package size. If no PZN is available, additional search filters in the web interface such as the drug name or dosage form make it possible to identify a total of over 90% of the drugs. If the search result does not contain the correct drug, the commonalities of the results can be copied, so that only the remaining data of the drug has to be entered.

Conclusion: The SHIPPIE drug module web application enables a distinct identification of participant's medication based on the PZN and allows for a recording as accurately as possible in the absence of a PZN. The application allows for precise calculations, e.g. of the total dose, dosage form or drug costs.

The authors declare that they have no competing interests.

The authors declare that an ethics committee vote is not required.