Development of a standardized knowledge base to generate individualized medication plans automatically with drug administration recommendations.

Aims: We aimed to develop a generic knowledge base with drug administration recommendations which allows the generation of a dynamic and comprehensive medication plan and to evaluate its comprehensibility and potential benefit in a qualitative pilot study with patients and physicians.

Methods: Based on a literature search and previously published medication plans, a prototype was developed and iteratively refined through qualitative evaluation (interviews with patients and focus group discussions with physicians). To develop the recommendations for safe administration of specific drugs we screened the summary of product characteristics (SmPC) of different exemplary brands, allocated the generated advice to groups with brands potentially requiring the same advice, and reviewed these allocations regarding applicability and appropriateness of the recommendations.

Results: For the recommendations, 411 SmPCs of 140 different active ingredients including all available galenic formulations, routes of administrations except infusions, and administration devices were screened. Finally, 515 distinct administration recommendations were included in the database. In 926 different generic groups, 29,879 allocations of brands to general advice, food advice, indications, step-by-step instructions, or combinations thereof were made. Thereby, 27,216 of the preselected allocations (91.1%) were confirmed as appropriate. In total, one third of the German drug market was labelled with information.

Conclusions: Generic grouping of brands according to their active ingredient and other drug characteristics and allocation of standardized administration recommendations is feasible for a large drug market and can be integrated in a medication plan.