medicX-KE : a knowledge graph for pharmacists’ drug information needs

Abstract

The contribution of pharmacists to patient care is evolving from the traditional role of dispensing medicine to that of providing pharmaceutical services within a multidisciplinary healthcare context. An important component for this transformation towards a holistic patient-centric approach is the availability of relevant and up-to-date information about medicinal products that leverages a robust data collection mechanism. Over the past decade, Knowledge Graphs (KGs) have rev-olutionised the organisation and utilisation of vast biomedical data. Powered by Artificial Intelligence (AI) and advanced inferencing techniques, KGs unlock the potential to derive valuable insights, discover hidden relationships, and enable data-driven decision making. In this paper, we present the medicX-KE knowledge graph that is tailored to the Maltese pharmaceutical landscape. The KG serves as a centralised repository, integrating data from several authoritative sources, including DrugBank and the British National Formulary (BNF). We present a specific use case for our KG that relies on information related to medicine provided by the Malta Medicines Authority (MMA). In addition, we conducted interviews with six pharmacists to understand their needs and to obtain feedback on the suitability of the KG. We describe data acquisition, mappings, ontology design, construction, and evaluation of medicX-KE, and discuss the challenges faced and potential future enhancements.