BACKGROUND
Herbal medicines have been used for centuries and are still widely used as alternative or complementary therapies for several conditions. However, the use of these products has raised concerns about their safety, particularly about interactions with conventional drugs, known as herb-drug interactions.
OBJECTIVE
In this paper, we showcase the ForPharmacy Decision Support System, designed to identify herb-drug interactions. With this system, pharmacists could effectively utilize scientific information, including randomized controlled trials, non-randomized trials, retrospective studies, and case reports, and convert it into practical knowledge for individual patients.
METHODS
The knowledge-based of ForPharmacy Decision Support System was built by conducting a literature review performed in English, using 3 electronic databases: PubMed, Scopus, and Web of Science, including studies conducted with human beings with no restriction of age, gender or disease. The search focused on herb/plant-drug interactions and included both studies with interactions and with no found interactions. The results allowedconstructing a corresponding knowledge database that feeds the ForPharmacy expert system. Additionally, a user-friendly interface was designed for an easy and efficient access and utilization of the system in an intuitive manner.
RESULTS
The ForPharmacy Decision Support System assists pharmacists in quickly and efficiently identifying herb-drug interactions. Its knowledge base is composed by a total of 187 interactions involving various herbs. The most frequent herb associated with drug interaction was St. John's Wort, accounting for 10.7% of the interactions, followed by Ginkgo (4.5%), Ginseng (3.6%), and Milk Thistle (3.6%). With respect to drugs, bupropion exhibited the highest probability of interactions, with a percentage of 13.37%, followed by dextromethorphan (11.8%) and warfarin (11.2%). This system offers pharmacists a user-friendly interface, which allows them to identify interactions between a particular drug and herb, or obtain a list of all potential interactions for a specific drug. Additionally, the system provides comprehensive details on the common symptoms associated with each interaction, along with the potential danger and significance value.
CONCLUSIONS
Our research provides a valuable contribution to the field of herb-drug interactions, as it offers a new Decision Support System that can effectively identify potential interactions and associated symptoms. The system's user-friendly interface and distinct use cases make it a practical tool for healthcare professionals and patients. Future work should focus on expanding the knowledge base to facilitate the application of machine learning models. This would allow autonomous identification of new interactions, thereby enhancing the system's ability to identify novel interactions. Our findings underscore the significance of ongoing research in this area to advance patient safety and prevent adverse events.