BetterChoice: A migraine drug recommendation system based on Neo4J

Migraine is a common disease throughout the world. Not only does it affect the life of people tremendously, but it also leads to high costs, e.g. due to inability to work or various required drug-taking cycles for finding the best drug for a patient. Solving the latter aspect could help to improve the life of patients and decrease the impact of the other consequences. Therefore, in this paper, we present an approach for a drug recommendation system based on the highly scalable native graph database Neo4J. The presented system uses simulated patient data to help physicians gain more transparency about which drug fits a migraine patient best considering her individual features. Our evaluation shows that the proposed system works as intended. This means that only drugs with highest relevance scores and no interactions with the patient's diseases, drugs or pregnancy are recommended.