Today, reports on drug safety are typically published in free text format, and are therefore computationally unusable. So, if somebody wants to search for information in these reports in a systematic way, he or she should formulate a verbal query using synonyms, perform the search, and finally synthesize the results based on his or her knowledge and experience. However, if these reports were transformed into the form of a Knowledge Graph, then queries could be executed based on the concepts defined in the Knowledge Graph (and not just the words in the text) providing the ability to interlink with other knowledge structures (for example ontologies and well-defined ”dictionaries”) as well as automatic inferencing capabilities.
OpenPVSignal is a semantic model, namely an ontology, developed in collaboration with researchers from the United States and France, aiming to facilitate the publishing of drug safety data in the form of a Knowledge Graph with the ultimate goal of facilitating their computational utilization.
In order to demonstrate the benefits of this approach, we built such a Knowledge Graph, focusing on drugs that are either the focus of undergoing clinical trials or have been suspected of worsening COVID19 patients.