Identification of Adverse Drug Events from the FAERS Database Utilizing the Low-Code KNIME Platform

Abstract

This thesis aims to tackle the challenges of utilizing the FDA Adverse Event Reporting System (FAERS) for extracting and comprehending drug safety data. FAERS is a repository of vast amounts of complex and unstructured data, which can be challenging to analyze without significant coding experience. To address this issue, this thesis proposes using a low-code platform, such as KNIME, to simplify the analysis of FAERS data. By employing KNIME, even users without coding knowledge can extract and visualize FAERS data, enabling more precise and efficient drug safety data analysis. This thesis utilized a hypothesis-generating approach to evaluate the potential association between Aspirin and bevacizumab regarding adverse events, specifically focusing on the theory that Aspirin could lower the incidence of bevacizumab-induced Hypertension. The aim was to compare the differences between individuals who experienced adverse events while taking Aspirin versus those who did not. The results demonstrate that the workflow successfully identified fewer adverse events among individuals taking Aspirin than those who did not. These findings provide significant insights into identifying the safety signals of any drug and answering clinical questions without the need for coding expertise but need further validation.