Risk management of antibacterial drugs: A decision tree classification method driven by real world data

Antibacterial drugs not only bring benefits to human health, but also pose serious risks. The purpose of this paper is to establish a classification model driven by real-world data to quantify and classify the risk level of antimicrobial drugs. The research data is sourced from Chinese spontaneous reporting database from 2010 to 2017, including 622 adverse reactions caused by 114 antibacterial drugs, totaling 237,715 reports. Firstly, we established three risk indicators to quantify the risk characteristics of antibacterial drugs, and labeled the drug risk levels based on the classification criteria of five provinces in China. Then, SMOTE was used to expand the sample for small category. Finally, J48 algorithm was used to construct a classification model. The experimental results based on the research data show that the proposed method achieves ideal classification performance. This study can provide decision-making support for the grading management and clinical use of antibacterial drugs.