The Role of Artificial Intelligence in Reducing the Risk of Adverse Reactions in Multiple Drug Interactions

Abstract

The available data on the role of polypragmasia in increasing the frequency of multiple drug interactions, when one drug interacts with two or more other drugs, increasing the risk of side effects associated with them, are considered. The application of network analysis and artificial intelligence to predict the development of clinically significant adverse reactions in conditions of polypharmacotherapy is described. The mechanisms of pharmacodynamic and pharmacokinetic interaction of drugs in the development of adverse reactions are considered and drugs potentially carrying an increased risk in multiple drug interactions are noted. The most dangerous drugs involved in drug interactions were psychotropic drugs, which accounted for about a third of all applicable medicines. The most common serious potential complications associated with this interaction were serotonin syndrome, seizures, QT prolongation, and bleeding. Graph probabilistic models, machine learning models for analyzing reliable sources of medical data, factor models that allow assessing the risks of taking two or more drugs together are proposed. These models are implemented in software and can be implemented in clinical decision support systems. It is concluded that the use of artificial intelligence can reduce the risk of adverse reactions during polypharmacotherapy, especially in elderly patients.