A Systemic Review of Machine Learning Approaches for Adverse Drug Reaction Detection: Novel Perspective and Challenges

Review article

Abstract

Medication errors significantly impact patient treatment outcomes, necessitating the integration of modern technologies for improved detection and prevention. This review amalgamates findings from multiple studies on medical decision support systems and machine learning to predict and mitigate prescribing errors. A systematic examination of 30 articles published between 2015 and 2023 reveals the utilization of various methodologies, including outlier detection testing, interruptive prescribing alerts, and probabilistic, machine learning-based clinical decision support systems. The review underscores the imminent need for sophisticated techniques to address the limitations of traditional Adverse Drug Reaction (ADR) detection methods. Notably, the incorporation and refinement of machine learning approaches emerge as promising strategies. The examination of these studies highlights the potential of machine learning to revolutionize patient safety and healthcare quality by enhancing efficiency and accuracy. In conclusion, this review emphasizes that machine learning represents a groundbreaking approach in detecting and preventing medication errors. The integration of advanced methods, coupled with a robust reporting system, is crucial for advancing the landscape of ADR discovery. This approach not only facilitates efficient and accurate healthcare delivery but also ensures a patient-centric focus, marking a significant stride towards improved patient safety and healthcare quality.