AI-enabled Clinical Decision Support System

Abstract

Each person’s life is extremely important and vital for the country’s development. Helping to limit the number of misdiagnoses can save many lives and strengthen families, as some families would lose their primary source of income as a result of misdiagnosis. Misdiagnosis is one of the significant errors in the medical field due to misjudgments by medical professionals eading to increased harm to patients. With 72 percent of errors occurring during the patient-practitioner encounter, it becomes increasingly important to reduce the error in real time. Lowering the mortality rate as a result of misdiagnosis would enhance and build the social well-being of the country. The area of general medicine is very vast and has more than 400 diseases and conditions under it. However, with differing and misleading symptoms for the diseases, it becomes rather confusing for a recent medical graduate to diagnose an individual within the limiting time frame for testing the patient. The developed solution is an artificial intelligence-based system that uses traditional machine learning algorithms and deep learning techniques to help new medical graduates and practitioners with limited experience reliably diagnose a patient’s medical condition based on the patient’s symptoms recorded during the clinical confrontation. The proposed solution is an AI-enabled Clinical Decision Support System in the form of a web application intended to assist medical graduates and healthcare professionals in accurately diagnosing a patient’s health condition based on the symptoms observed during doctor-patient encounters. The developed system has achieved the highest accuracy by using the Ensemble technique which is a combination of Support Vector Classifier, Random Forest, and Naive Bayes technique for textual data analysis and ResNet architecture for analyzing image data.