Neural Network Based Clinical Treatment Decision Support System for Co-existing Medical Conditions

摘要

Healthcare has always been considered as one of the most important determinant in improving the quality of life and preserving health of human beings around the whole world. However, millions of people have to put up with inaccurate medical care services due to lacking of experienced doctors, especially in the developing countries and areas. In this context, the diagnosis decision support systems are proposed, with which the physicians could interact to make a diagnosis decision. Meanwhile, the clinical treatment decision support systems are leveraged to help physicians choose appropriate clinical treatments. Recently there have been many researches focusing on providing clinical treatment advice on patients with single disease, but in reality there exist many patients suffering from co-existing medical conditions. In this paper, we propose a clinical treatment decision support system on co-existing medical conditions. The system automatically extracts features from the diagnosis by prestigious doctors and physicians. We then leverage a neural network to reveal the relationship between the diseases and the therapeutic regimen, providing evidence for clinical treatment decision making. Experiments over real-world dataset show that our approaches can achieve high precision in predicting therapeutic regimen based on electronic medical record data. Our system can help physicians find appropriate medical treatments for patients in some complicated conditions.