DesktopDx_JS: A lightweight web application helping clinical diagnosis in medicine

Jafar Alavy Toussy, Setareh Soltany, Parisa Alavi Toosi

Clinical approach is the main and most trusted tool in medicine for diagnosing diseases, but today medical staff are dealing with growing data in this approach. Companion applications, using medical models designed to help medical staff in this approach. We designed a lightweight web application, the DesktopDx_JS, with the goal of using medical prediction models to help in the diagnosing process of diseases. Desktop Dx_JS application uses JavaScript and AJAX as the coding language and NoSQL XML data. It uses modified/simplified object oriented MVC architecture. Model is composed of codes and data in the model XML files (in models directory), View is a simple HTML file, and Controller part is in the JS code of core engine. Core engine composed of subroutines for showing data in the start page and interacting with the user. Different models, both simple and complex, can used, including cutoff, regression, ANN, criteria and complex ones. Each model has a separate XML file that stores code and other data related to it. Now, the DesktopDx_JS database contains 19 prediction models. The system is very lightweight, needs no installation process can extend easily in the future and its maintenance is simple; but the main advantages of it are its flexibility and simplicity. Lightweight open source web applications, with flexible and simple coding platforms and user interface, are important for testing and deploying predictive medical models. The DesktopDx_JS system can do this task and may serve as a prototype for other similar applications and systems in the future.

Volume 12 | Issue 6

Pages: 2190-2203

DOI: 10.5373/JARDCS/V12I6/S20201181