A Decision Support System for Chronic Obstructive Pulmonary Disease (COPD)

Abstract

Obstructive chronic obstructive pulmonary disease (COPD) is a respiratory disease characterized by a chronic air flow limitation and associated with major economic and social problems. In an attempt to find a solution to these problems, numerous systems of clinical decision support for the management of patients with COPD have been developed in recent years. In particular, systems based on machine learning algorithms have been developed with the aim of monitoring the health status of patients and foreseeing and preventing exacerbations and hospital admissions. An in-depth research into scientific literature has shown that, in the state of the art, these goals have not yet been met and the performance of the current systems is not clinically acceptable. The aim of this work is the design and implementation of a new clinical decision support system that can at least partially fill the gaps present. The first step in the work was to try to replicate the performance of support systems for similar decisions, already present in scientific literature. Using the physiological parameters acquired by 414 patients using respiratory function tests, two predictive models were made using the same machine learning algorithms (neural network and support vector machine). The performance obtained was comparable to those of the scientific literature. The next step was to create a new predictive model, with superior performance to previous models. The machine learning algorithm chosen is C5.0. The performance obtained was significantly better than the two previous models. The new predictive model was implemented within a user interface, implemented in Java programming language, the COPD Management Tool. The software developed allows the evaluation and classification of the results of respiratory performance tests, with excellent performance, compared to the current state of the art and can therefore be used in many clinical applications.