Clinical decision support for screening, diagnosis and assessment of respiratory diseases: chronic obstructive pulmonary disease as a use case

In this thesis we propose a framework for designing, developing, a clinical decision support systems (CDSS) offering a suite of services for the early detection and assessment of chronic obstructive pulmonary disease (COPD), and then demonstrate how these services can be integrated into the work-flo...

ver descrição completa

Detalhes bibliográficos
Autor: Velickovski, Filip
Tipo de documento: tese
Estado:Versão publicada
Data de publicação:2016
País:España
Recursos:CBUC, CESCA
Repositório:TDR. Tesis Doctorales en Red
OAI Identifier:oai:www.tdx.cat:10803/457000
Acesso em linha:http://hdl.handle.net/10803/457000
Access Level:Acceso aberto
Palavra-chave:Sistemes d'ajuda a la decisió
Sistemas de soporte a la decisión
Decision support systems
Malalties respiratòries
Enfermedades respiratorias
Respiratory diseases
Funció pulmonar
Función pulmonar
Lung function
Espirometria
Espirometría
Spirometry
Aprenentatge supervisat
Aprendizaje supervisado
Supervised learning
Garantia de qualitat
Garantía de calidad
Quality assurance
616.2
62
Descrição
Resumo:In this thesis we propose a framework for designing, developing, a clinical decision support systems (CDSS) offering a suite of services for the early detection and assessment of chronic obstructive pulmonary disease (COPD), and then demonstrate how these services can be integrated into the work-flow of healthcare providers. Furthermore, we focus on supporting spirometry, one of the main diagnostic tools in respiratory disease assessment. We present two methods to offer decision support in assuring the quality of a spirometry test that can be easily embedded into the CDSS framework. The first method is a novel algorithm that relies on a set of rules operating on 23 new parameters to define a high quality test. The second is a machine-learning approach, where we optimise the distinction between a good quality spirometry test and a poor one using a set of supervised-learning classifiers and hyper-parameters