Download files
Abstract
The increasing use of telehealth technologies to remotely monitor patients with chronic obstructive pulmonary disease (COPD) has enabled pre-emptive management of these patients by clinical teams. However, the altered monitoring workload imposed on the clinical care team associated with using a telehealth management strategy has been given less attention.
This thesis describes the design of a decision support system (DSS) to assist clinical care teams in managing COPD patients. The development of an overall DSS framework for a prospective application is firstly described. An analysis of home telehealth data contained in retrospective databases is used to develop the DSS’s knowledge base (rules engine) to facilitate COPD management. Moreover, a preliminary exploration of the effect of data quality on DSS operation is also presented.
The proposed DSS design was implemented using a business process management system with a rules engine as the core component. The objective of the rules engine is to assist the clinical team with the detection of possible COPD exacerbation events, thus facilitating referral decisions. The rules were constructed with two separate clinical measurement databases (termed Database I and Database II); collected from COPD patients enrolled in home telehealth intervention groups as part of two randomised controlled trials. The data were pre-processed and features were extracted, then a classification and regression tree (CART) technique was used to generate the rules.
Four types of CARTs were constructed using four different reference standards, two from each database. The accuracies of the COPD exacerbation algorithms were 79.00% and 76.72%, and the referral recommendation CARTs kappa values were 0.52 and 0.45, for Database I and Database II, respectively. The results showed that the CARTs constructed using home telehealth data were capable of detecting COPD exacerbations as well as generating referral recommendations. In addition, data quality analysis that was performed on the data used by one of the CART algorithms confirmed that data quality issues did affect the reliability of the particular algorithm.
In summary, this thesis presents a DSS that specifically could be used to facilitate the remote monitoring and management of COPD patients. More generally it helps inform how similar DSSs linked to telehealth systems could improve the management of patients suffering chronic disease.