Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.709549
Title: Latent phase-type models for Italy's ageing population
Author: Mitchell, Hannah Jane
ISNI:       0000 0004 6058 9304
Awarding Body: Queen's University Belfast
Current Institution: Queen's University Belfast
Date of Award: 2016
Availability of Full Text:
Full text unavailable from EThOS. Please contact the current institution’s library for further details.
Abstract:
Quality of care is deemed a concept of immense importance, but also of great difficulty to define and analyse. This study proposes the development of a novel statistical approach to healthcare modelling which overcomes the need to define quality of care by treating it as a hidden layer in a special type of markov model. The study setting for this research is the Italian healthcare system, in particular admissions into geriatric wards of the Lombardy region of Italy during 2009. The Coxian phase-type distribution was applied to this dataset and shown to give the best representation of the flow of patients. Covariates were then incorporated into this distribution and applied to the data. A simulation study of Coxian phase-type distribution with covariates was also undertaken. The main purpose of this research was to develop the theory of the Coxian phase-type distribution by incorporating a hidden layer within it which can represent quality of care. In forming this model novel methodology was presented. A discrete-time and continuous-time version of the model were both applied to the data with the results analysed. A further extension of the continuous-time hidden Markov model with the Coxian phase-type distribution was developed whereby covariates where incorporated into the hidden element. The results of this model, with application to the Lombardy dataset was analysed followed by a simulation study of all the newly developed models presented. In addition to the hidden Markov model with Coxian phase-type distribution the model was extended to introduce a duration component within the hidden layer. This extension formed the hidden semi- Markov model which relaxes the strict Markov assumption. This model was also applied to the Lombardy dataset.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.709549  DOI: Not available
Share: