A statistical investigation of the risk factors for tuberculosis

Type of content
Theses / Dissertations
Publisher's DOI/URI
Thesis discipline
Statistics
Degree name
Master of Science
Publisher
University of Canterbury. Mathematics and Statistics
Journal Title
Journal ISSN
Volume Title
Language
Date
2013
Authors
van Woerden, Irene
Abstract

Tuberculosis (TB) is called a disease of poverty and is the main cause of death from infectious diseases among adults. In 1993 the World Health Organisation (WHO) declared TB to be a global emergency; however there were still approximately 1.4 million deaths due to TB in 2011. This thesis contains a detailed study of the existing literature regarding the global risk factors of TB. The risk factors identified from the literature review search which were also available from the NFHS-3 survey were then analysed to determine how well we could identify respondents who are at high risk of TB. We looked at the stigma and misconceptions people have regarding TB and include detailed reports from the existing literature of how a persons wealth, health, education, nutrition, and HIV status affect how likely the person is to have TB. The difference in the risk factor distribution for the TB and non-TB populations were examined and classification trees, nearest neighbours, and logistic regression models were trialled to determine if it was possible for respondents who were at high risk of TB to be identified. Finally gender-specific statistically likely directed acyclic graphs were created to visualise the most likely associations between the variables.

Description
Citation
Keywords
TB, Tuberculosis, NFHS-3, classification trees, nearest neighbours, logistic regression, Directed Acyclic Graphs
Ngā upoko tukutuku/Māori subject headings
ANZSRC fields of research
Rights
Copyright Irene van Woerden