Using statistical modelling approaches to explore the prevalence of cigarette smoking in the Global South

Abstract

The prevalence of tobacco use is one of the leading causes of preventable deaths. Most tobacco users reside in low- and middle-income countries (The Global South). Current estimations predict a surge in tobacco use and preventable deaths in these countries, which are more economically vulnerable to the variety of problems that tobacco use causes. Data must guide current and future national and international tobacco control measures. Data collection within Global South countries is often irregular, imperfect, or absent. Most prevalence estimates represent national rates, which obscure the variations between demographic and social groups at the small-area level over time. These heterogeneities of prevalence must be understood to plan effective future tobacco control.
This thesis investigates the heterogeneity of cigarette smoking prevalence across the Global South. It uses Demographic and Health Surveys and auxiliary datasets in binary logistic multilevel models to investigate the associations of selected independent variables with current smoking prevalence, the trajectories of these associations over time, and their links to country-level determinants of smoking. A Bayesian geostatistical model is used to predict small-area level smoking prevalence.
The thesis shows considerable variation between countries in the magnitude and direction of individual demographic and social determinants of smoking, varying trends in smoking over time, and the importance of country-level determinants of smoking, including the extent of commitment to tobacco control. Novel small-area predictions highlight areas needing to catch up on tobacco use reduction targets. Collectively, this thesis demonstrates the need to improve the targeting of tobacco control policies in the Global South.

More information

Identifiers

ORCID for Robert Steven Taylor: orcid.org/0000-0001-6308-8220

ORCID for Andrew Tatem: orcid.org/0000-0002-7270-941X

ORCID for Graham Moon: orcid.org/0000-0002-7256-8397

ORCID for Chris Jochem: orcid.org/0000-0003-2192-5988

Catalogue record

Export record