Work

Risk Factor Epidemiology and Predictive Modelling Approaches to Understanding Chronic Disease Incidence

Year: 2019
Type: dissertation
Study design: Observational Study
Abstract: The chronic disease burden is a public health priority with most Canadian adults having chronic disease. Addressing this burden is a complex task requiring multiple stakeholders, including researchers. While researchers have traditionally studied chronic diseases individually “in silos”, studying multiple chronic diseases together is important because the results provide a comprehensive understanding of the overall health of the individual and the population. This dissertation contributes to the field of chronic disease research by elucidating the relationships between a person’s first chronic disease with shared lifestyle risk factors and applying this knowledge to create a predictive tool. To enable fresh perspectives about chronic disease incidence, various survival analysis techniques were applied. The first objective examined the relationships between shared lifestyle risk factors and the incidence of first chronic disease at a given age over the life course. This study found that most individuals will develop chronic disease prior to death, and heavy, daily smoking was the most important lifestyle factor for chronic disease incidence at a given age. The second objective focused on the Royston-Parmar model and its application for prognostic modelling in health research. The study found that this relatively new model was suited for prediction but has not been widely adopted for health research. The third objective developed and validated the Chronic Disease Population Risk Tool (CDPoRT), a population risk algorithm that predicted the incidence of multiple chronic diseases simultaneously over a ten-year period in the general adult population using routinely-collected data. The end product was a tool that consistently demonstrated strong predictive performance in estimating chronic disease risk. Overall, the studies capitalized on representative health survey data linked to administrative data to study the relationships between shared lifestyle risk factors and multiple chronic diseases, and leveraged these findings to create CDPoRT using traditional and novel survival analysis methods.
Author: Ryan Ng
Language: English
Open Access status: green
OpenAlex ID: W3156573201
Cites: 436
Cited by:
Related to: 9