Introduction: This thesis examined a possible relationship between sleep and breast cancer. Only a small number of previous studies have investigated sleep and breast cancer and the results have been inconclusive with increased risks, decreased risks and no associations reported. The previous studies have had a number of limitations, most notably, the lack of reliable and valid sleep measurement tools and the small number of aspects of sleep being investigated. This thesis aimed to investigate ways to improve the measurement of sleep in epidemiological studies, and then apply these measurements to investigate the relationship between a number of aspects of sleep and breast cancer risk. Methods: This thesis undertook to improve the measurement of sleep by: Undertaking a reliability study of the 'usual' sleep questions that are commonly used in epidemiological studies. Undertaking a validation study to examine whether the 'usual' sleep questions commonly used correspond to objective measures of sleep as assessed using actigraphy. Investigating the development of novel 'sleep-exposure' metrics to overcome some of the limitations of 'usual' sleep questions, using occupational exposure assessment theory. This thesis undertook empirical investigation of the relationship between sleep and breast cancer by: Examining the relationship between sleep duration, sleep quality, sleep disturbance metrics and domestic exposure to light at night and risk of breast cancer in a population-based case-control study. Results: Methodological results: The results of the reliability and validity studies suggest that although the 'usual' sleep questions typically used in epidemiological studies are reliable, they do not show close correspondence with objective measures of current sleep as assessed using actigraphy. In investigating metrics of 'sleep-exposure', the metric 'average hours of sleep disturbance with a harmful effect of longer sleep' was found to be the best predictor of the data, relative to the other models. Empirical results: No associations were found between sleep duration, sleep quality, metrics of sleep disturbance or domestic exposure to light at night and risk of breast cancer. Analysis to investigate the potential impact of selection / misclassification bias did not indicate a strong influence of bias on the results. Conclusions: The results from methodological aspects of this thesis have highlighted the problems associated with reliance on self-report measures of 'usual' sleep, and the caution that may be required in interpreting studies that have used these types of questions. The novel sleep disturbance metrics developed in this thesis have provided an alternative way of conceptualising sleep that may serve as a useful reference for others looking to expand their scope of sleep measurement. The results of the empirical investigations in this thesis do not support any association between a number of different aspects of sleep, or exposure to light at night, and breast cancer risk. Analysis to quantify possible selection / misclassification bias did not indicate that the results were affected by bias. The results of this thesis may give reassurance to women experiencing short, poor quality or frequently disturbed sleep, that they do not appear to be at increased risk of breast cancer.
|Qualification||Doctor of Philosophy|
|Publication status||Unpublished - 2013|