Improving our understanding of the influences of the environment on physical activity

Abstract

Physical activity reduces the risk of many chronic diseases, and much of the population would benefit from being more active. Public health strategies increasingly identify the environment as an important influence on activity and therefore a target for intervention. However, it is difficult to draw conclusive evidence from the existing literature as the range of environments and spaces individuals are exposed to are rarely accounted for. This thesis aims to further develop our understanding of the ways in which environmental characteristics influence physical activity. The first part examines cross-sectional associations between environmental characteristics such as street connectivity, air pollution, and deprivation, and self-reported and objective measures of physical activity. Data are used from the national multicentre UK Biobank study. The findings suggest environmental characteristics have the potential to contribute to different physical activities but interventions which focus on a single environmental attribute may not have the greatest benefits. The UK Biobank study uses measures of environments around residential addresses which may not capture all locations where participants are active. The second part of my thesis provides a more representative picture of environments that people are exposed to by focusing on activity spaces: locations accessed by an individual as a result of their daily activities. I systematically review literature which uses the activity space concept, and discuss research questions that have been answered, the spatial and temporal methods used, and the implications for causal inference. Included studies used variable methods to assess the features of spaces themselves (such as shape or size) or features within spaces (such as density of destinations). Informed by the conceptual work of the review, the latter section of the thesis uses quantitative and qualitative data from the Commuting and Health in Cambridge study to understand how new transport infrastructure might give rise to changes in use of space. The development of a replicable process to clean and prepare GPS data is presented and findings show how the infrastructure provides a new space for physical activity. The final project explores the applicability of different spatial methods for assessing population levels of activity and how changes in the location of physical activity might contribute to overall levels of activity over time. By developing and applying scalable methods to show how the spatial patterning of behaviour and physical activity changes in the context of an intervention, this thesis provides methodological and scientific contributions to the field of physical activity and public health. Future research in this topic area should aim to strengthen the basis for causal inference and develop evidence to effectively inform public health policy and action.