Understanding urban-rural differences in cardiovascular disease risk factors across the life course

Background: Cardiovascular disease (CVD) is the leading cause of death and disability worldwide, contributing to 31% of all deaths globally in 2015. Similarly, in Australia CVD also accounted for approximately 30% of all deaths in 2014 and approximately 22% of Australia adults aged 18 years and over had one or more cardiovascular diseases. Australians living in regional, rural and remote areas are more likely to have some form of CVD and are more likely to die of CVD than those living in major cities. Modifiable risk factors including tobacco exposure, obesity, physical inactivity, unhealthy diets and harmful use of alcohol have been shown to increase the chance of developing CVD, yet there are significant gaps in knowledge about whether the distribution and clustering of these behaviours differs between those living in urban and rural areas. The existing literature fails to take a comprehensive life course perspective, meaning we do not currently know how these behaviours develop across the life course or how early life factors might contribute to the development of behaviour and disease over time and whether this differs among urban and rural populations. Furthermore, given that rurality and socioeconomic position (SEP) are highly inter-related, SEP must to be taken into consideration when investigating differences in CVD behavioural risk factors between urban and rural populations, but many studies have not properly considered SEP. The overall objective of this thesis was to compare the distribution and clustering of cardiovascular disease risk factors between Australians living in urban and rural settings from childhood (9-15 years) to mid-adulthood (31-41 years). The specific aims of the research presented in this thesis were: 1) To examine the distribution of CVD behavioural risk factors among young Australian adults (26-36 years) living in urban and rural areas and to establish the contribution of socioeconomic factors. 2) To identify CVD behavioural risk factor clusters among children and adolescents (9-15 years), and examine whether there are geographic or socioeconomic differences in cluster patterns 3) To determine the longitudinal relationship between childhood and adolescent CVD behavioural risk factor cluster patterns and adult cardio-metabolic risk factors. 4) To examine trends in body mass index (BMI), waist circumference and the prevalence of overweight and obesity among urban and rural children and adolescents (9-15 years) between 1985, 2007 and 2012. 5) To investigate whether trajectories of urban-rural area of residence from childhood (9-15 years) to adulthood (31-41 years) predicts BMI and weight status in mid-adulthood. Methods: Secondary analyses of data from three large population based studies: the 2007 Australian National Children's Nutrition and Physical Activity Survey (cross-sectional), the 2011-2013 Australian Health Survey (cross-sectional) and the Childhood Determinants of Adult Health study (longitudinal), a follow-up to the 1985 Australian School Health and Fitness Survey, were used. To address aim 1, data from children aged 9-15 years from the 1985 Australian School Health and Fitness Survey, the 2007 Australian National Children's Nutrition and Physical Activity Survey and the 2011-2013 Australian Health Survey, were used. To address aims 2-5 data from the Childhood Determinants of Adult Health study, were used. Participants aged 7-15 years in 1985 (n=8,498) were followed up in 2004-06 (n=3,999, aged 26-36) and 2009-11 (n=3,049, aged 31-41). Measurements included urban-rural area of residence, BMI, waist circumference, smoking status, alcohol consumption, diet, physical activity, depression and anxiety and socio-economic factors in both childhood and adulthood. Additional cardio-metabolic risk factors assessed in adulthood included fasting glucose, blood pressure and fasting lipids. A range of analytic methods were used, including log binomial, log multinomial and linear regression, a life course regression modelling framework and TwoStep cluster analysis. Key findings were: ‚Äö Among young adults (26-36 years), differences in CVD behavioural risk factors between urban and rural areas were identified. Young adults, particularly women, living outside of urban areas demonstrated poorer CVD behavioural risk factors than those living in urban areas. In general, socioeconomic position played a modest role but did not explain urban-rural differences. ‚Äö Among children and adolescents (9-15 years), four distinct cluster patterns of behavioural CVD risk factors were identified. These cluster patterns did not differ by urban-rural area of residence, but socioeconomic differences were apparent with unhealthier cluster patterns characterised by a higher proportion of participants of lower socioeconomic position. ‚Äö Unhealthier clusters of child and adolescent CVD behavioural risk factors predicted higher BMI, metabolic syndrome score and waist circumference in adulthood. These associations were independent of young adult CVD behavioural risk factors, socioeconomic position and urban-rural area of residence. ‚Äö There were no differences in BMI, waist circumference or the prevalence of overweight and obesity between urban and rural children and adolescents in 1985, 2007 and 2012. ‚Äö Greater cumulative exposure to rurality (over 25 years) and exposure during the 'sensitive period' of young adulthood (26-30 years) was associated with obesity in mid-adulthood (31-41 years). Conclusion: The research presented in this thesis addresses some crucial gaps in the literature and provides a fundamental first step in understanding geographic disparities in health. The findings have important implications for researchers, policy makers and health practitioners, and highlight a possible buffering effect for children living in rural areas, with selective migration (certain types of people, differentiated by factors such as age, socioeconomic position and health status including health behaviours, are more likely to move to certain types of areas) potentially contributing to the rural health disadvantage seen in adulthood. This suggests promising avenues for further research to disentangle how health status, health behaviours and socioeconomic factors affect complex social behaviour such as urban-rural migration. Doing so will be crucial for addressing the significant and unacceptable geographic disparities in health that are currently evident.