Understanding the patterns, levels and predictors of sedentary behaviour amongst socioeconomically disadvantaged women

Sedentary behaviour is any waking behaviour characterised by activity requiring low energy expenditure, such as sitting, reclining, or lying down. A high level of sedentary behaviour has been recognised as a risk factor for many poor health outcomes, independent of physical activity, including premature mortality, type-2 diabetes, cardiovascular disease, numerous cancers, obesity, and mental health problems such as anxiety, depression. Despite this, adults spend extensive amounts of time being sedentary at work, during leisure, and for transportation, spending more than 50% of waking time in sedentary activities. A rising trend of sedentary behaviour has been observed among adults, particularly television (TV) viewing, and computer use during leisure time, use of private automobiles for transportation, and occupational sitting. To develop strategies to prevent or decrease sedentary behaviour in groups at the highest risk, an in-depth understanding of the descriptive epidemiology (patterns, levels and determinants) of sedentary behaviour is required. However, little is known about the patterns and trajectories of sedentary behaviour or its longitudinal determinants. Social-ecological frameworks are useful for understanding behaviours such as sedentary behaviour and recognise that influences can occur at the individual, social and physical environmental levels. Individual level correlates of sedentary behaviour have mostly been examined in cross-sectional studies and include demographic (e.g., age), health (e.g., body mass index) or behavioural (e.g., smoking) factors. Factors less well understood are individual (e.g., cognitions such as self-efficacy, enjoyment), social (e.g., social cohesion,) and physical environmental (e.g., neighbourhood aesthetics, neighbourhoods walking, neighbourhood safety) factors. In addition, there are a lack of studies describing how sedentary behaviour is impacted over time during life events (e.g., the onset of parenthood, joining the workforce, cohabitation, marriage), which can de-stabilise behaviour in both negative and positive ways. The cross-sectional nature of previous studies limits the understanding of behaviour patterns over time, the temporal nature of associations, and the identification of potential causative insights. Lower socioeconomic position is a well-established risk factor for premature mortality, chronic conditions such as cardiovascular disease and type-2 diabetes, and poor health behaviours (e.g., physical inactivity, smoking). Although evidence around socioeconomic position and sedentary behaviour is in its infancy and currently equivocal, some evidence suggests leisure-time sedentary behaviour is highest amongst those who are most socioeconomically disadvantaged (e.g., living in deprived areas, low education, low income). Similarly, studies examining sex differences in sedentary behaviour have contrasting findings, yet there is some indication that women may spend more time in certain forms of sedentary behaviour (e.g., TV viewing) than men. Studies have also reported stronger associations between sedentary behaviour and poor health outcomes amongst women than men. Therefore, women of lower socioeconomic position are potentially a group at high level for sedentary behaviour and its negative health impacts. However, except one (socioeconomically disadvantaged mothers and their determinants of screen time) no longitudinal studies have examined patterns and determinants of different types of sedentary behaviour (e.g., TV viewing, computer use, overall sitting) over time in women of lower socioeconomic positions. Understanding patterns and determinants of sedentary behaviour among women with lower socioeconomic positions may help identify population subgroups at increased risk of chronic disease and identify potential targets for intervention. Objectives The overall aim of this thesis was to describe patterns and determinants of sedentary behaviour over time among women living in socioeconomically disadvantaged neighbourhoods. Therefore, this thesis investigated the following research objectives: 1. To describe patterns, levels and identify sociodemographic predictors of sitting over time in women living in socioeconomically disadvantaged neighbourhoods (Chapter 4). 2. To describe patterns, levels and identify sociodemographic predictors of TV viewing and computer use over time in women living in socioeconomically disadvantaged neighbourhoods (Chapter 5). 3. To determine the association between life events (change in parental status and change in employment status) and patterns of sitting, TV viewing, and computer use over time in women living in socioeconomically disadvantaged neighbourhoods (Chapter 6). 4. To determine the individual, social, and physical environmental factors associated with change in TV viewing over time (Chapter 7). 5. To examine level of education as a moderator of the longitudinal association between individual, social, and physical environmental factors and TV viewing. (Chapter 7). Methodology Four studies using data from the Resilience for Eating and Activity Despite Inequality (READI) study, were conducted to address these aims. Chapter 3 describes the methodology of that study in detail. In brief, women (n=4,347) aged between 18-46 years living in socioeconomically disadvantaged neighbourhoods across Victoria (Australia) self-reported at three-time points over five years (2007, 2010, 2012): sedentary behaviours (TV viewing, computer use, overall sitting); demographic (e.g., education, employment status, children living at home) and health (e.g. self-rated health, height/weight) characteristics; and individual (e.g. self-efficacy for physical activity), social (e.g. childcare access) and physical environmental (e.g. neighbourhood aesthetics) factors. For analyses, t-test and one way ANOVA were used to describe the patterns (Aim 1 & 2) and multilevel-mixed modelling was used to determine the predictors of sitting time (linear regression) (Aim 1 & Aim 3), TV viewing (negative binomial regression) (Aim 2, 3 & 4), and computer time (negative binomial regression) (Aim 2 & 3). A three-way interaction was used to examine the role of education as a moderator in Aim (5). Multiple imputations were used to address the missing values. Key findings Chapter 4 explored the patterns, levels and predictors (Aim 1) of change in sitting time over five years. A pattern of higher sitting time was observed in women who were young, living in urban areas, had poor to fair self-rated health, were employed full-time, were never married, and had no children at home. Longitudinally, baseline employment status and the number of children living at home were predictors of sitting time. Sitting time decreased in women working full-time and in those with no children, while sitting time increased in women who were not working and among those with two or more children living at home. Chapter 5 explored the patterns, levels and predictors (Aim 2) of change in TV viewing and computer time over five years. A pattern of higher TV viewing was found in those categorised as obese, with poor to fair self-rated health, current smokers, with low education levels, not working, with no income and with no children at home. A pattern of higher computer time was found in women who were young, living in urban areas, employed full-time, with higher education levels, with no partner and those with no children at home. Longitudinally, baseline employment status, education status and the number of children living at home were predictors of computer time. Computer time increased among those not working and with children living at home but decreased among those with higher education levels. No baseline sociodemographic and health predictors were identified for change in TV viewing over five years. Chapter 6 was informed by the findings of Chapters 4 and 5 and investigated the impacts of life events (change in parental status and employment status) on sedentary behaviour (Aim 3). These life events were associated with changes in sitting time, TV viewing, and computer time. Motherhood (having the same number of children, giving birth to a first or additional child/ren, fewer children (<18 years) living at home) appeared protective against high levels of sedentary behaviour. A decrease or increase in working hours was associated with a decrease in sitting time, while remaining in part-time work/not working and decreasing or increasing working hours were associated with a decrease in computer time. In addition, decreases in working hours were associated with an increase in TV viewing. Chapter 7 examined individual, social, and physical environmental factors and change in TV viewing over five years (Aim 4) and the role of education as a moderator of individual, social, and physical environmental factors and TV viewing (Aim 5). Cross-sectionally, individual (self-efficacy for physical activity, enjoyment for physical activity, outcome expectancies, and behavioural skills), social (interpersonal trust), and physical environmental (neighbourhood aesthetics, two or more TVs at home) factors were associated with TV viewing. No longitudinal individual, social and physical environmental predictors of TV viewing over time were found, and level of education was not a moderator between individual, social, and physical environmental factors and TV viewing. Conclusion This thesis provides important, comprehensive, and novel information to understand sedentary behaviour amongst women living in socioeconomically disadvantaged neighbourhoods, an underserved high-risk population. The longitudinal study design provides stronger evidence than before to understand the patterns and predictors of sedentary behaviour over five years amongst this population group. This thesis has identified a number of factors that may influence sedentary behaviour patterns: employment status, education level and parental statu...