Studies on the social and economic patterning of health behaviours in childhood and adolescence.

Abstract

The social patterning of harmful health behaviours (HHBs) is a major determinant of the social gradient in non-communicable diseases, and health generally. For this reason, more research is needed to understand why HHBs vary according to the social and economic position (SEP) of individuals in this structured manner. As middle-childhood and adolescence is recognised as a `critical' period for the formation of HHBs the data used for this thesis cover the period from age 9 to age 20. Previous research has often adopted an individualist paradigm that focuses on providing `bottom-up' explanations for health inequalities i.e. how individual choices and decisions aggregate to produce inequalities. In comparison, a structural paradigm considers the role of `top-down' processes. According to this paradigm, HHBs are a response to conditions imposed by social structures and must be understood in their social and economic context. Drawing on longitudinal data from the Growing Up in Ireland (GUI) 98 Cohort, this thesis sets out to contribute to both perspectives. The first paper examines the role of aspects of the young person's social environment in shaping their physical activity (PA) trajectories over a nine-year period, while controlling for established individual-level influences. We used multi-level linear spline modelling to estimate PA trajectories by SEP and compared trajectories of children from the least and most advantaged families. We then observed the reduction in this SEP differential with the addition of groups of variables (to our models) representing different aspects of the child's social environment family, neighbourhood and school. Family-level processes, namely role modelling and gatekeeping of PA, were the most effective in explaining the social patterning of PA followed by neighbourhood and school-level processes. This paper highlighted the important role of parents in shaping SEP differentials in PA. In the second paper, we theorised how family SEP in childhood might lead to smoking in late adolescence. Using path analysis, we established mediators of this relationship, and similar to paper one, we distinguished between individual-level and structural-level explanations. Our results indicate that there is no pathway via the variables that represent our individual-level explanations. There are, however, statistically significant mediating pathways between childhood SEP and smoking via the variables that represent our structural explanations. The role of the third paper was to identify HHBs that are likely to co-occur, and whether young people with similar patterns of behaviours have distinctive risk profiles. Using latent class analysis, we identified the clusters of HHBs: lifestyle behaviours (poor diet, low physical activity and high screentime) tend to occur together and appear to be relatively independent of substance use behaviours (tobacco smoking and underage drinking). Distinct clusters suggest distinct underlying processes, which we tested using multi-nominal logistic regression. According to our results, the experience of persistent poverty is associated with a higher risk of unhealthy lifestyle behaviours and appears to promote the development of `oppositional values', though only in a minority of cases. Where persistent poverty is accompanied by oppositional values, the probability of unhealthy behaviours and substance use rises significantly. The final paper examines why family economic hardship and deprivation is associated with a higher probability of smoking status among adolescents. The paper uses the Great Recession in Ireland as a `natural experiment' and quasi-experimental methods to examine the explanatory value of the `Family Stress Model'. We first adjust for the non-random probability of experiencing the effects of recession, using propensity score matching based on family characteristics before the onset of recession in 2009 when the child was aged 9. Second, we test whether economic strain, as a result of the recession, increased the probability of smoking initiation by the time the adolescent was aged 17/8. Our results show that increased economic strain results in a significantly increased likelihood of being a smoker. Moreover, we demonstrated that poor parental mental health and poor child social-emotional outcomes mediate this relationship, as proposed by the Family Stress Model (FSM). It could be argued that the central theoretical problem of sociology is the identification of the processes through which social structures shape individual behaviour, and the role of individual agency in this. This thesis teases out theoretical explanations for how social structures and agency interact to influence individual health behaviours that are tested empirically.