We have described three levels at which physical activity behaviour is influenced: the individual, the societal and the structural. The bulk of the evidence base has focused on the individual level, supported by theoretical physical activity behaviour change models. However, individuals do not exist in a vacuum. As indicated, factors operating at the societal and structural level mediate physical activity behaviour in older people. Li and colleagues have commented on the scarcity of research using methodologies that collect data across these levels simultaneously, whether cross-sectionally or longitudinally (Li, Fisher et al. 2005). It is important that we use multilevel models and multifaceted data collection methods in order to learn more about the inter-level influences. The conceptual model shown in Figure 5.1, constructed by Li and colleagues (2005), illustrates the multi-level nature of the relationship.
From a population health perspective, making structural changes to the environment is more likely to have a greater reach than programs targeting the individual. People of all ages and levels of activity within a neighbourhood will be influenced by changes to the area. Further, the changes are likely to be maintained for some time, compared to the limited impact of individual level interventions over time. Saelens and colleagues (2003) estimated that the increase of 15 minutes more walking per week seen in those living in areas of ‘high walkability’, whilst modest at an individual level, translates into significant public health gains.
Although studies empirically examining the link between structural factors and health are few, their numbers are increasing. Recently Frank and colleagues (2004) explored the relationships between urban form factors, physical activity and obesity in a sample of almost 11000 Atlanta US residents who completed a computer assisted telephone interview (CATI). Whilst there was no specific focus on older residents, the findings are relevant. Existing data were used to calculate street connectivity, residential density and land-use mix. Land-use mix was the strongest predictor of obesity (BMI >30 kg/m2). Land-use mix is measured on a scale of 0-1, where 0 equals single usage e.g. residential. An increase from 0.15 to 0.3 (reflective of other regions) would decrease the likelihood of obesity in a population by 5%. Very few reported walking for transport. Each additional km walked per day was associated with a 4.8% reduction in obesity likelihood. In a similar study in Sydney, Wen and colleagues (2006) found a strong relationship between driving to work and the prevalence of obesity. People who drove to work were significantly less likely to achieve recommended physical activity levels compared to non-car users. The risk of obesity increased with age (OR 1.95, 95% CI 1.43-2.65 in those age 60 and over) and was higher in those with lower socioeconomic status (as measured by SEIFA scores) (Wen, Orr et al. 2006). This research provides further level IV evidence that both individual and structural strategies are needed to produce health impacts.
SummaryIndividual, societal and structural factors all influence uptake and sustained participation in physical activity. Theoretical frameworks are increasingly being used in physical activity effectiveness studies. Studies incorporating distinct maintenance strategies and evaluation thereof are needed. Further, such strategies need to be multi-level, accommodating societal and structural influences on individual behaviour. Physical activity is increasingly being used in community capacity building strategies, to encourage social connectedness and, for older people, to reduce social isolation.
Top of page