Background: Efforts to reduce cardiometabolic disease often target individual-level risk factors. However, individual dietary behaviour and body mass index (BMI) may be influenced by the geographic patterning of food sources and local-area norms. Norms figure prominently in socio-behavioural theories yet spatial variations in norms have rarely been investigated in predicting individual health outcomes. This study assessed the contributions of fast-food availability and local-area norms for dietary behaviour and BMI to 10-year change in cardiometabolic risk.
Methods: Cardiometabolic risk, expressed using a validated continuous clinical index, was calculated using data from a population-based biomedical cohort in Adelaide, South Australia over three waves of follow up, 2000-2010. Local-area exposures were defined for participants (n=2228) using 1600m road-network buffers. Local-area norms (proportion residents with BMI ≥25 kg/m2 or not meeting fruit intake of ≥2 serves/day) were aggregated from individual-level self-reported data from a state-wide surveillance system, pooled across 2006-2010 for the study area. Fast-food availability (count of outlets) was extracted from a retail database. Separate multilevel models each featuring one local-area norm with fast-food availability as predictors of 10-year change in cardiometabolic risk were adjusted for area-level socioeconomic status and individual-level covariates.
Results: In separate models including fast-food availability, greater local-area proportion of residents with BMI ≥25 kg/m2 or not meeting fruit intake of ≥2 serves/day predicted worsening cardiometabolic risk across 10-years of follow up. Fast-food availability was not related to cardiometabolic risk. There was no evidence of interaction effects between local-area exposures.
Conclusions: Local-area norms (prevalence of unhealthful BMI and fruit consumption) reflect compositional population characteristics. Fast-food availability reflects built environmental context. In this study, only compositional norms, not fast-food context, predicted 10-year worsening of cardiometabolic risk. The implication of this longitudinal investigation is that the targeting of compositional norms might be more important than the targeting of fast-food context.