Residential proximity to urban centres, local-area walkability and change in waist circumference among Australian adults

Highlights

• Living far from city centre was associated with a greater adiposity increase.

• Living near suburban centre was associated with a greater adiposity increase.

• Walkability was not associated with adiposity increase.

Abstract

Consistent associations have been observed between macro-level urban sprawl and overweight/obesity, but whether residential proximity to urban centres predicts adiposity change over time has not been established. Further, studies of local-area walkability and overweight/obesity have generated mixed results. This study examined 4-year change in adults' waist circumference in relation to proximity to city centre, proximity to closest suburban centre, and local-area walkability. Data were from adult participants (n = 2080) of a cohort study on chronic conditions and health risk factors in Adelaide, Australia. Baseline data were collected in 2000–03 with a follow-up in 2005–06. Multilevel regression models examined in 2015 the independent and joint associations of the three environmental measures with change in waist circumference, accounting for socio-demographic covariates. On average, waist circumference rose by 1.8 cm over approximately 4 years. Greater distance to city centre was associated with a greater increase in waist circumference. Participants living in distal areas (20 km or further from city centre) had a greater increase in waist circumference (mean increase: 2.4 cm) compared to those in proximal areas (9 km or less, mean increase: 1.2 cm). Counterintuitively, living in the vicinity of a suburban centre was associated with a greater increase in adiposity. Local-area walkability was not significantly associated with the outcome. Residential proximity to city centre appears to be protective against excessive increases in waist circumference. Controlled development and targeted interventions in the urban fringe may be needed to tackle obesity. Additional research needs to assess behaviours that mediate relationships between sprawl and obesity.

Introduction

The relationship between local-area attributes and residents' obesity is the focus of an emerging body of research (Kirk et al., 2010, Sallis et al., 2012). A recent review on obesogenic environments found mixed associations, however, between environmental measures and obesity (Mackenbach et al., 2014). Walkability has been examined frequently, on the basis of its link with physical activity (Freeman et al., 2013, Van Dyck et al., 2010, Villanueva et al., 2014). However, among 19 studies that examined walkability in the review, fewer than half (8 studies) reported associations with measures related to obesity, and the rest reported either statistically non-significant associations or significant associations only for subgroups (Mackenbach et al., 2014). The most consistent relationships were found for urban sprawl (expansion of low-density residential areas at the urban fringe), with seven of nine studies reporting associations between sprawl and overweight/obesity, and the remaining two reported non-significant relationships (Mackenbach et al., 2014). More recent cross-sectional studies also attest to a relationship between urban sprawl and higher levels of obesity (Berrigan et al., 2014, Ewing et al., 2014). In addition, longitudinal studies indicate that moving to a new residential location with greater levels of sprawl is associated with subsequent weight gain (Arcaya et al., 2014, Plantinga and Bernell, 2007).

Sprawl is often operationalised as ‘county sprawl index’ (Arcaya et al., 2014, Berrigan et al., 2014, Ewing et al., 2014), a county-level measure calculated for US studies from population density and block size (Ewing and Hamidi, 2014). However, counties are a spatially large administrative unit with a median size of 1600 km² (United States Census Bureau, 2010). It would be quite possible that overweight and obese individuals are not evenly distributed within such a large spatial unit. It is thus arguably just as important to examine how sprawl measured within a metropolitan region relates to changes in weight status over time. Distance to city centre can be a reasonable measure in examining the relationship of sprawl and adiposity given that such development is often characterised as taking place at the periphery a city (Resnik, 2010), and car commuting, in particular long commutes, is known to be associated with greater levels of adiposity (Hoehner et al., 2012, McCormack and Virk, 2014, Sugiyama et al., 2016). It might be hypothesised that locations distal to city centre where residents are more likely to rely on cars for commuting might be conducive to weight gain. We are not aware, however, of any research that has examined the relationship between distance to city centre and adiposity changes over time. A similar urban-scale measure, distance to a suburban centre (shopping area with a transportation hub), which represents a local-scale access to various destinations, might also relate to changes in adiposity over time. Although this is not a measure directly corresponding to sprawl, living near such a centre (even if not close to a city) may promote active living, which could support maintaining healthful body weight.

This prospective observational study evaluated in a population-based cohort in Adelaide, Australia, how proximity to city centre, proximity to suburban centre, and local walkability were associated with change in waist circumference. We examined the independent and joint associations between these environmental measures and change in waist circumference to evaluate the unique and potential synergetic effects of proximity measures and walkability. In light of previous mixed findings regarding the associations between walkability and overweight/obesity, we also assessed whether the relationship between walkability and increasing waist girth was modified by individual demographic variables, area-level socioeconomic characteristics, and proximity measures.