Associations of built environment attributes with bicycle use for transport

An increasing number of studies have examined neighbourhood built environment attributes associated with cycling. Some of them suggest non-linear relationships between built environment attributes and cycling. This study examined the strength and shape of associations of cycling for transport with objectively measured built environment attributes. Data were from 9146 Australian adults who took part in the 2009 South-East Queensland Travel Survey. Participants (aged 18–64 years) completed a 24-hour travel survey, in which they reported modes of travel. Residential density, Walk Score and a Space Syntax measure of street integration were calculated at a neighbourhood level using geographic information systems. Multilevel logistic regression analyses examined associations of bicycle use with each built environment attribute, which was modelled continuously and categorically. All continuous measures of the built environment attributes were associated with bicycle use. Each one-decile increment in residential density, Walk Score, and street integration was associated with 13%, 16%, and 10% higher odds of bicycle use, respectively. However, the associations appeared to be non-linear, with significant odds ratios observed only for the higher categories of each built environment attribute relative to the middle category. This study found that adults living in high-density neighbourhoods with more destinations nearby and well-connected streets were more likely to cycle for transport. However, medium-level density, access to destinations and street connectivity may not be enough to facilitate bicycle use. Further studies are needed to investigate urban design threshold values above which cycling can be promoted.