Background Traffic and transport infrastructure can signal an area’s socially beneficial economic vitality. They can also cause harm by inhibiting pedestrian mobility of nearby residents. Living near busy roads is associated with increased cardiovascular mortality. The pathways of community severance (the negative impacts of traffic and transport on local people’s health) are unclear owing to inadequate metrics. We investigated the association between residential proximity to busy roads and perceived ability to walk locally, using a novel cross-sectional survey developed to measure community severance.

Methods Participants (N = 441; aged ≥18 y; 54% female) were randomly selected from addresses in three Southeast England neighbourhoods identified as likely to be affected by community severance, each bisected by a major road. A questionnaire was administered to measure participants’ demographic, socioeconomic and health characteristics, travel behaviours, and perceptions of neighbourhood traffic and transport infrastructure. Distance along the road and path network from each participant’s home address to their self-reported busiest local road was calculated using ArcGIS software. Associations between this distance and the extent to which traffic speed and volume affected participants’ perceived ability to walk locally were estimated by logistic regression.

Results Traffic speed was reported by 23% of participants, and traffic volume by 24%, as often/ always affecting their ability to walk locally, and both factors were significantly associated with worse self-rated health. Compared with participants who lived >400 m from their busiest road (the reference category), those who lived ≤100 m from it had three times higher odds of reporting that their ability to walk locally was often/ always affected by traffic speed (OR 3.5, 1.7–6.9). A similar result was found for traffic volume (3.4, 1.7–6.8). These associations were attenuated after adjustment for demographic and socioeconomic factors (2.7, 1.3–5.6; 2.9, 1.4–6.0 respectively). Using the same reference group, the odds of participants living >100 m to ≤200 m from their busiest road reporting impaired walking were around twice as high for traffic speed (2.3, 1.1–4.8) and for volume (2.5, 1.2–5.1) but adjustment rendered these associations non-significant (1.6, 0.8–3.5; 1.9, 0.9–4.0 respectively).

Conclusion Closer residential proximity to a road considered the busiest locally is associated with greater likelihood that local traffic speed and volume will often or always affect residents’ ability to walk in the area. Noise and air pollution likely explain much of the association between residential proximity to busy roads and poor health. A deterrent effect of busy roads on local walking, and therein health-promoting physical activity, warrants further investigation.