Background Despite strong evidence for health benefits from active travel, levels in the UK remain low. Changes to the physical and social workplace environment might encourage active travel but there is little high quality evidence for this.

Methods Data come from 419 participants in the Commuting and Health in Cambridge study, who completed postal questionnaires in 2011 and 2012. Participants lived and worked in the Cambridge area and were predominantly recruited through their workplaces. Each participant’s workplace environment was summarised using the number of physical characteristics from eight potential options (e.g. bicycle parking, shower facilities) and their level of agreement with five statements about workplace social norms around commuting (e.g. colleagues walk to work). We used a natural experimental approach to explore associations between changes in the physical and social workplace environment over time and changes in the proportion of commute trips i) exclusively by private motor vehicle, ii) exclusively by active modes and iii) including active modes, using fractional response logit regression in Stata 15, StataCorp. We additionally examined whether these associations differed between men and women.

Results In adjusted analyses, an increase of one physical characteristic was associated with a 2% (95% confidence interval 0 to 4) reduction in the proportion of commute trips by private motor vehicle and a 2% (95% CI 0 to 4) increase in the proportion of commute trips which included active modes. In sex stratified analyses these associations were only seen in males, with a 3% (95% CI 1 to 6) reduction in commute trips by private motor vehicle and an increase in commute trips including active travel of 5% (95% CI 3 to 8).

A change to more favourable social norms for walking or cycling among workplace management was associated with an increased percentage of commutes including active modes in women (4%, 95% CI 1 to 7) but not men. However, in both genders a change to more favourable social norms around colleagues’ cycling was associated with reduced commuting by exclusively active modes (-3%, 95% CI -5 to -1).

Conclusion This study provides robust longitudinal evidence for sex differences in the associations between workplace environment and commute mode. Physical factors were associated with more active commuting in men, while the social environment appeared to have more complex associations that were stronger among women. Although this study was small and geographically circumscribed, its findings propound larger studies in more diverse contexts.