Background Numerous cross-sectional studies have examined the associations between individual, psychological and environmental characteristics and active commuting, but few longitudinal studies exist to support more robust causal inference. Using two complementary approaches to longitudinal analysis, we aimed to build on this evidence by exploring whether (a) individual, psychological, workplace and route environmental characteristics, (b) changes in route environmental characteristics, were associated with changes in active commuting.

Methods 655 adult commuters completed postal questionnaires concerning past-week commuting trips at similar times of year in 2009 and 2010. We computed changes in time spent walking and cycling and the proportion of car trips, and identified those who took up or maintained walking, cycling or alternatives to the car as the usual mode of travel. Using multivariable logistic regression, we assessed the associations between uptake and maintenance of travel behaviours and baseline individual, psychological and perceived and objective environmental characteristics and between changes in active commuting and changes in perceived environmental characteristics of the route to work.

Results Mean within-participant changes in active commuting were relatively small (walking: +3.0 minutes/week; cycling: -5.3 minutes/week). A lack of free workplace parking predicted uptake of walking and alternatives to the car, while less favourable attitudes towards car use predicted continued use of alternatives to the car. Self-reported and objectively-assessed convenience of public transport respectively predicted uptake of walking (OR 2.47 [95% CI 1.44, 4.25]) and cycling (2.59 [0.99, 6.78]). Those who reported an increase in the convenience of public transport were more likely to take up alternatives to the car (3.24 [1.28, 8.14]) and those who reported convenient cycle routes at baseline were more likely to take up cycling (2.48 [1.04, 5.93]) and alternatives to the car (4.65 [1.45, 14.92]). Those who reported an increase in the perceived danger of cycling reported 8% more car trips, whereas those who reported a decrease were more likely to take up alternatives to the car (2.50 [1.07, 5.86]).

Discussion Individual, psychological and environmental factors were associated with uptake of active commuting. A few environmental characteristics relating to the convenience and safety of routes for active travel and the convenience of public transport were associated with changes in active commuting both as baseline predictors of change and as dynamic correlates of change, suggesting that these may be suitable targets for interventions to promote walking and cycling to work.