摘要：Objectives. We evaluated the effects of providing new high-quality, traffic-free routes for walking and cycling on overall levels of walking, cycling, and physical activity. Methods. 1796 adult residents in 3 UK municipalities completed postal questionnaires at baseline (2010) and 1-year follow-up (2011), after the construction of the new infrastructure. 1465 adults completed questionnaires at baseline and 2-year follow-up (2012). Transport network distance from home to infrastructure defined intervention exposure and provided a basis for controlled comparisons. Results. Living nearer the infrastructure did not predict changes in activity levels at 1-year follow-up but did predict increases in activity at 2 years relative to those living farther away (15.3 additional minutes/week walking and cycling per km nearer; 12.5 additional minutes/week of total physical activity). The effects were larger among participants with no car. Conclusions. These new local routes may mainly have displaced walking or cycling trips in the short term but generated new trips in the longer term, particularly among those unable to access more distant destinations by car. These findings support the potential for walking and cycling infrastructure to promote physical activity. Walking and cycling are important sources of everyday activity 1,2 and are independently associated with a wide range of health benefits. 3–6 The potential magnitude of such benefits may be particularly large in settings such as the United Kingdom, where most people are insufficiently active for health (only 5% of adults, as assessed using accelerometry 7 ) and where cycling in particular is rare (accounting for 2% of journeys 8 ). It is widely recognized that a supportive built environment may be needed to promote walking and cycling and to achieve an enduring increase in activity at the population level. 9–12 Nevertheless, multiple recent reviews 10,13–19 have highlighted the paucity of rigorous studies evaluating the effects of new walking and cycling infrastructure such as segregated cycle routes or traffic-free bridges. Drawing on such controlled, longitudinal studies as do exist, recent systematic reviews have provided limited evidence as to whether infrastructural improvements increase walking or cycling. 18–20 Five studies (from Denmark, England, the Netherlands, and the United States) reported increases in cycling after the implementation of fairly substantial infrastructure improvements such as building cycle parking, extending networks of on- and off-road cycle routes, or modifying junctions to create advance stop lanes for cyclists. 21–25 Interpretation is complicated, however, by the fact that these infrastructural modifications were sometimes accompanied by other cycling initiatives (e.g., media campaigns, cycle training, or community-based events). 22–24 Likewise, in many walking interventions the infrastructural improvements were small relative to other intervention components. For example, 3 studies (from Australia, Belgium, and the United States) evaluated whole-community walking programs that, among many other things, improved signage, made minor repairs to footpaths, or cleaned up walking trails. 26–28 One reported no overall change in population walking levels, 26 and a second reported modest increases. 27 The third focused on promoting walking trails and found that trail use increased but overall walking levels did not. 28 Another US study reported significant increases in walking and cycling around a newly built urban trail 29 but could not determine whether this reflected new walking and cycling or simply trips displaced from elsewhere. Finally, a recent Australian study reported that moving home to an area with greater access to local recreational or transport-related destinations predicted increases in walking 30 but not cycling. 31 In summary, the evidence has suggested that infrastructural interventions may increase walking or cycling when delivered at high doses, but at lower doses may be used without necessarily increasing total activity. In addition, few studies have examined whether any effects are observed equally across different population groups, and very few have examined equity impacts with respect to any characteristic other than gender. 18,19,32 Among those that have, 1 Australian study found a trend toward a larger increase in activity among women than men, 26 1 English study found comparable changes across all socioeconomic groups, 23 and 1 English and 1 US study found some suggestion of larger increases among socioeconomically disadvantaged groups. 24,28 None of these 4 studies included formal tests for interactions.
