Promoting cycling: a review of interventions.

objectiveTo determine which approaches promote cycling, the size of the effects, and whether there are associated benefits on physical activity or anthropometric measures, by means of a systematic review. DATA SOURCES: Intervention studies with a cycling outcome measure were identified from searches of 13 databases, previous systematic reviews, Web sites, and reference lists of selected articles. International experts were asked to identify further relevant published or unpublished studies. The search was completed in January 2010. STUDY SELECTION: The selection criteria were controlled trials and "before and after" studies of the effect of any intervention on cycling behavior; studies had to include a minimal or no intervention comparison group or a similar population; and a specific measure of cycling before and after the intervention was required. Studies using stationary bicycles were excluded. Of 32,916 records retrieved, 118 full-text articles were assessed, and 25 were included in the quantitative analysis. Each step in the selection process was independently reviewed by a second reviewer, and disagreements were resolved. DATA EXTRACTION: Details of the study setting, objective, intervention, sample, duration, outcome measures, and results were extracted and checked by 2 reviewers, who also assessed validity (evaluated in summary as yes/no on randomization, representativeness, comparability, measurement, and use of statistical test). Disagreements were resolved by discussion. Outcomes were summarized as absolute changes in cycling after adjustment for changes in the control group. Meta-analysis was not possible because of heterogeneity and statistical inadequacies in the studies. MAIN

resultsSix studies [2 randomized controlled trials (RCTs), 1 cohort study, and 3 controlled repeat cross-sectional studies, all meeting ≥3 validity criteria] included interventions primarily intended to promote cycling. An RCT using individual counseling and incentives as interventions found that Swedish women with abdominal obesity in the intervention group were more likely to report cycling >2 km per day after 18 months than the control group [odds ratio, 7.8; 95% confidence interval (CI), 4.0-15.0]. Increase in cycling was not associated with greater reductions in waist circumference in women in the intervention group than in the control group. A cluster RCT (Bike-Texas Safe Routes to School) using combined educational and promotional activities as interventions found no difference after 1 semester in cycling to school among fourth and fifth graders and their parents compared with control schools, but the prevalence of recreational cycling increased (net increase + 2.54 days per week; P = 0.02). In a controlled repeat cross-sectional study in the Netherlands, improving the connectivity of 1 area of Delft was associated with an increase from 40% to 43% in household cycle trips after 3 years, compared with a change from 38% to 39% in an unimproved area. In Odense, Denmark, promotion and improved infrastructure were associated, after 3 years, with a 3.4% net increase in bicycle trips (adjusted for regional trends). In a cohort study in Australia, residents in 1 area were encouraged by a wide-ranging promotional program, including cycle training and free bike hire, to use existing cycle paths. After 2 years, residents of the intervention area reported more use of the cycle paths than the comparison group (net change in prevalence, +5.1%; P < 0.01), and bicycle counters also showed greater cycle path use in the intervention area (net increase, 7.9%), but no increase in the prevalence of cycling in the population occurred. In English towns, media campaigns, services for cyclists, and improvements to infrastructure were associated with increased proportions of residents who reported cycling more frequently (net increase in proportions cycling ≥12 times per month in 2 areas, 0.97% and 1.65%, and net increase in proportions cycling for ≥30 minutes once per month, 2.78% and 1.89%) compared with control towns. There was a reduction in the proportion of residents in intervention areas who were classified as inactive (-2.6%; 95% CI, -3.7% to -1.5%), but no increase in the proportion who were more than moderately active. The effects of individualized marketing of walking, cycling, and public transit use as an alternative to car use were investigated in 16 studies (2 cohort studies and 14 controlled repeat cross-sectional studies; 13 studies met 2 validity criteria). Most studies included individualized information and incentives to households interested in changing their travel behavior and were associated with modest but generally consistent increases in self-reported cycling trip frequency (median, +8 trips per person per year; range, 0 to +21).

conclusionsIndividual and community intervention programs, improving infrastructure, and marketing to households all increased cycling (usually self-reported) to a small extent. Environmental changes combined with advice and support may be needed to increase cycling substantially and in a sustainable way in the population.