Effectiveness of interventions for modal shift to walking and bike riding: a systematic review with meta-analysis

Objective: To assess the efficacy of interventions aimed at increasing walking and cycling. Design: Systematic review with meta–analysis Study selection: The electronic databases MEDLINE, PsycINFO and Web of Science were searched from inception on 22nd May 2023. Eligible study designs included randomised and non–randomised studies of interventions with specific study design features that enabled estimation of causality. No restrictions on type of outcome measurement, publication date or population age were applied. Data extraction and synthesis: Two reviewers independently extracted data and conducted quality assessment with Joanna Briggs Quality Assessment tools. Studies were categorised by intervention types described within the Behaviour Change Wheel. Where possible, random–effects meta–analyses were used to synthesise results within intervention types. Main outcome measures: The main outcome of interest was modal shift to active modes (walking and cycling). Other outcomes of interest included cycling and walking duration, frequency and counts, active transport duration and frequency, and moderate to vigorous physical activity duration (MVPA). Results: 106 studies that assessed the impact of an intervention on walking, cycling or active transport overall were included. Findings demonstrate that physical environmental restructure interventions, such as protected bike lanes and traffic calming infrastructure, were effective in increasing cycling duration (OR = 1.70, 95% CI 1.20– 2.22; 2 studies). Other intervention types, including individually tailored behavioural programmes, and provision of e-bikes were also effective for increasing cycling frequency (OR = 1.33, 95% CI 1.23– 1.43; 1 study) and duration (OR = 1.13, 95% CI 1.02–1.22, 1 study). An intensive education programme intervention demonstrated a positive impact on walking duration (OR = 1.96, 95% CI 1.68 – 2.21; 1 study). An individually tailored behavioural programme (OR = 1.23, 95% CI 1.08 – 1.40; 3 studies) and community walking programme (OR = 1.15, 95% CI 1.14 – 1.17; 1 study) also increased the odds of increased walking duration. This body of research would benefit from more rigour in study design to limit lower quality evidence with the potential for bias. Conclusions: This review provides evidence for investment in high–quality active transportation infrastructure, such as protected bike lanes, to improve cycling and active transport participation overall. It also provides evidence for investment in other non–infrastructure interventions. Further research to understand which combinations of intervention types are most effective for modal shift are needed. Active transport research needs to include more robust trials and evaluations with consistent outcome measures to improve quality of evidence and provide evidence on which interventions (or combinations of interventions) are most effective. Study registration: PROSPERO CRD42023445982