Skip to main content
Download PDF
Download ePub

The effectiveness of new urban trail infrastructure on physical activity and active transportation: a systematic review and meta-analysis of natural experiments

International Journal of Behavioral Nutrition and Physical Activity volume 22, Article number: 36 (2025) Cite this article

Abstract

Background

Cities in Western countries are investing billions of dollars in new cycling infrastructure (urban trails) to support active transportation (AT) and leisure-type physical activity (PA). Little empirical evidence exists on the effectiveness of urban trails on changes in AT or PA.

Design and methods

We searched CINAHL, OVID, SPORTDiscus, Transport Research International Documentation (TRID), Web of Science Core Collection and Google Scholar for articles published from 2010 to 2023. We included controlled experimental studies that reported PA, AT or trail counts as outcome measures before and after construction of an urban trail. A modified risk of bias tool was employed to assess the methodological quality of each selected study (Prospero ID: CRD42023438891).

Results

Three independent reviewers screened abstracts from 3936 articles identified in the original search and identified 24 articles that met inclusion criteria: 11 studies (n = 11,464) that measured changes in PA, 8 studies (n = 92,001) that measured changes in cycling traffic and 5 studies (n = 4,958,203) that measured changes in rates of AT/cycling. Meta-analysis revealed that new trails increased PA levels among individuals in proximity to one, compared to those living in control areas (SMD = 0.12; 95% CI: 0.04, 0.20; I2 = 73%; n = 11,464). This effect was marginally stronger when data were restricted to individuals living in closest proximity to trails (SMD = 0.14; 96% CI: 0.06 to 0.25, I2 = 74%; n = 8234). Meta-analyses were not possible for measures of AT and cycling counts. All studies were at high risk of bias due to a failure to adhere to reporting guidelines for quasi-experimental studies.

Conclusions

There is limited but intriguing evidence that the addition of protected urban trails increases daily PA for individuals living in neighbourhoods that receive them. The strength of this evidence could be enhanced with the application of and adherence to principles of causal inference and increased diversity of individuals included in study designs.

Introduction

Implementing protected infrastructure to support cycling or walking (referred to herein as urban trails) is the most common governmental strategy for providing citizens with safe spaces for active transportation (AT) and physical activity (PA) [1,2,3]. Although the expansion of urban trails is the fastest-growing municipal strategy for reshaping the urban built environment [4], there is little empirical evidence on whether they increase AT or PA in urban populations.

Randomized controlled trials are impractical or not feasible to study the health impacts of changes to the built environment. As such, well controlled natural experiments offer the best empirical design to estimate causal effects of adding new urban trails on individual health outcomes or behaviours, including PA and AT [5, 6]. To date, several systematic reviews have tried to summarize the available evidence for the effectiveness of adding trails to urban environments on individual-level PA and walking/cycling behaviour for citizens living adjacent to them [7,8,9,10]. These reviews suggest that the implementation of urban trails leads to an increase in PA levels for individuals living in proximity to them. The causal nature of this evidence is limited however as some failed to include natural experiments that were published in transportation journals [9], others included natural experiments without control conditions and observational studies [7, 8] and some reviews did not meta-analyze results across multiple studies [7, 8]. Although journals have begun to adopt and adhere to guidelines for reporting natural experiments [11, 12] and scientists have called for more robust study designs that meet assumptions for causal inference [10], to the best of our knowledge no systematic review of urban trail natural experiments to date has adopted the most recent reporting guidelines for natural experiments when assessing the risk of bias for previously published studies [11, 13]. A more precise estimate of the effect of urban trails on both PA and AT and the overall quality of this evidence is needed to inform policy makers and expert guidelines for the promotion of PA through changes to the built environment [14,15,16,17].

The purpose of this systematic review and meta-analysis was to overcome the limitations of previous systematic reviews and provide a comprehensive estimate of the causal effects of implementing new urban trails on individual changes in PA and AT behaviours. We treated PA and AT as separate outcomes as urban trails can lead to increased leisure time PA alone, AT behaviours alone or both [18].

Research design and methods

Data sources, search strategy, and eligibility criteria

This review is reported according to the PRISMA guidelines [19]. An initial search strategy was developed and received a PRESS from another information specialist [20]. We then searched CINAHL with Full Text (EBSCO), EMBASE (Ovid), MEDLINE (Ovid), SPORTDiscus (EBSCO), Transport Research International Documentation (TRID), Web of Science Core Collection (Clarivate), and Google Scholar for grey literature. The systematic search of these databases was conducted to identify studies of natural experiments with a valid pre/post design and control condition published between January 1, 2010, and July 21, 2023. The year 2010 was chosen as the cut-off year as it serves as a salient time point when many governments began investing more significantly in creating cycling infrastructure and preliminary search efforts suggested most studies of natural experiments began to be published around 2010. The protocol was registered on PROSPERO (CRD42023438891). Institutional ethics approval was not required for this systematic review and meta-analysis as individual level data were not used for these analyses.

Classification of the intervention

Natural experiments of either high-comfort (cycle tracks, local street bikeways, bike paths) or medium-comfort (multi-use paths) trails according to the CAN-BICS Classification System [21, 22] were included in the review. Specifically, natural experiments of urban trails had to focus on the impact of high or medium comfort trails that were separated from the roadway via physical distance or a concrete barrier.

Inclusion criteria

We included studies that met the following inclusion criteria: (1) natural experiments of a new urban with a control arm/control area; (2) urban trails had to be cycling infrastructure that was protected from vehicular traffic; and (3) experiments had to report changes in either daily/weekly physical activity, rates of active transportation or cycling counts as a main outcome measure. We excluded (1) studies that did not have a control group; (2) studies that assessed a new trail that was not protected from vehicle traffic (i.e. low comfort trail) as the intervention, (3) studies that did not have a measure of PA, rates of AT/cycling or cycling or pedestrian traffic and (4) studies that were not published in English.

Outcomes of interest

The primary outcome of interest was daily PA, reported as minutes per day or week of physical activity, moderate to vigorous intensity PA or MET-mins per day or week. Secondary outcomes included rates of AT and cycling/pedestrian traffic along the trail.

Study selection

The systematic search yielded 3,936 abstracts (Table 1 appendix for results of the search strategy) that were screened independently by four reviewers (HSD, CN, IF, JL) to determine eligibility. Prior to the abstract screening, a preliminary screening of the first 30 articles yielded an agreement rate between the four reviewers of 100%. The same reviewers screened the remaining titles and abstracts, with conflicts being resolved collaboratively under the guidance of an independent reviewer (JMc). A flow chart of study selection from search to meta-analysis is provided in Fig. 1. Of the 3936 articles identified, 1583 were duplicates, and 2315 did not meet inclusion for full-text review. Of the 58 articles that were full-text screened for eligibility, four did not report an urban trail that met inclusion criteria, 16 did not have a control group, and nine did not report data for the outcomes of interest. A further five were excluded after passing the initial stages as they were either duplicate publications, did not present data correctly, improperly identified a control group or did not provide sufficient data for interpretation (i.e. reported Cohen’s D values only).

Table 1 Study characteristics for natural experiments with controlled per-post designs that Met inclusion criteria
Full size table
Fig. 1
figure 1

Flow chart of systematic selection of studies

Full size image

Data extraction

Data was extracted by one reviewer (IF) and confirmed by a second (JL). The data extracted included demographic characteristics of participants (measures of socio-economic status, age, education level, gender and race), the intervention (buffer distance that separated individuals into intervention or control conditions, length of the trail and trail type), timing of follow up measures and the outcomes of interest data. Studies varied in PA reporting methods, having both subjective and objective measures. Data extracted from self-reported PA and objective measures were extracted as they were reported in each manuscript. Corresponding authors were contacted for any missing data. Depending on the necessity of the missing information, if data were unable to be retrieved the study or data were excluded from the meta-analysis.

Risk of bias assessment

We used a modified risk of bias tool to rate the quality of the studies using the TREND reporting guidelines as a framework [11, 12]. The tool we created was categorized into 3 broad areas of potential sources of bias (study design, data analysis and report and reporting of intervention and participant characteristics). Key reporting metrics from TREND reporting guidelines, core elements of urban trail natural experiments (type of trail, length, surface type, buffer distance to categorize intervention and control areas) as well as key aspects of reporting PA-related outcomes (time of year, objective/subjective, adjusting for confounding) were used to assess the overall risk of bias. These varying metrics were reported on a binary scale with 1 for an element that was reported and 0 when it was not reported. Scores were summed and each study assigned a score out of a maximum of 35.

Statistical analyses

A DerSimonian and Laird random-effects model was used to compare standardized mean differences with 95% confidence intervals between changes in PA, treated as a continuous outcome between individuals living in areas that lived within the buffer area defined as receiving an urban trail (intervention) compared to individuals living in areas outside the buffer area (control). An additional subgroup analysis was performed to determine if the effect size was different for studies that reported a dose-response effect of the trail related to the proximity of individuals to the new trails. We planned to conduct a Mantel-Haenszel random effects odds ratio analysis to test for differences in rates of AT/cycling trips between individuals in intervention and control areas, however all but one study reported pooled, adjusted effect sizes without raw data for intervention and control areas before and after the construction of a new trail. Data were analyzed using Revman 5.4 (Cochrane Collaborative).

Results

After systemic searches of each database, 3,936 citations were included in a preliminary screen, 1,563 articles were deemed duplicates, leaving 2,373 articles to be screened (Table 1 appendix). At the end of the title and abstract screening, 2,238 papers were excluded, there was a 94.3% agreement rate between the four reviewers for the preliminary title and abstract screen and 135 were in conflict among four reviewers (IF, HSD, CN, JL) (Fig. 1). The 135 papers with conflicting decisions were resolved by the four screeners and one investigator (JMc) to achieve consensus for eligibility for full-text screening. After all conflicts were resolved, 58 papers were included in the full-text screen where 29 were deemed eligible for data extraction. Once the data extraction process began, 5 studies [23,24,25,26,27] were deemed ineligible, as they were either duplicates or did not meet inclusion criteria, either through the presentation of data, inability to contact authors for missing data, or missing control groups. This led to a final sample of 24 studies that met the inclusion criteria [28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51]. Of the 24 natural experiments studied, 11 measured PA as the primary outcome [28,29,30,31,32,33,34,35,36,37,38], 8 measured changes in cycling traffic [39,40,41,42,43,44,45,46], 5 measured the changes in AT [47,48,49,50,51] and 1 reported changes in both PA and AT [27].

Among the 11 studies that assessed PA [28,29,30,31,32,33,34,35,36,37,38] (n = 11,464) the proportion of females was 55% in the intervention arm, and 59% in the control arm. The median age was 50 years in the intervention arm (range: 10, 85), and 53 years (range: 10, 85) in the control arm. Six [28, 29, 31, 35, 37, 38] of the 24 studies reported on the ethnicity of the participants and within these studies > 75% of participants were white. For studies that assessed AT or cycling counts (n = 13)39–51 the proportion of females in the intervention arm was 49% with a mean age of 36 years (range: 10–85 years), and in the control arm, the female proportion was 47%, with a mean age of 38 years (range: 10–85 years). Only one study included outcome data for children and adolescents [47].

Intervention characteristics of the studies included in the analyses are provided in Table 1. In total, the 24 studies examined the construction of 439 km of new protected urban trails, with a median length of 41 km of new trail (range 2-102 kms). Within the 11 studies that measured changes in PA [28,29,30,31,32,33,34,35,36,37,38], the mean control group sample size was 269, with a mean intervention sample size of 329. The mean follow-up period was 3.1 years (range: 1 to 5 years). Among the 11 studies that assessed PA, only 1 used an objective measure [36]. Among the 8 studies that quantified changes in cycling traffic, 4 used objective measures (Eco-Counters, Montreal, Qc) [39, 41, 42, 45] and 4 used field observations [40, 43, 44, 46]. Among the 6 studies of AT/cycling frequency, 2 used census data [50, 51] and 4 used field surveys [27, 47,48,49] to measure outcomes. One study reported both PA and AT, using a self-reported survey.

Effectiveness of new trails on physical activity levels and cycling traffic

Among the 11 studies that provided 21 comparisons [28,29,30,31,32,33,34,35,36,37,38], PA increased in individuals who were exposed to the new trail compared to those not exposed [SMD = 0.12; 95% CI: 0.04, 0.20; I2 = 73%; n = 11,464] (Fig. 2). Restricting the analysis to individuals in closest proximity to the trail (n = 10 studies; 14 comparisons) modestly increased the effect size but reduced the precision of the estimate for adding an urban trail on individual PA [SMD = 0.14; 96% CI: 0.06 to 0.25, I2 = 74%; n = 8234; Figure s1 Appendix)]. A total of six studies [27, 47,48,49,50,51] provided 9 effect estimates with confidence intervals for the change in AT/cycling following the construction of a new trail (Table 2). Of the 9 comparisons, the mean odds ratio was 1.41 (range 1.01–3.52) and 4 comparisons from 4 studies [45, 48, 50, 51] found a positive effect of adding a new trail on rates of AT/cycling, and 5 comparisons from 3 studies [29, 49, 51] had confidence intervals that included the null. A total of eight studies [39,40,41,42,43,44,45,46] provided 10 comparisons for the effect of a new trail on changes in cycling traffic along intervention and control trails. Six comparisons from 5 studies [39,40,41, 43, 45] reported effect sizes from group by time interactions, while 4 comparisons from 3 studies [42, 44, 46] only reported pre-post effect sizes for intervention areas (Table 3). Three of six comparisons that reported a group x time interaction reported a significant increase in cycling counts. All four comparisons that only reported pre-post data, reported an increase in cycling counts.

Fig. 2
figure 2

Meta-analysis of 11 studies that reported changes in minutes of physical activity following implementation ofa new urban trail

Full size image
Table 2 Summary of studies examining effectiveness of urban trails on changes in active transportation
Full size table
Table 3 Effectiveness of urban trails on changes in cycling traffic
Full size table

Risk of bias

The overall quality of each study was evaluated on 29 different metrics (Table S2 of the appendix). Out of a maximal score of 35 for each study and the average score was 24 (range 15–28). Of the studies included in the data extraction, 0 reported intention to treat analyses, 1 of 24 studies [31] reported whether incentives were used as a form of recruitment aid, 1 study used an objective measure of PA [36], 5 of 2427,28,41,45,50 provided a power calculation, 3 of 2436,40,44 reported the surface of the new trail, and 7 of 2427,32,39,43–45 reported the preceding geography of the area. Most studies reported sources of confounding and strategies to overcome confounding bias (16/24 and 22/24, respectively). Only 12 out of 24 studies reported race for individuals surveyed, and none of those provided results disaggregated by race or sex.

Discussion

To the best of our knowledge, this is the most comprehensive systematic review and meta-analysis to date of the effects of natural experiments of urban trails on changes in individual PA levels, rates of AT and cycling traffic. The main finding of the meta-analysis was that there is a small but significant increase in PA among individuals living close to a new trail following trail implementation. These improvements in PA appear to be coupled with increased rates of AT and cycling traffic; however, data availability limited statistical comparisons. Lastly, we identified significant gaps that could be addressed in future studies including improvements in methodological design, particularly strategies to enhance causal inference and the use of objective measures to assess PA and AT outcomes. Additionally, the external validity of these studies is low due to a lack of data available for various segments of the population, particularly children and youth and individuals from structurally oppressed groups.

The expansion of cycling infrastructure is the fastest growing and most expansive change to the built environment that supports PA in cities in Asia, Europe and North America. Several research groups in Europe, China, Canada and the US have capitalized on these natural experiments to determine if they nudge citizens to engage in more PA or AT [28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51]. Two previous systematic reviews examined the pooled effects of these interventions on measures of PA for individuals living in proximity to them [7, 9]. Both studies suggested a positive effect of new trails on PA however either failed to meta-analyse results as they included observational studies without a control condition or failed to include studies of controlled natural experiments [7]. The other meta-analysis failed to include natural experiments published in transportation journals [9]. The current meta-analysis expands on these two studies with the addition of 5 natural experiments and supports previous conclusions. Specifically, with a nearly three-fold larger sample size (n = 11,343 vs. 4,081), we found an almost identical positive 12% increase in PA following the implementation of new urban trails [9]. Sub-group analyses suggest this effect is modestly higher for individuals living closer to the trail than for individuals further away. Additionally, we found that only 4 of 9 comparisons (6 studies) reported positive effects of implementing new trails on rates of AT or cycling frequency while 7 of 10 comparisons (8 studies) reported increased rates of cycling traffic following construction of a new trail. Collectively, these data provide evidence that the addition of protected spaces for cycling and walking, in the form of urban trails, is associated with modest increases in PA, AT and potentially cycling traffic for individuals living in areas 500 m-1 km from the trail.

Natural experiments provide the most rigorous methodological design to determine the causal effects of an intervention that is not suitable for a randomized control trial [6]. The credibility of the evidence emerging from a natural experiment is however dependent on the quality of the methods used to study it [52, 53]. Previous systematic reviews have identified the high risk of bias evident in the evaluations of natural experiments of urban trails [7,8,9,10]. These include (1) the risk of confounding bias if individuals in control and intervention areas are dissimilar at baseline (exchangeability assumption); (2) the risk of selection bias if all individuals do not have an equal chance of receiving the intervention (positivity assumption) and (3) the risk of spillover effects if additional interventions (simultaneous urban cycling policies) or unintended consequences of urban trail interventions (gentrification of neighbourhoods) occur in conjunction with the implementation of a new trail (consistency assumption) [54]. The majority of studies of natural experiments to date have failed to properly address these assumptions through design approaches. For example, few studies ensured baseline outcome data were similar prior to the intervention or matched control neighbourhoods with intervention neighbourhoods on baseline demographic information, or even reported demographic information for control arms [26, 35, 39, 41, 43, 44, 46, 47, 50]. In the absence of information on recruitment, it was unclear if individuals recruited were equally likely to use the trail in intervention and control areas. Lastly, few studies captured PA outcomes using valid tools, during the same season in the pre and post time periods [26, 35, 45] Accordingly, the risk of bias for these natural experiments is high and researchers should draw from the growing body of literature on the use of difference-in-difference designs [13, 52, 55] or interrupted time series with synthetic controls when designing the evaluation of future natural experiments of urban trails to more rigorously estimate causal effects of urban trails on health or behavioural outcomes.

Urban trails are not equally distributed within cities. Studies in Canada and the US reveal that investments in protected spaces for cycling and walking are often made in affluent areas with fewer individuals from racialized groups [56, 57]. Additionally, individuals from structurally oppressed groups (women, newcomers, racialized groups) are less likely to use existing urban trails [55]. The results from this systematic review add to this body of knowledge as few experimental studies recruited or document the impact of urban trails on individuals from structurally oppressed groups or other segments of the urban population, particularly children and adolescents. Exclusion of racially oppressed populations have several possible effects on the summary results presented here [58]. First, they could inflate the effect size if urban trails only meet the needs of certain racial groups [59]. Exclusion of oppressed groups could also diminish the effect size if individuals from racialized groups in some cities are more likely to rely on forms of active transportation (i.e. urban Indigenous populations in Canada). Overall however, the lack of data on the effectiveness of urban trails for increasing PA or AT among structurally oppressed limits the external validity of these findings and transferability to all urban populations [58, 59]. As cities implement policies for racial, gender and sexual equity, investigators studying natural experiments should make efforts to include diverse samples of individuals in their study designs and embrace equity and justice-informed race-based data collection [60].

The data presented here provide some insight into the design of studies focused on urban trail natural experiments. In brief, future studies should consciously consider measuring and reporting key variables outlined in the TREND statement reporting guidelines [11, 12], key assumptions for difference-in-difference designs [13], and best practices for collecting objectively measured PA [61]. With respect to the study of urban trails, future studies should clearly delineate the intervention and control areas using specific buffer areas, provide detailed descriptions of the urban trail being implemented (duration, type, width) and collect, report and balance individual- and area-level variables that could confound comparisons between intervention and control areas. Finally, researchers are encouraged to collect data from a diverse group of individuals, particularly structurally or racially oppressed populations that are often excluded from this type of research.

Strengths and limitations

To our knowledge, this is the most extensive systematic review and meta-analysis of controlled studies of natural experiments of new protected cycling/walking paths to date. The study was strengthened by the broad search strategy used to date and pre-registered methods. Despite these strengths, there are still several limitations of this review, particularly the lack of data available for meta-analysis within most published studies examining changes in rates of AT/cycling trips. Additionally, we restricted analyses to scientific papers written in English potentially influencing the effect estimates. Furthermore, there were a relatively small number of natural experiments available in the literature and most had limited sample sizes, limiting the generalizability and precision of our estimates. The risk of confounding and selection bias is high do a failure to collect and adjust for neighbourhood-level confounding. For example, most studies did not match or control for differences in demographic (age, gender, race) or environmental factors (trail connectivity, parks, fitness centres, socio-economic status) for individuals living within intervention and control areas. For example, nine studies did not provide sufficient information for participants at baseline, therefore it is unclear how comparable populations were prior to the implementation of new trail. Differences in age, activity level or health status could contribute to positive effects seen with trail implementation. The wide variation in reporting methods was also a limitation as not all the data could be pooled for meta-analysis. Few studies addressed the risk of spillover effects of gentrification of a neighbourhood following the construction of a new trail (part of the stable unit treatment value assumption [62]). Finally, the lack of data from countries outside North America, Europe, Asia and Oceania limit the generalizability of these findings to other countries with different climates, urban environments, socio-political cultures and resources. Despite these limitations, the estimates for the effects of urban trails on measures of PA and AT provided are the most robust to date. Future studies of natural experiments should avoid the methodological errors identified here to enhance the causal estimate of the effect of urban trails on PA, AT behaviours and cycling traffic.

Conclusions

The following systematic review and meta-analysis provides tentative empirical evidence that the implementation of urban trails increases levels of PA among individuals living in proximity to them. Data also suggest that new urban trails may increase cycling traffic and possibly rates of AT/cycling trips, however larger studies are needed to confirm these estimates. Finally, there is a need for future studies to use more robust epidemiological designs that address assumptions for causal inference and consider measures of equity and justice in research designs to determine if new trails are effective for increasing PA and AT for all urban residents.

Data availability

No datasets were generated or analysed during the current study.

References

  1. Mulvaney CA, Smith S, Watson MC, et al. Cycling infrastructure for reducing cycling injuries in cyclists. Cochrane Database Syst Rev Dec. 2015;10(12):CD010415. https://doiorg.publicaciones.saludcastillayleon.es/10.1002/14651858.CD010415.pub2.

    Article  Google Scholar 

  2. Pedal, Progress. Taking A Look At Europe’s Commitments to Cycling Infrastructure Post-Pandemic. 2023. https://www.welovecycling.com/wide/2023/07/27/pedal-progress-taking-a-look-at-europes-commitments-to-cycling-infrastructure-post-pandemic/

  3. Canada’s National Active Transportation Strategy. 2021–2026. 2021. https://housing-infrastructure.canada.ca/trans/nats-strat-snta-eng.html

  4. Dill J, Carr T. Bicycle commuting and facilities in major U.S. Cities: if you build them, commuters will use them. Transp Res Record: J Transp Res Board. 2023;1828(1):116–22.

    Google Scholar 

  5. Cousens S, Hargreaves J, Bonell C, et al. Alternatives to randomisation in the evaluation of public-health interventions: statistical analysis and causal inference. J Epidemiol Community Health Jul. 2011;65(7):576–81. https://doiorg.publicaciones.saludcastillayleon.es/10.1136/jech.2008.082610.

    Article  CAS  Google Scholar 

  6. Majumdar SR, Soumerai SB. The unhealthy state of health policy research. Health Aff (Millwood). Sep-Oct. 2009;28(5):w900–8. https://doiorg.publicaciones.saludcastillayleon.es/10.1377/hlthaff.28.5.w900.

    Article  Google Scholar 

  7. Stappers NEHCKD, Ettema D, Dr Vries NK, Kremers SPJ. The effect of infrastructural changes in the built environment on physical activity, active transportation and sedentary behavior– A systematic review. Health Place. 2018;53:135–49.

    CAS  PubMed  Google Scholar 

  8. Smith M, Hosking J, Woodward A, et al. Systematic literature review of built environment effects on physical activity and active transport - an update and new findings on health equity. Int J Behav Nutr Phys Act Nov. 2017;16(1):158. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12966-017-0613-9.

    Article  Google Scholar 

  9. Deng Y, Liang J, Chen Q. Greenway interventions effectively enhance physical activity levels-A systematic review with meta-analysis. Front Public Health. 2023;11:1268502. https://doiorg.publicaciones.saludcastillayleon.es/10.3389/fpubh.2023.1268502.

    Article  PubMed  PubMed Central  Google Scholar 

  10. Molenberg FJM, Panter J, Burdorf A, van Lenthe FJ. A systematic review of the effect of infrastructural interventions to promote cycling: strengthening causal inference from observational data. Int J Behav Nutr Phys Act Oct. 2019;26(1):93. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12966-019-0850-1.

    Article  Google Scholar 

  11. Haynes AB, Haukoos JS, Dimick JB. TREND reporting guidelines for Nonrandomized/Quasi-Experimental study designs. JAMA Surg Sep. 2021;1(9):879–80. https://doiorg.publicaciones.saludcastillayleon.es/10.1001/jamasurg.2021.0552.

    Article  Google Scholar 

  12. Des Jarlais DC, Lyles C, Crepaz N, Group T. Improving the reporting quality of nonrandomized evaluations of behavioral and public health interventions: the TREND statement. Am J Public Health Mar. 2004;94(3):361–6. https://doiorg.publicaciones.saludcastillayleon.es/10.2105/ajph.94.3.361.

    Article  Google Scholar 

  13. Wing C, Simon K, Bello-Gomez RA. Designing difference in difference studies: best practices for public health policy research. Annu Rev Public Health Apr. 2018;1:39:453–69. https://doiorg.publicaciones.saludcastillayleon.es/10.1146/annurev-publhealth-040617-013507.

    Article  Google Scholar 

  14. Lowe M, Adlakha D, Sallis JF, et al. City planning policies to support health and sustainability: an international comparison of policy indicators for 25 cities. Lancet Glob Health Jun. 2022;10(6):e882–94. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/S2214-109X(22)00069-9.

    Article  CAS  Google Scholar 

  15. Sallis JF, Bull F, Guthold R, et al. Progress in physical activity over the olympic quadrennium. Lancet Sep. 2016;24(10051):1325–36. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/S0140-6736(16)30581-5.

    Article  Google Scholar 

  16. Giles-Corti B, Vernez-Moudon A, Reis R, et al. City planning and population health: a global challenge. Lancet Dec 10. 2016;388(10062):2912–24. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/S0140-6736(16)30066-6.

    Article  Google Scholar 

  17. Salvo D, Kepper M, Hunter R, Jauregui A. Built environment and obesity prevention research: moving from niche to norm. Lancet Diabetes Endocrinol Sep. 2023;11(9):631–3. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/S2213-8587(23)00226-7.

    Article  Google Scholar 

  18. Miranda-Moreno L, Nosal T, Schenider RJ, Proulx F. Classification of bicycle traffic patterns in five North American cities. Transp Res Record: J Transp Res Board. 2013;2339:68–79.

    Google Scholar 

  19. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ Mar. 2021;29:372:n71. https://doiorg.publicaciones.saludcastillayleon.es/10.1136/bmj.n71.

    Article  Google Scholar 

  20. McGowan J, Sampson M, Salzwedel DM, Cogo E, Foerster V, Lefebvre C. PRESS peer review of electronic search strategies: 2015 guideline statement. J Clin Epidemiol Jul. 2016;75:40–6. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.jclinepi.2016.01.021.

    Article  Google Scholar 

  21. Winters M, Beairsto J, Ferster C, Labaree K, Manaugh K, Nelson T. The Canadian bikeway comfort and safety metrics (Can-BICS): National measures of the bicycling environment for use in research and policy. Health Rep Oct. 2022;19(10):3–13. https://doiorg.publicaciones.saludcastillayleon.es/10.25318/82-003-x202201000001-eng.

    Article  Google Scholar 

  22. Winters M, Zanotto M, Butler G. Sep. The Canadian Bikeway Comfort and Safety (Can-BICS) Classification System: A common naming convention for cycling infrastructure. Health Promot Chronic Dis Prev Can. 2020;40(9):288–293. Systeme de classification du confort et de la securite des voies cyclables canadiennes (Can-BICS): convention d’appellation commune des amenagements cyclables. https://doiorg.publicaciones.saludcastillayleon.es/10.24095/hpcdp.40.9.04

  23. Ogilvie D, Panter J, Guell C, Jones A, Mackett R, Griffin S. Health impacts of the Cambridgeshire guided busway: a natural experimental study. Public Health Res. 2016. Public Health Research; 1.

  24. Xie B, LY, Zeng Y. Casual evaluation of the effects of a large-scale greenway intervention on physical and mental health: A natural experimental study in China. Urban Forestry Urban Green. 2022;127419.

  25. Felix R, Cambra P, Moura F. Build it and give ‘em bikes, and they will come: the effects of cycling infrastructure and bike-sharing system in Lisbon. Case Stud Transp Policy. 2020;8:672–82.

    Google Scholar 

  26. Zhou Q, CM, Koh PP, Wong YD. Effects of improvements in non-motorised transport facilities on active mobility demand in a residential Township. J Transp Health. 2020;16(100834):1–14.

    CAS  Google Scholar 

  27. Xie B, Lu Y, Wu L, An Z. Dose-response effect of a large-scale greenway intervention on physical activities: the first natural experimental study in China. Health Place Jan. 2021;67:102502. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.healthplace.2020.102502.

    Article  Google Scholar 

  28. Aldred R, Croft J, Goodman A. Impacts of an active travel intervention with a cycling focus in a suburban context: One-year findings from an evaluation of London’s in-progress mini-Hollands programme. Transp Res Part A. 2019;123:147–69.

    Google Scholar 

  29. Aldred R, Woodcock J, Goodman A. Major investment in active travel in outer London: impacts on travel behaviour, physical activity, and health. J Transp Health. 2021;20(100958):1–20.

    Google Scholar 

  30. Crane M, Rissel C, Standen C, Ellison A, Ellison R, Wen LM, Greaves S. Longitudinal evaluation of travel and health outcomes in relation to new bicycle infrastructure, Sydney, Australia. J Transp Health. 2017;6:386–95.

    Google Scholar 

  31. Frank LD, Hong A, Ngo VD. Causal evaluation of urban greenway retrofit: A longitudinal study on physical activity and sedentary behavior. Prev Med Jun. 2019;123:109–16. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.ypmed.2019.01.011.

    Article  Google Scholar 

  32. He D, Lu Y, Xie B, Helbich M. Large-scale greenway intervention promotes walking behaviors: A natural experiment in China. Transp Res Part D. 2021;101(103095):1–12.

    Google Scholar 

  33. He D, Lu Y, Xie B, Helbich M. How greenway exposure reduces body weight: A natural experiment in China. Landsc Urban Plann. 2022;226(104502):1–9.

    Google Scholar 

  34. Hunter RF, Adlakha D, Cardwell C, et al. Investigating the physical activity, health, wellbeing, social and environmental effects of a new urban greenway: a natural experiment (the PARC study). Int J Behav Nutr Phys Act Oct. 2021;30(1):142. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12966-021-01213-9.

    Article  Google Scholar 

  35. Pazin J, Garcia LM, Florindo AA, et al. Effects of a new walking and cycling route on leisure-time physical activity of Brazilian adults: A longitudinal quasi-experiment. Health Place May. 2016;39:18–25. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.healthplace.2016.02.005.

    Article  Google Scholar 

  36. Stappers NEH, Schipperijn J, Kremers SPJ, et al. Tunneling a crosstown highway: a natural experiment testing the longitudinal effect on physical activity and active transport. Int J Behav Nutr Phys Act Aug. 2021;26(1):111. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12966-021-01180-1.

    Article  Google Scholar 

  37. West ST, Shores KA. The impacts of Building a greenway on proximate residents’ physical activity. J Phys Act Health Nov. 2011;8(8):1092–7. https://doiorg.publicaciones.saludcastillayleon.es/10.1123/jpah.8.8.1092.

    Article  Google Scholar 

  38. West ST, Shores KA. Does Building a greenway promote physical activity among proximate residents? J Phys Act Health Jan. 2015;12(1):52–7. https://doiorg.publicaciones.saludcastillayleon.es/10.1123/jpah.2012-0411.

    Article  Google Scholar 

  39. Auchincloss AH, Michael YL, Kuder JF, Shi J, Khan S, Ballester LS. Changes in physical activity after Building a greenway in a disadvantaged urban community: A natural experiment. Prev Med Rep Sep. 2019;15:100941. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.pmedr.2019.100941.

    Article  Google Scholar 

  40. Fitzhugh EC, Bassett DR Jr., Evans MF. Urban trails and physical activity: a natural experiment. Am J Prev Med Sep. 2010;39(3):259–62. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.amepre.2010.05.010.

    Article  Google Scholar 

  41. Garber MD, Flanders WD, Watkins KE, Lobelo F, Kramer MR, McCullough LE. Have Paved Trails and Protected Bike Lanes Led to More Bicycling in Atlanta? A Generalized Synthetic-Control Analysis. Epidemiology. Jul 1. 2022;33(4):493–504. https://doiorg.publicaciones.saludcastillayleon.es/10.1097/EDE.0000000000001483

  42. Hans S, Bredahl JK, Elizabeth VS, Nielssen TAS, Simon R. Effects of upgrading to cycle highways - An analysis of demand induction, use patterns and satisfaction before and after. J Transp Geogr. 2017;64:203–10.

    Google Scholar 

  43. Heesch KC, JB, Washington TL, Zuniga K, Burke M. Evaluation of the Veloway1: A natural experiment of new bicycle infrastructure in Brisbane, Australia. J Transp Health. 2016;3:366–76.

    Google Scholar 

  44. Nguyen P, Koh PP, Wong YD. Impacts of bicycle infrastructure: a case study in Singapore. Municipal Eng. 2015;168:186–98.

    Google Scholar 

  45. Xiao CS, Sharp SJ, van Sluijs EMF, Ogilvie D, Panter J. Impacts of new cycle infrastructure on cycling levels in two French cities: an interrupted time series analysis. Int J Behav Nutr Phys Act Jul. 2022;7(1):77. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12966-022-01313-0.

    Article  Google Scholar 

  46. Rissel C, Greaves S, Wen LM, Crane M, Standen C. Use of and short-term impacts of new cycling infrastructure in inner-Sydney, Australia: a quasi-experimental design. Int J Behav Nutr Phys Act Oct. 2015;6:12:129. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12966-015-0294-1.

    Article  Google Scholar 

  47. Keall M, Chapman R, Howden-Chapman P, Witten K, Abrahamse W, Woodward A. Increasing active travel: results of a quasi-experimental study of an intervention to encourage walking and cycling. J Epidemiol Community Health Dec. 2015;69(12):1184–90. https://doiorg.publicaciones.saludcastillayleon.es/10.1136/jech-2015-205466.

    Article  Google Scholar 

  48. Frank LD, HA, Ngo VD. Build it and they will cycle: causal evidence from the downtown Vancouver comox greenway. Transp Policy. 2021;105:1–11.

    Google Scholar 

  49. Brown BB, Smith KR, Tharp D, et al. A complete street intervention for walking to transit, nontransit walking, and bicycling: A Quasi-Experimental demonstration of increased use. J Phys Act Health Nov. 2016;13(11):1210–9. https://doiorg.publicaciones.saludcastillayleon.es/10.1123/jpah.2016-0066.

    Article  Google Scholar 

  50. Goodman A, Panter J, Sharp SJ, Ogilvie D. Effectiveness and equity impacts of town-wide cycling initiatives in England: a longitudinal, controlled natural experimental study. Soc Sci Med Nov. 2013;97:228–37. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.socscimed.2013.08.030.

    Article  Google Scholar 

  51. Patterson R, Ogilvie D, Laverty AA, Panter J. Equity impacts of cycling investment in England: A natural experimental study using longitudinally linked individual-level census data. SSM Popul Health Sep. 2023;23:101438. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.ssmph.2023.101438.

    Article  Google Scholar 

  52. Caniglia EC, Murray EJ. Difference-in-Difference in the time of cholera: a gentle introduction for epidemiologists. Curr Epidemiol Rep Dec. 2020;7(4):203–11. https://doiorg.publicaciones.saludcastillayleon.es/10.1007/s40471-020-00245-2.

    Article  Google Scholar 

  53. Tchetgen Tchetgen EJ, Park C, Richardson DB. Universal Difference-in-Differences for Causal Inference in Epidemiology. Epidemiology. Jan 1. 2024;35(1):16–22. https://doiorg.publicaciones.saludcastillayleon.es/10.1097/EDE.0000000000001676

  54. Hernan MA, Methods of Public Health Research - Strengthening Causal Inference from Observational Data. N Engl J Med Oct. 2021;7(15):1345–8. https://doiorg.publicaciones.saludcastillayleon.es/10.1056/NEJMp2113319.

    Article  Google Scholar 

  55. McGavock J, Hobin E, Prior HJ, et al. Multi-use physical activity trails in an urban setting and cardiovascular disease: a difference-in-differences analysis of a natural experiment in Winnipeg, Manitoba, Canada. Int J Behav Nutr Phys Act Mar. 2022;28(1):34. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12966-022-01279-z.

    Article  Google Scholar 

  56. Zhao Q, Winters M, Nelson T, Laberee K, Ferseter C, Manaugh K. Who has access to cycling infrastructure in Canada? A social equity analysis. Comput Environ Urban Syst. 2024;110:102109.

    Google Scholar 

  57. Cunha I, SC. Equity impacts of cycling: examining the spatial-social distribution of bicycle-related benefits. Int J Sustainable Transp. 2023;17(6):573–91.

    Google Scholar 

  58. Gilmore-Bykovskyi A, Jackson JD, Wilkins CH. The urgency of justice in research: beyond COVID-19. Trends Mol Med Feb. 2021;27(2):97–100. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.molmed.2020.11.004.

    Article  CAS  Google Scholar 

  59. Perry MJ, Arrington S, Freisthler MS, et al. Pervasive structural racism in environmental epidemiology. Environ Health Nov. 2021;17(1):119. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12940-021-00801-3.

    Article  Google Scholar 

  60. Breathett K, Spatz ES, Kramer DB, et al. The groundwater of Racial and ethnic disparities research: A statement from circulation: cardiovascular quality and outcomes. Circ Cardiovasc Qual Outcomes Feb. 2021;14(2):e007868. https://doiorg.publicaciones.saludcastillayleon.es/10.1161/CIRCOUTCOMES.121.007868.

    Article  Google Scholar 

  61. Pedisic Z, Bauman A. Accelerometer-based measures in physical activity surveillance: current practices and issues. Br J Sports Med. Feb 2015;49(4):219–23. https://doiorg.publicaciones.saludcastillayleon.es/10.1136/bjsports-2013-093407.

  62. Rubin J. Formal mode of statistical inference for causal effects. J Stat Plann Inference. 1990;25(3):279–92.

    Google Scholar 

Download references

Acknowledgements

Not applicable.

Funding

Funding for this project was provided by operating grants from the Heart and Stroke Foundation of Canada (G-17-0018638) and the Canadian Institutes of Health Research (PJT-153449; CPP-137910).

Author information

Authors and Affiliations

  1. Diabetes Research Envisioned and Accomplished in Manitoba (DREAM) Research Theme, Children’s Hospital Research Institute of Manitoba, Winnipeg, Canada

    Isaak Fast, Christie Nashed, Jack Lotscher, Hannah Steiman De Visser & Jonathan McGavock

  2. Faculty of Kinesiology, University of Manitoba, Winnipeg, Canada

    Isaak Fast, Jack Lotscher & Jonathan McGavock

  3. School of Medical Sciences, University of Manchester, Manchester, UK

    Christie Nashed

  4. School of Human Kinetics, Faculty of Health Sciences, University of Ottawa, Ottawa, Canada

    Jack Lotscher

  5. Neil John MacLean Library, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Canada

    Nicole Askin

  6. College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Canada

    Hannah Steiman De Visser & Jonathan McGavock

  7. Department of Pediatrics and Child Health, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Canada

    Jonathan McGavock

Authors
  1. Isaak Fast
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Christie Nashed
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Jack Lotscher
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Nicole Askin
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Hannah Steiman De Visser
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Jonathan McGavock
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

All authors contributed to the study and manuscript in alignment with current ICMJE guidelines. The study was conceived by JM. JM is the principal investigator on the original funded grant. NA designed and conducted the systematic search for manuscripts. JM, IF, JL and HSD participated in designing the study and submitting the original grant. JM and CN are involved in data cleaning and verification and conducted the statistical analyses. All authors contributed to the writing of the final document. CN, IF and JM drafted the original manuscript. All authors contributed to critically revising the manuscript for important intellectual content, gave their final approval and agreed to be accountable for all aspects of the work, and they will participate in future interpretation of the data and drafting of further manuscripts arising from this work. JM takes full responsibility for the work, conduct of the study, had access to the data and controlled the decision to publish.

Corresponding author

Correspondence to Jonathan McGavock.

Ethics declarations

Ethics approval and consent to participate

This study did not require ethics approval or consent for participation.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Role of funders/sponsors

Funding bodies were not involved in the study design, conduct, interpretation or manuscript preparation for this project. Scientists involved in this study had no relationship with funding agencies and conducted the study independent of funders.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Fast, I., Nashed, C., Lotscher, J. et al. The effectiveness of new urban trail infrastructure on physical activity and active transportation: a systematic review and meta-analysis of natural experiments. Int J Behav Nutr Phys Act 22, 36 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12966-025-01729-4

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12966-025-01729-4

Keywords

International Journal of Behavioral Nutrition and Physical Activity

ISSN: 1479-5868

Contact us