Skip to main content

Table 1 Characteristics and key findings of studies included in review

From: Systematic literature review of built environment effects on physical activity and active transport – an update and new findings on health equity

Author; year of publication; study location

Population description

Study design

Intervention description

Intervention delivery

Physical activity outcome

Key findings

Beenackers; 2012; Perth, Australia [70]

1427 adults participating in the RESIDE study - movers to one of 74 new residential housing developments in Perth

Uncontrolled longitudinal natural experiment

Change in objectively assessed built environment from original household to new residence. Objective neighbourhood variables were constructed in GIS using a 1600-m network service area buffer around the residential address. Street connectivity, residential density, land-use mix, and number of destinations relevant for transport or recreation were calculated.

Mechanism: moving from one neighbourhood to a new development. Medium: Liveable neighbourhood design, focusing on integrated mixed use and interconnected network.

Cycling for transport or recreation

After full adjustment increased residential density was associated with increased odds of taking up cycling for transport (OR 1.54, 95% CI 1.04, 2.26, p = 0.03), and increased street connectivity was associated with increased odds of taking up recreational cycling (OR 1.20, 95% CI 1.06, 1.35 p < 0.01).

Boarnet; 2005; California, USA [85]

Children attending one of nine intervention schools (3rd-5th grade), n = 1778 at baseline

Uncontrolled cross-sectional pretest-posttest

Types of intervention treatments investigated: sidewalk improvement projects (e.g., install pathway and signage; install sidewalk gap closures; install sidewalk, curb, gutter); traffic signal improvements (install traffic signal to replace four-way stop sign); crosswalk and crosswalk signal improvements (install in-pavement crosswalk signal system; install pedestrian-activated flashing warning system; add pedestrian count-down signals; install crosswalk and crosswalk signs); Install bike lanes.

Mechanism: school-neighbourhood infrastructural interventions. Medium: sidewalk improvements, traffic signal improvements, crosswalk improvements, bicycle path improvements.

Walking and cycling to and from school

Sidewalk, traffic signal, and crosswalk improvements were associated with greater increases in walking for children who passed the intervention on the way to school compared with those who did not (differences of 11.59%, t = 3.01; 15.63%, t = 2.43, and 21.16%, t = 3.15 across the three areas where gap closures were implemented; 11.34%, t = 2.12 for new sidewalks plus sidewalk gap closures; 6.67%, t = 1.04 for new pathway of decomposed granite; 10.94%, t = 2.80 and 14.43%, t = 2.52 for traffic signals replacing 4-way stop sign; 3.13%, t = 1.02 and 4.31%, t = 2.52 for in-pavement crosswalk lighting; and 13.33%, t = 1.85 for pedestrian count-down signals). There was insufficient evidence to determine effectiveness of the one bicycle path improvement project.

Brown; 2008; Salt Lake City, Utah, USA [86]

102 adult residents of households within one half mile of a new light rail stop

Uncontrolled longitudinal pretest-posttest

Intervention was a new light rail stop on the Salt Lake City TRAX line.

Mechanism: infrastructural interventions. Medium: increase in public transport accessibility

Moderate intensity activity bouts lasting at least 8 min per valid hour; number of leisure walks in last 2 weeks

Adjusted for income and employment, respondents who were rail users at baseline and follow-up had the largest number of moderate physical activity bouts (mean [SD] 3.68 [0.60] compared with 1.77 [0.83] for new riders at followup, and 1.07 [0.76] for non rail users at baseline and followup, F for ridership group effects = 3.89, p = 0.03). No difference observed between groups for leisure walks.

Clark; 2014; Southern Nevada, USA [61]

All users of intervention and comparison trails, n = 191 at baseline

Quasi-experimental cross sectional pretest-posttest

Intervention: 6 community trails which included a marketing campaign promoting trail use and the addition of way-finding and incremental distance signage to selected trails (October 2011–October 2012). The distance markings were embossed into the surface of the trails at 0.25 mile intervals. Way-finding signs were placed on the trails at major access points, as suggested by the local jurisdictions, and were mounted on square metal posts. Each side of the post was marked with a trail map, the name of the trail, the logo of the responsible jurisdiction, and icons for acceptable and unacceptable uses.

Mechanism: infrastructural interventions. Medium: presence of additional features.

Counts of trail users

Increases in mean users per hour were found for both the intervention and control trails (increased by 31%, p < 0.01 and 35%, p < 0.01, respectively). Increases did not vary between the treatment groups (Z = 0.9892, p = 0.32).

Cohen; 2012; Los Angeles, California, USA [57]

Users of 12 parks receiving outdoor exercise equipment interventions, and users of ten comparison parks, n = 7105 observed and n = 742 surveyed at baseline

Controlled cross-sectional pretest-posttest

“Fitness Zones” were implemented in 12 parks - comprising durable, weather and vandal resistant exercise equipment for strength training and aerobic exercise. Eight pieces of equipment were installed in each park

Mechanism: infrastructural interventions. Medium: provision of new equipment

Number of exercise sessions per week, active transport to park, METs

Intervention respondents reported engaging in more exercise sessions per week than comparison park respondents (2.9 versus 2.7; p < 0.0001) and getting to the park more often by walking than comparison park respondents (56% versus 35%, p = 0.002). A net gain of 1909 METs was observed in the 12 intervention parks, but this increase was not significant.

Cohen; 2014; Los Angeles, California, USA [87]

Users of pocket and comparison parks; 392 adults aged 18 years or older residing in houses within 0.5 miles of control and intervention parks

Controlled cross-sectional pretest-posttest

Three pocket parks were developed, two in previously vacant lots and the third in a former community garden site. All three had playground equipment and benches installed, and a walking path was developed around the largest park. All were fenced and enclosed by lockable gates.

Mechanism: development of new activity setting. Medium: retrofitting of existing green spaces to park settings, including installation of playground equipment

METs, MVPA, park visitation, leisure time exercise

At follow up, 25% of park users were engaged in MVPA in pocket parks compared with 36–41% in comparison parks. Estimated METs during observation times at follow up for pocket parks was 324, compared with 280–526 in comparison parks. Survey results showed a threefold increase in visits to any park at least once a week for residents living in pocket park areas (11.1% at baseline versus 33.9% at followup, p < 0.0001), an increase in engaging in leisure time exercise (25.8% versus 35.7%, p = 0.0025), and an increase in exercising in the park (9.6% versus 14.4%, p = 0.0395).

Cohen; 2015; San Francisco, USA [88]

Users of control and intervention parks; 922 residents of houses within 0.5 miles of control and intervention parks

Controlled cross-sectional pretest-posttest

Two parks underwent extensive renovations including installation of new play equipment, landscaping, and ground surfaces. One had added adult outdoor fitness equipment and a new recreation centre. Two parks were partially renovated and intervention components were not described.

Mechanism: infrastructural interventions. Medium: provision of new equipment and improved aesthetics and surfaces

Park use and METs expended in the park; frequency of park visitation; weekly exercise sessions

Increases in observed park use and METs occurred for the two parks with completed renovations only (233% and 255% respectively, both p < 0.001) versus decreases of 49% and 53% respectively in the control parks both p < 0.001), and no significant change for the parks under construction. Self-reported frequency of visitation increased for the control and renovated parks range 0.22–0.73, p < 0.01), and a significant decrease in visitation was reported for partially renovated parks (by approximately half, p < 0.001). The number of weekly exercise sessions increased for survey respondents in the partially renovated park areas only (0.29, p < 0.01), and decreased for respondents of park intercept surveys in renovated parks only (−0.42, p < 0.001)

D’Haese; 2015; Ghent (Flanders), Belgium [66]

167 children aged 6–12 years and their parents

Controlled longitudinal pretest-posttest

Twelve Play Street projects that lasted at least seven consecutive days were included in this study. A Play Street is organized between 1400 and 1900. The city council of Ghent also offers a box with play equipment that can be hired for free by the volunteers of the Play Streets. They can keep the box during the period of the Play Street intervention.

Mechanism: At least three volunteers living in the street have to sign an agreement with the city council to hold responsibility for the organization of the Play Streets. They are the contact persons between the city council and the other street inhabitants. The task of the volunteers is to inform the street inhabitants about the rules and timing of the Play Streets. The fences and traffic signs are delivered by the city council. Parents remain responsible for their children playing in the street. Medium: Each day the Play Street is organized, the volunteers enclose the Play Street with fences and a traffic sign, indicating that car traffic is forbidden in the streets. The council also offers a box with play equipment that can be hired for free and kept during the intervention period. Other larger materials are also available for free hire for one day. Inhabitants may also organise their own activities.

MVPA

Positive intervention effect was found for MVPA (β = 0.82 ± 0.43;χ2 = 3.6; p = 0.06). Between 1400 and 1900, MVPA of children living in Play Streets increased from 27 min during normal conditions to 36 min during the Play Street intervention, whereas control children’s MVPA decreased from 27 to 24 min. The intervention effects on MVPA (β = −0.62 ± 0.25; χ2 = 6.3; p = 0.01) remained significant when the effects were investigated during the entire day, indicating that children did not compensate for their increased MVPA during the rest ofthe day.

Dill; 2014; Portland, Oregon, USA [62]

490 adults residing within 1000 ft of intervention or control streets

Controlled longitudinal pretest-posttest

Bicycle boulevard installation in eight street segments (0.9–4.2 miles long).

Mechanism: Improved infrastructure for cycling. Medium: Installation of bicycle boulevards (0.9–4.2 miles long).

MVPA, active transportation

A significant decrease in number of minutes biked was observed in the intervention neighbourhood (where biking occurred for at least 10 min; β −0.09, p < 0.01). No other differences were observed for MVPA, walking, or biking (number of bike trips, made a bike trip, made a bike trip greater than 10 min duration).

Fitzhugh; 2010; Knoxville, Tennessee, USA [89]

Children, adolescents, and adults in free-living conditions within one experimental and two control neighbourhoods

Controlled cross-sectional, pretest-posttest

Retrofıtting a neighbourhood with an 8-ft-wide and 2.9-mile-long asphalt urban greenway/trail to connect the pedestrian infrastructure with nearby retail establishments and schools.

Mechanism: transport and recreation setting infrastructural interventions. Medium: new urban trail for walking and cycling.

Physical activity and active transport measured in the neigbhourhood and around schools (for school travel behaviours)

Total physical activity increased in the experimental neighbourhood only (median increase of 8 counts at follow-up, p < 0.001), and was greater in the experimental neighbourhood compared with the control neighbourhood at followup (median difference of 12 counts, p = 0.028). Changes in neighbourhood level walking (p < 0.001) and cycling (p = 0.038) were greater in the experimental neighbourhood than the control neihbourhood. There was no difference over time for the experimental neighbourhood for active transport to school, or for the difference in change in active transport to school between the experimental and control neighbourhoods (p = 0.2061).

Giles-Corti; 2013; Perth, Australia [69]

1808 adults (mean age 40.7 years) participating in the RESIDE study - movers to one of 74 new residential housing developments in Perth

Uncontrolled, longitudinal, repeated measures

At each time point, objective built environment measures were generated using geographic information systems. These measures included (standardized) neighbourhood walkability measures, such as street connectivity, residential density, and land-use mix, that were calculated for the areas accessible along the street network within 1600 m from the participants’ home.

Mechanism: Neighbourhood walkability. Medium: Street connectivity, residential density, land-use mix, number of bus stops, presence of railway station, number of types of services, number of types of convenience stores, number of types of public open spaces

Walking for transport or recreation

In fully adjusted models, transport-related walking increased by 5.8 min per week for each type of transport-related destination that increased (p = 0.045) and recreational walking increased by 17.6 min per week for each type of recreational destination that increased (p = 0.070). The association between the built environment and recreational walking was partially mediated by changes in perceived neighbourhood attractiveness: when changes in ‘enjoyment’ and ‘attitude’ towards local walking were removed from the multivariate model, recreational walking returned to 20.1 min/week (p < 0.040) for each type of recreational destination that increased.

Goodman; 2014; Cardiff, Kenilworth, and Southampton, UK [56]

3516 adults aged 18–89 residing within 5 km of the core Connect2 intervention

Uncontrolled, longitudinal, repeated measures

Cardiff: a traffic-free bridge was built over Cardiff Bay; Kenliworth: a traffic-free bridge was built over a busy trunk road; Southampton: informal riverside footpath was turned in to a boardwalk.

Mechanism: Major infrastructural interventions for improved walking and cycling. Medium: construction of bridges and boardwalks for non-motorised traffic

Use of the new infrastructure; self-reported walking or cycling for transport or recreation; self-reported time spent in moderate intensity leisure time physical activity and vigorous intensity leisure time physical activity

Overall 32% and 38% of participants reported using the infrastructure at years 1 and 2, respectively. Little evidence was found for whether proximity to the intervention predicted changes in the activity levels of residents at the year 1 follow-up. At year 2, individuals living closer to the intervention reported significant increases in walking and cycling relative to those living farther away (an effect of 15.3 min per week per km closer to the intervention; 95% CI 6.5, 24.2 min per week). Total recreational physical activity increased more for individuals living closer to the intervention than those living further away at 2 years (β = 12.5, 95% CI 1.9, 23.1).

Gustat; 2012; New Orleans, USA [63]

499 adults aged 18–70 residing in low-income, primarily African American neighbourhoods receiving infrastructural interventions and two comparison neighbourhoods

Controlled cross-sectional pretest-posttest

Path intervention neighbourhood: addition of an 8-ft-wide path of 6 blocks on a grassy, tree-filled median of a wide neighbourhood boulevard. The path connected a park outside the intervention area to a commercial corridor. Playground intervention: installation of a playground on the back lot of a local elementary school. The project paid supervisors to keep the fenced playground open after school hours and on weekends from summer 2007 through spring 2009.

Mechanism: site-specific infrastructural interventions. Medium: major streetscape improvements for walking and cycling and installation of a new playground accessible to the community outside school hours

Walking for transportation or leisure; MVPA

No difference in walking for transportation or leisure between control and intervention. Increase in proportion of people engaged in MVPA and vigorous activity only in the intervention area with the new path compared with the two control neighbourhoods (increase to 41% at followup compared with decreases to 24% and 38%, p < 0.001 for MVPA and increase of 3.2% compared with decreases in all other neighbourhoods, p < 0.001 for vigorous physical activity)

Harduar-Morano; 2008; Florida, USA [90]

243 adult residents of households in the intervention community, aged 18–64 years

Retrospective, uncontrolled

Community-wide improvements, including activity-specific improvements (streetlights installed and creating safe places to walk and exercise outdoors).

Mechanism: community-wide infrastructural interventions. Medium: installation of streetlights, creating safe places to walk and exercise outdoors

Exercising outside more than 2 years ago

Sixty-three percent of survey respondents indicated that improvements were made on streetlight installation and the creation of safe places to walk and exercise outdoors, which are both community improvement issues related to outdoor activities. Of these respondents, 95% provided a positive response when asked if they spent more time exercising outside than two years ago. When the issues related to outdoor activities were examined separately, respondents were more likely to exercise outdoors now compared to two years ago if they also indicated that the following issues had improved (vs. no change): installation of streetlights (OR = 18.33; 95% CI = 4.22, 79.64) or the creation of safe places to walk or exercise outdoors (OR = 14.18; 95% CI = 3.94, 51.03). Overall, 80% agreed they spent more time exercising outside than 2 years ago “because of the newly installed streetlights, park exercise equipment, and sidewalks”

Knuiman; 2014; Perth, Australia [71]

1813 adults (average age 39 years) participating in the RESIDE study - movers to one of 74 new residential housing developments in Perth

Uncontrolled, longitudinal, repeated measures.

At each time point, objective built environment measures were generated using geographic information systems. These measures included (standardized) neighbourhood walkability measures, such as street connectivity, residential density, and land-use mix, that were calculated for the areas accessible along the street network within 1600 m from the participants’ home.

Mechanism: Neighbourhood walkability. Medium: Street connectivity, residential density, land-use mix, number of bus stops, presence of railway station, number of types of services, number of types of convenience stores, number of types of public open spaces

Walking for transport

Both the connectivity and land-use mix walkability components (but not residential density), as well as neighbourhood access to public transit, were significantly related to transport walking in the neighbourhood in the fully adjusted model with fixed effects. Connectivity had an estimated subject-level OR of 1.13 (95% CI 1.01, 1.26), land use mix had an estimated subject-level OR of 1.33 (95% CI 1.16, 1.52). Participants who had 30 or more bus stops within 1600 m of their homes had odds of walking for transportation that were approximately double those of participants who had 0–14 bus stops, and the presence of a train station within 1600 m increased the odds of walking for transportation by approximately 50%. The objectively measured number of types of destinations was significantly related to walking for transportation in a dose-response manner (p for trend <0.05); however, when categorized into levels, the ORs comparing neighbourhoods with 8–15 destinations types within 1600 m to those with 0–3 was only approximately 1.3 (p = 0.04).

Lott; 1979; California, USA [91]

3364 cyclists in the intervention street and two parallel control streets

Controlled cross-sectional, pretest-posttest

Intervention involved installing bicycle lanes on both sides of the street. Control streets had existing bicycle lanes, but no description of these provided.

Mechanism: Improved facilities for cycling. Medium: Addition of new bicycle lanes.

Cycling

Significant increases in cyclists were observed in the intervention street compared with the control street, for adults aged 25 years and older only (x2 = 3.20, p = 0.08).

McDonald; 2013; Eugene, Oregon, USA [64]

Children attending one of 9 intervention or 5 control schools (K-8).

Controlled cross-sectional, uneven repeated measures

Groups: Education/Encouragement only: Education and crosswalks/sidewalks; Education and boltage; Education and covered bike parking; Education and covered bike parking and (crosswalks/sidewalks or Boltage). Boltage is a program which encourages biking by tracking frequency of biking to school and offering prizes based on participation.

Mechanism: school-neighbourhood infrastructural interventions. Medium: provision of improved and covered bike parking, crosswalk and sidewalk improvements.

Walking or cycling to school

Augmenting education programs with additional infrastructural improvements was associated with increases in walking and biking (only for education plus covered bike parking) of 4–18 percentage points above that gained from education/encouragement only (p < 0.05). Changes for education only were significant for proportion biking (β = 0.699, marginal effect [discrete change in outcome for each observation averaged over the sample, weighted by number of students surveyed] = 0.050 95% CI 0.019, 0.080) but not walking. Changes for education plus covered bike parking were β = 1.094, marginal effect = 0.188, 95% CI 0.095, 0.281 for walking, and β = 1.211, marginal effect 0.106, 95% CI 0.018, 0.195.

Morrison; 2004; Glasgow, Scotland [92]

244 adult residents of households within the intervention neighbourhood aged 15+ years

Retrospective (uncontrolled longitudinal pretest-posttest but physical activity outcome was retrospective report at follow-up)

The traffic calming scheme was built in the main road bisecting a deprived urban housing estate in Glasgow, Scotland. The scheme comprised five sets of speed cushions (raised platforms on the road to slow car drivers), two zebra crossings with adjacent railings, and creation of parking bays.

Mechanism: infrastructural interventions. Medium: presence of traffic calming features.

Walking or cycling in the area more

After the introduction of the traffic calming Scheme 20% (95% CI 14.1, 25.9) of respondents said that they walked in the area more as a result of it, and 3.8% said they cycled in the area more because of the scheme (95% CI 0.8, 6.8). With the exception of pensioners on one stretch of the road, the pedestrian count recorded substantial increases at most sites and in most age groups after the traffic calming scheme was built.

Parker; 2011; New Orleans, USA [58]

All users of a newly constructed bike lane

Uncontrolled cross sectional pretest-posttest

A 3.1 mile dedicated bike lane was installed. Bike lanes were striped on both sides of the road and were 5 ft wide.

Mechanism: New infrastructure for cycling. Medium: installation of dedicated bike lanes

Number of cyclists

A 57% increase in the average number of riders per day was observed; there was a 133% increase among adult female riders, and a 44% increase in adult male riders (all p < 0.001).

Parker; 2013; New Orleans, USA [93]

All users of a newly constructed bike lane and cyclists in two adjacent streets

Uncontrolled cross sectional pretest-posttest

A 1 mile dedicated bike lane was installed. Bike lanes were striped on both sides of the road and were 5 ft wide.

Mechanism: New infrastructure for cycling. Medium: installation of dedicated bike lanes

Number of cyclists

Mean number of cyclists increased significantly on the intervention street (average of 79 cyclists at baseline; 257 at follow-up), and decreased on the side streets (average of 54 at baseline; 36 at follow-up; location by time interaction Z = 24.27, p < 0.001). The increase in cyclists was greater among females than males.

Quigg; 2011; Dunedin, New Zealand [65]

184 children aged 5–10 years attending schools within the intervention and control neighbourhoods.

Controlled, longitudinal pretest-posttest

Playground upgrades in two playgrounds in the intervention community. In one playground, ten new components (including play equipment, seating, safety surfacing) were installed, and two existing components were removed. At the second playground, two new play equipment pieces were installed, and a small modification was made to another piece of equipment.

Mechanism: Improved quality of neighbourhood playgrounds. Medium: Addition of equipment, seating, and safety surfacing, removal or modification of existing equipment.

Total daily physical activity

Change in physical activity was associated with an interaction between body size (body mass index z-score) and intervention status (p = 0.006) whereby greater increases in activity were observed in the intervention neighbourhood for children with lower body mass index scores, but lower levels of activity for children with higher body mass index scores.

Ranchod; 2013; Multiple sites across USA [59]

6191 adult participants aged 45–84 from the Multi-Ethnic Study of Atherosclerosis, aged 45–84 years and free of clinical cardiovascular disease at baseline

Uncontrolled, longitudinal, repeated measures.

At each time point, an objective measure of recreation facility density was calculated using geographic information systems

Mechanism: Neighbourhood recreation facility density. Medium: Density of recreational facilities in individual participants’ neighbourhoods

Recreational physical activity

A greater increase in recreational density was associated with a less pronounced decline in physical activity (mean difference in annual change in physical activity for each 1-unit increase in recreational density over time = 10.3 (95% CI 0.7, 19.9). The association was stronger in older adults.

Roemmich; 2014; Grand Forks, North Dakota, USA [94]

245 users of the intervention setting (playground within a park) except teenagers aged 13–18. Classified as child (0–12 years), or adult (19+ years)

Uncontrolled cross-sectional pretest-posttest

Removal of seating around a playground. A-B-A design, where seating was reinstalled after the intervention period.

Mechanism: Community park infrastructural interventions. Medium: removal of seating around a playground

MVPA, park visitation

Physical activity intensity for children, adults, and families (combined children and adults) was significantly greater (p < 0.02) during the intervention period than for the pre or post measurements. In the first study The ORs of adults being in MVPA rather than sitting during the intervention period compared with the pre measurement and post measurement were 4.1 (95% CI 1.1, 15.1, p < 0.03) and 22.7 (95% CI 4.2, 122.0, p < 0.001), respectively. For children, the odds of being in MVPA rather than sitting were not associated with study condition (p ≥ 0.45). Significant differences in visitation were observed, with the intervention condition associated with lower visitation than other conditions (p < 0.0001)

Shu; 2014; Santa Monica, California, USA [95]

Users of Ocean Park Boulevard (intervention site)

Uncontrolled cross sectional pretest-posttest

Major street upgrade on a 1 km section of Ocean Park Boulevard. The segment of urban residential roadway originally featured raised sidewalks, dedicated on-street parking and bicycle lanes, and one vehicle lane in each direction. The intervention involved widening sidewalks, adding street furniture and over 100 new trees; improving existing crosswalks and bicycle lanes with more clearly marked flashing beacons and asphalt painting; raising and adding trees to the centre median; adding 75 pedestrian-scaled light poles were added; and many other improvements were made (e.g., storm-water management).

Mechanism: street-level infrastructural interventions. Medium: major streetscape improvements for walking and cycling.

Traffic volume of pedestrians and cyclists

The number of pedestrians increased by 37% compared to pre-retrofit conditions and the number of cyclists remained approximately the same. Except for the weekday morning session, all session-average increases were significant (p < 0.05).

Tester; 2009; San Francisco, CA, USA [96]

All users of two intervention parks and one control park setting, n = 1006 at baseline

Uncontrolled cross-sectional pretest-posttest

Artificial turf replaced uneven dirt fields, and new fencing, landscaping, lighting, and picnic benches were added. One park additionally had ‘soft’ programme initiatives such as professional training and skills development for park and recreation staff and expanded programs driven by community input.

Mechanism: Community park infrastructural interventions. Medium: improvements and additions to park infrastructure (turf grass, lighting, fencing, etc.)

Number of park users, proportion of users participating in MVPA

Average number of park users increased significantly across all age groups (p ≤ 0.008) except for teenage boys in both intervention parks (p = 0.813 and p = 0.931), and except for teenage girls only for the infrastructural plus programme intervention park (p = 0.116). Increases were more pronounced in the infrastructure plus programme park for children and adults, and the infrastructure intervention only for seniors. Mean number of people classified as moderately or vigorously active increased in both intervention parks (range of 0.29 for female vigorous activity in the infrastructure only park to 7.28 for male moderate activity in the infrastructure plus programme park; all p ≤ 0.05). With the exception of an increase in mean number of males classified as moderately active (increase of 2.39, p = 0.01), no changes in physical activity were observed for the control park.

Veitch; 2012; Melbourne, Victoria, Australia [97]

All park users aged 2 or more years, n = 235 at baseline

Controlled cross-sectional pretest-posttest

A park that was originally an open space area with few amenities. The intervention involved the establishment of a fenced leash-free area for dogs, an all-abilities playground, a 365 m walking track, a barbecue area, landscaping, and fencing to prevent motor vehicle access to the park.

Mechanism: infrastructural interventions. Medium: presence of additional features.

Walking and vigorous physical activity, number of park users

Significant difference in change in number of observed park users between the intervention park (increased from 235 to 985) and control park (decreased from 83 to 51, F = 14.99, p < 0.0005). Significant difference in change in number of people observed walking (increased from 155 to 369 in intervention park and decreased from 75 to 51 in control park, F = 11.70, p < 0.0005). Significant difference in change in counts of people being vigorously active (increased from 38 to 257 in intervention park, decreased from 5 to 0 in control park, F = 4.98, P = 0.008).

West; 2011; USA [98]

368 adult residents aged 30+ years living within 0.5 and 0.51–1.0 miles of a newly installed greenway (the intervention)

Uncontrolled longitudinal pretest-posttest

5 miles of greenway were developed and added to an existing greenway (an open space corridor reserved for recreational use or environmental preservation that connects urban centres).

Mechanism: recreation setting infrastructural interventions. Medium: additional greenway space for walking and cycling.

Walking, MVPA

Number of days spent walking increased for all respondents (by a mean of 0.48 days for residents living ≤0.50 miles to the new greenway section and 0.26 for those living 0.51–1.0 miles from the greenway). No significant interaction effect for time and proximity to the intervention was observed (p = 0.363). Number of days participating in MVPA increased for all respondents (by a mean of 0.63 days for residents living ≤0.50 miles to the new greenway section and 0.48 for those living 0.51–1.0 miles from the greenway). No significant interaction effect for time and proximity to the intervention was observed (p = 0.476). Number of days participating in vigorous activity increased by 0.46 days for residents living nearer and further from the intervention; no significant interaction effect for time and proximity to the intervention was observed (p = 0.962).

West; 2015; North Carolina, USA [60]

524 property owners who owned a single-family dwelling unit valued at more than $5000 and located within 1 mile of the greenway intervention or in the control area

Controlled longitudinal pretest-posttest

1.93 miles of greenway were developed and added to an existing greenway.

Mechanism: New infrastructure for walking and cycling. Medium: addition of 1.93 miles of greenway to an existing greenway

Walking, moderate physical activity, vigorous physical activity

No significant differences found between the experimental and control groups for any outcome variables (all p > 0.975).

  1. β beta coefficient, CI confidence interval, K Kindergarten, ages 5–6 years in USA, m metres, METs metabolic equivalents, MVPA moderate-to-vigorous physical activity, n number, OR odds ratio, UK United Kingdom, USA United States of America, Z Wilcoxon Signed Ranks test statistic