Social Science & Medicine 138 (2015) 22e30

Contents lists available at ScienceDirect

Social Science & Medicine journal homepage: www.elsevier.com/locate/socscimed

Review

Association of proximity and density of parks and objectively measured physical activity in the United States: A systematic review Carolyn Bancroft a, *, Spruha Joshi a, Andrew Rundle a, Malo Hutson b, Catherine Chong d, Christopher C. Weiss c, Jeanine Genkinger a, Kathryn Neckerman d, Gina Lovasi a a

Department of Epidemiology, Mailman School of Public Health, Columbia University, 722 W 168th St., New York, NY 10032, USA Department of City and Regional Planning, University of California at Berkeley, USA Department of Sociology, Columbia University, USA d Institute for Social and Economic Research and Policy, Columbia University, USA b c

a r t i c l e i n f o

a b s t r a c t

Article history: Available online 23 May 2015

One strategy for increasing physical activity is to create and enhance access to park space. We assessed the literature on the relationship of parks and objectively measured physical activity in population-based studies in the United States (US) and identified limitations in current built environment and physical activity measurement and reporting. Five English-language scholarly databases were queried using standardized search terms. Abstracts were screened for the following inclusion criteria: 1) published between January 1990 and June 2013; 2) US-based with a sample size greater than 100 individuals; 3) included built environment measures related to parks or trails; and 4) included objectively measured physical activity as an outcome. Following initial screening for inclusion by two independent raters, articles were abstracted into a database. Of 10,949 abstracts screened, 20 articles met the inclusion criteria. Five articles reported a significant positive association between parks and physical activity. Nine studies found no association, and six studies had mixed findings. Our review found that even among studies with objectively measured physical activity, the association between access to parks and physical activity varied between studies, possibly due to heterogeneity of exposure measurement. Self-reported (vs. independently-measured) neighborhood park environment characteristics and smaller (vs. larger) buffer sizes were more predictive of physical activity. We recommend strategies for further research, employing standardized reporting and innovative study designs to better understand the relationship of parks and physical activity. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Physical activity Parks Built environment Active travel Determinants Exercise Accelerometer

1. Introduction Over the past fifty years, the United States (US) has witnessed an increase in sedentary activity and a corresponding decrease in physical activity related to transportation, leisure-time and work (Brownson et al., 2005). Physical activity has been linked to improved mental and physical health (Warburton et al., 2006). While the US Centers for Disease Control and Prevention recommends at least 1 h of physical activity per day for children and adolescents, and a minimum of 225 min per week of moderate to vigorous physical activity for adults, most people in the US do not meet this recommendation (CDC, 2014; Tucker et al., 2011).

* Corresponding author. E-mail address: [email protected] (C. Bancroft). http://dx.doi.org/10.1016/j.socscimed.2015.05.034 0277-9536/© 2015 Elsevier Ltd. All rights reserved.

Research has shown that active transportation (primarily walking and biking) and access to exercise facilities can increase physical activity (Handy et al., 2002; Sallis et al., 1992). The built environmentddefined as human-made or modified environment, including transportation, food outlets, and parksdincreasingly has been recognized as a determinant of physical activity and population health (Lee and Moudon, 2004; Rao et al., 2007; Srinivasan et al., 2003). In particular, pedestrian-supportive built environment characteristics such as open space, mixed land use and walkability predict increased physical activity (Brownson et al., 2009; Durand et al., 2011; Papas et al., 2007; Sallis et al., 1998). Commercial facilities such as gyms may also promote physical activity, at least among persons with memberships (Kaufman et al., 2014). Parks are common locations for recreational physical activity and are accessible to a wider population (Giles-Corti et al., 2005; Godbey et al., 2005; Lee and Moudon, 2004). However, research on

C. Bancroft et al. / Social Science & Medicine 138 (2015) 22e30

the association of parks and physical activity has employed a mixture of self-reported and objective measurement approaches that may contribute to inconsistency in study findings. While previous research has reported mixed findings concerning the association of parks and physical activity, we focused more specifically on research using objectively measured physical activity through devices such as accelerometers. We propose that such device-based physical activity measurement is the best single measurement approach at this time for addressing our research question: does the current research literature support investments in local parks as a strategy to increase total physical activity of local residents? Previous systematic reviews have investigated the association of urban planning, recreational facilities, traffic, safety and parks with physical activity (Durand et al., 2011; Ferdinand et al., 2012; Kaczynski et al., 2008). One of these reviews suggested that the built environment was more likely to be associated with selfreported than objectively-measured physical activity (Ferdinand et al., 2012). In a review of the international research literature, fewer than one-quarter of studies (9 out of 41) included objective measures of physical activity, and only 3 of those papers found a positive association between activity and green space (Lachowycz and Jones, 2011). Despite the recent growth in research using accelerometers, pedometers and other portable devices, to our knowledge, no review has focused only on studies with objectively measured physical activity and measures of access to parks in the US. This review assesses evidence relevant to whether investments in creating, maintaining or improving parks would increase total objectively measured physical activity among area residents. It aims to identify limitations in current measurement and reporting practices for built environment characteristics and physical activity and to offer recommendations ways to standardize reporting and improve measurement strategies, providing a stronger evidence base for policy. 2. Methods 2.1. Information sources and eligibility criteria A systematic search of the published literature was conducted in PubMed, PsycInfo, TRIS, ALR Literature Database, and Web of Science, using similar methods as described previously (Lovasi et al., 2009). We searched PsycInfo and the ALR database for dissertations, in order to include gray literature. Previously published reviews and references from included studies were also screened. Studies were eligible for this review if: 1) conducted in the United States with a sample size of 100 individuals or more; 2) results were reported in English between January 1990 and June 2013; 3) the study included physical activity measured objectively (with a pedometer or accelerometer, sometimes used in combination with GPS tracking) as an outcome; and 4) the study included park-related built environment measures such as density of parks (number of parks per unit of land area such as buffer or square kilometer) or distance to nearest park (objective or self-reported) as predictors. 2.2. Search strategy The following search terms related to the outcome and exposure were used to identify relevant articles: accelerometer, pedometer, physical activity, green space, walkability, and recreational facilities. The full search terms for the meta-analysis project are included in Online Appendix 1. The search strategy was intended to have high sensitivity to identify a wide range of studies with built

23

environment measures as potential predictors of physical activity, regardless of whether parks were of primary interest. While we allowed a broad definition of the exposure (parks), we restricted our attention more narrowly to objective physical activity as an outcome. 2.3. Study selection Following automatic screening for duplicates (via Endnote), abstracts were screened by one team member for inclusion according to the eligibility criteria. A training set of fifty abstracts was used to harmonize screening across team members. If full text articles were consulted, a different team member evaluated the full article for final determination of inclusion. 2.4. Data collection process and data items For all articles included in the review, we abstracted information on study design, sample size, demographics of the study population, independent variables, outcomes, analytic approach and associations observed. A team member completed the initial data entry, which was verified by a different team member who made any necessary corrections. Discrepancies were discussed between team members and with the senior author (GSL), as needed, to reach consensus. 2.5. Methods of analysis/synthesis of results We first grouped the papers according to their built environment measure for parks and their objective measure of physical activity. Second, we summarized the study findings for specific park exposure categories according to the strength of the reported association between that measure and physical activity: significant in the hypothesized direction (e.g., more parks associated with more activity), significant in the opposite direction, no relationship or mixed findings (direction or statistical significance of association differed across analyses). Third, we noted potential sources of heterogeneity across the set of included studies. Finally, strengths and weakness were summarized to inform improved reporting standards and measurement practices as this research progresses toward a more credible and cohesive evidence-base to support informed action. 3. Results 3.1. Search and study selection The database search resulted in 15,739 abstracts of which 10,949 were unique (Fig. 1). These abstracts were then screened for inclusion. In addition, 133 articles from the Active Living Research Database and 187 systematic or narrative reviews were screened for potential additional references. Ultimately, 801 full text articles were reviewed for eligibility and 320 were abstracted into a database of US-based studies linking the built environment to physical activity or adiposity. Among these, 93 articles included a built environment exposure variable related to parks or green space. Of these 93 articles, 20 had objectively measured physical activity as an outcome and thus were included in this review. 3.2. Overview of studies Table 1 provides an overview of the 20 articles with the association of interest (parks and physical activity), study populations, location and sample size, outcomes and exposure measures and findings. The 20 publications were based on data from 16 unique

24

C. Bancroft et al. / Social Science & Medicine 138 (2015) 22e30

Fig. 1. Flowchart of article search and inclusion for systematic review on the association of parks and objectively measured physical activity in the United States.

study populations, most of which were recruited in urban or suburban settings. Multiple publications from the same study population were included only if they analyzed data that had different exposures or outcomes or were from different time periods. For dissertations, we used published versions when available.

3.3. Measures In most studies (16 out of 20), the outcome variable was amount of moderate-to-vigorous physical activity (MVPA) measured by accelerometer and tracked over a three to seven day period. Four studies used pedometers and tracked steps per day or walking trips per week (Galvez et al., 2013; King et al., 2005, 2003; Wieters, 2009). One study tracked only non-school MVPA (Cohen et al., 2006) and another study only weekend MVPA (Scott et al., 2007a). Physical activity was reported as a continuous outcome (average minutes of MVPA per day), a dichotomous outcome (e.g.
10,000 steps per day versus 10,000 steps/day vs. 10,000 steps/day recordings (n ¼ 128) specified vs. 10,000 steps/day (Jago et al., 2006a) None specified

a b c d e f g h i j k l m n

Crosssectional

Boys 10e14 years (n ¼ 210), Houston

Findings:a NS

M

NS

NS

M

NS

M

NS

Positive association (þ), mixed findings (M), No association (NS). Models included: school and site race/ethnicity, neighborhood racial diversity, SES, street connectivity, and participation in school free lunch program. Model included: race/ethnicity, SES (free lunch). Model included: age, ethnicity, education and intersection density. Model included: age, sex, BMI, household income. Model included: age, ethnicity, gender, interaction of parks with social support, self-efficacy and psychosocial barriers. Model included covariates representing individual, social and neighborhood characteristics using LASSO technique. Model included: age, race, free-and-reduced lunch, population density, jobsehousing ratio, road density, and income. Model included: age, gender, race/ethnicity, BMI, caregiver's education, block safety, and season. Model included: age, sex, education, residential density and sidewalk density. Model included: BMI, age, parental education and ethnicity. Model included: age and BMI. Model included: age, gender, BMI, years in home, and license/car ownership. Model included: age, income, gender, and education.

for increased moderate-to-vigorous physical activity (MVPA) given proximity to or density of parks (Cohen et al., 2006; Galvez et al., 2013; Jilcott et al., 2007; McConville, 2009; Rodríguez et al., 2012; Wieters et al., 2012) and percent change in MVPA for each unit change in parks (Carlson et al., 2012; Jago et al., 2006b; Norman et al., 2006; Ries et al., 2009; Saelens et al., 2012; Scott et al.,

2007a; Strath et al., 2012; Tappe et al., 2013; Young et al., 2014). The remaining five studies used multivariate analysis of variance (MANOVA) (Hall and McAuley, 2010), Wilcoxon rank sum test (King et al., 2003), t-test (King et al., 2005) or correlation coefficient (Jago et al., 2006a; Patnode et al., 2010) to measure the association between parks and physical activity. Due to heterogeneity in the

C. Bancroft et al. / Social Science & Medicine 138 (2015) 22e30

27

Table 2 Measures used for density and proximity in studies that measured association of parks with physical activity (n ¼ 20). Measure of park access Studies including measure

Number of studies

Perceived density Objective density

2 12

Perceived proximity Objective proximity

(Hall and McAuley, 2010; Ries et al., 2009) (Carlson et al., 2012; Cohen et al., 2006; Galvez et al., 2013; Hall and McAuley, 2010; Jago et al., 2006b; Jago et al., 2006a; McConville, 2009; Norman et al., 2006; Ries et al., 2009; Saelens et al., 2012; Scott, Cohen, et al., 2007; Strath et al., 2012; Young et al., 2014) (Hall and McAuley, 2010; Jilcott et al., 2007; King et al., 2003; Ries et al., 2009; Saelens et al., 2012; Tappe et al., 2013; Wieters et al., 2012) (Cohen et al., 2006; Hall and McAuley, 2010; Jago et al., 2006b; Jilcott et al., 2007; King et al., 2005; McConville, 2009; Patnode et al., 2010; Ries et al., 2009; Rodríguez et al., 2012; Saelens et al., 2012; Strath et al., 2012)

outcome measures and reporting timeframes, a quantitative metaanalysis was infeasible. 3.6. Findings from categorization of studies by park exposure 3.6.1. Park proximity and density Seven studies included measures of both park proximity and density. Two studies observed significant positive associations for both proximity and density measures with physical activity (Cohen et al., 2006; McConville, 2009); both of these studies used one mile, half mile and quarter mile buffers of participants' homes. Another study found significant positive associations for park proximity and density with quarter-mile buffers but not one-mile buffers (Jago et al., 2006b). An additional study using a small buffer (200 m) for proximity and density found no significant association (Strath et al., 2012). One study with both perceived and objective onemile proximity and density measures found a significant association for perceived but not objective measures of proximity (Ries et al., 2009). Two studies found no significant association between increased physical activity and park proximity or density within 1 km of home (Hall and McAuley, 2010; Saelens et al., 2012). Hall and McAuley hypothesized that the size of their buffer measure may have influenced the lack of association (Hall and McAuley, 2010). In sum, findings were mixed across the studies that included both objective and perceived measures, and smaller buffers appeared to yield stronger associations with physical activity than larger buffers. 3.6.2. Park density Six studies included only park density as the built environment exposure measure. Norman and colleagues found that the density of parks within one mile was significantly correlated with MVPA among girls but not boys (Norman et al., 2006). An additional study of boys found no correlation between parks within one mile and MVPA (Jago et al., 2006a). Scott et al. found that density of parks within one half mile had a marginally significant positive association with increased weekend MVPA (Scott et al., 2007a). Young found that density of parks within one mile was significant for a national sample but not for a sample in Maryland (Young et al., 2014). Two remaining studies that measured density of parks within a 500 m buffer of home or in the census block where children live found no association between parks (dichotomized into any versus none) and MVPA (Carlson et al., 2012; Galvez et al., 2013). Most studies using only density as an exposure found no association between parks and physical activity. 3.6.3. Park proximity Seven studies included only objective or perceived park proximity (usually operationalized as shortest distance from home to the nearest park) as the exposure measure. Wieters reported a significant positive association for individuals reporting parks within walking distance of work (Wieters, 2009) as did Tappe, who

6 11

measured presence of parks within a quarter mile buffer of home with GIS (Tappe et al., 2013). In a study by King, elderly women who reported a park within “walking distance of home” walked significantly more mean steps per day than those without such a park (King et al., 2003); however, objective measures of proximity were not associated with steps per day (King et al., 2005). In another study, objective and perceived proximity measures were correlated with each other but neither was significantly associated with physical activity (Jilcott et al., 2007). A study using GIS-based proximity measures found no association with physical activity (Patnode et al., 2010). One study found that objective proximity of individual participants to a park (within 50 m of a GPS/accelerometer data point) throughout the day was significantly associated with MVPA in one study site (Minneapolis) but not in another (Seattle) (Rodríguez et al., 2012). Among studies with a focus on park proximity, those with objective measures had fewer significant findings than studies with perceived measures.

3.6.4. Multiple potential explanations of heterogeneity Measurement choices may account for this heterogeneity in findings. First, there were multiple modes for measuring park proximity and density, including mapping, audits and participant surveys. Second, there were multiple buffer sizes within and across studies and varying measures of physical activity (e.g., steps per day, non-school or weekend MVPA). Finally, this review excluded thirteen studies that measured parks and open space but only included them as part of a combined measure or index and did not assess it as a single construct (see Online Appendix for overview of these papers). Potential sources of bias include “exposure reporting bias” in which authors may have coded exposures in multiple ways and then presented only the findings most consistent with their hypotheses. Conceptual models for parks are not specific about which measures are most meaningful with respect to buffer size. Three authors reported findings for multiple buffer sizes, for both objective and perceived measures or for both density and proximity. However, authors may not have reported findings for all exposure measures considered, making multiple comparisons problems difficult to assess. For those reporting comparable estimates of association across multiple measures, results sometimes differed across buffer size. Furthermore, selection bias may be a concern for those studies without systematic sampling or with low response rates. Uncontrolled confounding is a potential threat to validity across the studies. Some studies, but not all, adjusted for sociodemographic characteristics that seem strong candidates as potential confounders, such as age, race and ethnicity or socioeconomic status. Further, some studies adjusted for body mass index (BMI), which may have distorted the estimates if BMI was on the pathway of interest or functioned as a collider (i.e., if BMI was affected by both neighborhood context and physical activity). Furthermore, use of accelerometers for data collection on

28

C. Bancroft et al. / Social Science & Medicine 138 (2015) 22e30

physical activity is susceptible to measurement error, which is of greater concern if systematically different across subpopulations. Jilcott et al. reported differences in the MVPA for women wearing the accelerometer for seven days and those wearing it fewer days; more days of wear were associated with higher average activity levels (Jilcott et al., 2007). Many studies deploying objective measures of physical activity were not able to determine whether MVPA occurred in a park or recreational space, though increasing use of GPS tracking will provide new opportunities to do so. Measuring total MVPA may also contribute to conservative estimates of association as total MVPA may not be as sensitive to the built environment as specific domains of activity that can be distinguished in questionnaire data; nonetheless, total MVPA is of central interest for predicting health outcomes (Lovasi et al., 2012). 4. Discussion In this review of the past twenty years of US studies, we found no consistent pattern of results relating park exposure to objectively measured physical activity. Five studies found a statistically significant positive association between access to parks and physical activity; nine studies found no association; and six studies had mixed findings. Our results add to previous reviews that included objective and self-reported physical activity and called for consistency in measures of green space and physical activity (Ferdinand et al., 2012; Lachowycz and Jones, 2011). Future studies may benefit from comparing multiple measurement approaches (Davison and Lawson, 2006). In studies that did provide such comparisons, we noted stronger associations of perceived versus objective park environment measures with physical activity, and stronger park-physical activity associations for analyses with smaller buffer sizes (a specified radius distance surrounding a subject's home) versus larger buffer sizes. This review highlighted several limitations in current measurement and reporting practices for both built environment exposures and physical activity outcomes. Measures of park exposures were heterogeneous and these differences made it difficult to compare magnitudes across studies. Standardization of built environment measures remains a challenge for research in this area, as has been previously reported (Papas et al., 2007). The diversity of measures for built environment attributes included self-report and approaches generally labeled as objective (observational audits and GIS data) (Brownson et al., 2009). Studies used perceived and objective measures of park proximity and density to represent park access. Most studies had an objectively measured park exposure (usually proximity or density measured using GIS). However, buffers drawn around participant homes, as is typical, miss other environments experienced through regular travel to school, work or other destinations (Lovasi et al., 2012). There was modest agreement across perceived and objective measures of park access among the four studies that included both. However, studies using both self-report and objective park measures were more likely to find support for park-physical activity associations in analyses using self-reported measures. This pattern echoes findings from studies of perceived and objective measures of walkability (Gebel et al., 2009). It remains an open question whether the stronger associations of perceived environment with physical activity reflect a crucial role of perception as more proximal to behavior (perhaps along the pathway between objective environment and activity) or a vulnerability of perceived measures to bias (such that correlated errors could be distorting the true causal association when both exposure and outcome are selfreported) (Blacksher and Lovasi, 2012). Varying neighborhood definitions and categorization approaches for studies of parks and physical activity may have further

contributed to the heterogeneity in findings. Although most studies used a dichotomous proximity measure for parks (any vs. none within a certain distance) or a continuous density measure (count per area), the buffer size varied from 200 m to one mile and some studies included multiple buffer sizes. Most studies in this review did not include park size or measures of park disorder. Park size (in particular, larger parks which often have more facilities and amenities) has been shown to have a stronger association with physical activity and BMI than park proximity (Giles-Corti et al., 2005; Rundle et al., 2013). Parks may, however, need to be ‘discounted’ for neighborhood characteristics such as crime, and other characteristics that influence “social access” and may diminish the apparent benefit of spatial access (Weiss et al., 2011). In addition to park access measures, a few studies also included factors such as aesthetics and perceived quality (Ries et al., 2009), safety (Cohen et al., 2006; Wieters et al., 2012), amenities and facilities (Kerr et al., 2011; Scott et al., 2007b), and type of park (Jago et al., 2006b). A subset of studies combined GPS and accelerometer data to identify the location of physical activity (Dunton et al., 2013; Rodríguez et al., 2012). Research has investigated the role of age, safety and security, aesthetics and psychosocial factors in the association of parks with physical activity and found that these characteristics are important determinants of use (Lee and Maheswaran, 2011). We recommend that future studies incorporate data on potential effect modifiers such as neighborhood income and crime. While we were interested in the association of parks and physical activity, many factors potentially confound this seemingly basic relationship. Studies in this review included sociodemographic covariates that must be considered when investigating the relationship between physical activity and the built environment, but had limited ability to adjust for more elusive lifestyle and neighborhood preferences that may confound environmentbehavior associations. A few limitations of this review should be noted. First, non-US studies were excluded in order to focus on a particular context for policy decisions. Studies from other settings can offer useful insight. For example, an intriguing story from Chennai, India, suggests that social engagement can amplify the physical activity benefits of new parks (Mohan et al., 2006). Second, because studies meeting our inclusion criteria focused primarily on urban areas and selected age groups, our findings have limited generalizability. Finally, like most systematic reviews, this one could be affected by publication and reporting bias. In the future, obtaining original data from multiple studies for a pooling project may help to overcome limitations inherent to reviewing the results as published. This review also has a number of strengths. By using a broad, sensitive search strategy for the larger meta-analysis project, we captured physical activity studies in which parks were not the primary focus but for which results were reported. We also highlighted the numerous ways that parks are measured, using both perceived and objective measures, measures of facilities and type of parks. Finally, the review limits measurement bias and heterogeneity for the outcome (physical activity) by limiting inclusion to those studies with objective outcome measures. In conclusion, our review found that when physical activity is measured objectively, the observed association between access to parks and physical activity remains inconsistent across studies. Further improvement of reporting standards for physical activity and park access is needed. In particular, standardized exposure measures and comprehensive reporting for objective and perceived park access will make it more feasible to combine and compare results. We recommend at a minimum that studies use quarter, half and one-mile buffers (or the approximately corresponding 0.5, 1.0 and 1.5 km buffers) to measure density (number of parks) and

C. Bancroft et al. / Social Science & Medicine 138 (2015) 22e30

proximity (distance to nearest park). We also recommend that authors include measures of park size, features (e.g., trails, sport facilities), and aspects of appearance such as physical disorder when possible (see Online Appendix Table 2). In addition to total physical activity monitoring, researchers should consider GPS tracking or other objective measures that capture the context in which physical activity occurs, such as the System for Observing Play and Recreation in Communities (SOPARC) (McKenzie et al., 2006). In addition, if parks are included as part of a combined exposure measure, the study should also report the analysis with parks as a distinct exposure. Use of online supplemental materials may be helpful in reconciling these guidelines with journal space limitations. We also recommend improvements in study design and sampling. Further developments in longitudinal analyses, particularly natural experiments to evaluate park improvements (Cohen et al., 2009), may help clarify the relationship between access to parks and physical activity. Currently, studies linking parks to physical activity monitoring is happening mainly with children and elderly study populations and thus, inclusion of younger adult age groups, or multiple age groups in a single study, would support an understanding of how associations change over the life course. Clarifying the many factors that influence health benefits of park access, such as density, proximity, quality and safety, will help enrich our understanding of how parks might promote increased physical activity. Acknowledgments The authors would like to acknowledge financial support from Robert Wood Johnson Foundation Active Living Research program (grant #68507). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.socscimed.2015.05.034. References Blacksher, E., Lovasi, G.S., 2012. Place-focused physical activity research, human agency, and social justice in public health: taking agency seriously in studies of the built environment. Health Place 18 (2), 172e179. http://dx.doi.org/10.1016/ j.healthplace.2011.08.019. Brownson, R.C., Boehmer, T.K., Luke, D.A., 2005. Declining rates of physical activity in the United States: what are the contributors? Annu. Rev. Public Health 26, 421e443. http://dx.doi.org/10.1146/annurev.publhealth.26.021304.144437. Brownson, R.C., Hoehner, C.M., Day, K., Forsyth, A., Sallis, J.F., 2009. Measuring the built environment for physical activity: state of the science. Am. J. Prev. Med. 36 (4 Suppl. l) http://dx.doi.org/10.1016/j.amepre.2009.01.005. S99e123 e112. Carlson, J.A., Sallis, J.F., Conway, T.L., Saelens, B.E., Frank, L.D., Kerr, J., ..., King, A.C., 2012. Interactions between psychosocial and built environment factors in explaining older adults' physical activity. Prev. Med. 54 (1), 68e73. http:// dx.doi.org/10.1016/j.ypmed.2011.10.004. CDC, 2014. Physical Activity for Everyone. Retrieved February 20, 2014, from: http:// www.cdc.gov/physicalactivity/everyone/guidelines/index.html. Cohen, D.A., Ashwood, J.S., Scott, M.M., Overton, A., Evenson, K.R., Staten, L.K., ..., Catellier, D., 2006. Public parks and physical activity among adolescent girls. Pediatrics 118 (5), e1381e1389. http://dx.doi.org/10.1542/peds.2006-1226. Cohen, D.A., Golinelli, D., Williamson, S., Sehgal, A., Marsh, T., McKenzie, T.L., 2009. Effects of park improvements on park use and physical activity: policy and programming implications. Am. J. Prev. Med. 37 (6), 475e480. http://dx.doi.org/ 10.1016/j.amepre.2009.07.017. Davison, K., Lawson, C., 2006. Do attributes in the physical environment influence children's physical activity? A review of the literature. Int. J. Behav. Nutr. Phys. Act. 3 (1), 19. Dunton, G.F., Liao, Y., Almanza, E., Jerrett, M., Spruijt-Metz, D., Pentz, M.A., 2013. Locations of joint physical activity in parent-child pairs based on accelerometer and GPS monitoring. Ann. Behav. Med. 45 (Suppl. 1), S162eS172. http:// dx.doi.org/10.1007/s12160-012-9417-y. Durand, C.P., Andalib, M., Dunton, G.F., Wolch, J., Pentz, M.A., 2011. A systematic review of built environment factors related to physical activity and obesity risk:

29

implications for smart growth urban planning. Obes. Rev. 12 (5), e173e182. http://dx.doi.org/10.1111/j.1467-789X.2010.00826.x. Ferdinand, A., Sen, B., Rahurkar, S., Engler, S., Menachemi, N., 2012. The relationship between built environments and physical activity: a systematic review. Am. J. Public Health 102 (10), e7ee13. http://dx.doi.org/10.2105/AJPH.2012.300740. Galvez, M.P., McGovern, K., Knuff, C., Resnick, S., Brenner, B., Teitelbaum, S.L., Wolff, M.S., 2013. Associations between neighborhood resources and physical activity in inner-city minority children. Acad. Pediatr. 13 (1), 20e26. http:// dx.doi.org/10.1016/j.acap.2012.09.001. Gebel, K., Bauman, A., Owen, N., 2009. Correlates of non-concordance between perceived and objective measures of walkability. Ann. Behav. Med. 37 (2), 228e238. http://dx.doi.org/10.1007/s12160-009-9098-3. Giles-Corti, B., Broomhall, M.H., Knuiman, M., Collins, C., Douglas, K., Ng, K., ..., Donovan, R.J., 2005. Increasing walking: how important is distance to, attractiveness, and size of public open space? Am. J. Prev. Med. 28 (2 Suppl. 2), 169e176. http://dx.doi.org/10.1016/j.amepre.2004.10.018. Godbey, G.C., Caldwell, L.L., Floyd, M., Payne, L.L., 2005. Contributions of leisure studies and recreation and park management research to the active living agenda. Am. J. Prev. Med. 28 (2), 150e158. Hall, K.S., McAuley, E., 2010. Individual, social environmental and physical environmental barriers to achieving 10 000 steps per day among older women. Health Educ. Res. 25 (3), 478e488. http://dx.doi.org/10.1093/her/cyq019. Handy, S.L., Boarnet, M.G., Ewing, R., Killingsworth, R.E., 2002. How the built environment affects physical activity: views from urban planning. Am. J. Prev. Med. 23 (2 Suppl. l), 64e73. Jago, R., Baranowski, T., Baranowski, J.C., 2006a. Observed, GIS, and self-reported environmental features and adolescent physical activity. Am. J. Health Promot. 20 (6), 422e428. Jago, R., Baranowski, T., Harris, M., 2006b. Relationships between GIS environmental features and adolescent male physical activity: GIS coding differences. J. Phys. Act. Health 3, 230e242. Jilcott, S., Evenson, K., Laraia, B., Ammerman, A., 2007. Association between physical activity and proximity to physical activity resources among low-income, midlife women. Prev. Chronic Dis. 06e0049. Kaczynski, A.T., Potwarka, L.R., Saelens, B.E., 2008. Association of park size, distance, and features with physical activity in neighborhood Parks. Am. J. Public Health 98 (8), 1451e1456. http://dx.doi.org/10.2105/AJPH.2007.129064. Kaufman, T.S.D., Sheehan, D., Neckerman, K., Rundle, A., Lovasi, G., 2014. P430 Spatial Proximity and Membership to Gyms or Related Facilities Jointly Predict Objectively Measured Physical Activity Among Adults in New York City (Paper presented at the American Heart Association epidemiology council meeting, San Francisco, CA). Kerr, J., Carlson, J.A., Sallis, J.F., Rosenberg, D., Leak, C.R., Saelens, B.E., ..., Conway, T.L., 2011. Assessing health-related resources in senior living residences. J. Aging Stud. 25 (3), 206e214. King, W.C., Belle, S.H., Brach, J.S., Simkin-Silverman, L.R., Soska, T., Kriska, A.M., 2005. Objective measures of neighborhood environment and physical activity in older women. Am. J. Prev. Med. 28 (5), 461e469. http://dx.doi.org/10.1016/ j.amepre.2005.02.001. King, W.C., Brach, J.S., Belle, S., Killingsworth, R., Fenton, M., Kriska, A.M., 2003. The relationship between convenience of destinations and walking levels in older women. Am. J. Health Promot. 18 (1), 74e82. Lachowycz, K., Jones, A.P., 2011. Greenspace and obesity: a systematic review of the evidence. Obes. Rev. 12 (5), e183e189. http://dx.doi.org/10.1111/j.1467789X.2010.00827.x. Lee, Maheswaran, R., 2011. The health benefits of urban green spaces: a review of the evidence. J. Public Health 33 (2), 212e222. Lee, Moudon, A.V., 2004. Physical activity and environment research in the health field: implications for urban and transportation planning practice and research. J. Plan. Lit. 19 (2), 147e181. Lovasi, G.S., Grady, S., Rundle, A., 2012. Steps forward: review and recommendations for research on walkability, physical activity and cardiovascular health. Public Health Rev. 33 (4), 484. Lovasi, G.S., Hutson, M.A., Guerra, M., Neckerman, K.M., 2009. Built environments and obesity in disadvantaged populations. Epidemiol. Rev. 31, 7e20. http:// dx.doi.org/10.1093/epirev/mxp005. McConville, M., 2009. Defining Mixed-use: Which Land Uses Promote Walking? (University of North Carolina at Chapel Hill Master's Paper). McKenzie, T.L., Cohen, D.A., Sehgal, A., Williamson, S., Golinelli, D., 2006. System for observing play and recreation in communities (SOPARC): reliability and feasibility measures. J. Phys. Act. Health 3, S208. Mohan, V., Shanthirani, C., Deepa, M., Datta, M., Williams, O., Deepa, R., 2006. Community empowerment-a successful model for prevention of noncommunicable diseases in india-the chennai urban population study (cups17). J. Assoc. Physicians India 54, 858e862. Norman, G.J., Nutter, S.K., Ryan, S., Sallis, J.F., Calfas, K.J., Patrick, K., 2006. Community design and access to recreational facilities as correlates of adolescent physical activity and body-mass index. J. Phys. Act. Health 3, S118eS128. Papas, M.A., Alberg, A.J., Ewing, R., Helzlsouer, K.J., Gary, T.L., Klassen, A.C., 2007. The built environment and obesity. Epidemiol. Rev. 29, 129e143. http://dx.doi.org/ 10.1093/epirev/mxm009. Patnode, C.D., Lytle, L.A., Erickson, D.J., Sirard, J.R., Barr-Anderson, D., Story, M., 2010. The relative influence of demographic, individual, social, and environmental factors on physical activity among boys and girls. Int. J. Behav. Nutr. Phys. Act. 7, 79. http://dx.doi.org/10.1186/1479-5868-7-79.

30

C. Bancroft et al. / Social Science & Medicine 138 (2015) 22e30

Rao, M., Prasad, S., Adshead, F., Tissera, H., 2007. The built environment and health. Lancet 370 (9593), 1111e1113. http://dx.doi.org/10.1016/S0140-6736(07)612604. Ries, A.V., Voorhees, C.C., Roche, K.M., Gittelsohn, J., Yan, A.F., Astone, N.M., 2009. A quantitative examination of park characteristics related to park use and physical activity among urban youth. J. Adolesc. Health 45 (3), S64eS70. Rodríguez, D.A., Cho, G.-H., Evenson, K.R., Conway, T.L., Cohen, D., GhoshDastidar, B., ..., Lytle, L.A., 2012. Out and about: association of the built environment with physical activity behaviors of adolescent females. Health Place 18 (1), 55e62. Rundle, A., Quinn, J., Lovasi, G., Bader, M.D., Yousefzadeh, P., Weiss, C., Neckerman, K., 2013. Associations between body mass index and park proximity, size, cleanliness, and recreational facilities. Am. J. Health Promot. 27 (4), 262e269. http://dx.doi.org/10.4278/ajhp.110809-QUAN-304. Saelens, B.E., Salls, J.F., Frank, L.D., Cain, K.L., Conway, T.L., Chapman, J.E., ..., Kerr, J., 2012. Neighborhood environment and psychosocial correlates of adults' physical activity. Med. Sci. Sports Exerc. 44 (4), 637e646. http://dx.doi.org/10.1249/ MSS.0b013e318237fe18. Sallis, J.F., Bauman, A., Pratt, M., 1998. Environmental and policy interventions to promote physical activity. Am. J. Prev. Med. 15 (4), 379e397. Sallis, J.F., Hovell, M.F., Hofstetter, C.R., 1992. Predictors of adoption and maintenance of vigorous physical activity in men and women. Prev. Med. 21 (2), 237e251. Scott, M.M., Cohen, D.A., Evenson, K.R., Elder, J., Catellier, D., Ashwood, J.S., Overton, A., 2007a. Weekend schoolyard accessibility, physical activity, and obesity: the trial of activity in adolescent girls (TAAG) study. Prev. Med. 44 (5), 398e403. Scott, M.M., Evenson, K.R., Cohen, D.A., Cox, C.E., 2007b. Comparing perceived and objectively measured access to recreational facilities as predictors of physical activity in adolescent girls. J. Urban Health 84 (3), 346e359. http://dx.doi.org/

10.1007/s11524-007-9179-1. Srinivasan, S., O'Fallon, L.R., Dearry, A., 2003. Creating healthy communities, healthy homes, healthy people: initiating a research agenda on the built environment and public health. Am. J. Public Health 93 (9), 1446e1450. Strath, S.J., Greenwald, M.J., Isaacs, R., Hart, T.L., Lenz, E.K., Dondzila, C.J., Swartz, A.M., 2012. Measured and perceived environmental characteristics are related to accelerometer defined physical activity in older adults. Int. J. Behav. Nutr. Phys. Act. 9, 40. http://dx.doi.org/10.1186/1479-5868-9-40. Tappe, K.A., Glanz, K., Sallis, J.F., Zhou, C., Saelens, B.E., 2013. Children's physical activity and parents' perception of the neighborhood environment: neighborhood impact on kids study. Int. J. Behav. Nutr. Phys. Act. 10, 39. http://dx.doi.org/ 10.1186/1479-5868-10-39. Tucker, J.M., Welk, G.J., Beyler, N.K., 2011. Physical activity in U.S.: adults compliance with the physical activity guidelines for Americans. Am. J. Prev. Med. 40 (4), 454e461. http://dx.doi.org/10.1016/j.amepre.2010.12.016. Warburton, D.E., Nicol, C.W., Bredin, S.S., 2006. Health benefits of physical activity: the evidence. CMAJ 174 (6), 801e809. http://dx.doi.org/10.1503/cmaj.051351. Weiss, C., Purciel, M., Bader, M., Quinn, J., Lovasi, G., Neckerman, K., Rundle, A., 2011. Reconsidering access: park facilities and neighborhood disamenities in New York city. J. Urban Health 88 (2), 297e310. http://dx.doi.org/10.1007/s11524011-9551-z. Wieters, K.M., 2009. Integrating Walking for Transportation and Physical Activity for Sedentary Office Workers in Texas (Dissertation). Texas A&M University. Retrieved from: https://repository.tamu.edu/bitstream/handle/1969.1/ETDTAMU-2009-08-6988/WIETERS-DISSERTATION.pdf?sequence¼3. Wieters, K.M., Kim, J.H., Lee, C., 2012. Assessment of wearable global positioning system units for physical activity research. J. Phys. Act. Health 9 (7), 913e923. Young, D., Saksvig, B.I., Wu, T.T., Zook, K., Li, X., Champaloux, S., ..., Treuth, M.S., 2014. Multilevel correlates of physical activity for early, mid, and late adolescent girls. J. Phys. Act. Health 11 (5), 950e960.