You are here

Article Text

Article menu

Download PDFPDF

XML
Public health
Protocol
Built environment and physical activity in New Zealand adolescents: a protocol for a cross-sectional study
  1. Erica A Hinckson1,
  2. Scott Duncan2,
  3. Melody Oliver2,
  4. Suzanne Mavoa3,
  5. Ester Cerin4,
  6. Hannah Badland5,
  7. Tom Stewart2,
  8. Vivienne Ivory6,
  9. Julia McPhee2,
  10. Grant Schofield2
  1. 1Faculty of Health and Environmental Sciences, Auckland University of Technology, Auckland, New Zealand
  2. 2Human Potential Centre, Auckland University of Technology, Auckland, New Zealand
  3. 3SHORE & Whariki Research Centre, Massey University, Auckland, New Zealand
  4. 4Institute of Human Performance, University of Hong Kong, Hong Kong, Hong Kong
  5. 5McCaughey Centre, University of Melbourne, Melbourne, Victoria, Australia
  6. 6Department of Public Health, School of Medicine and Health Sciences, University of Otago, Dunedin, New Zealand
  1. Correspondence to Professor Erica A Hinckson; erica.hinckson{at}aut.ac.nz

Abstract

Introduction Built-environment interventions have the potential to provide population-wide effects and the means for a sustained effect on behaviour change. Population-wide effects for adult physical activity have been shown with selected built environment attributes; however, the association between the built environment and adolescent health behaviours is less clear. This New Zealand study is part of an international project across 10 countries (International Physical Activity and the Environment Network–adolescents) that aims to characterise the links between built environment and adolescent health outcomes.

Methods and analyses An observational, cross-sectional study of the associations between measures of the built environment with physical activity, sedentary behaviour, body size and social connectedness in 1600 New Zealand adolescents aged 12–18 years will be conducted in 2013–2014. Walkability and neighbourhood destination accessibility indices will be objectively measured using Geographic Information Systems (GIS). Physical activity and sedentary behaviours will be objectively measured using accelerometers over seven consecutive days. Body mass index will be calculated as weight divided by squared height. Demographics, socioeconomic status, active commuting behaviours and perceived neighbourhood walkability will be assessed using the Neighbourhood Environment Walkability Scale for Youth and psychosocial indicators. A web-based computer-assisted personal interview tool Visualisation and Evaluation of Route Itineraries, Travel Destinations, and Activity Spaces (VERITAS) and Global Positioning System (GPS) receivers will be used in a subsample of 300 participants. A qualitative research component will explore barriers and facilitators for physical activity in adolescents with respect to the built and social environment in a subsample of 80 participants.

Ethics and dissemination The study received ethical approval from the Auckland University of Technology Ethics Committee (12/161). Data will be entered and stored into a secure (password protected) database. Only the named researchers will have access to the data. Data will be stored for 10 years and permanently destroyed thereafter. The results papers will be submitted for publication in peer-reviewed journals.

  • Preventive Medicine
  • Epidemiology
  • Public Health

This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 3.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/3.0/

https://doi.org/10.1136/bmjopen-2013-004475

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Strengths and limitations of this study

    • Limited data exist on detailed and multilevel relationships of interaction between the social and physical environments specific to the New Zealand (NZ) adolescents.

    • The use of Global Positioning System/Visualisation and Evaluation of Route Itineraries, Travel Destinations, and Activity Spaces (GPS/VERITAS) will define the adolescents’ geographical context and will provide accurate estimates of location in which physical activity takes place.

    • The study forms the NZ arm of the international IPEN-Adolescents collaboration, whereby adolescents' physical activity and sedentary behavior data are collected using a common methodology across multiple countries.

    • Parents may self-select neighbourhoods; therefore, associations between built environment and walkability may in part be a reflection of neighbourhood self-selection bias. Parents’ neighbourhood preference and self-selection will be accounted for in the analysis.

    • When conducting spatial analyses on aggregated data, errors affecting the validity of results may be introduced. Geocoded data and other techniques will assist in gaining a more accurate understanding of neighbourhood boundaries for adolescents.

    Introduction

    The benefits of physical activity in youth are well documented.1–5 Regular moderate-to-vigorous physical activity (MVPA) is positively associated with musculoskeletal health, cardiovascular well-being (eg, healthy blood pressure, lipid and lipoprotein levels, cardiovascular autonomic tone), metabolic health, maintenance of a healthy weight, psychological well-being (eg, improved self-concept, reduced anxiety and depression) and reduced risk of type 2 diabetes.6 ,7 The accumulation of at least 60 min of MVPA per day is recommended for youth; however, accumulating physical activity below this threshold is still beneficial, especially for those whose health is at risk (eg, overweight or obese youth).6 Despite awareness of the well-established benefits of physical activity, rapid changes in technology and the habitual environment over the past 50 years may have caused an increase in sitting, passive travel and subsequently a reduction in incidental physical activity.8 Furthermore, over the course of adolescence, physical activity typically decreases by 60–70%,9 while sedentary behaviour remains high at 7–14 h/day.10–12 The latter trend is a matter of particular concern given that emerging evidence suggests that sedentary behaviour has negative effects on health that are independent of the beneficial effects of physical activity.13–15 In addition, levels of activity during school age years significantly predict activity levels16 and health outcomes17 into adulthood.

    Behavioural modification programmes have only achieved limited and mostly short-term physical activity improvements.18–20 For sustainable changes that optimise positive behaviours, it is important to understand that physical activity and sedentary behaviours occur within a broader ecological framework.21 It is recognised that in order to be effective, complex integrated interventions are required that include supportive policies and social and physical environments.22 ,23 Manipulating social and physical environments to be more health promoting will most likely have sustainable and far-reaching impacts on population health behaviours and outcomes. We have previously examined the relationship of objective built environment measures (ie, destination access, street connectivity, dwelling density, land use mix) with accelerometer-derived and self-reported physical activity in adults.24 The work was part of a larger international study (IPEN-International Physical activity and Environment Network) with 12 participating countries. The potential of walkable neighbourhoods for supporting health-enhancing increases in physical activity, at least for adults, was high.24 A 1 SD increase in neighbourhood walkability variables yielded a 7–13% increase in physical activity. This effect is likely to be much higher than effects achieved through behavioural intervention alone.25 ,26

    While the evidence base for associations between the built environment and physical activity in adults has been steadily accumulating,24 ,27 ,28 our understanding of this relationship in adolescents is at its infancy,29–38 and at times non-intuitive.33 ,39 Adolescents were consistently identified in our adult focus groups in our previous study as a subgroup whose changing needs for independent mobility and age and culturally appropriate forms of physical activity are less likely to be met, particularly in more suburban built environment forms.40 In a recent review, land-use mix and residential density were the most highly correlated built environment variables with overall physical activity in youth.41 However, the review did not find any environmental variables that consistently correlated with physical activity in adolescents. Nonetheless, latest research indicates that adolescents’ physical activity tends to occur close to their homes,35 ,42 and that strong associations exist between inactivity with lower neighbourhood walkability, amount of public open space and neighbourhood safety,43 as well as higher densities of cul-de-sac networks.36 MVPA is significantly lower for rural adolescents compared with those living in urban environments; however, these differences between neighbourhood type are not seen for body mass index (BMI).44 Geospatial data indicate that adolescent girls engage in higher intensity physical activity in places with parks, schools and higher population density, and that they accumulate lower levels of physical activity in places with more roads and food outlets.32 Low-income adolescents were physically active at fields/courts, indoor recreation facilities, small and large parks and swimming pools,42 but reduced accessibility of physical activity facilities and food outlets was associated with being overweight.34

    In the Built Environment and Adolescent New Zealanders (BEANZ) study, we seek to understand the relationship of physical activity, sedentary behaviour and body size with neighbourhood-level built environment features in New Zealand (NZ) adolescents. We hypothesise that neighbourhood walkability and neighbourhood destination accessibility indices will be positively associated with minutes of MVPA, and inversely associated with minutes of sedentary time and BMI. We will also investigate associations between the built environment and social connectedness to the community,45 the moderating effects of ethnicity and mediating effects of active commuting, neighbourhood mobility, and perceived neighbourhood walkability. A novel aspect of this study is the use of Portable Global Positioning System (GPS) receivers together with web-based interactive mapping and geocoding software to examine adolescents’ mobility and access to regular destinations. These ancillary data will enable the shape and scale of environmental exposure to be defined in considerable detail.

    Our study forms part of the IPEN-adolescent collaboration, using comparable data collection, management and protocol sharing across 10 countries. By comparing diverse countries, built environmental heterogeneity can be captured (and therefore generate robust estimates of the real effects) while facilitating intracountry and intercountry comparisons. The goal is to generate credible evidence to guide long-term town planning, policy change and redesign of existing urban environments to maximise physical activity and community connectedness and minimise sedentary behaviour and body size, all key determinants of human health.

    Method and analysis

    The standardised checklist for the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) recommendations was used to ensure that all the elements recommended were addressed within this section.46

    Design

    BEANZ will be based on an observational, cross-sectional design that examines the associations of objective and subjective measures of the built environment with physical activity, sedentary behaviour, body size and social connectedness in 1600 NZ adolescents aged 12–18 years from eight secondary schools (approximately 200 participants per school). Moreover, differences between non-Māori and Māori population groups will be explored. Demographics, socioeconomic status (SES), active commuting, psychosocial indicators and perceptions of the built environment will be measured in the full sample. Data will be collected in the 2013–2014 academic school years for the southern hemisphere. A GPS and interactive mapping substudy of approximately 300 participants will assess neighbourhood mobility by geo-locating participants’ destinations, modes of travel, activity locations, walking/cycling areas and perceived neighbourhood boundaries. Focus groups will explore barriers and facilitators to physical activity with respect to neighbourhood built and social environment in a subsample of approximately 80 participants. Data will be collected from two major cities in New Zealand: Auckland and Wellington. Auckland is the largest city in New Zealand with a population of approximately 1.4 million residents (one-third of the country's population),47 with a population density comparable to Los Angeles and Helsinki.48 Wellington, the capital city of New Zealand, is located on the southern part of the North Island and has a population density comparable to Vancouver and Honolulu.48

    Neighbourhood, school and participant selection

    Associations between exposure and outcome variables are estimated based on data collected using a multistage sampling strategy. This strategy maximises heterogeneity in the exposure variables (built environment) while allowing comparisons to be made between those of low and high SES. In the first instance, Geographic Information Systems (GIS) will be used to calculate three built environment measures—street connectivity, residential density and land use mix—for each meshblock (smallest census tract units available in New Zealand).40 Street connectivity will be calculated by dividing the number of 3-or-more-way intersections by the area in square kilometres. To avoid edge effects associated with meshblocks delineated by street centrelines, street connectivity will be calculated for 20 m meshblock buffers. Intersections will be extracted from 2013 street network datasets provided by territorial authorities. Residential density will be calculated by dividing the number of dwellings by the residential land area. The number of dwellings will be obtained from the 2006 census data provided at the meshblock level. Residential land area will be derived from 2013 zoning datasets provided by territorial authorities. Land use mix will be calculated using the area of five land use categories (residential, commercial, industrial, open space, other) in an entropy equation.49 Land uses will be determined using 2013 zoning datasets provided by territorial authorities. The raw scores for these three built environment measures will be normalised (converted to deciles) and summed to create a basic walkability index. This basic meshblock level walkability index will only be used in school and participant selection. The GIS-based built environment indices that will be created for each participant and used in analyses are described in a later section.

    The raw scores for these built environment measures will be normalised and summed to create a basic walkability index. Next, the basic walkability index and pre-existing deprivation data (NZ Dep 2006) will be used to classify all Auckland and Wellington urban meshblocks into one of four strata: (1) higher walkable, higher SES; (2) higher walkable, lower SES; (3) lower walkable, higher SES and (4) lower walkable, lower SES. Meshblocks with the top four walkability/SES deciles are classified as higher walkable/SES, and meshblocks with the bottom four walkability/SES deciles are classified as lower walkable/SES. Meshblocks with walkability or SES in deciles 5 and 6 are excluded.

    School selection will be based on convenience and proximity to large numbers of meshblocks in each of the four strata. Within each school, all potential participants will be sampled, regardless of the quadrant they reside in, and for each participant walkability will be calculated: all students will be assigned to the strata of the meshblock they primarily reside in. This procedure will take place prior to the consent process. Adolescents living in one of the four meshblock strata will be invited to participate in the study. Participation in the study will require written, informed consent from a parent or caregiver and written assent from the adolescent. At the time of consent, parents will be asked to rate the importance of a variety of reasons for choosing to live in their neighbourhood. Subsequent schools will be selected on the basis of the quadrants that need to be balanced. In addition to this approach, care will be taken to balance student numbers across the four strata within and across schools. A similar sampling strategy was used in our previous study of the environmental correlates of physical activity in adults; the heterogeneity generated by this technique permitted several meaningful associations to be detected.40 A subsample of approximately 40 participants will be randomly selected from each school for the GPS and interactive mapping measurements.

    Sample size

    In adjusted multilevel models, it has been estimated that a sample of 1600 adolescents recruited from two schools within each stratum (eight in total) would allow the detection of a small effect size (ie, 1.4% of explained outcome variance found in similar studies conducted elsewhere50) with 80% power. The calculated sample size assumes a two-tailed probability level of 5%, a conservative clustering effect equivalent to an intraclass correlation coefficient of 0.10, and a regression model with 25 background covariates explaining 25% of the outcome variance (comparable to what might be expected from the selected variables). With an anticipated sample size of 320 Māori adolescents (our smallest subgroup comparison), the corresponding detectable effect size will be 7% of the explained outcome variance (medium effect size).

    Exposure, outcomes and covariates

    Exposures

    GIS data provide multiple spatially-referenced layers that can be used to create meaningful and objective exposure measures of the built environment. They are used to objectively characterise the built environment surrounding the primary home address of each participant and can be applied across a range of road network buffers (eg, 500, 800, 1000 and 1600 m) in order to evaluate differences between various limits of exposure. Road network buffers can be created to define areas that can be reached on the street network system, but exclude areas that are not accessible due to major barriers (motorway, river and lake). Two main indices, each a composite function of 2–8 other variables, are used to assess physical environmental features: walkability index51 and neighbourhood destination accessibility index (NDAI).52 These are described in more detail below. All exposure measures (table 1) follow the common protocols established for the international IPEN-adolescents collaboration.

    View this table:
    Table 1

    Summary of study exposure, outcomes and covariates

    Detailed walkability

    The detailed walkability index is a summary score of five distinct variables calculated within GIS: net residential density, land use mix, retail density, street connectivity and street discontinuity. This protocol was created for the US-based Neighborhood Quality of Life Study project51 and has subsequently been implemented in the US-based TEAN study,53 the Australian PLACE study49 and all IPEN adult country study sites.54

    Neighbourhood destination accessibility

    Pedestrian access to destinations will be calculated using the NDAI.52 The NDAI is an objective measure of pedestrian access to neighbourhood destinations; it characterises the distribution of urban infrastructure within an 800 m street network distance from the residence. The NDAI has an advantage over most previous area-level measures of the urban environment in that it captures the range and intensity of everyday destinations such as schools, supermarkets and cafes, which may encourage active travel and enhance recreational physical activity at the population level. Also, the NDAI has been specifically designed for the New Zealand environment. The eight domains captured in the NDAI are education, transport, recreation and play, social and cultural, food retail, financial, health and other retail.

    Outcomes

    Physical activity

    Minutes of MVPA will be objectively measured using the hip-mounted triaxial accelerometers (Actigraph GT3X+) over seven consecutive days. The GT3X+ is a small, durable and water-resistant device worn on an elastic belt that records the frequency, duration and intensity of physical activity with a high level of accuracy and precision.55 Participants are asked to wear the Actigraph during all waking hours (except when bathing or swimming) for 7 days; however, at least five complete days (including at least one weekend day) will be required for analysis to ensure reliable estimates of MVPA.56 Consistent with previous research, a valid day will be defined as at least 10 h of data for weekdays and 8 h for weekend days; non-wear time will be defined as 60 min of consecutive zero counts.57 ,58 In addition, each participant will be given a 7-day compliance log to complete daily, which assists with identifying non-wear periods. On collection of the accelerometer, data are downloaded and screened for completeness and possible malfunction using the MeterPlus software (http://www.meterplussoftware.com). Accelerometer count data will be classified into minutes of light, moderate and vigorous activity using thresholds developed by Evenson et al59; these have performed well in a recent comparison of accelerometer count thresholds for youth.60

    Sedentary behaviour

    Minutes of sedentary activity will be objectively assessed using the GT3X+ accelerometer over the 7-day measurement period. The aforementioned cut-points established by Evenson et al59 will be used to define sedentary time (<100 counts/minute).

    Body size

    Height, weight and waist circumference of each participant will be measured by trained field researchers using a stadiometer, calibrated scales and a tape measure. These procedures occur immediately before the researchers distribute the accelerometers; participants wear light clothing and shoes are removed. BMI will be calculated as weight divided by squared height. Participants are classified into weight status categories using age-specific and sex-specific BMI thresholds.61

    Covariates

    Demographics and SES

    Age, sex, ethnicity and SES will be collected from the participants. Consistent with the IPEN-adolescents protocol, household income will be the preferred SES indicator, but the highest level of parental education will be used when income is unavailable.

    Active commuting

    The frequency, distance, duration and mode of all active commuting trips to or from the home address in the previous 6 months will be assessed with the computer-assisted personal interview (CAPI). The recall of each trip will be aided by a basic travel log (time, location and mode of transport only) to be completed nightly with the accelerometer compliance log. Participants will be asked to bring this information with them to the subsequent CAPI.

    Neighbourhood mobility

    The majority of studies investigating the built environment and health have focused exclusively on residential neighbourhoods as a predictor of exposure62 ,63 and overlooked the prospect that a large proportion of activity choices may be influenced by additional environments that are experienced during daily routines. This may reduce the accuracy of environmental exposure assessment64 and introduce errors that may confound research results. It has been suggested that investigating aspects of daily mobility (regular destinations and the movement between them) will be important to enhance the assessment of exposure65 and resolve the Uncertain Geographic Context Problem.66 Using GPS and interactive activity destination questionnaires, we aim to accurately capture the full extent of daily mobility and its mediating built environment effect on health.

    The Visualization and Evaluation of Route Itineraries, Travel Destinations, and Activity Spaces (VERITAS) is a web-based CAPI tool integrating interactive mapping capacities (based on Google Maps) and has the potential to explore destinations inside and outside the residential neighbourhood. VERITAS was initially developed and tested for the RECORD Cohort Study, a major longitudinal study of over 7200 French adults.67–70 The applicability and feasibility of this method to an adolescent population is detailed elsewhere (manuscript under review but available on request). While we will be using GIS to provide an objective assessment of the surrounding environment (ie, exposure measures), VERITAS will allow the research team to search and geolocate participants’ regular destinations (visited within the previous 6 months), activity locations, walking/cycling areas, routes and modes of travel between locations, travel companions, and perceived or experienced neighbourhood boundaries (ie, neighbourhood mobility). The VERITAS programme will run through an internet browser on a laptop computer, and will be designed to automatically upload all participant responses to our secure database when connected to a wireless network. Spatiotemporal data will be collected using the Qstarz BT-Q1000XT GPS receiver (Qstarz International, Taipei, Taiwan), which has been deemed to be one of the more accurate portable GPS receivers on the market.71 The GPS will be worn in a pouch alongside the accelerometer. GPS data will be cleaned, filtered and merged with accelerometer data using the Personal Activity Location Measurement System (PALMS, refer to: https://ucsd-palms-project.wikispaces.com).72 The merged data streams retrieved from PALMS will be disaggregated into discrete trips and imported into ArcGIS for further analysis. Data obtained from GPS and VERITAS differ both temporally (previous 1 week and 6 months, respectively) and spatially (a continuous sequential polyline compared with point data). Although VERITAS will be able to obtain data for extended periods, it lacks the temporal sequence of events available from GPS tracking. However, as short periods of GPS monitoring may not truly represent destinations visited over extended periods, the combination of both has been recommended to create complementary and more robust measures of environmental exposure.69

    The neighbourhood mobility data will allow the demarcation of the territorial range by active travel modes. A spatial ‘polygon’ will be created, consisting of a multisided geometric shape surrounding the home address that connects the various locations to which participants claim to have walked or cycled. The area (m2) within these polygons will be calculated and used to define separate shapes based on the travel modes. In situations where participants walk or cycle to only one location (eg, school), the polygon area will be the distance between the location and home addresses multiplied by 1 m. As with the active commuting assessment, the recall of visited locations and trips will be aided by the travel log that will be completed daily. Finally, using VERITAS, each participant will be able to map their perceived neighbourhood boundary, allowing us to isolate the effects of their self-defined neighbourhood environment on the outcome measures.

    Perceived neighbourhood walkability

    In order to understand the mediating effect of individual perceptions of the neighbourhood on the relationship between the objectively-measured built environment and physical activity behaviour, the Neighbourhood Environment Walkability Scale for Youth (NEWS-Y) will be administered as a self-completion hard copy survey. NEWS-Y is based on the NEWS, which has demonstrated good reliability and validity.73–76 In addition to the GIS-based walkability index variables (residential density, land use mix and street connectivity), NEWS-Y assesses pedestrian/cycle facilities, aesthetics, traffic safety and crime safety. The ten NEWS-Y subscales have acceptable test-retest reliability (ICC 0.56–0.87) and specific subscales were correlated significantly with physical activity for adolescents.77

    Psychosocial indicators

    A small number of psychosocial variables associated with adolescent physical activity will be measured in the study. These include: self-efficacy; perceived barriers to being physically active; family support; and peer support.77 These variables have shown the most consistent psychosocial correlations with adolescent physical activity in the literature.78 Further, by including such items, we are able to examine our findings within a multilevel framework, thereby accounting for and separating the various layers of influence (ie, individual, social and physical environments).79

    Self-reported physical activity, sedentary behaviour and commuting to school

    In addition to perceived neighbourhood walkability and psychosocial indicators, participants will be asked to report on commuting (to and from school, walking and biking, barriers to walking and cycling),80–82 physical activity (at and outside school, places for, barriers in the neighbourhood, decisions about, confidence about, enjoyment of, social support, workout equipment, activity rules and athletic ability)83–85 and sedentary behaviour (during school and weekend days, things in the bedroom and personal electronics).86 The scales have been shown to be reliable and valid in the adolescent population.80–82 ,86

    Weather

    We have previously demonstrated the significant impact of inclement weather conditions on physical activity in New Zealand children.87 To monitor these potential confounding effects, we will obtain hourly rainfall, mean temperature and hours of daylight statistics from the New Zealand Met Service for each data collection day and use these as covariates in the models.

    Procedures

    Data will be collected from participants within the school setting during school hours. During the measurement session, the NEWS-Y77 questionnaire will be administered, anthropometric measures will be taken and accelerometers and compliance logs will be distributed. Text messages will be sent to adolescents/parents before the data collection session as a reminder to attend. A random subsample of 40 adolescents per school will be allocated a GPS receiver to wear in conjunction with the accelerometer, and will complete the VERITAS interview. All participants will be instructed on the correct use, wear-time and care of the equipment. Participants will be issued with a $20 shopping voucher on completion of data collection and return of the monitors and compliance logs.

    Quantitative analyses

    The proposed dataset will have a hierarchical independent variable structure which consists of person-level observations nested within neighbourhoods and schools. The main aim of the study is to examine confounder-adjusted associations of environmental variables with physical activity and body size outcomes. For this purpose, cross-classified (by neighbourhoods and schools) generalised linear mixed models (MGLM) with random intercepts will be used. These can account for multiple sources of dependency (schools and neighbourhoods) and different types of data (eg, continuous or binary) following a normal distribution or other types of distributions (eg, negative binomial, Poisson).88 MGLMs perform well when the number of observations across areas is highly unbalanced,89 which will be relevant to this project as the number of participants may vary substantially across schools and neighbourhoods. Given the relatively small number of strata included in the study, MGLMs will be estimated using Restricted Maximum Likelihood (REML) or Bayesian Markov chain Monte Carlo (MCMC) methods with non-informative priors,90 the latter being appropriate for binary (eg, overweight/obese vs normal weight)90 ,91 or non-normally distributed outcomes.88 The non-linear relationship will be examined using restricted cubic splines.92 A probability level of 0.05 will be adopted.

    Qualitative methodology

    A total of 16 focus groups, with approximately 5–8 participants, will be conducted at eight participating schools. One researcher (VI) will conduct all the focus groups and at least one of the co-researchers will assist. Variability in walkability will be sought by recruiting two schools in Auckland (representing relatively low walkability) and Wellington (relatively high walkability). However, students within focus groups will be selected to represent a range of neighbourhood settings to facilitate discussion on differing experiences of the built environment. Participants will also take part in the quantitative component and complete all data collection. To aid open discussion and allow meaningful comparisons, separate focus groups will be conducted by age, with younger students (approximately 12–14 years) further stratified by sex, and older students (approximately 15–18 years) in mixed sex groups.42 Focus groups will be conducted using 40 min school periods to accommodate school timetables in a semistructured interview. The focus groups are designed to examine the enablers and barriers to being physically active, particularly with regard to active transport, engagement in formal and informal physical activity, safety and social drivers. Researchers will specifically seek discussion on activity within participants’ residential neighbourhood and school environments as well as alternative activity spaces in their everyday lives, including those outside of their geographical suburbs. Maps of local environments to prompt discussion on where youth are active (and where they avoid), types of activity and travel routes will be used. Interviews will be digitally recorded and transcribed by group, with all individual identifying information removed.

    Qualitative analyses

    Initially, two of the researchers will independently read the transcripts, code and extract themes. The themes will be presented to the team. Disagreement will be resolved through discussion and themes will be confirmed. A coding framework will be developed using NVivo software to organise data generated by the project research questions (deductive) and emergent topics (inductive).93 Analyses will be conducted across and within groups to examine commonalities and differences by built environment settings and individual factors (ie, sex, age and culture). Concurrent analyses of qualitative and quantitative data will allow insightful integration and triangulation of findings across the study components, allowing us to draw inferences about how youth interact with and manage their lived environments, and what that means for their physical activity and well-being.94 ,95

    Ethics and dissemination

    All adolescents will be required to provide assent to participate in the study. An information sheet will be designed specifically for adolescents in a manner that will be easy to understand. In addition, all parents of the assenting adolescents will be required to provide parental consent. Parents will also receive a detailed information sheet outlining the study and its requirements.

    Data will be entered and stored into a secure (password protected) database. Only the named researchers will have access to the data. Data will be stored for 10 years and permanently destroyed thereafter.

    It is unlikely that participants will experience discomfort or embarrassment during data collection. However, as body measures of weight and height will be assessed objectively, there is the potential of concern around body weight and size. The institution's counselling services will be accessed if a situation arises. All body measurements will be taken behind a portable screen with gender appropriate research officers. All data will be kept private and confidential.

    At the completion of the study, results will be provided to key stakeholders and organisations (eg, high schools, adolescents and parents). Results will be disseminated by means of a written report to schools that have participated in the study. Adolescents and/or their parents/legal guardians will receive a report detailing the individual results collected. Government organisations, health boards and councils will be able to access key findings and recommendations resulting from the project through seminar presentations and report distribution. Research findings will also be circulated to the scientific community in the form of publications in refereed journals.

    Discussion

    We have described the methods for the BEANZ study which seeks to estimate strengths of association between objective measures of the local environment with accelerometer-derived and self-reported physical activity and sedentary behaviour in youth. A novel aspect of this study is the exploration of detailed and multilevel relationships of interaction between the social and physical environments specific to the NZ adolescents. This will be achieved through additional measures (eg, GPS, VERITAS, focus groups, NDAI) which collectively serve to advance knowledge in this important area of health research, policy advocacy and ultimately youth health outcomes. In particular, the use of GPS/VERITAS to identify the locations that adolescents visit on a daily basis, defining their geographical context, will provide us with accurate estimates of location in which physical activity takes place.

    International evidence shows that the most consistent environmental attributes positively associated with reported physical activity in youth were land use mix and residential density, but inconsistent findings have been observed for parks, recreation facilities and street connectivity.41 Others found that proximity to parks, recreation facilities and proximity to school32 ,42 ,96 ,97 along with transport infrastructure were positively associated with physical activity in adolescents.97 Traffic hazards (number of roads to cross, traffic speed) and local conditions (crime, area deprivation) were negatively associated with physical activity.43 ,97 Obesogenic environmental attributes of homes, neighbourhoods and schools are believed to promote sedentary behaviour among youth,98 and there is growing evidence that being socially connected with others contributes to adolescent wellbeing.45 While some evidence exists to show the importance of the built environment for adolescent physical activity and wellbeing, the use of different methods and limited physical variability within any given environment may serve to consistently underestimate the associations observed. In this study, variance is maximised in two ways. Two major cities in New Zealand are sampled, and these data are subsequently combined with nine other countries through the IPEN-adolescent study. The larger study will improve our understanding of the nature of the relationships that exist between adolescent physical activity, sedentary behaviour and body weight with specific features of the built environment related to walkability, commuting and access to facilities for recreation.

    Individuals (or at least parents) may self-select neighbourhoods; therefore, associations between built environment and walkability may in part be a reflection of neighbourhood self-selection bias. Mixed results have been found when investigating neighbourhood self-selection and walkability.99–103 The relationship is a complex one and prospective studies are needed to study the effects of neighbourhood self-selection on neighbourhood walkability. When reviewing 38 empirical studies that used different approaches to explore the influence of self-selection, Cao et al104 established that all studies reviewed found a statistically significant influence of the built environment after accounting for self-selection. While exploring this particular relationship is not the focus of the present study, parents’ neighbourhood preference and self-selection will be accounted for in the analysis. As aforementioned earlier, parents will be asked to rank the importance of a variety of reasons for choosing to reside in the particular neighbourhood. The reasons (that address self-selection) could be: easy access to services, walkable environment and/or access to recreational and sporting facilities. This information will be used in the analysis.

    When conducting spatial analyses on aggregated data, errors affecting the validity of results may be introduced.105 The problem has been referred to as the Modifiable Areal Unit Problem defined as the ‘geographic manifestation of the ecological fallacy in which conclusions based on data aggregated to a particular set of districts may change if one aggregates the same underlying data to a different set of districts’.106 In other words, the way spatial data are aggregated may result in different findings. There has been disagreement in the literature on the best solution for this problem; however, it has been suggested that the only appropriate resolution is to use individual-level data that are geocoded based on residential location.107 Indeed, our selection strategy uses geocoded data and we are employing techniques (GPS and VERITAS) to gain a more accurate understanding of neighbourhood boundaries for youth. This will substantially advance our knowledge in this field.

    This study will contribute to national and international scientific knowledge by forming the NZ arm of the international IPEN-adolescents collaboration, whereby adolescents’ physical activity and sedentary behaviour data are collected using a common methodology across multiple countries (Australia, Belgium, Brazil, China, Denmark, Malaysia, New Zealand, Portugal and the USA). Furthermore, the larger study will improve our understanding of the nature of the relationships that exist between adolescent physical activity, sedentary behaviour and body weight with specific features of the built environment related to walkability, commuting and access to facilities for recreation.

    Ultimately, by showing the relationships between health outcomes and the neighbourhood built environment, we aim to influence and inform policy and city planning practices. City planners, policy makers and government agencies will be engaged early.108 ,109 Results will also be shared with other sustainable transport advocacy, urban planners and public health organisations. Dissemination of findings to NZ secondary schools and students themselves will maximise the potential impact of the findings.

    Acknowledgments

    The BEANZ study received a 3-year research grant from the Health Research Council of New Zealand (12/329).

    References

    1. ↵
      Department of health physical Activity Health Improvement and Protection. At least five a week: evidence on the impact of physical activity and its relationship to health. A report from the chief medical officer. London: Department of Health, 2004.
    2. ↵
      Department of Health Physical Activity Health Improvement and Protection. Start active, stay active: a report on physical activity for health from the four home countries’ Chief Medical Officers. London: Department of Health, Physical Activity, Health Improvement and Protection, 2011.
    3. ↵
      Ministry of Health. DHB toolkit: physical activity. Wellington: Ministry of Health, 2003.
    4. ↵
      US Department of Health and Human Services. Physical activity and health: a report of the surgeon general. Atlanta, GA: Center for Disease Control and Prevention, 2002.
    5. ↵
      World Health Organization. Global recommendations on physical activity for health. Geneva: World Health Organization, 2010.
    6. ↵
      1. Janssen I,
      2. LeBlanc AG
      . Review Systematic review of the health benefits of physical activity and fitness in school-aged children and youth. Int J Behav Nutr Phys Act 2010;7:40.
      OpenUrlCrossRefPubMed
    7. ↵
      1. Strong W,
      2. Malina R,
      3. Blimkie C,
      4. et al
      . Evidence based physical activity for school aged youth. J Pediatr 2005;146:732–7.
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      1. Brownson RC,
      2. Boehmer TK,
      3. Luke DA
      . Declining rates of physical activity in the United States: what are the contributors? Annu Rev Public Health 2005;26:421–43.
      OpenUrlCrossRefPubMedWeb of Science
    9. ↵
      1. Dumith SC,
      2. Gigante DP,
      3. Domingues MR,
      4. et al
      . Physical activity change during adolescence: a systematic review and a pooled analysis. Int J Epidemiol 2011;40:685–98.
      OpenUrlAbstract/FREE Full Text
    10. ↵
      1. Matthews CE,
      2. Chen KY,
      3. Freedson PS,
      4. et al
      . Amount of time spent in sedentary behaviors in the United States, 2003–2004. Am J Epidemiol 2008;167:875–81.
      OpenUrlAbstract/FREE Full Text
    11. ↵
      1. Mitchell J,
      2. Pate R,
      3. Beets M,
      4. et al
      . Time spent in sedentary behavior and changes in childhood BMI: a longitudinal study from ages 9 to 15 years. Int J Obes 2012;37:54–60.
      OpenUrl
    12. ↵
      1. Maddison R,
      2. Turley M,
      3. Legge N,
      4. et al
      . A National Survey of Children and Young People's Physical Activity and Dietary Behaviours in New Zealand: 2008/09. Auckland: Clinical Trials Research Unit, The University of Auckland, 2010.
    13. ↵
      1. Hamilton MT,
      2. Healy GN,
      3. Dunstan DW,
      4. et al
      . Too little exercise and too much sitting: Inactivity physiology and the need for new recommendations on sedentary behavior. Curr Cardiovasc Risk Rep 2008;2:292–8.
      OpenUrlCrossRefPubMed
    14. ↵
      1. Hamilton M,
      2. Hamilton D,
      3. Zderic W
      . Exercise physiology versus inactivity physiology: an essential concept for understanding lipoprotein lipase regulation. Exerc Sport Sci Rev 2004;32:161–6.
      OpenUrlCrossRefPubMedWeb of Science
    15. ↵
      1. Hamilton MT,
      2. Hamilton DG,
      3. Zderic TW
      . Role of low energy expenditure and sitting in obesity, metabolic syndrome, type 2 diabetes, and cardiovascular disease. Diabetes 2007;56:2655–67.
      OpenUrlAbstract/FREE Full Text
    16. ↵
      1. Telama R,
      2. Yang X,
      3. Viikari J,
      4. et al
      . Physical activity from childhood to adulthood: a 21-year tracking study. Am J Prev Med 2005;28:267–73.
      OpenUrlCrossRefPubMedWeb of Science
    17. ↵
      1. Hancox RJ,
      2. Milne BJ,
      3. Poulton R
      . Association between child and adolescent television viewing and adult health: a longitudinal birth cohort study. Lancet 2004;364:257–62.
      OpenUrlCrossRefPubMedWeb of Science
    18. ↵
      1. Reilly JJ,
      2. McDowell ZC
      . Physical activity interventions in the prevention and treatment of paediatric obesity: systematic review and critical appraisal. Proc Nutr Soc 2003;62:611–20.
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      1. Dobbins M,
      2. De Corby K,
      3. Robeson P,
      4. et al
      . School-based physical activity programs for promoting physical activity and fitness in children and adolescents aged 6–18. Cochrane Database Syst Rev 2009;1:1–106.
      OpenUrl
    20. ↵
      1. Summerbell C,
      2. Waters E,
      3. Edmunds L,
      4. et al
      . Interventions for preventing obesity in children. Cochrane Database Syst Rev 2005;3:1–86.
      OpenUrl
    21. ↵
      1. King AC,
      2. Stokols D,
      3. Talen E,
      4. et al
      . Theoretical approaches to the promotion of physical activity: forging a transdisciplinary paradigm. Am J Prev Med 2002;23(2S):15–25.
      OpenUrlCrossRefPubMedWeb of Science
    22. ↵
      1. Giles-Corti B,
      2. Timperio A,
      3. Bull F,
      4. et al
      . Understanding physical activity environmental correlates: increased specificity for ecological models. Exerc Sport Sci Rev 2005;33:175–81.
      OpenUrlCrossRefPubMedWeb of Science
    23. ↵
      1. Hinckson EA,
      2. Garrett N,
      3. Duncan JS
      . Active commuting to school in New Zealand Children (2004-2008): a quantitative analysis. Prev Med 2011;52:332–6.
      OpenUrlCrossRefPubMed
    24. ↵
      1. Witten K,
      2. Blakely T,
      3. Bagheri N,
      4. et al
      . Neighborhood built environment and transport and leisure physical activity: findings using objective exposure and outcome measures in New Zealand. Environ Health Perspect 2012;120:971–7.
      OpenUrlCrossRefPubMedWeb of Science
    25. ↵
      1. Conn VS,
      2. Hafdahl AR,
      3. Mehr DR
      . Interventions to increase physical activity among healthy adults: meta-analysis of outcomes. Am J Public Health 2011;101:751–8.
      OpenUrlCrossRefPubMedWeb of Science
    26. ↵
      1. Metcalf B,
      2. Henley W,
      3. Wilkin T
      . Effectiveness of intervention on physical activity of children: systematic review and meta-analysis of controlled trials with objectively measured outcomes (EarlyBird 54). BMJl 2012;345:1–11.
      OpenUrl
    27. ↵
      1. De Bourdeaudhuij I,
      2. Sallis JF,
      3. Saelens BE
      . Environmental correlates of physical activity in a sample of Belgian adults. Am J Health Promot 2003;18:83–92.
      OpenUrlCrossRefPubMedWeb of Science
    28. ↵
      1. Sallis JF,
      2. Bowles HR,
      3. Bauman A,
      4. et al
      . Neighborhood environments and physical activity among adults in 11 countries. Am J Prev Med 2009;36:484–90.
      OpenUrlCrossRefPubMedWeb of Science
    29. ↵
      1. Sallis JF,
      2. Glanz K
      . The role of built environments in physical activity, eating, and obesity in childhood. Future Child 2006;16:89–108.
      OpenUrl
    30. ↵
      1. Ding D,
      2. Sallis JF,
      3. Norman GJ,
      4. et al
      . Community food environment, home food environment, and fruit and vegetable intake of children and adolescents. J Nutr Edu Behav 2011;44:634–8.
      OpenUrl
    31. ↵
      1. Davison KK,
      2. Lawson CT
      . Do attributes in the physical environment influence children's physical activity? A review of the literature. Int J Behav Nutr Phys Act 2006;3:19.
      OpenUrlCrossRefPubMed
    32. ↵
      1. Rodríguez DA,
      2. Cho G-H,
      3. Evenson KR,
      4. et al
      . Out and about: association of the built environment with physical activity behaviors of adolescent females. Health Place 2012;18:55–62.
      OpenUrlPubMed
    33. ↵
      1. McDonald K,
      2. Hearst M,
      3. Farbakhsh K,
      4. et al
      . Adolescent physical activity and the built environment: a latent class analysis approach. Health Place 2012;18:191–8.
      OpenUrlPubMed
    34. ↵
      1. Casey R,
      2. Chaix B,
      3. Weber C,
      4. et al
      . Spatial accessibility to physical activity facilities and to food outlets and overweight in French youth. Int J Obes 2012;36:914–19.
      OpenUrl
    35. ↵
      1. Yin L,
      2. Raja S,
      3. Li X,
      4. et al
      . Neighbourhood for playing: using GPS, GIS and accelerometry to delineate areas within which youth are physically active. Urban Studies 2013:1–18.<PE: References 35 and 42 are same. Please renumber the references.
    36. ↵
      1. Laxer RE,
      2. Janssen I
      . The proportion of youths’ physical inactivity attributable to neighbourhood built environment features. Int J Health Geogr 2013;12:31.
      OpenUrl
    37. ↵
      1. Lake A,
      2. Townshend T
      . Exploring the built environment, physical activity and related behaviours of young people attending school, college and those not in employment. J Public Health 2013;35:57–66.
      OpenUrlAbstract/FREE Full Text
    38. ↵
      1. Kasehagen L,
      2. Busacker A,
      3. Kane D,
      4. et al
      . Associations between neighborhood characteristics and physical activity among youth within rural–urban commuting areas in the US. Matern Child Health J 2012;16:258–67.
      OpenUrlPubMed
    39. ↵
      1. Hobin E,
      2. Leatherdale S,
      3. Manske S,
      4. et al
      . A multilevel examination of factors of the school environment and time spent in moderate to vigorous physical activity among a sample of secondary school students in grades 9–12 in Ontario, Canada. Int J Public Health 2012;57:699–709.
      OpenUrlPubMed
    40. ↵
      1. Badland HM,
      2. Schofield GM,
      3. Witten K,
      4. et al
      . Understanding the relationship between activity and neighbourhoods (URBAN) study: research design and methodology. BMC Public Health 2009;9:224.
      OpenUrlCrossRefPubMed
    41. ↵
      1. Ding D,
      2. Sallis JF,
      3. Kerr J,
      4. et al
      . Neighborhood environment and physical activity among youth: a review. Am J Prev Med 2011;41:442–55.
      OpenUrlCrossRefPubMed
    42. ↵
      1. Slater S,
      2. Fitzgibbon M,
      3. Floyd MF
      . Urban adolescents’ perceptions of their neighborhood physical activity environments. Leisure Sci 2013;35:167–83.
      OpenUrl
    43. ↵
      1. Kann L
      . Obesity and other correlates of physical activity and sedentary behaviors among US high school students. J Obes 2013;2013:276318. http://dx.doi.org.ezproxy.aut.ac.nz/10.1155/2013/276318published
      OpenUrl
    44. ↵
      1. Moore JB,
      2. Brinkley J,
      3. Crawford TW,
      4. et al
      . Association of the built environment with physical activity and adiposity in rural and urban youth. Prev Med 2012;56:145–48.
      OpenUrl
    45. ↵
      1. Jose PE,
      2. Pryor J
      . New Zealand youth benefit from being connected to their family, school, peer group and community. Youth Stud Aust 2010;29:30–7.
      OpenUrl
    46. ↵
      1. Vandenbroucke JP,
      2. Von Elm E,
      3. Altman DG,
      4. et al
      . Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration. PLoS Med 2007;4:e297.
      OpenUrlCrossRefPubMed
    47. ↵
      Statistics New Zealand. Subnational population estimates at 30 June 2012. Wellington: Statistics New Zealand, 2013.
    48. ↵
      Demographia. Demographia World Urban Areas: 9th Annual Edition (2013.03). Belleville, IL: Demographia, 2013.
    49. ↵
      1. Leslie E,
      2. Coffee N,
      3. Frank L,
      4. et al
      . Walkability of local communities: using geographic information systems to objectively assess relevant environmental attributes. Health Place 2007;13:111–22.
      OpenUrlCrossRefPubMedWeb of Science
    50. ↵
      1. Frank L,
      2. Kerr J,
      3. Chapman J,
      4. et al
      . Urban form relationships with walk trip frequency and distance among youth. Am J Health Promot 2007;21:S1–7.
      OpenUrl
    51. ↵
      1. Frank LD,
      2. Sallis JF,
      3. Saelens BE,
      4. et al
      . The development of a walkability index: application to the Neighborhood Quality of Life Study. Br J Sports Med 2010;44:924–33.
      OpenUrlAbstract/FREE Full Text
    52. ↵
      1. Witten K,
      2. Pearce J,
      3. Day P
      . Neighbourhood destination accessibility Index: a GIS tool for measuring infrastructure support for neighbourhood physical activity. Environ Plann A 2011;43:205.
      OpenUrl
    53. ↵
      1. Sallis J,
      2. Kerr J,
      3. Conway J
      . TEAN (Teen Environment and Neighborhood) study (Unpublished). San Diego: Unversity of California San Diego, 2007–2011.
    54. ↵
      1. Kerr J,
      2. Sallis JF,
      3. Owen N,
      4. et al
      . Advancing science and policy through a coordinated International study of physical activity and built environments: IPEN methods. J Phys Act Health 2012;10:581–601.
      OpenUrl
    55. ↵
      1. de Vries SI,
      2. Bakker I,
      3. Hopman-Rock M,
      4. et al
      . Clinimetric review of motion sensors in children and adolescents. J Clin Epidemiol 2006;59:670–80.
      OpenUrlCrossRefPubMed
    56. ↵
      1. Trost SG,
      2. Pate RR,
      3. Freedson PS,
      4. et al
      . Using objective physical activity measures with youth: how many days of monitoring are needed?. Med Sci Sports Exerc 2000;32:426–31.
      OpenUrlCrossRefPubMedWeb of Science
    57. ↵
      1. Oliver M,
      2. Witten K,
      3. Kearns R,
      4. et al
      . Kids in the City Study: research design and methodology. BMC Public Health 2011;11:587.
      OpenUrlPubMed
    58. ↵
      1. Choi L,
      2. Liu Z,
      3. Matthews CE,
      4. et al
      . Validation of accelerometer wear and nonwear time classification algorithm. Med Sci Sports Exerc 2011;43:357–64.
      OpenUrlCrossRefPubMedWeb of Science
    59. ↵
      1. Evenson KR,
      2. Catellier DJ,
      3. Gill K,
      4. et al
      . Calibration of two objective measures of physical activity for children. J Sports Sci 2008;26:1557–65.
      OpenUrlCrossRefPubMedWeb of Science
    60. ↵
      1. Trost SG,
      2. Loprinzi PD,
      3. Moore R,
      4. et al
      . Comparison of accelerometer cut points for predicting activity intensity in youth. Med Sci Sports Exerc 2011;43:1360–8.
      OpenUrlCrossRefPubMedWeb of Science
    61. ↵
      1. Cole T,
      2. Flegal KM,
      3. Nicholls D,
      4. et al
      . Body mass index cut offs to define thinness in children and adolescents: international survey. BMJ 2007;335:194.
      OpenUrlAbstract/FREE Full Text
    62. ↵
      1. Inagami S,
      2. Cohen DA,
      3. Finch BK
      . Non-residential neighborhood exposures suppress neighborhood effects on self-rated health. Soc Sci Med 2007;65:1779–91.
      OpenUrlCrossRefPubMedWeb of Science
    63. ↵
      1. Cummins S
      . Commentary: investigating neighbourhood effects on health—avoiding the ‘local trap’. Int J Epidemiol 2007;36:355–7.
      OpenUrlFREE Full Text
    64. ↵
      1. Kwan M-P
      . The uncertain geographic context problem. Ann Assoc Am Geogr 2012;102:958–68.
      OpenUrlCrossRefWeb of Science
    65. ↵
      1. Chaix B,
      2. Méline J,
      3. Duncan S,
      4. et al
      . GPS tracking in neighborhood and health studies: A step forward for environmental exposure assessment, a step backward for causal inference? Health Place 2013;21:46–51.
      OpenUrl
    66. ↵
      1. Golledge RG,
      2. Stimson RJ
      . Spatial behavoir: a geographical perspective. New York: Guilford Press, 1997.
    67. ↵
      1. Chaix B,
      2. Bean K,
      3. Leal C,
      4. et al
      . Individual/neighborhood social factors and blood pressure in the RECORD cohort study which risk factors explain the associations? Hypertension 2010;55:769–75.
      OpenUrlCrossRef
    68. ↵
      1. Chaix B,
      2. Billaudeau N,
      3. Thomas F,
      4. et al
      . Neighborhood effects on health: correcting bias from neighborhood effects on participation. Epidemiology 2011;22:18–26.
      OpenUrlCrossRefPubMedWeb of Science
    69. ↵
      1. Chaix B,
      2. Kestens Y,
      3. Perchoux C,
      4. et al
      . An interactive mapping tool to assess individual mobility patterns in neighborhood studies. Am J Prev Med 2012;43:440–50.
      OpenUrlCrossRefPubMed
    70. ↵
      1. Havard S,
      2. Reich BJ,
      3. Bean K,
      4. et al
      . Social inequalities in residential exposure to road traffic noise: an environmental justice analysis based on the RECORD Cohort Study. Occup Environ Med 2011;68:366–74.
      OpenUrlAbstract/FREE Full Text
    71. ↵
      1. Duncan S,
      2. Stewart TI,
      3. Oliver M,
      4. et al
      . Portable Global Positioning System receivers: static validity and environmental conditions. Am J Prev Med 2013;44:e19–29.
      OpenUrl
    72. ↵
      1. Kerr J,
      2. Duncan S,
      3. Schipperjin J
      . Using Global Positioning Systems in health research: a practical approach to data collection and processing. Am J Prev Med 2011;41:532–40.
      OpenUrlCrossRefPubMed
    73. ↵
      1. Brownson RC,
      2. Hoehner CM,
      3. Day K,
      4. et al
      . Measuring the built environment for physical activity: State of the science. Am J Prev Med 2009;36:S99–S123. e12.
      OpenUrlCrossRefPubMedWeb of Science
    74. ↵
      1. Leslie E,
      2. Saelens BE,
      3. Frank LD,
      4. et al
      . Residents’ perceptions of walkability attributes in objectively different neighbourhoods: A pilot study. Health Place 2005;11:227–36.
      OpenUrlCrossRefPubMedWeb of Science
    75. ↵
      1. Saelens BE,
      2. Sallis JF,
      3. Black JB,
      4. et al
      . Neighbourhood-based differences in physical activity: an environment scale evaluation. Am J Public Health 2003;93:1552–58.
      OpenUrlCrossRefPubMedWeb of Science
    76. ↵
      1. Cerin E,
      2. Conway TL,
      3. Cain KL,
      4. et al
      . Sharing good NEWS across the world: developing comparable scores across 12 countries for the neighborhood environment walkability scale (NEWS). BMC Public Health 2013;13:309.
      OpenUrl
    77. ↵
      1. Rosenberg D,
      2. Ding D,
      3. Sallis JF,
      4. et al
      . Neighborhood Environment Walkability Scale for Youth (NEWS-Y): reliability and relationship with physical activity. Prev Med 2009;49:213–18.
      OpenUrlCrossRefPubMed
    78. ↵
      1. Van Der Horst K,
      2. Paw MJCA,
      3. Twisk JWR,
      4. et al
      . A brief review on correlates of physical activity and sedentariness in youth. Med Sci Sports Exerc 2007;39:1241–50.
      OpenUrlCrossRefPubMedWeb of Science
    79. ↵
      1. Norman GJ,
      2. Schmid BA,
      3. Sallis JF,
      4. et al
      . Psychosocial and environmental correlates of adolescent sedentary behaviors. Pediatrics 2005;116:908.
      OpenUrlAbstract/FREE Full Text
    80. ↵
      1. Grow HM,
      2. Saelens BE,
      3. Kerr J,
      4. et al
      . Where are youth active? Roles of proximity, active transport, and built environment. Med Sci Sports Exerc 2008;40:2071.
      OpenUrlCrossRefPubMedWeb of Science
    81. ↵
      1. Sallis J
      . The active where? Questionnaire 2005. http://www.drjamessallis.sdsu.edu/measures.html.
    82. ↵
      1. Frank L,
      2. Leerssen C,
      3. Chapman J,
      4. et al
      . Strategies for metropolitan Atlanta's regional transportation and air quality (SMARTRAQ). Atlanta, GA: Georgia Institute of Technology, 2001.
    83. ↵
      1. Prochaska JJ,
      2. Sallis JF,
      3. Long B
      . A physical activity screening measure for use with adolescents in primary care. Arch Pediatr Adoles Med 2001;155:554–9.
      OpenUrlCrossRefPubMedWeb of Science
    84. ↵
      1. Norman GJ,
      2. Sallis JF,
      3. Gaskins R
      . Comparability and reliability of paper-and computer-based measures of psychosocial constructs for adolescent physical activity and sedentary behaviors. Res Q Exerc Sport 2005;76:315–23.
      OpenUrlCrossRefPubMedWeb of Science
    85. ↵
      1. Sallis JF,
      2. Nader PR,
      3. Broyles SL,
      4. et al
      . Correlates of physical activity at home in Mexican-American and Anglo-American preschool children. Health Psychol 1993;12:390–8.
      OpenUrlCrossRefPubMedWeb of Science
    86. ↵
      1. Marshall S,
      2. Biddle S,
      3. Sallis J,
      4. et al
      . Clustering of sedentary behaviours and physical activity among youth: a cross-national study. Pediatr Exerc Sci 2002;14:410–17.
      OpenUrl
    87. ↵
      1. Duncan JS,
      2. Hopkins WG,
      3. Schofield G,
      4. et al
      . Effects of weather on pedometer-determined physical activity in children. Med Sci Sports Exerc 2008;40:1432.
      OpenUrlCrossRefPubMedWeb of Science
    88. ↵
      1. Goldstein H,
      2. Browne W,
      3. Rasbash J
      . Multilevel modelling of medical data. Stat Med 2002;21:3291–315.
      OpenUrlCrossRefPubMedWeb of Science
    89. ↵
      1. Burton P,
      2. Gurrin L,
      3. Sly P
      . Tutorial in biostatistics. Extending the simple linear regression model to account for correlated responses: an introduction to generalized estimating equations and multi-level mixed modeling. Stat Med 1998;17:1261–91.
      OpenUrlCrossRefPubMedWeb of Science
    90. ↵
      1. Browne WJ,
      2. Draper D
      . A comparison of Bayesian and likelihood-based methods for fitting multilevel models. Bayesian Anal 2006;1:473–514.
      OpenUrlCrossRef
    91. ↵
      1. Austin PC
      . Estimating Multilevel Logistic Regression Models when the number of clusters is low: a comparison of different statistical software procedures. Int J Biostats 2010;6:1557–4679.
      OpenUrl
    92. ↵
      1. Marrie RA,
      2. Dawson NV,
      3. Garland A
      . Quantile regression and restricted cubic splines are useful for exploring relationships between continuous variables. J Clin Epidemiol 2009;62: 511–17. e1.
      OpenUrlCrossRefPubMed
    93. ↵
      1. Attride-Stirling J
      . Thematic networks: an analytic tool for qualitative research. Qual Res 2001;1:385–405.
      OpenUrlAbstract/FREE Full Text
    94. ↵
      1. Östlund U,
      2. Kidd L,
      3. Wengström Y,
      4. et al
      . Combining qualitative and quantitative research within mixed method research designs: a methodological review. Int J Nurs Stud 2011;48:369.
      OpenUrlCrossRefPubMed
    95. ↵
      1. Tashakkori A,
      2. Teddlie C
      . Handbook of mixed methods in social & behavioral research. Sage Publications, Inc, 2003.
    96. ↵
      1. Cohen DA,
      2. Ashwood S,
      3. Scott M,
      4. et al
      . Proximity to school and physical activity among middle school girls: the trial of activity for adolescent girls study. Pediatrics 2006;118:e1381.
      OpenUrlAbstract/FREE Full Text
    97. ↵
      1. Davison K,
      2. Lawson C
      . Do attributes in the physical environment influence children's physical activity? A review of the literature. Int J Behav Nutr Phys Act 2006;3:19.
      OpenUrlCrossRefPubMed
    98. ↵
      1. Giles-Corti B,
      2. Kelty SF,
      3. Zubrick SR,
      4. et al
      . Encouraging walking for transport and physical activity in children and adolescents. Sports Med 2009;39:995–1009.
      OpenUrlCrossRefPubMedWeb of Science
    99. ↵
      1. Owen N,
      2. Cerin E,
      3. Leslie E,
      4. et al
      . Neighborhood walkability and the walking behavior of Australian adults. Am J Prev Med 2007;33:387–95.
      OpenUrlCrossRefPubMedWeb of Science
    100. ↵
      1. Boarnet MG,
      2. Sarmiento S
      . Can land-use policy really affect travel behaviour? A study of the link between non-work travel and land-use characteristics. Urban Studies 1998;35:1155–69.
      OpenUrlAbstract/FREE Full Text
    101. ↵
      1. Cao X,
      2. Handy SL,
      3. Mokhtarian PL
      . The influences of the built environment and residential self-selection on pedestrian behavior: evidence from Austin, TX. Transportation 2006;33:1–20.
      OpenUrlCrossRefWeb of Science
    102. ↵
      1. Handy S,
      2. Cao X,
      3. Mokhtarian PL
      . Self-selection in the relationship between the built environment and walking: Empirical evidence from Northern California. J Am Plann Assoc 2006;72:55–74.
      OpenUrlCrossRefWeb of Science
    103. ↵
      1. Frank LD,
      2. Saelens BE,
      3. Powell KE,
      4. et al
      . Stepping towards causation: do built environments or neighborhood and travel preferences explain physical activity, driving, and obesity? Soc Sci Med 2007;65:1898–914.
      OpenUrlCrossRefPubMedWeb of Science
    104. ↵
      1. Cao X,
      2. Mokhtarian PL,
      3. Handy SL
      . Examining the impacts of residential self-selection on travel behaviour: a focus on empirical findings. Transp Rev 2009;29:359–95.
      OpenUrl
    105. ↵
      1. Yang T-C
      . Modifiable areal unit problem. GIS Resource Document 2005.
    106. ↵
      1. Waller LA,
      2. Gotway CA
      . Applied spatial statistics for public health data. John Wiley & Sons, 2004.
    107. ↵
      1. Weeks JR
      . The role of spatial analysis in demographic research. Spatially Integr Soc Sci 2004:381–99.
    108. ↵
      1. McCann B
      . Making physical activity research relevant to policy makers. J Phys Act Health 2006;3(Supp1):S267–72.
      OpenUrl
    109. ↵
      1. Brownson RC,
      2. Jones E
      . Bridging the gap: translating research into policy and practice. Prev Med 2009;49:313–15.
      OpenUrlCrossRefPubMed

    Footnotes

    • Contributors EH developed the first draft of the manuscript. EH, SD, MO, SM, EC, HB, VI, JM, and GS contributed to the conception and design of the study. All authors provided feedback during manuscript development. Each author has read and approved the final manuscript.

    • Competing interests None.

    • Patient consent Obtained.

    • Ethics approval Ethical approval was received by the Institution's Ethics Committee (AUTEC, 12/161).

    • Provenance and peer review Not commissioned; externally peer reviewed.