Active design of built environments for increasing levels of physical activity in adults: the ENABLE London natural experiment study

Physical activity level/pattern and location

Objectively measured physical activity (daily steps) was the primary outcome and was assessed over a consecutive 7-day period using hip-mounted ActiGraph GT3X+ accelerometers, combined with an assessment of physical activity location using Global Positioning System (GPS) travel recorders (BT-1000XT). Accelerometers provided daily measures of steps and different intensities of activity, including moderate to vigorous physical activity (MVPA) (both overall and in 10-minute bouts, in accordance with UK physical activity recommendations, although levels in bouts were too low to be usefully used in analyses). 5 The approaches to analysing ActiGraph data are detailed in Chapter 3. Simultaneous use of ActiGraph accelerometers and GPS travel recorders allowed active travel components of physical activity (i.e. walking), as well as indoor and outdoor activities, to be identified using methods previously described by the investigators48,49 and the methods detailed in Chapter 5. In addition, GPS data allowed the geographical location at which different levels of physical activity occurred to be identified (from sedentary to vigorous, using established cut-off points in accelerometer data), both at baseline and at follow-up. Together, these measures allowed accelerometry data to be interpreted in depth, allowing the nature and location of recorded activities, particularly active forms of transport such as walking and cycling, to be identified. Moreover, it allowed the contribution of active transport local to place of residence to be quantified and compared between those living in East Village and control areas (see Chapter 5).

Environmental exposures

Geographical information systems (GIS) were used to extract objective data on features of the local environment (see Chapter 4). In combination with ActiGraph and GPS data from study participants, this allowed the location of different levels of physical activity (including both high and low levels of activity) to be accurately identified. This method has been used previously by the investigators to establish the important contribution of walking to school and location (including land use type) to MVPA levels in children. 49,50 A number of data sources were used to identify environmental and activity-permissive features within East Village and control areas, including Ordnance Survey (OS) MasterMap Tomography Layer (versions December 2018, May 2014, June 2015) (Ordnance Survey, URL: www.ordnancesurvey.co.uk/business-government/products/mastermap-topography; accessed September 2019), Integrated Transport Network and Transport for London (TfL)51 sources, Olympic Delivery Authority and local authority data, as well as other online resources. In particular, OS data were used to derive indices, such as land use mix, street connectivity, residential density, walkability and connectivity indices, including walking distance to particular features of the built environment, including green space. 52

Anthropometric measurements

Height was measured to the last complete millimetre with a portable stadiometer (Leicester Stadiometer; Seca, Birmingham, UK) at baseline and follow-up. Both weight and leg-to-leg bioimpedance were assessed using an electronic Tanita SC-240 body composition analyser (Tanita Inc., Tokyo, Japan) to provide measures of fat mass (kg) and fat-free mass (kg); body mass index (BMI) was calculated as weight/height² (kg/m²). In total, eight Leicester Stadiometers and Tanita SC-240 body composition analysers were used to measure the participants. The Tanita devices were operated using factory default settings and were regularly checked in accordance with recommended review procedures.

Questionnaire data

Questionnaires were converted into electronic format using SNAP Surveys software (version 11, SNAP Surveys, London, UK), and completed by study participants using dedicated laptops. The questionnaires used established validated methodologies to collect detailed information on patterns and types of activity that were local to the place of residence. In particular, the ‘Neighbourhood Physical Activity Questionnaire’ provided data to examine walking within the neighbourhood,53 and the ‘Neighbourhood Environment Walking Scale’ provided data on perceptions of the neighbourhood environment. 54,55 Information on self-defined ethnic origin (based on the 2011 Census45) and a range of social markers were recorded (including employment status, income, duration and location of work), together with home address and postcode of residence, allowing GIS-determined distance to local amenities to be measured. Questions about general health/health status,45 well-being, anxiety and depression, including both clinical and subclinical forms of assessment suitable for use in community settings, were also used (see Chapter 6). 22,56–58 Physical activity was assessed using an adaptation of the short-form, self-reported International Physical Activity Questionnaire (IPAQ)59 to provide perceived levels of physical activity in addition to objective measures. Adults were asked about their attitudes to physical activities (including both sedentary, such as screen time, and physically active forms) and factors that influence their physical activity behaviour. Participants were asked about perceived personal, social and environmental influences on physical activity, their use of recreational space (particularly walkways and cycle paths) and facilities in their residential neighbourhood (including costs incurred). Participants were also asked about the availability, accessibility (method of travel and journey times) and usage of local amenities (walkways, cycle paths, parks, swimming pools, etc.); their perceptions of the safety of these amenities and the degree to which they permit their child independent or supervised use were also elicited. The questionnaire also included sections to ascertain levels of social participation, support, cohesion and trust. 60 These items were particularly relevant to gauge how the use and perceptions of the local area by others had an impact on individual use and how this might differ from objectively measured features of their neighbourhood. The main questionnaire is included in Appendix 1. Questions were repeated at follow-up, but further questions were included to capture use and frequency of use of the Olympic Park and shopping facilities in the East Village area. Additionally, given the 2-year follow-up, a further question was included to establish whether or not participants had ever lived in East Village between baseline and follow-up interviews.

Qualitative data

In addition to the rich quantitative data, focus groups among study participants in the social housing sector, who were the first to move into East Village, were carried out to identify issues of importance, particularly about perceptions and use of their local environment, and to confirm whether or not the ENABLE London study questionnaire was capturing a suitable and accurate insight of constructs that would identify environmental factors influencing health behaviour (see Chapter 7). GIS and GPS data were also combined with qualitative spatial narratives among study participants to explore differences in active travel associated with moving to East Village. These narratives used individual participant maps to assess how the built environment influences travel behaviour and physical activity: examining change in travel patterns, reasons for the change, purpose of travel and choice of route (see Chapter 8).

Sample recruited and examined

A total of 1819 households (749 social, 738 intermediate and 332 market-rent) expressed an interest in taking part in the study, of which 1006 (392 social, 421 intermediate and 193 market-rent) were examined at baseline from January 2013 to January 2016 (Figure 1). More than one person per household was invited to take part. In total, 1497 individuals took part: 1278 adults (520 social, 524 intermediate and 234 market-rent) and 219 children (209 social, eight intermediate and two market-rent). Figure 2 shows the geographic home locations of study participants at baseline, which highlights the Newham focus of those in social housing, and Greater London geographic diversity of participants seeking intermediate and market-rent accommodation. The small number of children recruited was not expected given that the accommodation in East Village was largely designed for family occupation. Most of the children recruited (95%) were from the social housing sector, and very few children were recruited from intermediate and market-rent groups, which were largely adult-only households. The overall number of children was too small to provide sufficient power to detect change in physical activity; therefore, only adults were considered from here on. The smaller number of adult participants recruited from the market-rent sector reflected the limitations placed on the extent and duration of access to applicants seeking this type of accommodation in East Village. The ENABLE London cohort was predicated on recruiting 1200 adults from 1200 households and succeeded in recruiting 1278 adults from 1006 households. Given the modest imbalance between movers and non-movers, the compliance and the follow-up rate observed, the study was powered to detect a 750-step change [0.3 standard deviation (SD)] at 90% power and with a probability of 0.01 among those who move to East Village. 61

FIGURE 1.

Flow diagram of households and participants examined at baseline and follow-up in the ENABLE London study. Reproduced from Ram et al. 61 Published by the BMJ Publishing Group Limited. This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/.

Reproduced from Ram et al. 61 Published by the BMJ Publishing Group Limited. This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/

FIGURE 2.

Baseline locations of social, intermediate and market-rent households participating in the ENABLE London study. Reproduced from Ram et al. 61 Published by the BMJ Publishing Group Limited. This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/.

Reproduced from Ram et al. 61 Published by the BMJ Publishing Group Limited. This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/

The characteristics of the baseline sample by housing sector are summarised in Table 1. Participants seeking social housing in East Village were older, mainly female, and more likely to be from ethnic minority groups than the intermediate and market-rent sectors. The social housing sector was largely residing in households of four or more people (58% vs. intermediate 29% and market-rent 27%), which included children (82% vs. intermediate 18% and market-rent 10%). Only one-quarter (24%) of the social housing sector had a degree qualification compared with 82% of intermediate and 80% of market-rent participants. Half of the social housing sector were employed, compared with 94% of the intermediate and 87% of the market-rent participants. There were also social gradients in the type of employment, with 12% of the social housing sector participants in higher managerial or professional occupations compared with 72% of the intermediate sector participants and 66% of the market-rent sector participants. Fewer participants in the social housing sector reported very good or good overall health (76%) than did intermediate (88%) and market-rent (91%) participants.

Follow-up participation rates

Of the 1278 adult participants from 1006 households examined at baseline, 877 (69%) adults from 710 (71%) households were examined at follow-up (see Figure 1). A total of 401 adult participants seen at baseline were not followed up.

Participants

Examining Neighbourhood Activities in Built Living Environments in London (ENABLE London) study participants were recruited from those seeking or who had applied for new accommodation in East Village. There were three types of housing tenure being sought based on level of income: social (provided by the local authority or housing association at subsidised rates), intermediate (a mixture of shared ownership, shared equity and affordable rent) and market-rent. The inclusion criteria was broad and included anyone interested/applying for single or multiple occupancy accommodation in East Village. There was no explicit exclusion criteria; adults of any age, gender, ethnic group, with or without handicap, were invited to participate. Current housing status was strongly linked to aspirational housing status; those seeking social accommodation were currently in social housing or on social housing waiting lists, and those seeking intermediate and market-rent accommodation were largely in privately rented housing. Aspirational housing tenure is integral to the design of ENABLE London, and we have shown that this provides a clear socioeconomic marker of study participants. For example, those seeking social housing in East Village are more likely to be unemployed, less educated and more likely to represent ethnic minorities (a classic marker of socioeconomic vulnerability), compared with those seeking affordable and market-rent accommodation (see Chapter 2). 61 We have also shown key differences in mental health and well-being between housing sectors; those seeking social housing were more likely to be depressed, anxious and have poorer well-being than other housing sectors (see Chapter 6). 64 Moreover, this is entirely consistent with earlier studies that found both current housing tenure and aspirational housing tenure are associated with a variety of health outcomes, including mental health and measures of general health. 21,65

Assessments of participants were carried out in their place of residence before any potential move to East Village, and 2 years later (February 2015 to October 2017) for those who had moved to East Village, moved elsewhere or not moved. 66

Measures

Statistical analysis

All analyses were carried out using Stata^® Special Edition version 15 for Windows^® (StataCorp LP, College Station, TX, USA).

To obtain average physical activity variables, we fitted a multilevel linear model for each outcome to allow for repeated measurements of daily physical activity, by fitting participant as a random effect and adjusting for day of the week, day order of recording and month as fixed effects. Raw level one residuals were obtained from the model and a within-person average value of each outcome variable was obtained by averaging these raw residuals. The average of these raw residuals for each participant was added to the sample mean for that particular physical variable to derive an unbiased average level of each physical activity variable at baseline and at follow-up for each person.

At baseline

Overall participation rates and participation rates by housing sectors were outlined earlier (see Chapter 2). For the main outcomes of interest, complete data on adiposity were available for 1240 participants (97%); of these, a subset of 1107 participants (89%) met the inclusion criteria for analyses of objectively measured physical activity. Participant characteristics (age, sex, ethnicity) and levels of adiposity were similar among those who did and did not provide physical activity data; however, participants from black and Asian ethnic groups were less likely to provide physical activity data. Report Supplementary Material 2 shows characteristics of participants at baseline for the 1240 adults with measurements of adiposity at baseline, which are similar to those with physical activity data.

Adjusted mean levels of physical activity, and adiposity outcomes by housing sector and participant characteristics, are shown in Report Supplementary Material 3. Table 3 shows the association of housing sector and other participant characteristics with objectively measured physical activity (steps, time spent in MVPA, time spent in MVPA in ≥ 10-minute bouts) and BMI and percentage fat mass. Participants seeking social housing had markedly lower levels of steps, MVPA and MVPA in ≥ 10-minute bouts and markedly higher BMI and percentage fat mass than those seeking intermediate housing, although there were no differences between those seeking market-rent and those seeking intermediate accommodation.

All physical activity measures were lower among females. Steps and MVPA were slightly higher in 25- to 34-year-olds and steps were also higher among 35- to 49-year-olds than among 16- to 24-year-olds; however, there were no age group differences for MVPA in ≥ 10-minute bouts. Participants of black and Asian ethnicities had lower levels of steps, MVPA and MVPA in ≥ 10-minute bouts than people of white ethnicity. Participants who reported having a LLI had lower levels of steps and MVPA, but not MVPA in ≥ 10-minute bouts. Percentage fat mass was higher in females than males, although there was no difference in BMI (see Table 3). BMI and percentage fat mass were higher among all older age groups than among 16- to 24-year-olds. Participants of black ethnicity had higher levels of BMI and percentage fat mass than people of white ethnicity; there were no differences in BMI and percentage fat mass between Asian or other/mixed ethnic groups and people of white ethnicity. Those with a LLI had higher levels of both BMI and percentage fat mass. Educational attainment level was not associated with any of the outcomes once housing sector had been adjusted for, and adjustment for educational attainment did not materially alter housing sector differences in physical activity or other outcomes (data not presented).

Sensitivity analyses for physical activity outcomes were carried out in 931 participants who wore an ActiGraph for ≥ 4 days with ≥ 540 minutes of recording per day at baseline (see Report Supplementary Material 4). There were no differences between market-rent and intermediate groups (consistent with the main analysis presented in Table 3). Differences between social and intermediate groups were broadly similar with the results presented in Table 3 for the main analysis.

Baseline differences in physical activity variables between housing sectors were examined by day of the week to explore whether or not differences between sectors were consistent across the week (Figure 3). Levels of physical activity [steps (Figure 3a), MVPA (Figure 3b) and MVPA in ≥ 10-minute bouts (Figure 3c)] were generally consistent across weekdays (Monday to Friday) among all groups. In the intermediate group, steps were higher on Saturdays and lower on Sundays; MVPA and MVPA in ≥ 10-minute bouts were lower on Sundays, but there was no difference on Saturdays compared with weekday activity. In the market-rent sector, steps, MVPA and MVPA in ≥ 10-minute bouts were higher on Saturdays and similar to weekdays on Sundays. In the social group, steps, MVPA and MVPA in ≥ 10-minute bouts were on average lower on Saturdays and lower still on Sundays. Registered time of recorded physical activity (Figure 3d) was lowest on average in the social group during weekdays, decreasing on Saturdays and Sundays. The intermediate and market-rent groups had higher levels of registered time during weekdays than the social group, which decreased on average on Saturdays and Sundays (despite recording more steps and minutes in MVPA, suggesting a higher intensity of activity). Mean levels of steps, MVPA, and MVPA in ≥ 10-minute bouts on weekdays, and differences on Saturday and Sunday compared with weekdays, are shown by housing sector in Report Supplementary Material 5. The marked differences in activity between weekdays and weekend days in the social group were not explained by differences in registered time.

FIGURE 3.

Daily physical activity by day of the week and housing sector: N = 6206 days from 1107 participants. Means and 95% CIs are adjusted for sex, age group, ethnic group, LLI, month of recording, day order of recording, day of week, housing sector, an interaction between housing sector and day of week and random effects to allow for multiple days of measurement and clustering of participants within households. Reproduced from Nightingale et al. 70 © Author(s) (or their employer(s)) 2018. Re-use permitted under CC BY. Published by BMJ. This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/licenses/by/4.0/.

Reproduced from Nightingale et al. 70 © Author(s) (or their employer(s)) 2018. Re-use permitted under CC BY. Published by BMJ. This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/licenses/by/4.0/

Associations between perceived neighbourhood quality and crime scales, and physical activity and adiposity outcomes are shown in Table 4, adjusted for the participant characteristics (see Table 3). Participants with the most positive perceptions of neighbourhood quality (highest quintile) had higher steps, recorded longer durations of MVPA and had a lower BMI than those who had the most negative perceptions of neighbourhood quality (lowest quintile). There were no significant associations between perceptions of neighbourhood crime and physical activity or adiposity.

The effect of adjustment for perceived neighbourhood quality on the differences in physical activity and adiposity between housing sector groups is presented in Table 5. All associations between perceived neighbourhood quality and crime-free neighbourhood and outcome variables were approximately linear and were therefore fitted as continuous variables in the model. In addition, associations between perceived neighbourhood quality and crime-free neighbourhood and outcome variables were similar across the three housing sectors (all p > 0.05). Adjustment for perceptions of neighbourhood quality reduced the differences in steps (by 10%), MVPA (by 10%) and MVPA in ≥ 10-minute bouts (by 7%), BMI (by 10%) and percentage fat mass (by 6%) between the social and intermediate groups. Differences between market-rent and intermediate groups in adiposity and physical activity variables were not statistically significant before or after adjustment. Larger proportions of the social–intermediate group differences in steps, MVPA and MVPA in ≥ 10-minute bouts on weekends were explained by adjustment for perceptions of neighbourhood quality (10%, 16% and 16%, respectively) than that for weekdays, which were reduced by 10%, 8% and 3% respectively (data not shown).

At follow-up

In total, 877 adults (69%) from 710 households were examined after 2 years; 441 (50%) had moved and were living in East Village at follow-up. Data were available at baseline and follow-up for 762 participants for physical activity outcomes and 822 participants for adiposity variables. There were no differences in age, sex or ethnic group between those followed up and not followed up (see Report Supplementary Material 6), although those followed up had slightly higher socioeconomic status and recorded more sedentary time at baseline than those not followed up.

Baseline characteristics for the 877 adults who were seen at follow-up are shown in Table 6, by housing sector and overall. In the social housing sector, age, sex, socioeconomic status and children in the household were similar for those living in East Village and for controls, although the East Village group had a higher proportion of participants of black African/Caribbean ethnic origin and a lower proportion of participants of Asian ethnic origin than the control group. In the intermediate sector, the East Village group were younger, less likely to be female, and more likely to be of white ethnicity, be economically active and have no children in the household than controls. In the market-rent sector, age, sex, ethnicity, socioeconomic status and children in the household were similar in the East Village and control groups. For housing sectors combined, the proportion of females in the two groups was similar, but participants in the East Village group were younger, more likely to be of black African–Caribbean ethnicity, less likely to be of Asian ethnicity and less likely to be in higher managerial/professional occupations than the control group. Overall, baseline daily steps and daily levels of MVPA were greater in the control group than in the East Village group, although differences by housing sector were less apparent. Although there was no overall difference in baseline adiposity levels, controls in the intermediate group had higher levels of BMI and fat mass percentage than those living in intermediate East Village accommodation. Those who had moved to East Village reported a sizable increase in scores for the perceptions of crime-free neighbourhood and higher quality neighbourhood at follow-up (Table 7 and Report Supplementary Material 7). Compared with baseline data, participants living in East Village at follow-up lived closer to their nearest park, had better access to public transport and lived in a more walkable area (see Table 7 and Report Supplementary Material 7).

The effect of moving to East Village on physical activity (daily steps, daily minutes of MVPA, daily sedentary time) and adiposity (BMI and percentage fat mass) is shown in Table 8, both overall and stratified by housing sector. Overall, there was weak evidence that moving to East Village was associated with a small increase in daily steps of 154 [95% confidence interval (CI) –231 to 539] after adjusting for sex, age group, ethnic group and housing sector. This appeared more pronounced in the intermediate sector (433 steps, 95% CI –175 to 1042 steps). There was also weak evidence of an increase in daily minutes of MVPA overall, both in total MVPA and in ≥ 10-minute bouts, in participants in the intermediate and market-rent sectors, and a decrease in daily minutes in participants in the social sector, although none of the effects were statistically significant. There was no evidence of a change in daily sedentary time, BMI or percentage fat mass, both overall and by housing sector (see Table 8). Inclusion of an interaction term between the East Village/control group and housing sector to allow for potential differential effects was not statistically significant (p > 0.1). Restricting analyses to 652 (86%) participants who recorded at least 4 days of ≥ 540 minutes accelerometry wear at baseline and at follow-up gave a larger effect for moving to East Village on daily steps (324 steps, 95% CI –93 to 741 steps; see Report Supplementary Material 8), but CIs still spanned the null value. There was weak evidence that the changes were larger at weekends (428 steps, 95% CI –288 to 1144 steps) than on weekdays (199 steps, 95% CI –223 to 620 steps), but CIs included the null value.

Separate analyses comparing change in daily steps in East Village participants with (1) controls who remained at their baseline address and (2) controls who moved elsewhere, allowed the effect of any move to be examined (see Report Supplementary Material 8). Although none of the differences were statistically significant, there was a suggestion of larger differences in daily steps when those in East Village were compared with those who had moved elsewhere [particularly in the intermediate housing sector where a difference of 677 steps (95% CI 15 to 1339 steps) was observed], with a smaller difference when compared with those who did not move. It is noteworthy that daily steps were marginally lower among those in the social sector who moved to East Village than those who moved elsewhere, but effect sizes were not statistically significant. There was no evidence of a difference after removing 21 women who reported being pregnant at baseline or follow-up. The missing data imputation analyses (adding a further 46 individuals at follow-up) gave similar results to the complete case analysis (see Report Supplementary Material 9).

Baseline findings

The baseline findings from this study showed that, in the social housing sector, levels of physical activity were particularly low at weekends compared with weekdays, possibly reflecting higher occupational physical activity and lower leisure-time physical activity; weekday–weekend differences in physical activity were less marked among those seeking intermediate and market-rent housing. However, the lower registered time at weekends but higher MVPA and steps suggests more intense activity at weekends in the intermediate and market-rent housing sectors. Participants seeking social housing in East Village also had lower levels of physical activity and higher levels of BMI and percentage fat mass than those seeking intermediate and market-rent housing, even when adjusted for demographic factors. 70 Positive associations between perceived neighbourhood quality and physical activity, BMI and percentage fat mass at baseline were also shown. Adjustment for differences in perceived neighbourhood quality reduced the differences in physical activity and BMI by approximately 10% between social and intermediate housing sectors, equivalent to a reduction of 111 daily steps, 0.5 minutes of MVPA and 0.5 kg/m² BMI. However, a larger proportion of the difference in physical activity was apparent at weekends; equivalent to a reduction of 222 daily steps and 2.2 minutes of MVPA. These baseline findings provide further insight into the physical activity patterns of the ENABLE London cohort, and how moving to East Village might affect physical activity and adiposity levels in different socioeconomic groups.

Follow-up findings

Compared with pre-move residential areas, East Village was assessed as a more walkable place with greater access to public transport and closer proximity to parks, and was perceived as having marked improvements in neighbourhood quality. Despite this, at 2-year follow-up there was weak evidence that moving to East Village was associated with an overall increase in daily steps compared with not moving. This increase was the largest in those who moved into intermediate housing, whereas those in social housing showed a small decline in their physical activity. Overall increases in time spent in MVPA daily and decreases in daily sedentary behaviour were also observed, but again effects were small. There were no changes in markers of adiposity associated with moving to East Village, either overall or by housing tenure.

Conclusions

Teasing out the causal relationship between the built environment and health behaviours is complex,96 and the reasons for failing to observe clear effects in this longitudinal study when cross-sectional findings are less equivocal remain unclear. 33 Some systematic reviews have avoided this issue by simply excluding longitudinal or experimental studies, which raise concerns that such reviews lead to greater certainty about biased findings,97–100 especially when those living in better neighbourhoods, with built environments more conducive to physical activity, may be fundamentally different from those who do not. However, it is notable that longitudinal studies may also be prone to selection bias, whereby those who choose to move to better neighbourhoods have different health behaviours from those who choose to remain. The ENABLE London study sought to minimise these effects by recruiting a cohort that was seeking to move before any move to East Village, where half of whom moved to East Village and the other half of whom did not. Further consideration of the strengths and limitations of the ENABLE London study and the potential implications of the findings are considered in Chapter 9.

Baseline data

Of the 1278 study participants, we excluded those who lived outside Greater London (n = 81) and those without sufficient accelerometer data (at least 9 hours for at least 1 day is sufficient) on either weekdays or weekends (n = 144) (some participants fit multiple exclusion categories). Those who were excluded from the analytical sample had similar characteristics to those included in the sample in terms of housing status (p = 0.17), age (p = 0.27), sex (p = 0.06) and ethnicity (p = 0.18) (see Report Supplementary Material 10). Of the 1064 participants retained for analyses, 442 were seeking relocation into social, 436 into intermediate and 186 into market-rent accommodation; 1053 participants had physical activity data on weekdays, 848 on weekends and 837 on both weekdays and weekends. To maximise power to detect associations and reduce selection bias, analyses were performed with the maximum number of participants available (i.e. 1053 individuals in analyses on physical activity on weekdays and 848 individuals in analyses on physical activity at weekends). Sensitivity analyses were performed on the subset of 837 participants who had data on both weekdays and weekends.

Follow-up data

Of the 877 participants followed up (69%), those who lived outside Greater London at either baseline or follow-up and those who did not have physical activity data at baseline or follow-up (n = 190) were excluded. Those excluded from the analytical sample had similar characteristics to those included with regard to sex (p = 0.36) and ethnicity (p = 0.76), but were younger (p = 0.009) and had more intermediate and less market-rent seekers compared with those included (p = 0.003) (see Report Supplementary Material 11). A total of 687 participants were retained for analyses; 680 participants had physical activity data on weekdays and 517 on weekend days, both at baseline and at follow-up. Of the 687 participants, 283 were seeking relocation into social housing, 301 into intermediate housing, and 103 into market-rent accommodation.

Primary outcomes

Environmental exposures

Participants were geocoded to the centroid of the footprint of their building of residence, using OS AddressBase Premium (versions 2015 and 2017 for geocoding at baseline and at follow-up, respectively) to match participants’ declared residential addresses with XY coordinates. Residential locations were then used to derive a range of built environment factors that were hypothesised to be associated with physical activity, as follows.

Covariates

Covariates included sex (female, male), age group (16–24, 25–34, 35–49 and ≥ 50 years), ethnicity (white, black, Asian, mixed/other) and aspirational housing tenure being sought (social, intermediate, market-rent). Seekers of intermediate accommodation were used as the reference group, as this group had the largest number of participants.

Household income, occupational status, presence of children in the household and education level were highly correlated to aspirational housing status, hence, these variables were not adjusted for in the final regression models to avoid overadjustment (see Report Supplementary Material 15). Other hypothesised covariates, including car use and attitude towards physical activity, were excluded after initial consideration, as they were not found to be associated with exposures and outcomes in bivariate analyses or their addition to fully adjusted models did not appreciably alter coefficients.

Statistical analyses

All analyses were carried out using Stata Special Edition version 15 for Windows.

Associations between the built environment and physical activity at baseline

Associations between built environment and physical activity

The associations between built environment factors and the number of daily steps and amount of MVPA taken on weekdays and at weekends are presented in Table 10, and displayed graphically in Report Supplementary Material 16.

TABLE 10 - Regression estimates for the baseline associations of daily steps and MVPA with residential built environment factors

Model 1 adjusts only for clustering at household level (referred to as ‘minimally adjusted model’ in the text).

Model 2 adjusts for sex, age group, ethnic group, aspirational housing tenure and clustering at household level.

Model 3 adjusts for sex, age group, ethnic group, aspirational housing tenure, clustering at household level and environmental variables entered simultaneously.

β coefficient is statistically significant at p < 0.05.

Adapted with permission from Clary et al. 125 © 2020 Clary et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. See: http://creativecommons.org/licenses/by/4.0/. The table includes minor additions and formatting changes to the original.

Residential built environment factors	Model 1a β (95% CI)	Model 2b β (95% CI)	Model 3c β (95% CI)	Model 1a β (95% CI)	Model 2b β (95% CI)	Model 3c β (95% CI)
	Daily steps on weekdays (n = 1053)			Daily steps on weekends (n = 848)
Walkability	52 (–30 to 134)	–5 (–86 to 76)	38 (–57 to 132)	255 (139 to 371)d	135 (28 to 242)d	144 (19 to 269)d
Street connectivity	48 (–132 to 228)	–39 (–215 to 137)	–	322 (63 to 582)d	156 (–79 to 391)	–
Residential density (1000 habitants/km2)	41 (5 to 76)d	5 (–31 to 41)	–	129 (78 to 179)d	51 (3 to 99)d	–
Land use mix	498 (–644 to 1641)	–291 (–1419 to 837)	–	3520 (1884 to 5156)d	1547 (34 to 3061)d	–
Distance to closest metropolitan park (km)	–294 (–446 to –143)d	–206 (–354 to –58)d	–264 (–422 to –107)d	–363 (–582 to –144)d	–125 (–323 to 74)	–144 (–354 to 66)
Distance to closest district park (km)	–93 (–265 to 79)	7 (–160 to 174)	–64 (–240 to 113)	–370 (–617 to –124)d	–180 (–402 to 42)	–138 (–372 to 96)
Distance to closest local park (km)	170 (–174 to 514)	–84 (–420 to 253)	–185 (–529 to 158)	1147 (669 to 1624)d	597 (161 to 1032)d	598 (155 to 1040)d
Accessibility to public transport
Low	442 (–362 to 1247)	767 (–12 to 1546)	1186 (296 to 2076)d	–1352 (–2516 to –187)d	–608 (–1657 to 442)	60 (–1136 to 1256)
Intermediate	–481 (–948 to –15)d	–152 (–609 to 305)	8 (–487 to 503)	–1128 (–1803 to –453)d	–364 (–981 to 252)	–8 (–670 to 654)
High	–	–	–	–	–	–
	Daily MVPA (minutes) on weekdays (n = 1053)			Daily MVPA (minutes) on weekends (n = 848)
Walkability	0.7 (–0.0 to 1.3)	0.1 (–0.6 to 0.6)	0.4 (–0.3 to 1.2)	2.2 (1.2 to 3.1)d	1.2 (0.3 to 2.1)d	1.4 (0.4 to 2.5)d
Street connectivity	0.6 (–0.9 to 2.0)	–0.3 (–1.7 to 1.0)	–	2.7 (0.6 to 4.8)d	1.3 (–0.6 to 3.2)	–
Residential density (1000 habitant/km2)	0.4 (0.1 to 0.7)d	0.0 (–0.2 to 0.3)	–	1.1 (0.7 to 1.5)d	0.4 (0.0 to 0.8)d	–
Land use mix	9.1 (0.0 to 18.2)d	1.3 (–7.5 to 10.1)	–	32.1 (18.9 to 45.2)d	16.8 (4.6 to 29.0)d	–
Distance to closest metropolitan park (km)	–2.5 (–3.7 to –1.3)d	–1.8 (–2.9 to –0.5)d	–2.1 (–3.4 to –0.9)d	–2.9 (–4.7 to –1.2)d	–1.1 (–2.7 to 0.4)	–1.2 (–2.9 to 0.5)
Distance to closest district park (km)	–0.6 (–2.0 to 0.7)	0.3 (–1.0 to 1.6)	–0.2 (–1.5 to 1.2)	–2.4 (–4.4 to –0.4)d	–0.9 (–2.7 to 0.9)	–0.4 (–2.3 to 1.4)
Distance to closest local park (km)	2.0 (–0.8 to 4.7)	–0.3 (–3.0 to 2.3)	–1.0 (–3.7 to 1.6)	9.0 (5.2 to 12.9)d	4.7 (1.2 to 8.2)d	4.9 (1.3 to 8.5)d
Accessibility to public transport
Low	2.7 (–3.7 to 9.1)	5.8 (–0.3 to 11.9)	9.7 (2.7 to 16.6)d	–9.5 (–18.9 to –0.1)d	–4.0 (–12.5 to 4.5)	2.7 (–7.0 to 12.4)
Intermediate	–4.5 (–8.2 to –0.8)d	–1.7 (–5.3 to 1.8)	–0.2 (–4.1 to 3.6)	–8.2 (–13.7 to –2.7)d	–2.7 (–7.6 to 2.3)	0.7 (–4.7 to 6.0)
High	–	–	–	–	–	–

Weekdays

In models adjusted for sociodemographic characteristics, both daily steps taken and MVPA accumulated were negatively associated with the distance from home to the closest metropolitan park [mean difference in daily steps (–206 steps, 95% CI –354 to –58 steps); mean difference in daily minutes of MVPA (–1.8 minutes, 95% CI –2.9 to –0.5 minutes) per kilometre of distance to the closest metropolitan park], indicating that the more distant the metropolitan park, the smaller the overall number of steps taken and the shorter the time spent in MVPA (Figure 4), particularly among those from intermediate and market-rent tenured groups (see Figure 4). Associations between accessibility to public transport and daily steps and MVPA were borderline statistically significant in models adjusted for sociodemographic characteristics (mean differences in daily steps: 767 steps, 95% CI –12 to 1546 steps; mean difference in daily minutes of MVPA: 5.8 minutes, 95% CI –0.3 to 11.9 minutes for those with low accessibility compared with those with high accessibility), and reached statistical significance after further adjustment for the other residential built environmental factors (mean differences in daily steps: 1186 steps, 95% CI 296 to 2076 steps; mean difference in daily minutes of MVPA: 9.7 minutes, 95% CI 2.7 to 16.6 minutes). Adjustment for other residential built environmental factors, however, led to an inflation of the regression estimate and a widening of the CI compared with models adjusting only for sociodemographic characteristics, raising potential multicollinearity issues.

FIGURE 4.

Daily steps and MVPA (mean and 95% CI) by quintile of selected residential built environmental factors, overall and by housing sector. (a) Daily steps at weekends and walkability; (b) daily steps at weekends and distance to closest local park; (c) daily steps on weekdays and distance to closest metropolitan park; (d) daily MVPA at weekends and walkability; (e) daily MVPA at weekends and distance to closest local park; and (f) daily MVPA on weekdays and distance to closest metropolitan park. Notes: mean daily steps (95% CI) at weekends or on weekdays are plotted against median value of quintile of built environment variable, for all housing sectors combined and by housing sector. Daily steps are adjusted for sex, age group, ethnic group as fixed effects and household as a random effect in a multilevel model. Daily steps for all housing sectors are additionally adjusted for housing sector as a fixed effect. Footnotes: mean daily MVPA (95% CI) at weekends or on weekdays is plotted against median value of quintile of built environment variable, for all housing sectors combined and by housing sector. Daily MVPA is adjusted for sex, age group, ethnic group as fixed effects and household as a random effect in a multilevel model. Daily MVPA for all housing sectors is additionally adjusted for housing sector as a fixed effect.

On weekend days

In minimally adjusted models, both daily steps and MVPA accumulated were positively associated with residential walkability (mean difference in daily steps and daily minutes of MVPA per unit of walkability: 255 steps, 95% CI 139 to 371 steps, and 2.2 minutes, 95% CI 1.2 to 3.1 minutes, respectively), and with the distance from home to the closest local park (mean difference in daily steps and daily minutes of MVPA per kilometre of distance to the closest local park: 1147 steps, 95% CI 669 to 1624 steps, and 9.0 minutes, 95% CI 5.2 to 12.9 minutes, respectively). Conversely, steps taken and minutes of MVPA were negatively associated with the distance (km) to the nearest metropolitan park (–363 steps, 95% CI –582 to –144 steps, and –2.9 minutes, 95% CI –4.7 to –1.2 minutes, respectively), the distance (km) to the closest district park (–370 steps, 95% CI –617 to –124 steps, and –2.9 minutes, 95% CI –4.7 to –1.2 minutes, respectively) and accessibility to public transport (–1352 steps, 95% CI –2516, to –187 steps, and –9.5 minutes, 95% CI –18.9 to –0.7 minutes, respectively, for those with low compared with those with high accessibility). After adjustment for sociodemographic characteristics, associations of residential walkability with daily steps taken (135 steps, 95% CI 28 to 242 steps) and with minutes of MVPA (1.2 minutes, 95% CI 0.3 to 2.1 minutes), and of distance from home to the nearest local park with daily steps taken (597 steps, 95% CI 161 to 1032 steps) and with minutes of MVPA (4.7 minutes, 95% CI 1.2 to 8.2 minutes) remained statistically significant. This suggests that the more walkable the residential environment the higher the number of steps taken and the greater the time spent in MVPA, and that overall the further the closest local park, the higher the number of steps taken and the greater the time spent in MVPA on weekend days (Figure 4); this appeared more evident in the social and intermediate housing sectors (Figure 4). Associations remained consistent after further adjustment for the other residential built environmental factors (model 3).

Interaction analyses

An interaction between low accessibility to public transport and social versus intermediate housing seekers in relation to mean daily steps on weekdays was observed (p = 0.042). However, a Wald test for the interaction term between accessibility to public transport and housing suggested that the interaction term could be omitted from the model (p = 0.06).

Change in built environment and change in physical activity

Baseline findings

Participants took a greater number of steps at the weekend when living further away from a local park. This may arise because physical activity associated with the use of local parks relies on the journey to the park (assuming a physically active commute), rather than on the activities undertaken within them. Conversely, living further away from the closest metropolitan park was associated with fewer steps taken on weekdays. Metropolitan parks are sparser and therefore further away on average than local parks, and this may reflect decreased interest in reaching this type of park with increasing distance. Overall, our findings highlight the importance of distinguishing between different park types. This may help to explain inconsistent findings in the literature on the access to green space and physical activity, which reports both null112,122 and positive32,111,113,128–130 associations. Exploring individuals’ mobility behaviours, for instance through the combined use of GPS and accelerometer data,131 would help unpick the use made of these different parks and refine the mechanisms by which they could be associated with weekday and weekend physical activity.

In line with other studies, we found that with poorer accessibility to pubic transport, fewer steps were taken and time spent in daily MVPA was lowest at the weekends. 32,113 This suggests that on weekends people’s interest in the use of public transport may decrease with decreasing accessibility; however, there was weak evidence for an association after adjustment for sociodemographic factors. Conversely, on weekdays participants with a low accessibility to public transport were taking more daily steps than those with high accessibility. One explanation could be that participants rely on public transport for weekday activities such as commuting to work. As a result, those with poor accessibility to public transport walk more to reach transit points. Overall, differing patterning of associations on weekend days and weekdays suggest that the use of public transport depends on utility and travel function: a commute to work on weekdays and leisure and recreation at the weekends. Further exploration of the driving forces for the use of public transport on weekdays versus at weekends would be valuable to better understand the overall contribution of public transport accessibility to physical activity level.

We explored two plausible pathways through which the built environment may partly contribute to the observed aspirational housing tenure differences in physical activity: differences in the way the three groups interact with the physical activity-supportive features of their residential environment and differences in the availability of these features. We found little evidence for the former explanation, with weak evidence for an interaction. As for the latter explanation, we found that social housing seekers lived in less walkable environments, closer to local parks and further away from metropolitan parks than intermediate and market-rent housing seekers, which, in turn, were all associated with a decreased number of daily steps and daily minutes of MVPA. However, disparities in these three environmental features accounted for little of the variability in daily steps (1.5%, 4.7% and 9.9%, respectively) and in daily minutes of MVPA (2.6%, 5.9% and 11.9%, respectively) between social and intermediate participants. Our findings support previous work that suggests that housing tenure might relate to health behaviours by sorting disadvantaged people into neighbourhoods provided with fewer health-promoting amenities. 132,133

Longitudinal findings

The longitudinal findings strengthen previous cross-sectional evidence that more walkable environments are associated with higher levels of physical activity. 110,121,134–136 The significant associations observed were mostly driven by the two components of walkability, residential density and land use mix, which were both strongly and positively associated with increased physical activity levels, consistent with other longitudinal studies. 89,95,137–140 Greater land use mix is posited to support walking by offering accessibility to a wider range of services and employment. 141 As for higher residential density, it is assumed to provide more of a critical mass of walkers seen by other people, who may, in turn, be encouraged to walk by safety in numbers142 and/or by a desire to comply with the social norm. 143 Traffic congestion associated with higher residential density may also promote active travel. 144 Unlike others,145,146 we did not find that a change in street connectivity was associated with a change in the number of steps taken or the amount of MVPA accumulated. This may be partly because our street connectivity metric relied on street network data and therefore fell short in capturing the permeability of newly built active design environments, like East Village, composed of large pedestrianised areas and informal footpaths. Hence, the limitation of our street connectivity measure to reflect the permeability of the residential neighbourhood as experienced by East Village movers may have underestimated the strength of the effect between walkability and physical activity.

Although the relationships between walkability (and its component variables) and physical activity on weekend days were all in the same direction, none reached statistical significance. This may be because the built environment is more likely to be associated with transportation walking than other types of physical activity, including recreational walking. 147 Alternatively, the limited number of participants who had physical activity data available at weekends compared with weekdays may have resulted in underpowered analyses for models fitting physical activity at weekends as an outcome.

We found evidence that increasing accessibility to public transport is associated with a statistically significant decrease in physical activity on weekdays among social housing participants and, conversely, an increase in weekdays physical activity among market-rent housing participants. Among possible explanations, different patterns in the distance between home and workplace and/or in the availability of individual modes of transport across socioeconomic groups may trigger different responses to increased accessibility to public transport. Overall, socioeconomic differences in the relationship between accessibility to public transport and physical activity may explain the mixed findings outlined by some literature reviews. 147

Unlike others, we did not find evidence that living closer to a park at follow-up than at baseline was associated with a change in physical activity level. 89 Previous studies have highlighted the importance of distinguishing between different types of park, as size and attractiveness are associated with physical activity. 128 However, since half of the ENABLE London participants were relocated to the edge of the Olympic park, a large metropolitan park with many sport facilities, such a classification became meaningless. It was difficult to disentangle the different relationship that individuals may have with different types of parks, which may partly explain the null findings.

Those who have lived in their current home for < 2 years have been found to take longer to orientate themselves within their neighbourhood and have a greater mismatch between objectively measured and perceived neighbourhood features. 148 This may partly explain the absence of findings between changes in exposure to some of the built environment features (i.e. distance to parks) and changes in physical activity behaviours. It also suggests that the associations found may strengthen over time.

Strengths

The strengths of this study include the use of validated objective measures of physical activity,149 considering weekday/weekend variations in physical activity undertaken,113 and exploring the contribution of the residential built environment in explaining socioeconomic differences in physical activity levels, which has been little assessed. Sensitivity analyses further strengthened the findings by showing consistent results in both the inclusive analytical sample and the more restricted sample of participants with physical activity data on both weekdays and weekends. This study is also one of very few longitudinal studies to have enabled examination of how change in GIS-derived residential built environment features is associated with change in accelerometer-based physical activity levels. Its design provided a great variability in changes of exposures to various features of the environment, facilitating the detection of associations. Moreover, a relatively large number of participants were enrolled compared with other longitudinal studies (e.g. Wells and Yang145 and McCormack et al. 150).

It is of note that these analyses looking at physical activity in relation to change in the built environment showed statistically significant increases in physical activity, whereas the increases in physical activity were not statistically significant when comparing the East Village group with those who did not move into East Village. There are a number of possible explanations that we are exploring further. In particular, ‘change in walkability’, for example, is a continuous measure and therefore has more power to detect changes in physical activity than using the dichotomous variable East Village or control.

Limitations

Limitations also warrant consideration. The cross-sectional study design of the baseline data does not rule out the selection of more active people into neighbourhoods supportive of physical activity, and therefore restricts interpretations about the direction of effects. Moreover, not considering physical activity facilities available in routinely visited settings other than place of residence (e.g. workplace) may have led to an overestimation of the association between the residential environment and health behaviours. 151 Because the sample was not randomly selected from the population at large, that is the study was targeted at those seeking to move to East Village, findings may not be more broadly generalisable, which may have implications for external validity.

Conclusions

The findings indicated that the residential built environment is associated with physical activity behaviours differently at weekends and on weekdays and that changes in the residential built environment are associated with changes in physical activity, and both of these differ by weekday and weekend. There was limited evidence that socioeconomic disparities in residential built environment factors contribute to socioeconomic differences in physical activity. These findings provide strong support for considering more walkable urban designs, composed of more mixed land use and higher residential density, as levers to increase physical activity. In addition, this work also suggests that policy-makers should be sensitive to the possibility that environmental interventions might have differential impacts on physical activity at weekdays and weekends.

Methods

The method to identify travel modes from accelerometry and GPS data was developed using data collected at baseline in the ENABLE London study. Full details of this are published166 and are summarised below. Parts of this section are adapted or reproduced from Procter et al. 166 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. The text below includes minor additions and formatting changes to the original text.

Accelerometry has been widely used to measure physical activity, with many different devices deployed. The most common devices in the literature are the ActiGraph, Actiheart (CamNtech, Cambridge, UK), Actical (Philips Respironics, Amsterdam, the Netherlands), activPAL (PAL Technologies Ltd, Glasgow, UK) and GeneActiv (Activinsights Ltd, Huntingdon, UK), with over half of published studies up to 2015 using the ActiGraph. 167 Many of these different devices convert raw acceleration (measured in g) into a form of activity count variable, which has then been used to classify physical activity intensity and energy expenditure. 67,168–170 However, the methods used by instrument manufacturers to convert raw acceleration into counts are often unclear. For this reason, and because the raw acceleration data contain much more information to train an algorithm than derived count variables, we have made use of the raw data in this study.

Previous work on travel mode identification has been developed from the transport perspective rather than that of physical activity, where segmentation into journeys is important to assess travel behaviour. 160,171–173 A focus on journeys often results in short periods of physically active transit behaviour, such as walking between bus stops, being identified as part of a non-active travel mode. However, for physical activity researchers, quantifying the volume and intensity of physical activity when actively travelling is an essential component of the overall purpose of a journey. Consequently, it is important to identify all data points that denote active travel, so that all time in active travel modes can be quantified. As a result, we identify the travel mode of each GPS data point (recorded every 10 seconds) without prior segmentation into journeys.

Participants in the ENABLE London study were asked to wear an accelerometer (ActiGraph GT3X+) and a GPS receiver on an elasticated belt around their waist for 7 consecutive days. Participants also completed a questionnaire to describe their travel patterns to work/place of study. They reported the specific days on which they would be travelling to work or study during the ActiGraph and GPS wear period and whether or not they commuted at the same time on each day. Reported travel modes for these journeys to and from work were underground (tube), train (over ground), bus/minibus/coach, taxi, motorcycle/moped, driving a car or a van, passenger in a car or a van, bicycle, walk, jog and other. These travel modes were recategorised into walk, cycle, vehicle (taxi, motorcycle, car/van driver, car/van passenger and bus/minibus/coach) and train (underground and overground rail). Insufficient participants consistently jogged to work for us to be able to separate ‘jog’ as an additional mode. In addition, time of leaving and arriving for each journey to and from work were collected.

Data preparation and cleaning

The raw accelerometer data were extracted as .csv files using ActiLife 6 software (ActiGraph, Florida, USA) summarised in 10-second epochs and merged by time stamp to GPS data. For the GPS data, measures of satellite signal quality were derived using the sum of the signal to noise ratio from each satellite the GPS device was connected to at each epoch. This takes into account possible obstructions to the GPS signal caused by the participant being indoors and gives a single measure of signal quality. Variations in speed and accelerometer data across a 10-second epoch were smoothed out using a 4-minute moving window calculating mean, SD, 10th and 90th percentiles of data from each accelerometer axis and speed from the GPS device. Distance from GPS time points to train lines was calculated using a combination of Meridian 2 (Ordnance Survey, Southampton, UK) rail network data for the UK and OS OpenMAP (Ordnance Survey, Southampton, UK) data for central London. 174 Distance travelled over the previous and next minute were also calculated. All variables were chosen as they were likely to differ between travel modes: vehicles and trains have higher speed than walking and cycling; walking shows greater accelerometer activity than other modes; both vehicles and trains have metal structures that may obstruct GPS signal; distance over the previous or next minute would be low if the participant is stationary.

Development of the algorithm

A training data set was created using a subset of participants (n = 326) identified from the baseline data who commuted to and from work using the same mode of transport every day and specified the time that they usually travelled for both journeys. All combined accelerometer and GPS data were extracted and exported to a Geographical Information System [ArcGIS 10.4 (ESRI, CA, USA)], and all points during the commute of each participant were manually identified to confirm the stated mode of commute. Vehicle and train journeys were confirmed by location of appropriate features (e.g. roads or rail tracks). Points in the commute clustered around a single location were classified as ‘stationary’. It should be noted that ‘stationary’ does not imply inactive, as walking within a building will appear stationary in terms of GPS signal and be classified as such using our method. Any sequences of points for which a mode of travel could not be clearly identified were excluded, for example repeated GPS signal loss. If the participant appeared to be travelling by a method other than that stated, for example they usually commute by train but on one day they do not move along a train line, then these points were not assigned a travel mode and were excluded from the training data set.

The training data set was then used to develop a model. We tested different moving window sizes (i.e. 1-, 2-, 3-, 4-, and 5-minute moving windows) and assessed how accurate each was for predicting travel mode on a subset of the training data. To test the generalisability of our fitted model to other data sets, we compared manually identified data from 10 participants from the Sedentary Time and Metabolic Health in People with type-2 diabetes mellitus study (STAMP-2) (Professor Ashley R Cooper, University of Bristol, 2018, personal communication) data set with those predicted using our model. The STAMP-2 study was a cross-sectional observational study of sedentary behaviour in adults with newly diagnosed type 2 diabetes mellitus, conducted in two English NHS foundation trusts in south-west England. Participants in one centre wore an accelerometer and GPS monitor concurrently for 7 days, providing similar data to the ENABLE London study. The advantages of testing our model on the STAMP-2 participants are that they were independent of the ENABLE London study and recruited from a city with different travel options from that of central London. They were also recently diagnosed with type 2 diabetes mellitus, were generally older and had a higher prevalence of obesity than the ENABLE London participants. A good predictive performance on this data set would demonstrate the generalisability of our algorithm to other populations.

Training data

The training data set consisted of 326 participants providing 131,573 data points (365.5 hours): 66 (20%) who walked to work (12,791 points, 35.5 hours), 34 (10%) cyclists (11,607 points, 32.2 hours), 94 (29%) vehicle users (29,407 points, 81.7 hours), 132 (40%) train users (18,269 points, 50.7 hours) plus an additional 59,499 ‘stationary’ data points (165.3 hours). Approximately one-quarter of the training data (33,529 points) were used to test moving window sizes of 1, 2, 3, 4 and 5 minutes and the 4-minute moving window resulted in the highest predictive accuracy for active travel modes. The remaining 98,387 points were used to build a cross-validated model.

Model prediction

In the model cross-validation, we correctly predicted 97.3% of points and all five travel modes were predicted with high accuracy. In comparison with manually identified data (n = 21), overall accuracy was still high, with 96% of predictions correct, although this was driven by the fact that most people spent most of their time stationary (83.7% of time stationary). Cycling was least well predicted. When our model was tested on the STAMP-2 data set, we correctly predicted 96.5% of points compared with manual identification of travel modes, demonstrating that our predictions performed well considering the different participants and context. However, this was again driven by our high accuracy on stationary points, which was the dominant mode (86.8% of time stationary), and the poorest-performing predictive mode was, again, cycling.

Methods

The algorithm developed on a subset of the ENABLE London baseline data set was applied to the full baseline and 2-year follow-up accelerometry and GPS data.

Outcomes

The primary outcome was minutes of ‘active’ travel, that is, daily minutes of walking and cycling. Secondary outcomes were vehicle, train, stationary and total minutes.

Statistical methods

The data collection and processing procedures were identical at baseline and at follow-up. Daily GPS data were processed to classify each 10-second epoch into one of five travel modes: walking, cycling, vehicle, train and stationary. Owing to the loss of GPS signal when travelling underground by tube, thus leading to gaps in the data, further work was carried out to classify some of these gaps as ‘underground’, when all of the following conditions were true:

• GPS signal was lost within 200 m of an underground station or had the predicted travel mode ‘train’.

• GPS signal was regained within 200 m of an underground station or had the predicted travel mode ‘train’.

• The time gap between loss and reacquisition of the signal was between 2 minutes and 2 hours.

• The loss and regaining of GPS signal did not take place within 200 m of the same underground station.

It should be noted that underground trains in the London transport system also run above ground, so there is potential for overlap between ‘underground’ and ‘train’ as methods of travel in our data.

The daily 10-second epoch data were summed to provide daily minutes for each travel mode and total daily GPS minutes. Walking and cycling minutes were also combined to provide a measure of ‘active travel’. This provided a data set of up to 7 days of data for each participant, with daily time spent in each travel mode. Daily data were processed to provide an average summary daily figure for each participant, using identical methods to those for the ENABLE London accelerometer data, as described in Chapter 3. Participants were asked to wear the accelerometer at the same time as the GPS monitor, and in line with accelerometry analyses, days where accelerometer wear time was less than 540 minutes (9 hours) were dropped from further analyses. For each mode of travel, average daily minutes at baseline were derived using multilevel linear regression models (level 1 was day within individual and level 2 was individual), regressing daily time on day order of wear, day of week and month of wear. This was repeated for the follow-up data. The change in average daily minutes from baseline to follow-up for each travel mode was then examined using multilevel linear regression models, where level 1 was individual and level 2 was household. For each travel mode, average daily minutes at follow-up was regressed on average daily minutes at baseline, East Village/control group, sex, age group and ethnic group as fixed effects and household as a random effect. The model coefficient for East Village/control group is thus the additional change in the East Village group adjusted for any change in the control group. This method of analysis measures change while controlling for regression to the mean, and has been shown to both minimise bias and maintain power. Checks were carried out to confirm that the distribution of residuals from the models were normally distributed. Sensitivity analyses included limiting to those who were working or studying at baseline. Imputation analyses for the main outcome of interest, ‘active travel’ (walking and cycling minutes), were carried out using multiple imputation methods using Stata Special Edition version 15 for Windows.

The ENABLE London baseline cohort

At baseline, 1063 out of 1278 (83%) participants provided GPS data, of whom 991 provided ≥ 1 corresponding day with ≥ 540 minutes accelerometer wear time. Baseline characteristics for these 991 participants are shown in Table 15, overall and by housing sector. Across all the housing sectors, two-thirds were aged < 35 years, 56% were female and 52% were of white ethnic origin. Participants from the social housing sector were more likely to be female (74%), older and of black or Asian ethnic origin. Intermediate and market-rent participants were younger, equally likely to be male or female and more likely to be of white ethnic origin. These patterns are similar to those seen in the full ENABLE London cohort (see Chapter 2). Overall, the majority of the cohort (85%) were either working or studying at baseline, although this fell to two-thirds in the social housing sector and was higher (95%) in the intermediate and market-rent sectors. This is reflected in the proportions that were classified as economically inactive: 47% in the social sector and 7–9% in the intermediate and market-rent sectors. Three-quarters of the cohort used public transport to travel to or from their work or study and 50% reported some walking, cycling or jogging, which can be classified as ‘active travel’. The proportions of ‘active travel’ varied between the housing sectors: the social sector participants were less likely to walk, cycle or jog and more likely to use private transport. Table 15 also shows the time spent in different travel modes at baseline from the GPS monitors. Participants in the intermediate and market-rent sectors were broadly similar in their average daily time spent in each travel mode. In contrast, social sector participants recorded fewer walking, train and underground minutes, and higher vehicle and stationary minutes. Differences between housing sectors for time spent in different travel modes were statistically significant for all travel modes.

Changes in mode of travel baseline to follow-up

At follow-up, 714 out of 877 (81%) participants provided GPS data, 681 with ≥ 1 corresponding day with ≥ 540 minutes of accelerometry wear time. Longitudinal analyses were restricted to 578 participants who had valid GPS data at both baseline and follow-up.

Baseline characteristics for the 578 adults with GPS data at baseline and follow-up are shown in Table 16 by East Village/control group and housing sector. Reassuringly, age, sex, ethnicity and NS-SEC patterns were similar to those for the full 877 who were followed up (see Chapter 2). In the social housing sector, the East Village and control group were similar in age, sex and socioeconomic distributions, but the East Village group were more likely to be of black ethnic origin. In the intermediate sector, the East Village group were more likely to be younger, male, of white ethnic origin and economically active. In the market-rent sector, age, sex, ethnic group and socioeconomic status were similar in the East Village and control groups. There were no differences between the East Village group and the control group in the proportions of households who owned a car at baseline. Slightly fewer of the East Village group were working at baseline and 24% were classified as economically inactive compared with 18% of the control group. At baseline, there was greater use of public transport for travel to work or study among those who moved into East Village (p = 0.004 for all housing sectors). Use of private transport and walk/cycle/jog were similar in the East Village and control groups, although social housing sector participants were more likely to use private transport and less likely to walk/cycle/jog than intermediate and market-rent participants.

Table 17 shows the change in the neighbourhood perceptions and built environment variables for the East Village and control groups. Compared with the baseline data, those participants who had moved to East Village lived closer to their nearest park, had better access to public transport and lived in a more walkable area. They also reported more positive perceptions of their local area. These differences were most marked for social housing sector participants.

The effect of moving to East Village on the time spent in different travel modes is shown in Table 18. Overall, there was little change in participants’ walking or cycling minutes. However, vehicle minutes decreased overall by 7–8 minutes per day, with greater effects in the intermediate housing sector (10 minutes’ decrease, 95% CI –17 to –2 minutes’ decrease; p = 0.01), and time spent travelling by underground increased by 3–4 minutes, particularly in the market-rent sector (12 minutes, 95% CI 4 to 19 minutes; p = 0.001). There were large decreases in both stationary and total minutes; however, this was as a result of issues with the GPS signal being blocked by East Village housing, discovered after follow-up had been completed.

Strengths of the algorithm developed

Our study has a number of important strengths. Applying the algorithm to the STAMP-2 data set resulted in similar levels of accuracy of travel mode prediction within a very different group of participants living in a different environmental setting. This finding suggests that our method may be generalisable to other data sets and could be used by other researchers without the time-consuming steps of creating new training data. However, until more robust tests have been completed, we would recommend a manual identification checking stage similar to our methods to verify the generalisability of the method.

We exhibit similar levels of accuracy to the personal activity location measurement system (PALMS),173 which is a freely available method to process physical activity data. The purpose of the output is somewhat different though, with PALMS identifying journeys and our method identifying each data point. The preferred method will depend on the research question. One advantage of our R package that the server-based system in PALMS does not have is that it can be run on a researcher’s own machine. Data used for the present (and other similar) analyses are subject to strict data privacy and ethical conditions. Running the analyses on a researcher’s own machine, rather than uploading to a server for remote processing, can help avoid problems related to data privacy. Furthermore, our method is open source, which means that all code is freely available online. 165 As a result, other researchers can suggest edits and improvements and contribute to the development of our method, and its utility to the research community.

The prediction of active travel modes should complement existing analyses of physical activity using accelerometers. Accelerometers have been used with great effect to objectively quantify activity, but they are not without limitations. A well-known problem of traditional physical activity analyses using accelerometers is that cycling is not recognised as a form of MVPA because cycling generates relatively low readings on a waist-worn accelerometer compared with other active modes. Identification of cycling from combined accelerometer and GPS data will allow better quantification of cycling as a form of physical activity. Traditional physical activity analyses will still miss cycling as a form of activity whereas our method will quantify it, albeit with moderate precision. Furthermore, assessment of active travel using our method will help to understand how people are active. For example, activity at a single location, such as at home or at the gym, will not be classed as active travel using our method but as a stationary travel mode; therefore, a participant may show high levels of overall physical activity but low levels of active travel. Incorporation of this extra information will help to understand participants’ overall physical activity patterns. For example, if a participant shows increased physical activity but not active travel, it is likely that they have increased their activity at locations such as at home or at the gym. If we see no change in physical activity but an increase in active travel, then participants have replaced some of their physical activity at a location with active travel.

Limitations of the algorithm developed

It is unsurprising that our accuracy scores are lower when compared with manual identification of both participants within this study and from elsewhere than in cross-validation of the training data set. Identification is most likely to be accurate during journeys of a consistent mode, which is how the training data were defined. In full days of data there are likely to be other more ambiguous forms of movement that may be short in duration and, therefore, difficult to identify, or which do not fully represent one of the forms of travel we have included. Our model performs least well for the detection of cycling and, therefore, is probably overfitted to our training data. As a result, we would recommend manual checking of cycling data to improve accuracy. In this study, cycling represents a relatively small amount of total time; therefore, manual checking of this subset represents a much smaller time investment than in the full analysis.

Visual inspection of the data in the GIS revealed that much of the disagreement between prediction and manual classification was found at the start and the end of journeys. This highlights the challenge of identifying modal shift, that is when to switch from one mode to another. This is a limitation of the current method and all other indirect methods. However, the strength of identifying individual data points rather than full journeys is that the imprecision in identifying modal shift leads to small numbers of misclassified points, rather than entire misclassified journeys. Small numbers of misclassified points will only have a small effect on total time in each travel mode. Our inaccuracy at the start and end of trips means that any prediction to a data set where many short-duration trips are expected would be predicted to yield lower accuracy. Conversely, within a data set containing longer journeys, travel mode should be able to be identified with greater precision.

It is also worth highlighting that some disagreement within our test data sets may not be true errors. For example, if a participant is stationary but on a train it is questionable whether they should be classified as using a train or stationary; this may not matter as long as the rule is applied consistently. However, when comparing the predicted and manually derived methods this causes disagreement because during manual classification stationary points on train lines were termed stationary (e.g. standing at a train station), yet the algorithm identifies them as ‘train’. A number of the misclassifications between our manual identification and predicted data sets may, therefore, be open to interpretation and may not be true misclassifications.

Using the model

First, we were only able to identify the active modes of walking and cycling: there is no consideration of running or any other activity. This limitation is based on our study sample, where no commuters consistently used these modes so we could not include them in our training data. However, walking and cycling represent the most frequently used active transport modes and other modes were rarely reported by our participants. Any form of running will most likely be identified as walking using our method because of the high acceleration that running causes on an accelerometer, and so will still be identified as active travel. We therefore feel that this limitation will have little impact on our study, although we would recommend caution in applying our method to a data set where a large quantity of running is expected. Second, we have demonstrated our capacity to identify major travel modes but were not able to discriminate car travel from public transport. Consequently, if a study is attempting to quantify the use of public transport then the current method is inadequate. To address this, we have developed a second model that discriminates bus travel. However, this leads to reduced overall accuracy, because the pattern of speed/activity can be confused between buses and other vehicles. Further work in this area would need to assess the generalisability of our methods. Third, although we assess different geographical contexts, both are in the UK. We do not know how well the method would perform elsewhere in the world, where other transport options may be available. Furthermore, we have only tested the current model on adults, and further research could assess how well our model performs on older adults or children, to potentially be of use in a wider group of studies.

Applying the model to the ENABLE London baseline and follow-up data

Using the algorithms developed on a subset of the baseline ENABLE London GPS and accelerometry data sets, we have been able to characterise the travel use of a large group of adults in London and identify the time spent in different modes of travel. Furthermore, we have examined changes in the time spent in different travel modes following a move to East Village. However, despite the built environment variables showing that East Village is a more walkable area, we found no change in the time spent walking or cycling compared with those living elsewhere. However, there was some suggestion that vehicle travel had decreased, particularly in the intermediate housing sector, and that underground travel had increased, more so in the market-rent sector.

Strengths of applying the model to the ENABLE London data

The East Village neighbourhood was designed to encourage active living. Restrictions are placed on parking facilities in East Village but it has good public transport links, being close to Stratford station with both overground links and underground links. Half of the cohort were living in East Village at 2-year follow-up and half had either stayed at their baseline address or moved elsewhere. Following up the same individuals 2 years later enabled us to examine change in behaviour at an individual level with participants acting as their own controls. The GPS monitors were worn at the same time as the accelerometers, which provided direct and time-stamped comparisons of physical activity and location. GPS monitors rely on having good signal links with satellite receivers and the signals can be blocked by some buildings and when travelling underground, but the concurrent use of accelerometry data helped to identify different modes of travel and daily wear time. The eligibility criteria for moving into East Village, set externally by the housing association and companies, meant that we could examine differences between three housing sectors: social, intermediate and market-rent. Eligibility for these different housing sectors was based on income, which provided a measure of socioeconomic status. Finally, our method of analysis, regressing follow-up-values on baseline values, allows direct comparison of change in outcomes between the East Village and the control groups and minimised the issues of regression to the mean.

Limitations

Lack of adequate wear time led to a reduced sample of 548 participants with GPS data at both baseline and follow-up. However, baseline demographic characteristics were similar to the full cohort. After data collection had been completed and the GPS data were being processed, it was discovered that the East Village housing blocked the transmission of the GPS signal leading to low GPS-recorded time in the East Village group. However, as the accelerometers were worn at the same time as the GPS monitors and provided an estimate of wear time, we were able to use this as a proxy for wear time for the GPS monitor. Additionally, the decrease in total GPS minutes was almost identical to the decrease in stationary time in the East Village group, confirming that the GPS time lost was most likely to be spent stationary by participants in their East Village homes.

Conclusions

In summary, we have developed a method to identify travel modes from accelerometer, GPS and GIS data for the ENABLE London study, which successfully predicts > 90% of points tested in a range of contexts. This method can be of use to complement existing analyses of physical activity data and assess active travel alongside physical activity. All code necessary to replicate the analysis, apply the method to other data sets or predict from our models to other data sets is provided to facilitate usage by other researchers.

Outcomes

Statistical methods

Baseline mental health and well-being

Logistic regression analyses showed appreciably higher levels of depression among participants seeking social housing than in participants from the intermediate group [odds ratio (OR) 2.5, 95% CI 1.6 to 3.9] (Table 19, model A). The ORs were similar after adjustment for age, sex, ethnicity (model B) and marital status (model C), but marginally attenuated after adjustment for educational level (model D) and health status (model E). Levels of depression were marginally lower in the market-rent group than in the intermediate group for model A but almost identical in models B to E (see Table 19). The social housing sector were more likely to be anxious than those seeking intermediate housing (OR 1.3, 95% CI 0.9 to 1.8; model A) and ORs were slightly stronger after adjustment for age, sex, ethnicity (model B) and other covariates (models C to E). Levels of anxiety were also higher in the market-rent group than in the intermediate group, with similar ORs (OR 1.3, 95% CI 0.9 to 2.0; model A), which remained after adjustment for covariates (models B to E). Those in the social housing sector were more likely to report lower levels of life satisfaction than those participants seeking intermediate housing (OR 1.7, 95% CI 1.2 to 2.3), although this was progressively weakened by adjustment (models B to E) and was no longer statistically significant. Similarly, there were no significant differences in low levels of satisfaction between market-rent and intermediate groups. Low feelings such as ‘things you do in life are worthwhile’ were marginally higher in social housing and market-rent groups than in those seeking intermediate housing, but differences were not statistically significant and not materially affected by adjustment (models A to E). Low levels of happiness were higher in the social group than in the intermediate housing sector (model A, OR 1.3, 95% CI 1.0 to 1.8), and in the market-rent group compared with the intermediate housing sector (model A, OR 1.5, 95% CI 1.0 to 2.2), which remained similar after adjustment for sociodemographic factors (models B to D) and health status (model E).

Neighbourhood perceptions

There were consistent inverse associations between neighbourhood perceptions and all mental health and well-being outcomes, where better perceptions were associated with poorer mental health and well-being scores (see Report Supplementary Material 21). Hence, we examined the impact of further adjustment of baseline housing sector differences in depression, anxiety and well-being for neighbourhood perception scores (Table 20). Additional adjustment for neighbourhood perceptions of crime (see model 2, Table 20) and quality (see model 3, Table 20) weakened the higher levels of depression and other marginally higher adverse mental health/well-being outcomes among social housing participants compared with intermediate participants, suggesting that neighbourhood perceptions are a partial explanation for these housing sector differences. However, adjustment for neighbourhood perceptions had no effect on the mental health/well-being associations between market-rent and intermediate participants (see Table 20).

A comparison of the baseline characteristics of those who subsequently moved to East Village with those of people who did not (Table 21) showed little difference in mental health and well-being outcomes, overall and by housing sector. The exception was that there were higher median anxiety scores in the control group than in the East Village group among social housing participants and conversely higher levels in the East Village group than in the control group among the intermediate group. There were also baseline differences in neighbourhood perceptions by subsequent control and intervention groups. The overall baseline crime-free and quality scores were markedly higher in the control group than in those who subsequently moved to East Village; for quality, these differences were also evident by housing sector.

Follow-up

Moving to East Village was associated with small improvements in depression and anxiety scores (except for the social housing sector who showed a slight increase in anxiety), although none of these effects was statistically significant (Table 22). Although moving to East Village was also associated with modest improvements in well-being scores, none of the effects was statistically significant, except for life satisfaction among the intermediate sector (0.34, 95% CI 0.04 to 0.65). However, moving to East Village was strongly associated with improved neighbourhood perceptions of a crime-free and better quality neighbourhood, both overall and by housing sector. Notably, the largest effect size for improved perceptions of a crime-free neighbourhood associated with moving to East Village was observed among the social housing group (see Table 22).

Relating 2-year change in neighbourhood perceptions to mental health and well-being scores (see Report Supplementary Material 22) showed that improved crime-free and quality scores were associated with decreased levels of depression and anxiety (all p-values ≤ 0.01). Although there was less of an effect of neighbourhood perceptions on well-being scores, there was some evidence that improved perceptions of quality were related to reduced reporting of low levels of life satisfaction (p = 0.05) and feeling worthwhile (p = 0.01).

Baseline findings

At baseline, those seeking social housing were more likely to be depressed, anxious and less satisfied with life than those from intermediate and market-rent housing sectors. These housing sector differences in mental health and well-being outcomes appeared to be partially or wholly explained by demographic factors, including age, sex, ethnicity and partner/marital status, which are all well-recognised factors associated with psychological distress and well-being. 199,200 However, increased levels of depression among the social group persisted after adjustment for these potential confounders. This finding is consistent with a body of literature showing that those in social housing are more socioeconomically disadvantaged and at greater risk of poorer mental health outcomes, which is associated with increased stress exposures related to their surroundings. 22,201–203 Poorer well-being and neighbourhood characteristics, such as perceived levels of increased crime and decreased safety and access to green space, have all been shown to have adverse effects on depression, anxiety and well-being;46,188,194,204 it has been suggested that prolonged exposure to poor-quality neighbourhoods may have an amplifying effect. 205 Our results show that positive perceptions of neighbourhoods being crime-free and quality appeared to reduce levels of depression (as well as moderating non-significant associations with anxiety and poorer well-being) among those in the social sector compared with those seeking intermediate housing, suggesting that the negative neighbourhood perceptions observed among those seeking social housing may partly contribute to the higher levels of depression within this group. It was noteworthy that the sociodemographic characteristics of those seeking market-rent and intermediate accommodation were similar, and there were little differences in mental health and well-being except in having a marginally lower prevalence of feeling happy the previous day among the market-rent group. Allowing for neighbourhood perceptions attenuated these modest differences in levels of happiness and, unsurprisingly, did not alter the similarities in other mental health and well-being-related outcomes.

The differences observed with the social sector demonstrate the potential importance of the built environment on mental health among the more disadvantaged, which is entirely consistent with a body of literature showing the effect of the built environment on depression, anxiety and well-being. 153,187,204 Hence, it remains entirely plausible that the mental health outcomes of those affected by poorer surroundings might be improved by moving people into environments that encourage positive well-being,206 which longitudinal follow-up of the ENABLE London cohort was able to examine.

Follow-up findings

Despite observing sizeable improvements in objective measures of the built environment (see Chapter 3) and neighbourhood perceptions associated with moving to East Village, particularly in the social sector where there were marked improvements in perceptions of a crime-free neighbourhood, there was little evidence of improved mental health and well-being outcomes after 2 years (although there was a modest improvement in life satisfaction among those living in intermediate East Village accommodation).

Unfortunately, there is very limited literature to put these study findings in context, as few longitudinal studies have examined the long-term effects of urban regeneration schemes on health, and even fewer have examined effects specifically on mental health and well-being-related outcomes. 207 One notable exception is a large-scale experimental study carried out in five US cities, the Moving to Opportunity (MTO) study, which examined the long-term effect of being randomised to receive or not receive housing vouchers to encourage moving from high- to low-poverty neighbourhoods on the physical and mental health of 4606 minority low-income families. 208–210 Although improvements in living conditions among the intervention group were demonstrated over a 10- to 15-year period, there was little difference in adult mental health-related outcomes and economic self-sufficiency between groups, although a small improvement in subjective well-being associated with moving to less deprived neighbourhoods was observed. 206,208 However, there were mixed effects among the participants’ offspring. There was evidence of increased rates of depression, stress and conduct disorders at follow-up among teenage boys who moved to less deprived areas but, conversely, reduced rates of conduct disorders among girls. 210 Other more direct evaluations of urban regeneration schemes in Spain211 and the Netherlands212 have shown little effect on mental health outcomes from after 3211 to 5 years. 212 From a UK perspective, a large long-term housing improvement programme in Glasgow, Scotland, showed a small positive effect on mental health scores after 2 years213 and found that mental health scores in the most deprived areas that received higher levels of investment improved more after 5 years than lower investment areas. 214 However, effects were small and less favourable findings showed that the scheme may have actually increased fear of crime, as relocation may have disturbed established social networks. 215 Other UK studies that have examined the effect of urban regeneration programmes, including change in the built environment, have also shown little effect on mental health and well being outcomes, although change in the built environment has not always been well defined. 216 The closest study to ours geographically also used a natural experiment to examine the effect of London Olympic regeneration among 2254 children attending secondary schools across the London Borough of Newham, compared with pupils attending schools in other east London boroughs. They found no effect of urban regeneration on self-reported physical activity, mental health and well-being outcomes after 18 months, and repeated cross-sectional surveys among 995 parents suggested that levels of anxiety and depressive symptoms might have increased as opposed to decreased in Newham, compared with control areas. 217

In summary, the effects of urban regeneration on mental health-related outcomes appears modest at best and mixed, not applying equally across population groups. Our finding that moving to East Village did not have any sizeable effect on mental health/well-being-related outcomes is entirely consistent with these studies, suggesting that more needs to be done to demonstrate the effect of change in the local environment on mental health and well-being, and to understand why greater differences are not observed. 96

The ENABLE London study was carried out to provide robust evidence of the effect of the built environment on health (particularly in a UK context). However, a key limitation is that the mental health and well-being outcomes were secondary to the main hypothesis of the study (where the primary outcome was physical activity) and lacked statistical power. Moreover, the staged recruitment, whereby those in social housing were moved in before those in other housing types, before the East Village development was fully complete, may have dampened exposure effects and not allowed sufficient time for social networks to become established. Although no appreciable changes in physical activity and adiposity outcomes were observed (see Chapter 3), which could have plausibly affected mental health and well-being, it remains possible that longer-term follow-up may demonstrate significant effects. However, this seems unlikely given the modest effects on mental health and well-being outcomes observed after 2 years and that continued development of the area, including high-rise accommodation blocks and reductions in green space,218 could potentially lead to adverse effects (see Chapter 9).

Despite the growing need for more housing, particularly in major cities, opportunities to examine the potential health impact of such developments are limited. It is widely accepted that area regeneration programmes should be designed to have positive impacts on its residents, as well as reducing health inequalities. 219,220 However, it is challenging to create high-density urban environments with appropriate local facilities to promote positive health behaviours while also protecting residents from the potentially adverse effects of high-density housing. 221 There is little understanding of what is considered to be the optimum density to encourage social contact while moderating other exposures, especially among the more disadvantaged. 153 Further research is needed to establish how people from different socioeconomic backgrounds living in the same neighbourhood interact with each other and their local built environment to identify specific aspects that may be contributing to poorer health. Such studies will inform the planning and design of housing developments for improved mental health and well-being.

Cost of living

Participants commented on issues around the expense of living in an area of London that is undergoing rapid redevelopment and gentrification. They were concerned about the higher costs associated with living in the new development, including higher costs of basic utilities and council tax, which meant they were unable to save:

My first thing is the expense. I can’t save. Whereas in my other house, I could save. I can’t save at all. Obviously, it’s a band D area, I wasn’t in a band D, so council tax has gone up, the rent has gone up. I’m paying for water. I’m paying to heat my water. I’m paying the electricity. I’m paying to clean my communal area. It’s too much. I can’t save. I have to pick and choose the brands I buy in the supermarket to try and keep my costs down. I’m more financially strained now.

A, female

And to compare with what we were paying before and here . . . my water bill has gone by, I’d say about 70%, so that’s a lot . . . We were told it was going to be a decrease. One of the reasons I chose the house is that it would be 25% less living cost compared to where we were living. So I thought, ‘Oh! 25%. That’s quite a bit.’ But actually it’s about 50% increase.

B, female

The high costs were cited by residents as a major reason why they would not consider living in East Village long term:

I don’t see myself living here. With my children. It’s a lot of expensive. You can’t save money.

C, female

. . . more and more expensive, I just think, oh right, because they say it’s the East Village, I don’t know if you’re paying because of all the re-development they’ve done, we’re now paying for it, umm, but it wouldn’t be feasible to live here long term, if it carries on.

G, female

Amenities such as shopping facilities were limited for the social housing sector who were the first to move to East Village. The residents did not really use the Westfield Shopping Centre, which forms a gateway to the Olympic Park, citing expense as the main barrier. Many reported going to shops near their former residence:

And it’s funny to say, I don’t shop at Westfield, I do go around, just to browse around. I’ll probably go to some of my familiar shops to do you know like clothes shopping and stuff, but I don’t buy my groceries . . .

F, female

Waitrose did until umm Sainsbury’s opened. It did get used but I used to find– I’ve got grown-up adults children, who are eating for the world, I would go in there and get like two bags and it would be like £20/£30 and that was nearly enough every day . . . Oh my God, I thought, I can’t live here.

B, female

Perceptions of new home

East Village residents were generally happy and reported a sense of pride in their new homes. Many noted that there was a greater sense of calm in their new homes associated with better family space and family members having more of their own space. Better ventilation, no damp and brighter homes were all noted as positives:

The positive is because my children are older they’ve got their own rooms now. We’ve got a bigger kitchen . . . and now the whole family can sit down around the table. We have a family meal every day, whereas before it was once every so often. So it’s brought us closer . . . This place is cleaner. The area is cleaner compared to where I lived before. . . . there are a lot of positives. Everyone’s calmer. It makes the family a little bit more calmer. Also, they’re very proud: it’s a nice place to live. They feel happy.

B, female

I’m quite happy here. It’s bigger and the children are a lot more happier here . . . I feel secure in here because the security is 24 hours . . . We have a bigger kitchen now and we have family meals every day now, we eat together. They’ve all got their own rooms. So we [are] calm.

D, female

Anyway, I live in a townhouse and I found the place very good. As you said the ventilation is excellent in the house. The rooms are big and everything. About the house I’ve got no issues with the house, it’s just excellent.

D, female

Despite this, the families also had some complaints about the new homes. Individuals in the flats complained that there were limited storage options and strict rules regarding where they were allowed to store items. A further complaint centred around rules for the social housing tenants, which stated that they were not allowed to personalise their flats, including adding curtains to the windows, which had an impact on privacy:

Most of the proportions are– once you put a wardrobe and a chest of drawers in it’s really tight.

E, male

I must say that what we’ve said about storage, the storage was better in my old house because I did– I’m not necessarily a hoarder, but I have things that I know that, for example, we will use in the summer, whereas now I had to buy storage, I had to buy a big storage to put on my balcony. I’ve had to buy storage containers to put in my hallway. I’ve had to buy another big shoe storage to put– I’ve had to spend loads of money, hundreds of pounds, just to– because I’m a believer that umm everything should have a home and I don’t like clutter, so if it’s out of order, I throw it in the bin, and obviously– so as I said storage is a big problem. Can’t store my buggies anywhere whereas in my old house we had like a downstairs sort of like cupboard where we could store stuff.

A, female

Perceptions of the neighbourhood

Overall, participants were very positive about their new neighbourhood, in particular commenting on the cleanliness, order and greater security compared with where they lived before:

That’s really nice actually. I think the greenery is lovely. The greenery is lovely. The way that like we’re close to transport, close to Westfield, all of that’s nice as well. The streets are clean. I like the way that they have tried to upkeep the cleanliness of the area as well, you don’t hardly see any garbage on the streets. Even down to– well I haven’t experienced it, like unruly kids on the street, like– you know you get normally like people hanging around, because that was my fear here, was coming here and having like little kids hanging around, gangs hanging around, or whatever, because of the way that the community is laid out, but I haven’t experienced any of that, I haven’t seen any of that yet.

A, female

However, participants also reported certain frustrations with the development. In particular, many of the social housing tenants had young children and they were concerned about the number of restrictions on children’s play areas. Despite having access to the playing areas in the Olympic Park, many parents felt that those areas were too far away for their young children to go to alone or to play in unsupervised:

I feel that there is a lot of restriction, so children cannot, for example, play ball games, ride bikes in the terrace. If they have to they have to go to different parts, even though the parks are not so far, but it requires leaving everything.

F, female

The issue of limited facilities to keep teenagers engaged was also noted by the residents. Participants were concerned that as children grow up, there would be little for them to do, which could lead to antisocial behaviour. This was a major source of worry for these parents, and would be very likely to have an impact on their decision to stay on in the development:

I think also like for my daughter, who is a teenager, what we were saying before, there’s not much in the area for them to do, apart from going to Drapers Fields and I think there should be like a little community thing where they can do things like table tennis and– under supervision, you know, like after school, because my daughter does get bored. She just– in my area, in my block, in the community, there’s a lot of young children, not much children her age, and I’ve noticed that children who come to the school that do live in the area umm I don’t know, it’s like they are forced to argue. Because they don’t really have much to do, a lot of the time my daughter’s on her tablet, with people in the area, umm, abusing each other, and I think if they– yes, no, but seriously, there’s no sense of cohesion with the teenagers in the community. There’s a sense of cohesion with the small ones, but me personally– and a lot of the time I see them sitting on the steps and stuff like that, as I was saying about finding the condom in the sand pit area, they are doing things like that, whereas if there was something positive that they can go to . . .

A, female

People hanging around on the staircases, it’s going to get worse and worse and especially if they don’t build a communal area for the teenagers, it’s going to get worse and worse. They’ve only built it up for young children. So that’s one thing that I’m dreading in the future and that’s one reason, another reason, why I’d want to leave as well is . . .

A, female

Sense of community

Participants’ relationships with their neighbours were generally positive. The management was also active in organising a range of activities for residents to meet each other and managers. However, it appeared that only a few were aware of or had made use of facilities organised by the management:

Well my neighbours are really really nice. My neighbours on my floor are really really nice because we send each other Christmas cards, even when I had my daughter, they sent me gifts, etc. That’s really good. I must say that I don’t– I know there’s a lot of community things going on and they display it on the boards in the flats, but I don’t make use of them personally, so I can’t really.

A, female

Encouraging healthy living

East Village was designed to encourage healthy living and has access to gym and sports facilities in the Village and nearby Olympic Park, as well as the swimming facilities in the Aquatics Centre. Parents with young children felt that facilities for children to play should be made available closer to the houses. A few concerns were also raised regarding security in the available play areas:

And we’ve got so much open space, I think for children, they should do something– like a proper park, with swings, a slide, things like that, because they don’t want you to use the communal gardens, so they should provide a little local park for them.

G, female

In comparison to my old house we had like, as we said, a little park, where my child could go to. Here you’ve got little bits of parks, here and there, which is a bit of a bummer and I don’t think some of the parks are controlled very well around here, whereas in my old house, they were very very controlled . . .

A, female

Overall, experiences with increasing physical activity were mixed. Participants were often unaware of the facilities available to them and many reported not using them. In certain cases the lack of local shops made it necessary for people to travel further away to buy things, which meant they were more likely to drive or use public transport rather than walk, therefore reducing their physical activity:

Umm I would say mine is the opposite. Umm– because where I used to live before, umm, you could walk to like the local shops, etc. I know I’ve got Sainsbury’s now but sometimes if you want something elsewhere, I used to like umm– because of my culture we buy certain foods that we can’t buy in Sainsbury’s, but I used to have shops close to me that I could just walk to, but now I have to drive to work to get those things, so I won’t walk, I will just take the car, so it’s restricting me, umm, I don’t walk as much in that sense now . . . The only difference I would say is walking from my house to the car park, that’s probably the only thing that’s changed, because in my other house, I could just walk outside my house and my car would be there, so in terms of physical activity, no.

A, female

I think I’ve gained some weight since I have moved from my old address to this new address and that’s partly because, what you just said, I drive around as well, I don’t– because I find things a bit far, I mean, where I was in my old home before, I could go to my work, because I work in the school as well, that’s part time. I used to take a walk to the children’s centre but now I can’t take a walk, it’s quite far, so if I have to get to my children’s centre, I have to drive.

F, female

Umm I’m less active. Because where I lived before I used to walk to work, walk home, now I have to get a bus.

G, female

Others reported using the facilities to a greater extent and encouraging their children to do the same:

I do make use of the– because I go to the gyms as well there and the swimming. I get the kids involved. I’ve got a 12-year-old who has put on a bit of weight since we moved in here. He actually put on weight when we moved here, so I said, right, swimming, basketball, table tennis, these are all free for him, so I said right, I’m registering you in to everything and then what we do we go there, weather’s good, soon as I’m finished from here that’s what I’m getting ready to go, we will sit and spend the whole day there.

B, female

Well for me, my activity is because I have to walk more. Before I was not. Because the parking (. . .) to leave your car at work because when you move your car, when you come back, there won’t be parking, so why don’t you leave it here, walk, just across the other side of Stratford and go to work, so health-wise I’ve seen a lot concerning my life and the children and then, umm, the rest I don’t know the rest but for me it works for me.

C, female

As East Village was designed to encourage active living, there are restrictions on parking and limited availability of parking spaces. Although residents were made aware of this, some participants, particularly those with large families or with health problems, found this restriction limiting:

The only problem is the parking because I can leave my car and go to work, but when it comes to shopping, as a family of five children, there’s no way you can do shopping by bus or train, so you need that car, but sometimes I have to wait to go shopping, you can’t have a place to park, even though you have a free parking along there, outside, or– so you have to pay the inside one, which is £60 a month, which is very, very expensive for a mother of five, so that’s the only problem, apart from that I’m really enjoying the place.

C, female

I needed my car, but coming here, because I never had two children under the age of 10, I wasn’t allocated free parking, which was horrible, I mean I’ve had parking tickets, I mean the parking guidelines are not clear here, as well . . . I wasn’t allocated underground parking, which was obviously a thing, and now I’ve got another daughter as well, so it’s a big problem, so I have to pay £65 Stratford International to park my car every month.

A, female

I gave my car up just as we moved in, umm, because I wasn’t allowed– I applied for the parking, I didn’t get it, and you know, the one that you get with the underground, and at that time, I actually didn’t know about the council parking, because I did ask about that, and all they told me about was the parking that you know the Triathlon– the East Village parking so I wasn’t aware of [whether] we was going to get a council parking, afterwards, because the first car is free, so I got rid of my car, so I do find shopping is a little bit hard, because I’ve got umm early-onset arthritis and to carry shopping from anywhere it’s a bit difficult, but I mean, I suppose we do online shopping and get them in somehow.

B, female

Design

This was a qualitative study based on an analysis of semistructured interview data generated with individuals living in East Village.

Participants and sampling

Purposeful recruitment was used to select participants with regard to change in travel behaviour over a 2-year period based on measures derived via an algorithm developed for the ENABLE London study, which used GPS, accelerometry and GIS measures (see Chapter 5). 166 Participants eligible for recruitment were selected on the following basis:

• They all were living in East Village at the follow-up interview.

• They met the ActiGraph compliance threshold at both phases (including ≥ 1 day at both time points).

• Participants from all housing sectors were included (social, n = 40; intermediate, n = 92; and market-rent, n = 14 adults).

• Information on age, sex and ethnicity was also provided.

From this sampling frame, three groups were targeted based on change in travel behaviour:

1. Those who were in the lowest quartile for the proportion of time spent in active travel at baseline and who had increased the proportion of time spent in active travel (walking and/or cycling) by ≥ 5%.

2. Those who were in the lowest quartile for active travel but did not change the proportion of their time spent in active travel over the 2-year period.

3. Those who decreased the time spent in active travel by ≥ 5% over the 2-year period.

Participants from the control group were not invited, as comparing perceptions of new and old neighbourhoods and the change in travel behaviour in response to the move was a key element of data generation. Exclusion criteria were limited to anyone who was non-English-speaking and all participants were offered a voucher incentive. All participants were given information about the study and gave their written consent before interviews took place, including for the recording of the interview. Focusing on those in the lowest activity levels is important because they potentially have the most to gain; for example, Kelly et al. 248 found in their systematic review that reduced risk of all-cause mortality was greatest in individuals with the lowest levels of activity (i.e. ≤ 120 minutes of walking or ≤ 101 minutes of cycling per week).

Qualitative data generation

Semistructured interviews were conducted in participants’ homes between November 2017 and January 2018 by two experienced qualitative researchers (Rachel Hahn and Bina Ram). The responses were recorded on a digital recorder, transcribed verbatim, anonymised and stored on a University of Bristol secure server before being imported into NVivo10 (QSR International, Warrington, UK) for data organisation. The interview schedule was based around comparative perceptions of individuals’ old and new neighbourhoods, aspects of everyday travel and leisure and how moving to the East Village had influenced this (see Appendix 2 for the questionnaire used). This study provided a unique opportunity to use individual travel patterns, which were generated by a programme using GPS, accelerometry and GIS data as part of the ENABLE London study166 to provide the framework for a spatial narrative whereby residents could describe their routines, influences and decision-making within a spatial framework. Individual maps were generated for the interviews, which described patterns and places before and after moving to East Village. The interview schedule was developed in discussion between a researcher from the ENABLE London study (AP), a PhD student who was analysing interviews as part of her PhD (ER) and the researcher who conducted the majority of the interviews (RH). Interviews were structured to take no longer than 45 minutes (see Appendix 2 for the interview schedule). One-to-one interviews were selected as the method of inquiry because of the highly variable activity patterns of individuals, which meant that the interviews could be tailored more effectively.

Data analysis

For the purposes of this analysis, inductive thematic analysis was used. 235 Inductive thematic analysis followed the six steps outlined in Figure 7 and themes were generated from the data inductively rather than by imposing a predetermined coding framework.

FIGURE 7.

Thematic analysis steps. Information sourced from Braun and Clarke. 235

Perceptions of the environment and relationships with physical activity

The majority of participants generally expressed more positive perceptions of their new built environment in the East Village irrespective of gender, type of housing or whether or not their travel had changed since moving to East Village. The initial preliminary themes generated from the interviews are extracted and summarised below in relation to change in four key features of the built environment: Functional features, Feeling safer, Aesthetics and Destination features. This structure allows interpretation in relation to existing frameworks249 and the most contemporary systematic review on qualitative studies on the perceptions of the built environment and physical activity in adults. 250 Examples of functional features include direct routes, street and path design and maintenance, traffic control and vehicle parking. Safety features include surveillance, crossing aids and lighting. Aesthetic features include cleanliness, interesting sights, maintenance, greenery, architecture, noise and pollution. Destination features include proximity, accesses and availability of local facilities, green spaces, shops, health care, sports facilities, parking facilities, public transport and other destinations.

Functional features

The main functional feature that emerged from the interviews was parking and its impact on regular access to a car. Perceptions on the feasibility of owning a car varied. Negative perceptions included:

Now if you have a car, it’s a problem, the biggest problem is having a car here.

F (female), SH (social housing), NC (no change activity group)

It’s too expensive to drive.

F, SH, D (decreased activity)

Some residents had a more positive view, and indeed owned a car, mainly because of the option to purchase a dedicated car parking space in or close to the building in which they lived. Although a significant cost, this was seen as attractive, even if it was not used that often, and affordable relative to other costs and income, largely for those in the intermediate housing sector. Car ownership was reported less by those in the social housing sector, who tended to rely on friends and taxis for access. Other transport options influenced the decision whether to own a car or not, which was expressed as a weighing-up of options, for example the cost of car parking versus the cost of train travel at weekends:

From this side of London you’ve got access to quite a few different sort of bigger places and the train fare [sic] sometimes are just ridiculously high for a spur-of-the-moment kind of thing that you want to do.

F, IH (intermediate housing), D

Cars were rarely used during the week if a participant was commuting for work into London, as public transport was a more viable option. Good public transport infrastructure was frequently cited as a key reason for not needing a car:

There’s no point having a car.

F, SH, D

Furthermore, the view was expressed that it would be different if public transport infrastructure was not available:

If I lived somewhere with minimal transport . . . I’d get a car.

F, SH, D

Other than differences between housing sectors (e.g. less likely to own a car in the social housing sector), car ownership did not vary particularly between the activity change groups (increase, no change, decrease).

Feeling safer

A consistently reported change across all groups since moving was an improved perception of safety to be out at night. This was largely a result of improved lighting and the presence of other people, and led to more reported walking at night both for leisure and to get to destinations than was reported in their old neighbourhood:

It’s going to be 9:30 or 10, so there’s still activity and there’s lighting, whereas where I lived before, that whole side I had to walk it was a very quiet road, there’s no lighting there, so that makes a difference for me . . . you can even walk at 11 p.m. There’s always activity . . .

F, SH, I (increased activity)

This was most frequently reported by women in relation to personal safety and removed what was a recognised barrier to activity:

. . . the perception is it’s safer and then when you actually see like the crime stats and things it is safer, so that kind of makes you think, right, I’ll go out for a run, I’ll be fine.

F, IH, D

This was also reported in relation to the general design, which had wider open spaces to increase visibility:

I’ve never felt unsafe . . . feel more unsafe in areas where there are more cars because there’s more potential for me not to see something coming up on me, whereas here . . . you can tell someone’s intention when they’re on foot, so in that respect, I feel safer here.

F, IH, D

Therefore, the combination of good lighting, other people being around and space to observe those around were key to the improved perceptions of safety at night.

Aesthetics

Participants reported the advantages of wider pavements and diverse buildings, with the combination of more green spaces amid the built spaces seen as a positive compared with previous neighbourhoods. Noise, although not traditionally included as an aesthetic feature, is included here as it was mentioned by many participants in several contexts both positive and negative.

A reduction in traffic-related noise was consistently reported as a positive aspect of the residents’ new environment; however, noise within East Village was also reported as a potential negative. This was dependent to some extent on familial status, with residents without children commenting on children’s noise from playing outside whereas those with children commented on loud parties/groups sometimes in the communal areas, such as the barbecue areas. This affected the use of space and the choice of spaces to go to (e.g. to avoid identified ‘play spaces’ or those that attract crowds, such as barbecue spaces). This led to clear patterns in use of space with residents matching their needs and led, to some extent, to spatial segregation. However, the large amount of space and diversity of options meant that this did not necessarily restrict the use of outdoor space although to some extent it limited social interaction between some groups:

I don’t use the central courtyard at all. I just find that really awkward. It’s a nice space to have but . . . it doesn’t feel very adult to me, so there’s lots of different sized families and people that live here and the kids use that space quite a lot, but it doesn’t really feel very grown-up to me, so I just don’t bother with it.

M (male), IH, D

Destination features

Positive perceptions not enough

Although, as highlighted above, the residents’ perceptions of environmental features in the East Village were generally very positive, these positive perceptions were not in themselves enough to promote increased physical activity overall. Other factors mediated the relationship between the environment and the participants’ behaviour. Some of these were individual factors related to motivation to be active for health reasons and a sense of being active to take responsibility for your own health:

. . . if I have to lose weight, let me discipline myself and walk up and down [I don’t need a gym].

F, SH, I

The environment was a prompt for participants with a lot of previous activity experience to reintroduce old practised behaviours, such as re-joining a gym or taking up running again. A change in circumstances also superseded environmental influences, for example new work patterns (e.g. working long hours), change in capacity (e.g. because of surgery) or age, which was a reported barrier to taking up new activities owing to high levels of perceived risk, even within a supportive environment:

No, not going to ride a bicycle, not at my age now. Even if I did it would be in a very safe place. I’m not Lara Croft!

F, IH, I

Early stages of neighbourhood development

It was also acknowledged by participants that East Village was still developing and did not feel as vibrant as more established areas:

I think it’s because we . . . could kind of walk down . . . which was kind of cool, a bit snazzy and then make our way back through the back streets or something, whereas here it’s not quite the same, like you haven’t quite got the same vibes.

F, IH, I

However, it was also noted that the East Village was changing and evolving into a more interesting space as new facilities were introduced, albeit they were relatively limited:

. . . it feels like its grown into itself so there’s much more going on, like you said, the café’s downstairs, there’s some restaurants . . .

F, SH, I

Part of this evolution that was particularly slow was the social development between residents within the village. Many residents reported having positive experiences with visitors to East Village and maintaining contact with their old neighbourhood by visiting or by using social media, but there were few examples of social interactions between residents apart from among children within families and for structured activities, for example playing football. Residents were aware of group activities but did not report participation in activity. There was, however, participation in more static social events, such as when football was screened outside or in the barbecue area.

The level of social connection was to some extent influenced by a sense of connectedness to East Village. Some were less connected as a result of strong ties to their previous neighbourhood; this was particularly true for younger adults for whom the move was more of a family decision than their own decision. In most cases this was the result not necessarily of a negative view of the new neighbourhood but of a specific connection to their old neighbourhood. Others had little connection to their old neighbourhood and were keen to explore their new neighbourhood, extending their range and sense of place as time went on. They were keen to invest in their new neighbourhood environment marked by, for example, walking and/or testing new neighbourhoods, appreciating new activities and shops, and having a clear sense of pride in where they lived.

Explaining the effects of the intervention

Understanding the mechanism through which the built environment might influence health behaviour is not straightforward and may well be multifaceted. 96 Reasons for the lack of clear effects of the East Village development, the ‘intervention’, on the physical activity-related behaviour of residents remain to be fully understood,33 especially when findings from cross-sectional studies are less equivocal. 93 One potential explanation is that those seeking to move are inextricably different from those who do not. The ENABLE London study sought to minimise this bias by recruiting a cohort of individuals seeking to move to East Village, which we believe, in the absence of randomisation, offers the best possible study design. It is also possible that the magnitude of change in the built environment was insufficient to induce a major change in physical activity. However, the East Village development showed marked changes in objective GIS-derived markers of the built environment and sizeable improvements in neighbourhood perceptions (particularly in perceptions of crime and quality) among residents moving to the area compared with those who did not move or moved elsewhere. So why did the change in the built environment not result in a bigger change? This report ventures a number of explanations, including whether exposure to the environment was long enough or sufficient enough to induce the change expected; the environment was incomplete when the first social housing residents moved in and was compromised by further high-rise development and loss of green space more latterly. 218 In the absence of more discernible effects on health and behaviour, it is difficult to establish clearer explanations for the lack of more definitive findings. However, it appears that the built environment alone is insufficient to induce the change in health behaviour expected by passive means and that more interactive strategies, perhaps in conducive environments such as East Village, are needed (i.e. perhaps higher-agency programmes combined with low-agency environments are needed to induce change). However, although high-agency population interventions are often used, particularly by governments to improve diet and reduce obesity, they are often ineffective254 with time constraints and low levels of motivation often being cited as reasons for poor uptake. 255 Moreover, high-agency interventions that require more effort from the individual may reinforce or even worsen socioeconomic inequalities in health. 254 Hence, low-agency interventions that require less investment from the individual may be more effective, having a small effect in greater numbers and hence may be more equitable by having greater reach. 254 In the context of East Village, greater agency could be achieved through greater awareness among the residents of the activity-inducing features and facilities that exist (such as walkways, cycleways and sporting venues) and fostering the belief, particularly among social tenants, that these are available to all. However, the premise of this study was to examine the effect of the built environment per se, without these additional ‘nudges’ to promote use. 256

Generalising the effects of the intervention

In terms of the representativeness of the ENABLE London cohort, we compared our physical activity data at baseline to a nationally representative study, the Health Survey for England,257 that used a similar methodology: the same waist-worn accelerometer (ActiGraph) worn for an equivalent wear time (1 week). Adults aged 16–34 years from this study recorded 40 minutes per day in MVPA, of which, on average, 15 minutes was in 10-minute bouts. Our baseline data suggest comparable levels of activity among those of a similar age in the social sector with 47 minutes of daily MVPA and 7 minutes in bouts (with an IQR between 1 and 15 minutes), but higher levels among those in the intermediate and market-rent sectors with 65 minutes of MVPA and > 20 minutes recorded in bouts. Although this suggests differences in baseline physical activity levels across the housing sectors in the ENABLE London cohort, there was no evidence of a trend across other social markers (i.e. income groups) in the Health Survey for England study. Moreover, the greater change in physical activity associated with moving to East Village in the intermediate and market-rent tenured groups compared with those in social housing, does not suggest that these higher levels (albeit not significantly higher) limited scope for change (i.e. participants from these non-social tenured groups had reached the capacity for further increases). In terms of geographic patterns in physical activity, reanalysis of the Health Survey for England7 data did not suggest that self-reported higher levels of physical activity in London were unduly higher or lower than other government office regions. 258 Hence, we believe that null findings from the ENABLE London study do not reflect bias within our cohort and are transferable beyond our study population. However, it is acknowledged that a cohort of ‘movers’ seeking to move to East Village, an active permissible space being built on active design principles, were targeted for the study and their activity patterns may differ from the population at large. Hence, this may have some implications for the generalisability of the findings.

In terms of specific features of the East Village development that could have influenced health and well-being (such as the use of pathways, cycle paths, links to public transport, open spaces and leisure facilities) are (or could be) features of many built environment developments. Hence, we believe that findings from this study have substantial potential for wider generalisability to other urban areas. An ultimate goal of the project was to identify evidence-based design features of the built environment that encourage physical activity and improve health behaviours. It was hoped that the identification of these environmental features would provide architects, urban designers and planners with evidence-based urban design elements, which are required for future developments.

We hope that findings from the ENABLE London study will be generalisable to other urban residential housing developments, especially given the urgent need for higher-density housing, and will help inform evidence-based urban planning in future.

Methodological investigations and development

We believe that the ENABLE London study design that we have developed offers a robust test of the effect of the built environment on physical activity and health behaviour, in the absence of randomisation. We did explore carrying out randomisation of accommodation with Triathlon Homes, a partner organisation of housing associations, which manages social and intermediate homes in East Village, but unfortunately this was not feasible. Hence, the approach used offers the next-best alternative. The richness of the data recorded, although making participation and data collection more onerous, provided an excellent opportunity for novel exploration. The development of an automated method of recording travel mode journeys from combined accelerometry and GPS data allowed change in active forms of travel to be examined in an objective automated way, which would have been prohibitively lengthy using previous manual approaches. 48,49 The tool developed (outlined in Chapter 5) has been validated in the cohort (as well as in a separate study) and although it showed little change in active forms of travel (i.e. walking and cycling), consistent with no overall change in physical activity levels (see Chapter 3), it showed modest reductions in car travel and an increased use of public transport (i.e. underground), which was associated with moving to East Village. These methods have been published and made available for use by the wider scientific community, offering an ongoing legacy from the work.

Engagement with patient and public involvement/stakeholders

The ENABLE London study was developed in partnership with a network of both local and regional stakeholders identified through our collaborator links to agencies, which were involved with the design, planning and management of large-scale accommodation developments. Locally, these included local authorities (particularly Newham) and a number of housing associations, in particular the East Thames Group and Triathlon Homes (which continue to be responsible for social and intermediate tenured housing in East Village). Open meetings held by the East Thames Group and Triathlon Homes were attended by the investigators at an early stage of the project, to directly interface with those contemplating moving to East Village, to establish the drivers and reasons for considering moving to East Village, and to ensure that these were fully captured in the forms of assessment. Moreover, qualitative work was also convened early on in the project to gauge experiences among those who were the very first to move to East Village, so that follow-up examinations could be adapted to ensure that all relevant areas were covered, and that assessments and participation remained engaging and enjoyable, and to ensure the continued significance and potential generalisability of the work. Another key aspect of engagement/stakeholder involvement was attendance and presentation at national and international conferences and special stakeholder workshops, particularly those convened by the Town and Country Planning Association at an early stage of the project, to encourage further stakeholder participation and to share experiences with other researchers carrying out similar work; this provided the opportunity for ENABLE London study methods to be refined and optimised. Similarly, publications to date, particularly those describing study methods and baseline data, also provided opportunities for further dissemination and potential input from those working within the field. The investigators continue to be in contact with representatives of the Town and Country Planning Association, Public Health England and the London Borough of Newham not only to disseminate findings from the ENABLE London study but, more importantly, to inform the future direction of research in this area, with the goal of providing stakeholders, architects and planners with the evidence base needed to inform future housing developments.

Challenges of the study

The ENABLE London study provided a unique opportunity to observe a ‘natural experiment’ to assess the impact of a rapidly changing built environment, specifically designed to encourage walking and cycling, on the physical activity patterns of residents. Natural experiments are associated with a number of challenges that are often out of the control of the researchers. Our original plan, as specified in our bid for funding, was to carry out the baseline recruitment between January 2013 and mid-2014 and then to follow up participants after 1 year, between January 2014 and mid-2015. However, there were marked delays in the release of East Village accommodation, which resulted in a prolonged period of recruitment from January 2013 to January 2016. In addition, the delayed opening of East Village facilities (a crucial element of the East Village environment and the natural experiment under evaluation) led to a postponement of the follow-up assessment from 1 year to 2 years. The scientific case for delayed follow-up was strong in that (1) East Village was not fully open for use in 2014 when first follow-up was initially envisaged and (2) previous studies have suggested that it takes 2 years as opposed to 1 year for habitual levels of physical activity to change in response to a changed built environment. 44 The scientific rationale for 2-year follow-up was approved by our Trial Steering Group. As a result, the timing of the follow-up examination (originally envisaged to take place between January 2014 and mid-2015) began in February 2015 and was completed in October 2017. We remained adaptive to these challenges throughout the study by modifying the working hours of field workers in response to demand (particularly making use of part-time as opposed to full-time appointments) to ensure that the research funding that was awarded covered the extended period of fieldwork. However, the prolonged period of fieldwork inevitably resulted in difficulties in consistency of the exposure to the built environment, which, as outlined above, was underdeveloped at first follow-up, becoming fully open and then more latterly overdeveloped from what was initially envisaged (see Limitations of the study). Moreover, the high levels of rental tenure in East Village resulted in a more mobile population, and many of the intermediate and market-rent occupants who took part in the study have since moved away from the area.

Strengths of the study

The building of East Village provided an important novel opportunity to evaluate a ‘natural experiment’ based on the major and focused change of an inner-city urban built environment that has been specifically designed to encourage walking, cycling and healthy living. It is also unique because the study included residents from widely differing socioeconomic backgrounds. Although many developments of this scale are under way or are being planned in cities globally, few have been (or are being) evaluated and most are less easily evaluated given that the timescale of their development is much longer than in the case presented by East Village. The rapid occupancy of this development was a major strength, providing the opportunity for pre and post assessment and to compare 2-year change in health outcomes among those who did and did not move to East Village. The different housing tenures within East Village also allowed for the evaluation of socioeconomic position as an effect modifier, as the impact of the built environment may vary by socioeconomic position. The focus on increasing levels of accessible and low-cost forms of physical activity remains particularly relevant to individuals and households of lower socioeconomic status and has the potential to inform efforts to reduce health inequalities. 259 This allowed us to examine whether or not the built environment favourably influenced higher levels of physical activity (particularly walking and cycling), as well as reducing time in sedentary activities, particularly among low-income groups with fewer opportunities for recreational activities. 259 Moreover, the co-location of this diverse population of differing housing tenures to one community also constituted a social experiment, providing opportunities for residents to observe and learn from the behaviour of others. 260

Another key strength is that the study sought to recruit a cohort who were seeking to move, half of whom succeeded and half of whom did not, which avoided potential biases in health behaviour. In the absence of randomisation, we believe that this study design offers the next-best alternative, providing less biased and more robust evidence about the potential effect of the built environment on health-related behaviours and outcomes. Following the same individuals before and after any move to East Village also offers statistical efficiencies in that individuals act as their own controls, which eliminates confounding factors that do not change within subject. 261,262 Moreover, the consistency of effects using 1 or 4 days of objectively measured physical activity allays fears about potential selection effects associated with using fewer days of recording, which were needed to maximise the number of study participants, particularly among the harder-to-reach social housing sector. The use of accurate objective assessment of physical activity is another strength, including assessments of MVPA that underpin current physical activity recommendations;4 this allows the potential public health importance of the findings to be gauged. However, effects on MVPA were small, suggesting that the change in the built environment had limited impact in achieving current physical activity recommendations, which focus on higher levels of activity. 4 Objective assessment of physical activity assessment reduces measurement error and potential biases arising from self-reported levels of physical activity,88 which could be artefactually associated with moving to a new environment being advocated for healthy active living. Furthermore, we have recently published a paper comparing the effect of a successful pedometer-based walking intervention on change in physical activity observed from accelerometry and IPAQ forms of assessment, and have shown that the change in physical activity was estimated with far less precision compared with the hip-worn accelerometry form of assessment. 263 Hence, use of the IPAQ data would lead to greater uncertainty about the findings, which already show no clear evidence of effect. Similarly, objective measures of adiposity using validated measures of bioimpedance minimise measurement error and limit any potential biases in ascertainment, particularly among a multiethnic population. 264,265 The ability to examine different socioeconomic groups allows social inequalities in effects and potential determinants to be examined. 266 Although there was the suggestion that change in physical activity was greater in the intermediate sector than in the social housing and market-rent groups, there were no clear differences observed by housing sector, emphasising that the study was underpowered to formally examine impacts on health inequalities.

Limitations of the study

A key limitation was that the study was powered to detect a 750-step difference. 61 Hence, the study was underpowered to show that the change in steps observed associated with moving to East Village compared with the control group was statistically significant overall (i.e. 154 steps, 95% CI –231 to 539), let alone in housing tenure subgroups with fewer participants. Effects on MVPA associated with moving to East Village (particularly in ≥ 10 bouts, commensurate with current UK physical activity guidelines)4 were small and uncertain (i.e. effect estimates were associated with wide CIs). Limitations in sample size were a result of the modest participation rates (50–60%) and unforeseen restrictions on recruitment, particularly among the market-rent sector. 61 Moreover, the staged recruitment in which those in the social sector were moved in first when East Village was not fully completed, may have resulted in partial exposure to the physical activity promoting environment and before residential density had reached capacity, which may have limited the full impact of the exposure. 61 This was partly a reason for follow-up after 2 years rather than 1 year, but also because previous work has suggested that longer durations are needed for habitual health behaviours to evolve and to avoid early ‘honeymoon’ effects44,267–269 or, conversely, for residents to become fully familiar and make optimal use of their residential area. 148 The RESIDE study found stronger differences over time, as suburbs evolved and new facilities were built and habits became more established. 95 Further follow-up could plausibly demonstrate beneficial effects but given the mobility of the cohort and the fact that the study is underpowered these are unlikely to yield statistically significant findings. Moreover, East Village is evolving and the building of high-rise accommodation (> 30 storeys) among the existing 10- to 12-storey accommodation, and subsequent loss of green space, may dampen exposure effects even further. 218

Information on social environmental factors, including qualitative research exploring the lived experience of residents, may help to explain why residents of this new area did not respond to an environment designed to be health promoting. It is plausible that the facilities provided did not meet the needs of the new residents and/or they did not feel welcome to use those facilities, especially social housing participants. 233 This requires further exploration. However, another reason why effects were not more evident might be that London is largely a ‘green city’, especially in relation to other capital cities,270 limiting the opportunities to detect change if controls moved to an equally good space. However, the effect was in fact marginally strengthened when comparing East Village with controls who moved elsewhere and the difference was reduced when comparing East Village with controls who remained at their baseline address, suggesting that this is not the case. There was also no effect of moving to East Village on adiposity levels, and no effect on mental health and well-being outcomes of public health importance.

Another limitation was the adult focus of the study, as East Village did not attract as many families as anticipated, leading to a limited number of children being recruited to participate in the study. This was also evidenced by the Chobham Academy, originally designated to educate children moving to East Village, having to increase the catchment area to fill its pupil capacity. This was an unforeseen circumstance, out of the control of the researchers, which resulted in the study being underpowered to determine any change in physical activity and health behaviours associated with moving to East Village among this age group. Hence, this limits the generalisability of the findings, particularly to younger age groups.

Implications for future research

Although it remains plausible that the built environment alters physical activity patterns and other health behaviours, this study suggests that the effects may be small, particularly, as in this case, when people relocate into new high-density neighbourhoods. There are many discussions globally about the health impacts of high-density urban living, particularly high rise, and there is an urgent need to mitigate potentially adverse consequences. 26 However, this does not mean that a small effect across a whole population is unimportant. In contrast to other community-wide interventions, which have shown limited effects on physical activity,29 change in the built environment potentially has greater reach. Even though there were modest effects on MVPA, potentially a 400 step per day change (observed in the intermediate group) across the whole population could have appreciable effects on health outcomes at a population level, potentially reducing all-cause mortality by 2% (95% CI 1% to 4%). 271 However, further evidence is required, particularly about the factors that resulted in social housing residents who relocated actually decreasing their level of physical activity. Such studies would have to be large in scale, larger than the ENABLE London study, to demonstrate change. As availability and affordability of housing is a pressing concern, particularly in major conurbations such as London, and the environments in which we live and the accompanying social context have been shown to be related to physical activity levels (including higher levels of physical activity),104,272 more studies evaluating the effect of large-scale housing redevelopment and/or new development programmes are needed to inform the debate about future housing policy. These studies should be diverse in nature to encompass high-density as well as low-density housing. Moreover, suitability and attractiveness of these housing types to different populations (young and old, different tenures and income groups) should be considered to improve the generalisability of potential findings, whether these be favourable or otherwise.

Although there is a strong scientific case for such large-scale studies, challenges of funding, especially in the current economic climate, will limit opportunities. Hence, novel approaches using simulation studies or alternative low-cost forms of population surveillance may be needed to make such large-scale studies viable.