UK Biobank Urban Morphometric Platform (UKBUMP)


Research Centre: Healthy High Density Cities Lab

UK Biobank Urban Morphometric Platform (UKBUMP): A Big data platform for evidence-based healthy city design

Project Team:

Dr Chinmoy Sarkar1 (Concept Lead, Developer & Principal Investigator), Assistant Professor (GIS, Urban Health & Environment)
Professor Chris Webster1 (Co-Principal Investigator), Dean – Faculty of Architecture & Chair Professor
Professor John Gallacher1,2 (Co-Investigator), Professor & Director – UK Dementias Platform
Ms Sarika Kumari1, Research Assistant

1Healthy High Density Cities Lab, HKUrbanLabs, University of Hong Kong, Hong Kong Special Administrative Region, China.

2Department of Psychiatry, Oxford University, Warneford Hospital, Oxford, UK

Research Centre: Healthy High Density Cities Lab (HealthyHDCities), HKUrbanLabs, University of Hong Kong

Project Funder: The University of Hong Kong Research Assistant Professorship grant, UK Economic & Social Research Council Transformative Research grant, UK Biobank (UKB Research application 11730)

 “Understanding the urban factors that are risk or protective factors for health can capitalize on the positive aspects of urban living and lead to the development of appropriate interventions and preventive measures. Given the growing predominance of the urban living, interventions that take into account features of the urban environment have the potential to be widely applicable and to influence the health of vast number of people”.

Vlahov and Galea, 2003

The Guardian. 6th Oct. 2017. Inner-city living makes for healthier, happier people, study findshttps://www.theguardian.com/society/2017/oct/06/inner-city-living-makes-for-healthier-happier-people-study-finds

Daily Mail. 6th Oct. 2017. Cities make healthier, happier people. http://www.dailymail.co.uk/wires/reuters/article-4953838/Cities-make-healthier-happier-people–report.html

International Business Times. 6th Oct. 2017. We’re better off living in cities – and here are three reasons why http://www.ibtimes.co.uk/were-better-off-living-cities-here-are-three-reasons-why-1642132

Thompson Reuters. 5th October 2017. Cities make for healthier, happier people – report. https://news.trust.org/item/20171005223645-zgrp8/

ACKNOWLEDGMENTS

The study was conducted using the UK Biobank resource (approved UK Biobank Research application: 11730). The authors thank the Ordnance Survey (UK national mapping agency) for providing access to its UK-wide spatial data for use in this study.

 

<a href='http://www.sciencedirect.com/science/article/pii/S1877343516301002' style='color: #00ff00'>ABSTRACT</a><br><br>

In an era of rapid urbanization and significant demographic shifts, human health has emerged as a primary focus of sustainable development. According to the environmental model of public health, the environment is one of the first causes of disease, injury and mortality. In this paper we discuss the complexities involved in urban environment–human health interactions; provide an overview of the environmental determinants of health; and highlight future directions and challenges. A deeper understanding of the relationships between urban environment and human health will help devise effective preventive interventions towards minimizing/offsetting environmental risk exposures reducing resulting health burdens, lead to healthy lifestyle and behaviour and thereby fulfil the goals of sustainable development.
Urban health niche model of multilevel risk clustering, interactions and pathways. (Modified from the Health Niche model of Sarkar, Webster & Gallacher, 2014). Note: The superscripts in the above figure refer to the related research-evidence (denotes citation number in the Reference section of this paper).
Current Opinion in Environmental Sustainability 2017, 25:33–44.
<a href='https://link.springer.com/article/10.1007/s11524-016-0122-1' style='color: #00ff00'>ABSTRACT</a><br><br>
Global demographic shifts in the form of increasing urbanization and ageing populations means that the attainment of healthy urban environments will remain a primary challenge for human society. This paper discusses methodological challenges inherent in the current breed of built environment–health studies; and highlights recent advances in state-of-the-art big data spatial modelling, data linkage, anonymization, and large scale prospective cohorts.  Such advances, on the boundaries between city planning, urban design, public health, epidemiology and big data geo-computation hold the prospect of generating an evidence base for creating, designing, and managing the healthy cities of tomorrow. In big data era, we foresee that large scale prospective built environment–health studies enabled through multidisciplinary global collaborations can provide sufficiently robust evidence for preventive environmental interventions in the form of health-specific planning and design of neighbourhoods and cities as well as create decision supports systems for facilitating healthy cities. We discuss some of current issues and challenges and ways forward.  
<strong><a href='https://hub.hku.hk/bitstream/10722/209764/1/Content.pdf?accept=1' style='color: #00ff00'>ABSTRACT</a></strong><br><br>

The built environment (BE) has emerged as one of the ‘first causes’ of chronic disease, capable of explaining its socio-spatial variation in our cities. There is an increasing need for objective, detailed and precise measurements of attributes of built environment that may influence our lifestyle, behaviour and hence physical and mental health.<br><br>

The <a href='http://biobank.ctsu.ox.ac.uk/crystal/label.cgi?id=100115' style='color: #00ff00'>UK Biobank Urban Morphometric Platform (UKBUMP)</a> is the first ever very large sample size high resolution spatial database of health-specific urban morphological metrics (morphometrics), developed for half-a-million participants of the <a href='http://www.ukbiobank.ac.uk/' style='color: #00ff00'>UK Biobank Prospective study </a>spatially distributed across 22 UK cities. Large scale objective assessment of the BE has been automated employing state-of-the-art spatial and network analyses performed upon multiple national-level spatial datasets including UK Ordnance Survey Mastermap (OSM) and AddressBase Premium databases, UKMap, UK Land Registry, colour infrared imageries, digital terrain models, National Public Transport Access Node, among others. The resultant spatial built environment database led to the modelling and compilation of more than 750 health-specific individual-level urban morphometrics of density, street-level design, greenspace, destination accessibility, building-level attributes all measured in reference to geocoded  dwelling locations of participants of the UK Biobank cohort.<br>

The highly characterized built environment exposure database (UKBUMP) aims to identify specific associations between exposures to built environment attributes such as urban density, greenspace and configuration and specific outcomes of behaviour, health risks, health and wellbeing after adjusting for a range of individual-level covariates including prevalent disease. The baseline UKBUMP acts as a national resource that can be accessed by urban planning and public health research communities, providing a platform for evidence-based healthy city planning and interventions for the first half of the twenty-first century.
Spatial locations of 22 UKB assessment centres with number of participants.Overview of BE morphological metrics of the UK Biobank Urban Morphometric Platform (UKBUMP).Illustration of potential applications of UKBUMP database in BE–health studies.Network modelled street-level movement density within 800m street catchment.Terrain variability around UK Biobank participants’ dwellings modelled as standard deviation in slope (measured from 5m resolution digital terrain model).<br>

Annals of GIS 2015, 21:2, 135-148.<a href='http://www.thelancet.com/journals/lanplh/article/PIIS2542-5196(17)30119-5/fulltext' style='color: #00ff00'>ABSTRACT</a><br><br>
Obesity is a major health issue and an important public health target for urban design. However, the evidence for identifying the optimum residential density in relation to obesity has been far from compelling. We examined the association of obesity with residential density in a large and diverse population sample drawn from the UK Biobank to identify healthy-weight-sustaining density environments.<br><br>
For this full-data, cross-sectional analysis, we used UK Biobank data for 419 562 adult men and women aged 37–73 years from 22 cities across the UK. Residential unit density was objectively assessed within a 1-km street catchment of a participant's residence. Other activity-influencing built environment factors were measured in terms of density of retail, public transport, and street-level movement density, which were modelled from network analyses of through movement of street links within the defined catchment. We regressed adiposity indicators of body-mass index (BMI; kg/m2), waist circumference (cm), whole body fat (kg), and obesity (WHO criteria of BMI ≥30 kg/m2) on residential density (units per km2), adjusting for activity-influencing built environment factors and individual covariates. We also investigated effect modification by age, sex, employment status, and physical activity. We used a series of linear continuous and logistic regression models and non-linear restricted cubic spline models as appropriate.<br><br>
The fitted restricted cubic spline adiposity-residential density dose–response curve identified a turning point at a residential density of 1800 residential units per km2. Below a residential density of 1800 units per km2, an increment of 1000 units per km2 was positively related with adiposity, being associated with 10% increased odds of obesity (odds ratio [OR] 1·10, 1·07 to 1·14). Beyond 1800 units per km2, residential density had a protective effect on adiposity and was associated with 9% decreased odds of obesity (OR 0·91, 0·90 to 0·93). Subgroup analyses identified more pronounced protective effects of residential density among individuals who were younger, female, in employment, and accumulating higher levels of physical activity.<br><br>
Housing-level policy related to the optimisation of healthy density in cities might be a potential upstream-level public health intervention towards the minimisation and offsetting of obesity; however, further research based on accumulated prospective data is necessary for evidencing specific pathways. The evidence points to the value of housing and land-use planning policy related to densification as an upstream-level candidate for public health intervention against adiposity. Suburban densification provides a public health opportunity to be embraced via creation of multi-functional places in our suburbs. <br><br>
Association between adiposity and housing density, allowing for non-linear effects. The continuous line represents the estimated mean adiposity outcome and shaded areas represent 95% CIs. The barcode shows the distribution of the analytic sample across the residential density continuum. Separate models were fitted for BMI, waist circumference, and whole body fat, with restricted cubic splines with Harrell's knots, adjusting for age, sex, education, employment, car ownership, housing tenureship, smoking, processed meat intake, mother's illness, medication use, physical activity, retail, public transport, street-level movement density, and neighbourhood deprivation. Point A indicates the detected turning point of the curve (at which the first derivative or slope changes sign), observed at a density of 1800 units per km2. Point B represents the point in the curve corresponding to the residential density of 3200 units per km2, which is the density of newly developed housing in the UK over the past 2 years. Association between physical activity (log-transformed MET h/week) and residential density, allowing for non-linear effects.<br>
Shaded areas represent 95% CIs. The barcode shows the distribution of the analytic sample across the residential density continuum. Restricted cubic splines with Harrell's knots were fitted, adjusting for age, sex, education, employment, car ownership, housing tenureship, smoking, processed meat intake, mother's illness, medication use, physical activity, retail, public transport, street-level movement density, and neighbourhood deprivation. Point A indicates the detected turning point of the adiposity–residential density curve (figure 1), observed at a density of 1800 units per km2. The odds ratios for doing low physical activity (<7·5 MET h/week) are reported on either side of point A.
Association between BMI and housing density with effects modification by age, sex, employment status, and physical activity.<br>
The Lancet Planetary Health 2017, 1(7), e277-e288.
<a href='http://www.sciencedirect.com/science/article/pii/S0160412017302416?via%3Dihub' style='color: #00ff00'>ABSTRACT</a><br><br>

With the rapid urbanization and prevailing obesity pandemic, the role of residential green exposures in obesity prevention has gained renewed focus. The study investigated the effects of residential green exposures on adiposity using a large and diverse population sample drawn from the UK Biobank.<br><br>
This was a population based cross-sectional study of 333,183 participants aged 38–73 years with individual-level data on residential greenness and built environment exposures. Residential greenness was assessed through 0.50-metre resolution normalized difference vegetation index (NDVI) derived from spectral reflectance measurements in remotely sensed colour infrared data and measured around geocoded participants' dwelling. A series of continuous and binary outcome models examined the associations between residential greenness and markers of adiposity, expressed as body-mass index (BMI) in kg/m2, waist circumference (WC) in cm, whole body fat (WBF) in kg and obesity (BMI ≥ 30 kg/m2) after adjusting for other activity-influencing built environment and individual-level confounders. Sensitivity analyses involved studying effect modification by gender, age, urbanicity and SES as well as examining relationships between residential greenness and active travel behaviour.
Residential greenness was independently and consistently associated with lower adiposity, the association being robust to adjustments. An interquartile increment in NDVI greenness was associated with lower BMI (βBMI = − 0.123 kg/m2, 95% CI: − 0.14, − 0.10 kg/m2) as well as a 3.2% reduced relative risk of obesity (RR = 0.968, 95% CI: 0.96, 0.98). Residential greenness was beneficially related with active travel, being associated with higher odds of using active mode for non-work travel (OR = 1.093, 95% CI: 1.08, 1.11) as well as doing >30 min walking (OR = 1.039, 95% CI: 1.03, 1.05).<br><br>
Residing in greener areas was associated with healthy weight outcomes possibly through a physical activity-related mechanism. Green allocation and design may act as upstream-level public health interventions ameliorating the negative health externalities of obesogenic urban environments. <br><br>
Residential greenness measured as Normalized Difference Vegetation Index from 0.5m resolution colour infrared imagery.Association of residential green with body mass index and obesity; effect modification by gender, age, urbanicity and SES adjusting for all other factors.<br>
Bars show 95% confidence intervals.<br>
Q1, Q2, Q3 and Q4 represent 1st, 2nd, 3rd, 4th quartiles of index of urbanicity; Qn1, Qn2, Qn3, Qn4 and Qn5 represent 1st, 2nd, 3rd, 4th, 5th quintiles of Townsend index (TI) respectively.<br>
Environment International 106 (2017) 1–10.<br>
Association of residential green with body mass index and obesity; effect modification by gender, age, urbanicity and SES adjusting for all other factors.<br>
Bars show 95% confidence intervals.<br>
Q1, Q2, Q3 and Q4 represent 1st, 2nd, 3rd, 4th quartiles of index of urbanicity; Qn1, Qn2, Qn3, Qn4 and Qn5 represent 1st, 2nd, 3rd, 4th, 5th quintiles of Townsend index (TI) respectively.<br>
Environment International 106 (2017) 1–10.<br><strong>BOOK</strong><br><br>
Sarkar, C., Webster, C., and Gallacher, J. (2014) Healthy Cities: Public Health through Urban Planning. Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing.<br><br>
<a href='http://www.e-elgar.com/shop/healthy-cities' style='color: #00ff00'>http://www.e-elgar.com/shop/healthy-cities </a>
UKBUMP RESEARCH IN THE MEDIA<br><br>
Fast Company. 7th Nov. 2017. Living in a dense city makes citizens healthier. <br><br>
<a href='https://www.fastcompany.com/40492045/living-in-a-dense-city-makes-citizens-healthier' style='color: #00ff00'>https://www.fastcompany.com/40492045/living-in-a-dense-city-makes-citizens-healthier</a>
UNIVERSITY OF HONG KONG
FACULTY OF ARCHITECTURE