How healthy is the built environment? Challenges and paradigms for measuring urban health

ABSTRACT

Health evidence and health data are not widely used to shape healthier cities through urban planning. Indeed research has not produced an extensive standardised assessment able to engage the transdisciplinary nature of the relationship between population health and urban settings. This paper aims to introduce and review how the built environment and its relation to population health are assessed in the literature, identifying methodological shortcomings and research potential to be addressed jointly. First, the article outlines the gap between the plural nature of urban health and sectorial assessments as a lack in blending multiple bio-medical fields to guide planning, when on the other hand, extensive urban health studies often lack appropriate measures. It is preferable to investigate the built environment combined with population health investigation outlined in space and time. Second, to integrate health measurements, considerations of the spatio-temporal context to embed health in the built environment study are presented. Indeed, the searches for prospective approaches at adequate smaller scales deepen the residence history and heterogeneity in the spatial extension of individual space behaviour and urban features. The authors intend to promote a transdisciplinary discussion upon a broad diagnosis of the healthiness of the built environment to better inform decision-makers.

KEYWORDS:

• Built environment
• transdisciplinary
• urban health
• spatial polygamy
• residential history
• health inequities

Introduction

An urbanisation process able to tackle the increasing population in urban areas needs to address population health (CSDH 2008, United Nations 2015); therefore preventing, protecting and improving health and wellbeing; with equity and sustainability (Rydin et al. 2012, Wong 2015, Healthy People 2020, WHO 2020). While the features needed to make cities healthier are widely recognised, how to deliver the potential health benefits equally within cities and between the citizens is not well understood (Rydin et al. 2012). Addressing health in relation to urban environments means understanding, measuring, assessing, contextualising population health and the built environment in their variety of impacts, geography and temporal evolution (Entwisle 2007, WHO 2020). Cities are considered in urban health research as a place of the built environment defined by Roof and Oleru as ‘the human-made space where people live, work, and recreate on a day-to-day basis’ (Roof and Oleru 2008). Multiple conceptualisations of the built environment can be found depending on the discipline which studies it (Moffatt and Kohler 2008) since the built environment includes many features besides a geographic setting, not detangled by nature, culture and society (Lang and Rayner 2012). The context of the relationship between the built environment and health encompasses the urban physical space and measures of built and natural environment, transport networks, or the social environment (de Leeuw and Simos 2017).

Framing population health is challenging without including the built environment. For example, the epidemiologic transition postulates that urban development makes chronic diseases supersede infective diseases (Omran 2005). This transition can cease when quick and uncontrolled urban growth is not managed, instead of delivering a triple burden to the population’s health: cities become centres of diffusion of epidemics, places of exposure to unhealthy built environments leading to chronic diseases and place where injuries take place systematically (Rydin et al. 2012, Mberu et al. 2016). On the contrary, the built environment can impact health by preventing exposures to hazards and by promoting healthy behaviour as an outcome of the process of planning, dealing with both communicable and non-communicable diseases (Corburn 2004, Rydin et al. 2012, WHO 2020).

When planning drives interventions ‘to change the physical form or physical management of the city or part of it, with the intention of a positive health outcome; and an increase in health equity is considered a positive health outcome’, it is termed Healthy Urban Planning (HUP) (Grant 2015), alternatively City Health Development Planning (Green et al. 2009) or Healthy Urban Environment and Design (Grant 2015). HUP research aims to measure the impact of urban environments on population health, adapt planning to promote health, and spread the findings to policymakers and stakeholders (Sallis et al. 2016).

Acknowledging that the built environment is required to become healthier (Corburn 2004, Rydin et al. 2012), it is of primary interest to understand how to assess the impact of urban environments on population health (WHO 2020). The extent to which the built environment is shaping population health remains unclear. Planners of New York suggest that planning can regulate the built environment and socio-economic factors in such a way that determines 80% of community health outcomes, as opposed to the 20% shaped by the healthcare system (RPA 2016). In the Healthy Cities project meta-theory by de Leeuw, few core determinants regulate the health determinants: genetics, environment, and lifestyle, which contributes hinges on the contexts (de Leeuw 2009, Vrijheid 2014). The healthcare system is based on the idea that if a core determinant fails, it must intervene. Indeed, health promotion and prevention cannot be restricted to the healthcare system alone, but for example, it can be integrated into planning whose potential contribution is underestimated (Corburn 2004, 2015, WHO 2020). A small share of the total spending is used for prevention, i.e. 12% health worldwide in 2016 (WHO 2018), considering that percentages in high -income countries are below 5% (de Bekker-Grob et al. 2007, WHO 2018, 2019) despite some underestimations of prevention spending (de Bekker-Grob et al. 2007).

Promotion of healthiness and prevention of illness through the management of urban settings is targeted to all inhabitants, with no inequities in delivering health benefits and hazards through urban planning. According to the founding principles of the 2030 Agenda for Sustainable Development and the Healthy People 2030 initiative, equity in health should be attained for everyone with equity through sustainable development of cities and communities (Healthy People 2020, United Nations 2015). In essence, considerable and persistent differences can be observed between countries, within countries and even within cities (Elsey et al. 2016, WHO 2016). To target inequity in health, WHO launched the Urban Health Equity Assessment and Response Tool to help member countries achieve equity in health in all sectors, including urban planning (WHO Centre for Health Development of ‎‎‎Kobe, Japan 2010). After selecting a group of health indicators to compare 57 cities, Stauber et al. concluded that environmental improvements have an essential influence on population health and its disparities (Stauber et al. 2018).

To frame and improve population health in urban contexts and address inequities, a better understanding of how and where the built environment impacts health within cities is needed (WHO Centre for Health Development of ‎‎‎Kobe, Japan 2010, Elsey et al. 2016). This paper aims to introduce and review how the built environment and its relation to population health are assessed in the literature, identifying methodological shortcomings and research potential to be addressed jointly.

Strengths and weaknesses of theoretical and applied studies in the many fields of urban health, such as spatial epidemiology or geographic medicine, were identified when dealing with the assessment of the healthiness of the built environment. Although both non-communicable and communicable diseases are considered for their relationship with the built environment, the health components of the built environment identified from the literature are current in non-communicable diseases. Unfortunately, systematic approaches for reviewing sectoral assessments of the healthiness of the built environment were not possible since the trans-disciplinary nature of the topic includes separate sectors that do not share similar aims, methods or even vocabulary (Gatzweiler, Lawrence and Gatzweiler 2017, WHO 2020). Instead, the few publications dealing with extensive approaches are found to be published in grey literature, thus directly by official institutions in reports, tools, regulations and websites which does not allow a systematic and explicit selection of sources, compared to the research through peer-reviewed literature through scientific databases, i.e. PubMed.

The paper is structured in two main parts. First, the assessment of the healthiness of the built environment is introduced as an informing tool able to drive urban planning. Assessments in the biomedical field have produced a large body of evidence on the multiple impacts of the built environment on health. Despite the multidisciplinary nature of urban health depicted by theoretical studies, extensive approaches to assess the healthiness of the built environment remain scarce. Assessment approaches that collect multiple sectors are depicted to analyse potentials and shortcomings in including health as a primary objective and in capturing synergies and conflicts among urban health issues. Therefore we review the various tools and framework for assessing the built environment to address the importance of broad approaches and equity in urban health.

Second, HUP research frequently mentions the need for longitudinal study and smaller adequate scales of spatial analysis. The embedding of health measurements with the assessment of the built environment is discussed. The investigation of causality and loops that characterise the binding of individual health and exposures and experiences to the built environment are introduced. Then, the spatio-temporal contextualisation of health and built environment data is discussed, as it is fundamental in relating health with urban features and therefore address equity in HUP. Novel concepts of the exposome, residential history and spatial polygamy are reviewed to detangle the spatio-temporal contextualisation in urban health studies.

The authors aim to stimulate discussion on a partial view of the vast topic of urban health: the nature of the models, protocols, and tools able to deliver a coherent, exhaustive and sound understanding of the impact on the health of the built environment. Thereby, the diagnosis of the healthiness of the built environment can inform decision-making on a constant temporal basis and a precise spatial scale, for example, through a standardised protocol of assessment.

Assessment of healthiness of the built environment

Four levels of health integration

Health is mainly investigated and monitored through measures of health outcomes. Health outcomes are not easily collected since they need to be validated by professionals, require expensive tests and treat sensitive information. The diseases can be represented by indirect measurements such as self-reported health data and health determinants such as behaviour. While the latter is usually outlined as an individual characteristic, it also shaped by the built environment, social environment, natural environment and the economy (Barton and Grant 2006). Therefore, the investigation of health can be composed of additional levels to draw a picture of population health about the built environment or the related risk intake (Künn‐Nelen 2016, Frank et al. 2019).

The broad method used to study the impact of commuting on health made by Künn-Nelen is here suggested for general urban health investigation (Künn‐Nelen 2016). Health can be concurrently represented through four levels: the objective health (the measurement, the real health outcome), the subjective health (e.g. the health perception or satisfaction), the health behaviour (e.g. the lifestyle and habits), and healthcare consumption (e.g. utilisation of healthcare services) (Künn‐Nelen 2016).

The large body or urban health literature frequently investigate health using one or two of the four levels. Between levels, health consumption investigation is scarce in the literature and also it is difficult to interpret. For example, a low rate of healthcare service uptake could indicate reduced accessibility (physically, economically or socially) as well as an absence of illness. Health satisfaction is considered an indicator of the quality of life and not frequently discussed in urban health studies. The ‘embodiment’ process of place in health (Krieger 2001) should be based more on mixed methods (Amaratunga et al. 2002): studies that are mainly based on quantitative approaches which collect biometrics, without including perceptions and stories of participants (Brownson et al. 2009, Miller and Tolle 2016, Petteway et al. 2019). For example, Paine et al. audited the living experience of residents and users in four developing residential areas in New South Wales, Australia, reporting the necessity of monitoring, guidance and promotion of the adequate use of built environment besides the physical intervention (Paine et al. 2018). The NHS Scotland created a participatory tool based on a broad framework to understand perception and use of the daily living places, therefore collecting data about perception and stimulating urban health knowledge and population engagement (NHS 2019).

Overall, the analysis of health measurements requires further consideration about the space-time contextualisation of exposure and experience to the built environment, discussed in “Putting health into a place” section. Existing health databases regarding multiple illnesses and health determinants are represented at large spatial scales only, mainly at the country level, which does not allow an adequate comparison with the characteristics of urban settings.

Assessment tools

An assessment of the built environment is a procedure to identify the geographic location of specific characteristics of an urban setting and characterise the built environment for its role in accomplishing a specific task, e.g. in delivering health benefits or hazards. The built environment is often represented inadequately. Generally, it is a context where its composing features are resumed by three outlines: availability (as the existence of a built environment feature); accessibility (as the location or spatial organisation of an object); and the quality (conceived as the characteristics linked with their practical utility or impact). A large body of quantitative HUP studies does not consider the quality of urban features related to their practical use (Gatzweiler et al. 2017) and their effects on health, including their perception (An et al. 2018).

The basis for the health assessment can be evidence from urban health studies, gathering and listing the characteristics of the built environment that are related to health, eventually including investigation of population health. Evidence about inference between built environment features and illness has been created for a large number of urban health issues (Northridge et al. 2003, Barton et al. 2013, Grant 2015), such as air pollution, active mobility, thermic stress, road injuries, and sanitation among the many. However, other urban health issues are challenging to address due to a lack of research, scientific consensus, the large-time lag between exposure and illness, multi-causality or because of individual and local characteristics; e.g. electromagnetic pollution, noise pollution, climate change, social cohesion, urban agriculture or the food environment (‘Yotti’ Kingsley et al. 2009, Pirrera et al. 2010, Caspi et al. 2012, Orsini et al. 2013, Halperin 2014, Morris et al. 2017, Bandara and Carpenter 2018, Jennings and Bamkole 2019). While the impact of the built environment on health is commonly accepted, it is not supported by evidence globally (Grant 2015). The only extensive assessment of the effects of planning interventions took place in the field of infectious diseases and sanitation for short time-lags only (Grant 2015), the main reason being that for infectious diseases, environmental determinants of diseases are in small numbers and time lags between exposure and illness are small (Rothman et al. 2008). These characteristics allow easier identification of a ‘dose-effect’, and therefore governance endorses effective direct interventions investing more in preventive measures (WHO 2018, 2019). Determinants of health that are hardly traceable to their cause in space or time scale, or causality, otherwise called distal according to (Morris et al. 2017), are likely not to resolve. Instead, determinants of health in non-communicable diseases are many and heterogeneous within the population. Furthermore, multi-causality in population health does not allow easy identification of the causal pathway from illness to urban environments (Tesh 1994, Grant 2015).

The collection of urban health issues deals with a large number of diseases, a large number of characteristics of cities and multiple sectors of urban planning (Elsey et al. 2016, Lawrence and Gatzweiler 2017, WHO 2020). Urban planning is the tool considered in this article to address urban health by managing the built environment. Urban planning is a transdisciplinary process that spatially develops and design the land use of the built environment (Levy 2016). This process implements optimal solutions to solve conflicts among multiple sectors and harness synergies among them (Levy 2016). Similarly, HUP encompasses multiple sectors of planning, such as transportation, housing and environmental protection, to promote health promotion and prevention (WHO 2020).

In contrast to the multidisciplinary nature of urban health, monodisciplinary approaches adopt biomedical models based upon symptom and treatment, rather than integrating multidisciplinary interpretations able to confront the impact levels of different built environment features on health (Diez Roux 2007, Reis et al. 2015). Disciplinary approaches to assessing single determinants of health miss the synergic and conflictual effects of multiple features of urban environments on population health and its spatial coherence. Besides, health cannot be disconnected by other dimensions, such as the natural and social environment, in which cities are deeply tangled within the same ecosystem (CSDH 2008). As the actual state of research, assessments of the healthiness of the built environment are widely performed on multiple urban health issues separately, and they are not geographically contextualised (Northridge et al. 2003, Barton and Grant 2013, Barton et al. 2013, Yang et al. 2013, Gomez et al. 2015, Schulz et al. 2018, Wierzbicka et al. 2018).

For example, Health impact assessment (HIA) may be considered a broad assessment for its potential application on a wide range of policies, programs and projects (Harris-Roxas and Harris 2011). HIA is defined as ‘a combination of procedures, methods and tools by which a policy, program or project may be judged as to its potential effect on the health of a population, and the distribution of those effects within the population’ (European Center for Healthy Policy 1999). HIA serves as a fundamental approach for health protection and promotion to support planning as an active agent and an integrated part of planning in all its phases through six steps: screening, scoping, assessing, recommending, reporting, monitoring and evaluating (Fehr et al. 2014, WHO 2020).

Despite being initially conceived with methodological shortcomings (Krieger et al. 2003, Winkler et al. 2013), HIA has been applied to a broader range of sectors with success (Rogerson et al. 2020). For example, in the Healthy Cities network, HIA initially introduced in Phase IV followed a wide range of application in Phase V (de Leeuw and Simos 2017). Reviewing HIA and equity, Buse et al. reported a limited but increasing number of HIAs addressing inequalities (Buse et al. 2019). In particular, 53 of the 89 items reviewed included multiple risks, hazards and impacts for cases of healthy cities and urban planning issues, cumulative health risk assessment, corporate health impact assessment, climate change-focused HIA, intersectionality-based assessment, and global health HIA (Buse et al. 2019). On the one hand, HIA has been voluntarily implemented to climb over barriers related to decision-making acceptance, from the other hand, HIA is not usually seen as a pro-active, preventive and supportive actor in urban planning: health promotion and prevention is added as a condition and not regularly as a final objective (Carmichael et al. 2012). HIA is a pragmatic and powerful approach to integrate HiAP systematically (Harris-Roxas and Harris 2011, de Leeuw and Simos 2017), yet HIA faces multiple bureaucratic, methodological and political barriers needing support from external agents, such as the Healthy City network or NGOs (Carmichael et al. 2012, Winkler et al. 2013). HUP could profit from HIA to integrate health as a primary goal and address health equity, prevention and promotion through a wide range of health issues that can stratify disproportionally within the city area (Buse et al. 2019, Rogerson et al. 2020).

Beyond HIA, other examples of urban health research display various advantages and shortcomings in building knowledge to inform planners and decision-makers that protocols of assessment such as HIA could undertake. For example, the protocol created by the Canadian Urban Environment Health Research Consortium sets a broad assessment characterised by spatial contextualisation (Brook et al. 2018). Environmental measures are merged to provide a protocol for exposure analysis of air pollution, transportation, noise pollution, greenness, climate and neighbourhood factor. Postal codes aggregate metrics to match the geospatial information of existing health datasets. Despite the large size of the dataset involved, health outcomes are investigated as a product of the built environment within the area covered by postal codes containing residence. However, when measurements are based on postal codes of rural and urban areas, the size of a spatial unit of analysis is highly variable, and spatial unit of analysis may be not small enough to represent the relation of the built environment with population health (Gomez et al. 2015, Elsey et al. 2016, Verma et al. 2017).

A final example is the Built Environment Tool created by the Center for Disease Control and Prevention to study obesity and its determinants only (Eyler et al. 2015). The Built Environment Tool is made to assess the spatial characteristics of the environment able to promote active mobility, recreational activities and healthy nutrition following measurements of selected features of the built environment. The Built Environment Tool is restricted to the study of obesity but remains multi-sectorial and based on measures of the urban environment, such as street pattern characterisation of crossings and sidewalks pathway (The built environment assessment tool manual 2019).

Many other studies are listed in the geospatial assessment for urban health in the fields of heat exposure, cardiovascular diseases, air quality, food environment, health care accessibility, active mobility, among the many (Lu and Delmelle 2019). WHO itself offers multiple open-source analytical tools to assess urban environments, such as GreenUr for urban green spaces, AirQ+ for air pollution, HEAT for walking and cycling and also the Health Impact Project’s cross-sector toolkit, which collects HIA experiences and guides (WHO 2020). Also, indicators are overused in defining healthy urban forms (Pineo et al. 2018, Rothenberg et al. 2015, Stauber et al. 2018) and are rarely spatially explicit within the city area. Representation of composite indicators to represent the healthiness of the built environment broadly has been criticised, because it summarises complex information or because the adequate weighting of the composing indicators has not been addressed yet (Pineo et al. 2018). While overall composite scores are not clear, the joint representation of multiple indicators or multiple tools could have a meaningful and valuable utilisation when spatially explicit. When proposing a global urban environment and health index representing 10 categories and 58 indicators generated by a causal pathway framework through iterative stakeholder engagement, Pineo et al. (2018) called for neighbourhood-scale data gathering to enable smaller-scale analysis.

As opposed to sector-specific assessments based on measurement protocols seen before, extensive approaches are primarily theoretical. While the former is not able to capture synergies and conflicts between urban health issues, the latter is not prone to environmental and individual health measures. Among extensive frameworks that outline urban health, Galea et al. proposed four generic urban health characteristics: population characteristics, physical environment, the social environment and health and social service system (Galea et al. 2005). Grant (2015) identified eight important issues for HUP: climate change and public health emergencies, exposure to noise and pollution, healthy urban planning, healthy transport, healthy urban design, housing and regeneration, safety and security, creativity and liveability. Otherwise, the ‘Social determinants of health and environmental health promotion’ scheme from Northridge et al. display the multiple causal pathways from environmental features at different scales to health outcomes through intermediate steps (Northridge et al. 2003). The previous conceptualisations of the relationship between health and the built environment, like many others, do not set a basis for spatial analysis and measurements. However, conceptual models, i.e. eDPSEEA, are already used to engage stakeholders in facing a broader conceptualisation of urban health (Reis et al. 2015). Still, the previous theoretical examples are instead meant to embrace multiple urban health issues, interdisciplinarity, and causal path and feedback loops to inform and guide the planning process. In summary, future approaches could benefit converging measurements of the built environment and transdisciplinarity, thus tackling multiple health determinants based on what has been diagnosed in the city to address health prevention and promotion (Northridge et al. 2003, Barton et al. 2013, Urban design for health 2022, WHO 2020).

Putting health into a place

Currently, the assessment of the healthiness of the built environment can be based on existing evidence in similar contexts, thus assuming that similar urban environments lead to similar health outcomes. However, this can introduce a bias that can lead to a similar response and prevent new hypotheses and solutions. HUP research should investigate the built environment combined with local population health inquiry contextualised in space and time. The latter is also tangled with the four levels of health (Four levels of health integration), which completes the pathway from place to illness or healthiness. The following paragraph discusses the challenges in capturing the relationship between health and places of exposure and experience of the built environment.

Challenges to detect causality

The study of health and built environment relations means researching pathways of illness or health promotion. First, urban health research suggests that causality may fade away due to feedback loops or social selectivity (Eid et al. 2008, Duncan et al. 2014, Jokela 2014). An example of a feedback loop can be identified in the study of obesity. In essence, car use leads to air pollution, which discourages physical activity and then increases the risk of being obese, which boosts car use (An et al. 2018). Generally, loops are either negative or positive, depending on the need for resources to be maintained. Negative feedback loops can be subsidised to maintain stability, e.g. public transport has good viability and low fares thanks to public funds, while positive loops are often self-nourished to unstable growth, as periphery expansion led by the construction of new motorways (CSDH 2008).

Uncertainties carried by feedback loops do not easily allow the identification of frequencies of response in population health (i.e. a ‘dose-response’ measure) which is a great challenge for HUP (Saarloos et al. 2009). The response to interventions in urban settings is not linear but exhibits adaptation and resistance to interventions at larger scales than the individual one instead (Saarloos et al. 2009, Tozan and Ompad 2015). Likewise, urban systems can cope with breakdowns and dissipate overloads (Rydin et al. 2012). In a socio-ecological system, a higher complexity is brought by human agency, namely ‘the ability of individuals and groups to act knowingly based on what they value in order to maintain or improve their wellbeing’ (Lawrence and Gatzweiler 2017). Individual behaviour is self-organising, non-linear, path-dependent, dynamic and influenced at different hierarchical levels by interrelated factors (Saarloos et al. 2009). In addition, individual behaviour is incorporated into the multi-causal web that is formed between health outcomes and urban environments (Tesh 1994, Thomas 2006).

However, following the concept that individuals are also active agents in shaping health, HUP research investigates impacts on population health, missing the opportunity to understand whether individuals likelihood to select unhealthy urban environments or choose unhealthy behaviours (Eid et al. 2008, Gomez et al. 2015). In theory, on the one hand, HUP should not let populations live in unhealthy urban environments (Michael 2006); on the other HUP could better understand whether unhealthy life human agency copes with intervention for health promotion or protection to address any urban health issues at the local area and prevent adverse health outcomes (Gatzweiler, Lawrence and Gatzweiler 2017).

Spatio-temporal contextualisation

Besides the complexity in defining the concept of causality in urban health research, in practice, studying impacts on overall health involves both exposures and experiences of hazards or health benefits, which have multiple spatial and temporal outreach. A broad assessment and adequate geographic contextualisation are required to identify inequities within the city and synergies and conflicts between interventions in space (WHO Centre for Health Development of ‎‎‎Kobe, Japan 2010).

First, time has not been widely accounted for in the relationship between the built environment and population health (Saarloos et al. 2009, Spring et al. 2012, Brook et al. 2018). In order to draw causal inferences with built environment characteristics, prospective studies are needed to capture the time lag between cause and effect (Michael 2006, Næss 2016), such as in the study of air pollution (Bind 2019). Only a small proportion of studies were prospective or focused on follow-up data (Brook et al. 2018, Northridge et al. 2003, Schulz et al. 2018, Verma et al. 2017, Yen and Syme 1999). A large body of urban health research is based on cross-sectional studies, while longitudinal studies are still rare (Elsey et al. 2016, Petteway et al. 2019). Population health investigation could be based on the concept of the exposome to study the complexity of urban health. Wild (2005) defines exposome as the measure of the totality of exposures in a lifetime span. The study on total exposome to the built environment could be based on a contextual collection of exposures and experiences across time and space (Matthews and Yang 2013, Vrijheid 2014). Following an approach based on the concept of exposome also means dealing with multiple sectors simultaneously, i.e. adopting an extensive approach (Vrijheid 2014). Nevertheless, to demonstrate that health is a product of multiple exposures, multiple measures that did not exist or were not collected in the past would be needed. Instead, ongoing exposome studies can gather multiple measures during the expected time lag between the beginning of exposure and the illness diagnosis (Louis and Sundaram 2012, Vineis 2019).

Furthermore, research cannot efficiently address exposome because the time lag between intervention and health outcomes can be more significant than the changes in the built environment, such as the city plan itself (Grant 2015). In contrast with different time lags, separate intervention and health outcomes, cities applying a HUP intervention document a systematic change in human response, such as the involvement of key stakeholders and a better understanding of the social environment (Grant 2015).

Assessments of urban health require spatial contextualisation, which couples individuals and space (Elsey et al. 2016, Kanaroglou and Delmelle 2016, Lu and Delmelle 2019). However, health measurements are frequently spatially un-explicit during the planning process as well as during planning interventions (WHO 2010, Rydin et al. 2012, Matthews and Yang 2013, Grant 2015). The lack of spatial data on the built environment and health mainly affects the urban most deprived area as it leads to a strong bias in the analysis of health inequities within cities and comparative studies between rural and urban health (WHO Centre for Health Development of ‎‎‎Kobe, Japan 2010, Verma et al. 2017). Afterwards, spatial contextualisation requires adequate scales to identify differences within cities, and finally, optimise health benefits delivery through targeted interventions (Verma et al. 2017). Indeed HUP needs to be spatial explicit at smaller scales to tackle urban health issues with equity (Elsey et al. 2016, Verma et al. 2017). Rothenberg et al. reviewing urban health indices, highlight the lack and the need for local data on small areas (Rothenberg et al. 2015), although, in Europe, HUP has seen an increase in small-scale approaches in the last decades (Grant 2015). Geospatial individual data can provide the best level to investigate health and built environment. Only with individual-level variables it is possible to sort out the effects of behaviour and circumstances. Collecting individual geospatial data is not only highly time-consuming and costly but may show lower participatory rates when it is seen as an invasion of privacy (Cottrill and “Vonu” Thakuriah 2015).

Furthermore, the spatial dimension can be coupled with spatial statistics to identify significant spatial patterns of health outcomes (Bailey 2001, Guessous et al. 2014, Kanaroglou and Delmelle 2016, Joost et al. 2018). Spatial clustering of health determinants could be used to identify the areal unit of exposure or experience to the built environment (Chaix et al. 2009). However, when the casual pathway is misunderstood, and covariates are not taken into account, the spatial dimension of health datasets allows a simplistic association with built environment features rather than causality (Kanaroglou and Delmelle 2016). Besides a better understanding, spatial information also allows the direct visualisation of health outcomes and places of urban environments to boost the translation of research into practice (Andrienko et al. 2007).

The contextualisation of the health of individuals in space and time is then affected by two constraints at different temporal scale: the study of individual residential history and the spatial unit to consider daily exposure to and experience of diverse places.

Residential history

Assuming that health is a product of the built environment where people live, the analysis should include the amount of time spent in each place. This consideration is valid when the assessment encompasses non-communicable diseases, which may occur many years after the exposure or being a result of exposure over a long time. In terms of lifespan time and spatial dimension, relocation introduces a significant variation. Spatial epidemiological studies on non-communicable diseases commonly filter participants choosing individuals with fixed residence for several years. Altogether studying chronic diseases and exposome demands longitudinal approaches. Accordingly, places of exposure (or experience) are multiple depending on the residence history (Ben-Shlomo and Kuh 2002, Matthews and Yang 2013). Therefore, individual health could be stratification related to different places. Relocation has been only recently considered as a change of approach for health spatial contextualisation (e.g. to study active mobility and health determinants after relocation) (Beenackers et al. 2012, Giles-Corti et al. 2013, Hirsch et al. 2014, Foster et al. 2015, Braun et al. 2016). Wu et al. on blood pressure and air pollution compared built environments after and before relocation, using both movers and permanent individuals (the latter as control population) (Wu et al. 2013). Generally, relocation habits have been studied jointly with health, such as stressors (e.g. during childhood (Jelleyman and Spencer 2008), in aged adults (Sanders et al. 2004)) or as positive indicators of self-reported health in adult movers (Lin et al. 2012). In order to assess the healthiness of the built environment, a residence history study could allow the comprehension of the stratification of exposure and experience of different settings.

Furthermore, the relocation study may answer whether causality between illness and urban forms or self-selection of unhealthy environments is observed. For example, after investigating the relationship between sprawl and obesity, Eid et al. concluded that individuals who are more likely to be obese select residences in a sprawling neighbourhood showing some influence by the urban environment (Eid et al. 2008). The relevant areas of exposure and experience of urban features depend on the geographic location of residence and frequent primary destination, such as workplace and schools (Chaix et al. 2009, Kramer and Raskind 2017). The neighbourhood of individuals can be ‘fluid’, varying with time, space and other driving factors (Giles-Corti et al. 2013), including individual determinants, such as age, proximity, utility, and destination density.

Spatial units of analysis and spatial polygamy

Changes of exposure and experience of the built environment are represented by residential history and their spatial extent. In order to ‘put people into place’ (Entwisle 2007), and therefore to relate built environment and population health, the spatial extent of exposures and experiences within cities has to be investigated.

Kwan defined the Uncertain Geographic Context Problem (UGCoP) (Kwan 2012) as ‘the spatial uncertainty in the actual areas that exert contextual influences on the individuals being studied and the temporal uncertainty in the timing and duration in which individuals experienced these contextual influences’. The concept initially applied in social sciences has been used to study urban health about its spatio-temporal contextualisation of individual health measures and urban environmental characterisation (Lu and Delmelle 2019). Indeed, research in healthy urban planning needs to ‘address the individual space-time behaviour that underlies the interaction between the built environment and health’, addressing the question of ‘where does the built environment impact health of one individual?’ (Saarloos et al. 2009). Therefore geospatial information about the built environment as well as individuals is needed at smaller scales.

First, the assessment of the healthiness of the built environment can be performed on selected spatial units (territorial units or neighbourhood) or individual-based units (ego-centred neighbourhood) (Chaix et al. 2009, Gomez et al. 2015). Within each spatial unit, characteristics of urban forms can be represented with their physical extension (points, lines, polygons or 3D objects) or for a spatial unit of representation. Studies on place and health are widely based for convenience on aggregated data on geopolitical units, such as administrative boundaries, postal codes or census districts (Yang et al. 2013, Gomez et al. 2015, Rothenberg et al. 2015, Elsey et al. 2016, Kanaroglou and Delmelle 2016, Schulz et al. 2018). Indeed, neighbourhoods have been used as the activity space and spatial unit, despite the inconsistency in the perception of its spatial extension (Chaix et al. 2009, Matthews and Yang 2013, Saarloos et al. 2009). Meanwhile, we know that ‘neighbourhoods do not exist in isolation’ (Martin and Michael 2004). Attributes describing urban environments do not shift immediately and discreetly but display different gradients of variation between neighbourhoods (Chaix et al. 2009). Otherwise, the spatial unit is defined by the perceived neighbourhood by residents (Diez Roux 2007, Coulton et al. 2013) rather than objectively experienced neighbourhood (Chaix et al. 2009).

Furthermore, the geographic location of residence or other frequent destinations may allow assessing the built environment using multiple distance buffers to multi-modal transport buffers. Attributes of urban environments may show different spatial extension and vary within a given population (Bhat and Guo 2007, Schulz et al. 2018). For example, public green areas may affect potential physical activity when accessible, therefore around its location following a decay proportional to travel time (Brownson et al. 2009, Zhang et al. 2011). Instead, night-time noise pollution can be modelled by expected sources and built environment design within a buffer around the residence (Khan et al. 2018). On the contrary, indoor characteristics have no spatial extension out of the residence location (Badland et al. 2017). Accordingly, the spatial extension in shape and size depends on the element of interest, and it cannot be easily approximated (Kramer and Raskind 2017).

The concept of ‘spatial polygamy’ encloses this variability in individual mobility and the spatial extension of urban features that influence health differently. Matthews and Yang defined spatial polygamy as ‘the simultaneous belonging or exposure to multiple nested and non-nested, social and geographic, real, virtual and fictional, and past and present contexts’ (Matthews and Yang 2013). Knowledge of individual space-time behaviour is fundamental to fill the gap between experience/exposure to built environment and health outcomes (Entwisle 2007), which could be studied directly using location-aware technologies (Matthews and Yang 2013, Miller and Tolle 2016), such as geo-trackers, smartphone application or activity diaries (Saarloos et al. 2009, Roswall et al. 2017, Oh et al. 2018, Zhang et al. 2018); or indirectly by models like activity-based modelling (Saarloos et al. 2009, Wang et al. 2018).

Conclusions and perspectives

The need for tackling non-communicable diseases and inequalities in health has recently triggered the demand for healthy planning rather than curative approaches within healthcare. Health prevention and promotion could be attained through urban planning, reducing exposures to hazards, stimulating healthy behaviours, and building a vibrant neighbourhood for everyone.

However, a better understanding of how and where urban forms impact health is essential to guide health promotion. Assessing and contextualising the urban form able to shape population health could inform the planning process and policies able to guide the multiple agents that drive urban development. The object of assessment is defined as the built environment, which is widely accepted as a determinant of population health and the product of planning. While it is widely accepted that the built environment has both positive and negative impacts, the health evidence and data are not universally used in urban planning yet. More interventions should arise from the postulated conclusion suggested by the analysis of the built environment. Therefore, research should produce valid and reliable methods to monitor and assess the healthiness of the built environment to support urban planning.

Initially, we outlined the concept of assessing the built environment, the evidence about its impacts on health and identified different tools used to assess the relationship between health and the built environment.

Firstly, in contrast with the transdisciplinary nature of urban health, the actual assessments of the healthiness of the built environment are not broad, transdisciplinary, and able to tackle multiple health determinants jointly. In general, regarding the analysis of urban forms, Clifton et al. showed how disciplines (and scales) draw different and conflictual answers in urban planning (Clifton et al. 2008). Hence synergies and conflicts between coexisting urban health issues are missed. Such a task cannot be assigned to the classic categories of urban form analysis but rather a mix of them. Additionally, the assessment procedures are triggered subsequently by different planning goals or following the detection of adverse health outcomes and interventions are often designed only to be neutral for population health. Planning for health should go beyond the concept of ‘geographic neutrality’, which intends to regulate activities to neutralise hazards with little or no regard to geographic clustering and cumulative effects on population health of multiple hazards (Corburn 2004). Instead, healthy urban planning can conceive health as a continuum, where built environment lead to pathogenesis or salutogenesis, rather than a dichotomy between illness and absence of illness (Gatzweiler et al. 2017).

Even though assessments can be based on evidence of the impacts of the built environment on health, assessment is preferable to be supported by population health investigation, which is expected to be characterised by an adequate spatio-temporal contextualisation to explore its relation with urban environments. healthy urban planning research is looking for more participatory, mixed methods to achieve integrated measurements of perception and satisfaction and therefore shift to transdisciplinary knowledge on urban health. Consequently, urban health investigation is required to be spatially explicit within cities in its health and built environment data, allowing the identification of local priorities, and ergo, urban health issues are also addressed equitably through planning.

After discussing the contents of the assessment of healthiness of the built environment, this work puts light on how research in healthy urban planning has only begun to understand exposure to the built environment and experience of the built environment, discussing the spatio-temporal contextualisation to overcome methodological shortcomings. The search for causality in urban health itself is affected by the implication of analysing a complex system such as cities. From one side, the built environment features and health data should be spatial explicit and then represented through multiple scales with a better precision putting less emphasis on aggregated data (Corburn 2004, Elsey et al. 2016). On the other side, population health investigation looks for a more longitudinal study and on the larger temporal span toward studying an individual’s exposome.

Overall, the article addressed the limitations of urban health research to assess the built environment to inform urban planning jointly, despite the variety of quantitative analysis of urban forms at hand for healthy urban planning (Clifton et al. 2008, Rydin et al. 2012, WHO 2020).

Among the challenges to face to assess the built environment and its relationship with health, data collection is considered the main one; since lifespan, geospatial, individual, and health measures are highly costly and time-consuming. Improvement in data collection should make health and environmental data widely available, enabling more extensive, integrated longitudinal studies, e.g. exposome studies (Vrijheid 2014, Maitre et al. 2018), as well as employing social networks, social networks, smartphones application or biometric sensors; with particular attention to the protection of privacy (Gostin et al. 2018, Ghermandi and Sinclair 2019). For example, the Place Standard tool of NHS Scotland combines population participation and engagement with a broad framework to capture the perception and experience of the built environment (NHS 2019). Besides, open-source datasets and sharing platforms could facilitate interdisciplinary approaches, notably through a partnership between healthcare and non-health settings (Verma et al. 2017). Another barrier is the interpretation of complex datasets. While datasets are increasingly available and extensive, urban health research should improve data interpretation through transdisciplinary knowledge, such as integrating ecological approaches able to couple individual space behaviour with healthy behaviours and the built environment to target population health (Matthews and Yang 2013).

Although complex system dynamic modelling techniques are already applied to urban systems, the challenge of assessing the healthiness of the built environment at a local scale will require a tailored application of these models and deeper insights. In the end, local characteristics of urban health may be a constraint to the translation of evidence for standardised assessments, therefore requiring screening by health measurements and transdisciplinary approaches of assessment (Rydin et al. 2012).

Since conceptual models (Reis et al. 2015), international initiatives (Barton and Grant 2013, Sallis et al. 2016) and institutions (Healthy People 2020, WHO 2020) already engage planners and stakeholders to tackle broadly urban health, they could benefit from the variety of quantitative and qualitative analysis of urban forms at hand for healthy urban planning. Still, standardisation and convergence of measurement protocols are needed (Clifton et al. 2008, Rydin et al. 2012). For example, environmental risk assessment often adopts weight-of-evidence (Linkov et al. 2009), but further research is needed to address the measures and assessments of amalgamation in healthy urban planning research.

The challenge of assessing the healthiness of urban environments can be achieved. Studies in spatial epidemiology and geographic medicine, behavioural sciences, urban design, and spatial analysis, et cetera, should gather to set a protocol of transdisciplinary analysis and consequently collect, share and interpret ‘big’ environmental and health data. The outcome will be assessments able to inform and guide further healthier, equitable policy within cities. A new protocol of assessment would not replace Health Impact Assessment but rather support it, therefore diagnosing regularly the healthiness of the built environment in all its forms and on the whole area of the city. The creation of a protocol of urban health assessment is only one of the conditions together with economic means and political orientations to guarantee the integration of health as a primary goal in policies that govern urban settings.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Andrea Salmi

The Communauté d’Études pour l’Aménagement du Territoire (CEAT) is a group of researchers whose works address questions related to the built environment from a multidisciplinary perspective, carrying therefore advanced research on spatial planning. Moreover, the Community’s objectives include teaching (Bachelor, Master and PhD students) and providing high level expertise services for public authorities. By focusing on the creation of synergies, CEAT favors collective, networked actions, aiming to promote the multiple connections between people, institutions, disciplines and the social space (“territories”), at different levels: local, regional, transnational and global.