Invited Perspective: Studying Walkability and Cancer Incidence—A Step in the Right Direction

Walkability is a mixture of features of the built environment that provide opportunities to improve population health in myriad ways. First defined in the 1960s based on observations of sidewalks and safety in U.S. downtown areas,1 walkability, or features that increase our propensity to walk, has become a cornerstone of urban planning. At the population level, living in more walkable neighborhoods is associated with higher levels of physical activity and lower adiposity.2 Jeff Speck, a celebrated city planner who has worked with hundreds of American city mayors to improve the pedestrian experience, developed the General Theory ofWalkability,3 which suggests that walks should be useful, safe, comfortable, and interesting. Tobin et al.4 quantified these attributes using geospatial data on land use mix, residential and business density, public transit, street connectivity, traffic calming measures, street trees, population density, and more. Walkable neighborhoods also tend to support active transportation and recreational physical activities, as well as social interactions. With the strengthening of evidence linking walkability with healthier body weight, epidemiologists have begun asking whether walkability reduces the risk of obesity-related chronic diseases (e.g., cancer) andmortality.5,6 In this issue of Environmental Health Perspectives, IndiaAldana et al.7 report findings that support the link between walkability and lower risks of obesity-related cancers. Their analysis used data from the New York University Women’s Health Study (NYUWHS), a richly contextualized prospective cohort of women who were regularly followed over approximately three decades. The authors calculated annual walkability scores using censustract population density and destination accessibility; they found that higher walkability was inversely associated with total obesityrelated cancer after adjustment for covariates. Similarly, higher walkability was associated with a lower risk of postmenopausal breast cancer, endometrial cancer, multiple myeloma, and ovarian cancer and was suggestively associated with lower incidence rates for all other sites of obesity-related cancer except thyroid cancer. The inverse association with postmenopausal breast cancer was stronger among participants who were hormone receptor–positive rather than negative, further supporting the hypothesis that environmental factors affect risk of breast cancer, especially for those breast cancers influenced by physical activity.8 Notably, this study was the first to assess the risk of incident cancers in relation to a timeweighted averagewalkability index based on regularly updated residential histories, over nearly 30 years of follow-up. The variety in walkability exposures used in epidemiological studies reflects the complexity and lack of consensus on a conceptual definition.4 Much of the operationalization ofwalkability exposure draws on the 3 D’s—density (of population), diversity (of land use), and design (of street networks)—with varying levels of integration and diversity in data sources.9–11 An ongoing trend in walkability exposure assessment is expansion of what the term encompasses. Notably, Ewing and Cervero12 proposed expanding the 3 D’s to 5 D’s to include destination accessibility and distance to transit. Researchers acknowledge that walkability represents not only physical access but also a mixture of social and esthetic aspects that encompass the experience of traversing an environment.13,14 Some measures have been added into studies to account for these complexities, such as the incorporation of vegetation data to assess “green walkability” and integration of perceived walkability measures that additionally draw on subjective assessments of traffic and crime safety levels.15,16 An ongoing challenge of walkability exposure assessment is the lack of data that enable parameterization of each component of walkability, with the spatial and temporal range to support prospective study designs. In the context of cancer epidemiology, assessing walkability exposure over etiologically relevant timewindows (e.g., decades) requires historical data sets. In the new work by India-Aldana et al.,7 the walkability score captures two components—the density of people and of businesses—thus excluding other components known to predict walking, such as density of intersections and public transit, which were included in a four-item walkability score that was validated within the NYUWHS by the authors17 and used in their analysis of walkability and mortality.5 The time-variant two-item score was correlated with the time-invariant four-item score assessed at study baseline (1990), but the utility of the findings for urban practitioners may be limited. In addition, risk estimates could be biased by residual confounding of the association between walkability and cancer incidence by unmeasured environmental and socioeconomic confounders not captured by the individual-level covariates or time-invariant neighborhood poverty confounder assessed only at study baseline. In comparison with a California study on obesogenic environments that assessed multiple environmental components in relation to postmenopausal breast cancer incidence,18 this study lacked detailed quantification of neighborhood socioeconomic status, racial residential segregation, ethnic enclaves, distance to hospitals, green space, and street connectivity. To avoid overly simplistic translation into policy, it is crucial for walkability to be assessed holistically and rooted in context, which can be achieved by adequately examining socioeconomic and racial/ethnic neighborhood composition, environmental coexposures, and safety of walking spaces. Incorporation of context is question-specific (e.g., as modifiers or confounders), but context is indispensable for tackling not only overall cancer risk but also cancer disparities.19 Address correspondence to Charlotte Roscoe, Harvard T.H. Chan School of Public Health, Department of Environmental Health, Harvard University, 181 Longwood Ave., Floor 3, Channing Division of Network Medicine, Boston, MA 02115 USA. Email: croscoe@hsph.harvard.edu The authors declare they have nothing to disclose. Received 24 July 2023; Revised 28 August 2023; Accepted 6 September 2023; Published 4 October 2023. Note to readers with disabilities: EHP strives to ensure that all journal content is accessible to all readers. However, some figures and Supplemental Material published in EHP articles may not conform to 508 standards due to the complexity of the information being presented. If you need assistance accessing journal content, please contact ehpsubmissions@niehs.nih.gov. Our staff will work with you to assess and meet your accessibility needs within 3 working days.

cancers influenced by physical activity. 8Notably, this study was the first to assess the risk of incident cancers in relation to a timeweighted average walkability index based on regularly updated residential histories, over nearly 30 years of follow-up.
The variety in walkability exposures used in epidemiological studies reflects the complexity and lack of consensus on a conceptual definition. 40][11] An ongoing trend in walkability exposure assessment is expansion of what the term encompasses.Notably, Ewing and Cervero 12 proposed expanding the 3 D's to 5 D's to include destination accessibility and distance to transit.Researchers acknowledge that walkability represents not only physical access but also a mixture of social and esthetic aspects that encompass the experience of traversing an environment. 13,14ome measures have been added into studies to account for these complexities, such as the incorporation of vegetation data to assess "green walkability" and integration of perceived walkability measures that additionally draw on subjective assessments of traffic and crime safety levels. 15,16An ongoing challenge of walkability exposure assessment is the lack of data that enable parameterization of each component of walkability, with the spatial and temporal range to support prospective study designs.In the context of cancer epidemiology, assessing walkability exposure over etiologically relevant time windows (e.g., decades) requires historical data sets.
In the new work by India-Aldana et al., 7 the walkability score captures two components-the density of people and of businesses-thus excluding other components known to predict walking, such as density of intersections and public transit, which were included in a four-item walkability score that was validated within the NYUWHS by the authors 17 and used in their analysis of walkability and mortality. 5The time-variant two-item score was correlated with the time-invariant four-item score assessed at study baseline ( 1990), but the utility of the findings for urban practitioners may be limited.In addition, risk estimates could be biased by residual confounding of the association between walkability and cancer incidence by unmeasured environmental and socioeconomic confounders not captured by the individual-level covariates or time-invariant neighborhood poverty confounder assessed only at study baseline.In comparison with a California study on obesogenic environments that assessed multiple environmental components in relation to postmenopausal breast cancer incidence, 18 this study lacked detailed quantification of neighborhood socioeconomic status, racial residential segregation, ethnic enclaves, distance to hospitals, green space, and street connectivity.To avoid overly simplistic translation into policy, it is crucial for walkability to be assessed holistically and rooted in context, which can be achieved by adequately examining socioeconomic and racial/ethnic neighborhood composition, environmental coexposures, and safety of walking spaces.Incorporation of context is question-specific (e.g., as modifiers or confounders), but context is indispensable for tackling not only overall cancer risk but also cancer disparities. 19dress correspondence to Charlotte Note to readers with disabilities: EHP strives to ensure that all journal content is accessible to all readers.However, some figures and Supplemental Material published in EHP articles may not conform to 508 standards due to the complexity of the information being presented.If you need assistance accessing journal content, please contact ehpsubmissions@niehs.nih.gov.Our staff will work with you to assess and meet your accessibility needs within 3 working days.

Invited Perspective
Exposure assessment is a critical issue in epidemiology.Whereas evidence synthesis requires consensus around how an exposure is defined and consistency in the use of that definition, understanding mechanisms typically requires assessment of the joint and individual effects of well-defined exposure components.In a systematic review of 40 health impact assessments of walkability, Westenhöfer et al. 20 noted that the evidence was predominantly on active transport and public transport, which highlighted the lack of quantitative data on the relationship between other components of walkability (e.g., density, connectivity, land use, traffic safety, experience, community, greenspace) and increased physical activity, social interaction, and perceived safety and stress.Establishing clear connections between the components of walkability used in epidemiological studies and the implementation tools used by local governments is invaluable.For example, higher-density development, mixed land use, and transportation alternatives can be captured as walkability components and can also be captured in regulatory items in zoning ordinances. 21Composite and component-specific walkability analyses pave the way for rapid translation of epidemiological research into policy.
Multicomponent definitions of walkability can inform policy on walking, cycling, and public transit infrastructure, 22 which can help reduce vehicle emissions and mitigate climate change, with extensive concomitant health benefits. 23Future triangulation of evidence on the association of walkability with obesity-related cancers from prospective cohort studies (with negative controls), interventions, and natural experiments, as well as detailed exploration of interactions of components, will further strengthen policy-relevant evidence on walkability.][26] Each city includes walkable land that has evolved from the socioeconomic conditions and decisions of the people who live, and have lived, there.This mutability of cities offers opportunities for epidemiological research and motivates the translation of research findings into actionable interventions for shaping built environments with the aim of population-level promotion of health.Within the role of the built environment in the etiology of cancer, studies linking walkability exposure to key risk factors for cancer, such as physical activity, provide a mechanistic rationale.The natural next step is to begin disentangling the mechanisms that drive associations between walkability and cancer, with a future vision of leveraging population-wide walkability exposure to contribute to cancer prevention.Urban practitioners have been debating optimal density for decades. 27,28It is therefore crucial that epidemiologists tackle the complexity of walkability, which can be informatively assessed as joint and individual components when placed in context.