Defining gentrification for epidemiologic research: A systematic review

Neighborhoods have a profound impact on individual health. There is growing interest in the role of dynamic changes to neighborhoods–including gentrification–on the health of residents. However, research on the association between gentrification and health is limited, partly due to the numerous definitions used to define gentrification. This article presents a systematic review of the current state of literature describing the association between gentrification and health. In addition, it provides a novel framework for addressing important next steps in this research. A total of 1393 unique articles were identified, 122 abstracts were reviewed, and 36 articles published from 2007–2020 were included. Of the 36 articles, 9 were qualitative, 24 were quantitative, and 3 were review papers. There was no universally accepted definition of gentrification; definitions often used socioeconomic variables describing demographics, housing, education, and income. Health outcomes associated with gentrification included self-reported health, preterm birth, mental health conditions, alcohol use, psychosocial factors, and health care utilization, though the direction of this association varied. The results of this review also suggest that the impact of gentrification on health is not uniform across populations. For example, marginalized populations, such as Black residents and the elderly, were impacted more than White and younger residents. In addition, we identified multiples gaps in the research, including the need for a conceptual model, future mechanistic studies, and interventions.


Introduction
Neighborhoods are an integral component of the social determinants of health, with a substantial body of literature reporting associations between the neighborhood in which an individual lives and their risk for multiple health conditions. Prior research suggests a positive, linear relationship between neighborhood wealth and better health outcomes [1], with residents of wealthier neighborhoods generally receiving the health benefits of greater access to green space [2], healthier food options [3], less crime [4], and other amenities, as compared to residents of poorer neighborhoods. Increasingly, researchers have been interested in understanding the population health impacts of changes in neighborhoods over time [5].
As population trends indicate that the neighborhoods of major cities are experiencing rapid demographic and socioeconomic shifts [6], there is a growing need to understand the differential impact that these changes have on neighborhood residents and community dynamics. One such change occurs when a neighborhood becomes upwardly mobile through an influx of younger and/or higher SES individuals, resulting in increased housing prices and a higher cost of living-a process generally referred to as "gentrification [7,8]." Gentrification has been examined extensively in economics and sociology literature, broadly encompassing themes around the influx of higher-income households, changes to the racial composition of neighborhoods, and the displacement of lower-income individuals [9][10][11]. Outcomes have focused on the impact that gentrification has on crime, labor markets, and education opportunities. Our analysis focuses specifically on what is arguably downstream of these other outcomes: the impact of gentrification on health. As epidemiological research increasingly studies the impact of social determinants of health, it is critical to examine both the ways in which processes such as gentrification are defined and how a given definition may affect its association with the health of neighborhoods and the individuals residing within them.
Researchers have not reached a consensus as to whether gentrification has a net positive or negative impact on the health of communities, largely due to variation in the literature regarding both how gentrification is defined and the wide range of outcomes examined. To date, there has been minimal effort to forge a consensus on the optimal definition of gentrification. This lack of a common definition has further stymied a comprehensive understanding of gentrification's potential health impacts on residents of changing neighborhoods. This systematic review aims to fill that gap, critically evaluating research that has examined relationships between gentrification, health, and determinants of health, while also identifying the quantitative definitions used by articles to phenotype gentrification. We summarize studies that look at different proximal and distal health outcomes related to gentrification. Based on our comprehensive review of this literature, we also propose a conceptual model to guide future research on gentrification and health.

Methods
A recent rule change (as of October 2019) prevents reviews that have begun the process of data extraction from registering their protocols on the International Prospective Register of Systematic Reviews (PROSPERO). We were, therefore, unable to register our protocol. However, both our protocol and review, described below, follow the 2009 PRISMA Statement [12], as outlined in the supporting information (S1 Table).
include articles discussing the mechanisms and moderators that may affect the relationship between gentrification and health, our search included terms such as "inequity," "equity," "social capital," "race", and "socioeconomic status." A complete list of terms used in in our search can be found in the supporting information (S1 Appendix).
Literature searches were rerun every week to ensure inclusion of relevant and newly published articles. References cited by included articles were also screened for relevance. In addition, all original studies from included reviews were examined for inclusion into our review.

Inclusion and exclusion criteria
Articles included in this review focus on the relationship between gentrification and constructs of health. Peer-reviewed qualitative, quantitative, and review articles, both within and outside the United States, were considered for inclusion. This review aims to identify the phenotypes used to define gentrification, and therefore, did not formally define gentrification a priori; we included articles examining the process in relation to its impact on health. Although associated terms such as "urban renewal" and "urbanization" were included in the literature search, articles using these terms were ultimately excluded, as these terms were too broad and/or did not adequately capture the process of gentrification. Additional exclusion criteria included: non-English language, not peer-reviewed, health not measured as a dependent variable, and no explicit definition of gentrification.
The dependent variable of interest was health and was broadly defined. We focused on one or more of the following health measures and outcomes: self-reported health, sub-clinical conditions (e.g., hypertension, diabetes), health-related behaviors (e.g., alcohol consumption), environmental changes that impact health (e.g., park construction), and social determinants of health (e.g., housing insecurity) with implications for health. Peer-reviewed literature examining the mechanisms by which gentrification may affect health were also included.

Study selection
The initial searches yielded 1,484 articles. After removing duplicates (n = 91), articles with unrelated titles (n = 760), and those that did not discuss gentrification or health specifically (n = 511), 122 articles were selected for abstract review. After eliminating articles that did not meet inclusion criteria, 31 articles were selected for a full review. Of these, 21 met the inclusion criteria. Repeated searches following the initial review date and utilizing the same screening process yielded 21 additional articles for full review and 14 for inclusion, for a sum of 35 peerreviewed articles (Fig 1). One additional article was found by reviewing articles included in systematic reviews and references cited within original research. This resulted in a total of 36 peer reviewed articles.
All three co-authors participated in the full text review and discussed all articles that were included in this manuscript. Potential disagreements were resolved with a majority consensus on the study's final inclusion.

Data extraction
For data extraction, included articles were categorized into qualitative, quantitative, and review articles. The following study characteristics were extracted and recorded onto a pre-defined data abstraction form: author(s), title, manuscript type, study population, variables used to define gentrification, definition of gentrification, sources of data, analytical methods, and a summary of article findings. Additionally, all articles were organized into one or more of the following health categories: self-reported health, health related behaviors, environmental changes that have implications for health or health behaviors, moderators/group differences, potential mechanisms (social capital and psychosocial stress), and health outcomes.
For articles quantitatively defining gentrification, we summarized the variables used to define gentrification, geographic level used to identify gentrification, and the time between the baseline and follow-up assessment to measure neighborhood change.

Quality assessment
To assess the quality of studies included in the review we used three tools. For quantitative studies, we used an assessment tool developed jointly by the National Institute of Health and Research Triangle Institute International [13]. For qualitative studies, we used an assessment tool developed by the National Institute for Health and Care Excellence [14]. For systematic reviews, we used the guidelines of the respective protocols the reviews adhered to (PRISMA and RAMESES) [12,15]. All three co-authors reviewed the studies and dual review was conducted on 20% of the studies. Disagreements between two co-authors reviewing a study were adjudicated by the third co-author.

Risk of bias
We assessed all 36 studies for potential bias that could have impacted their findings. Both qualitative and quantitative studies were assessed against 14 metrics of high-quality studies [13,14]. In general, studies met most criteria and were deemed of relatively high quality with a low risk of bias. All studies explicitly stated the study question, adequately defined the study population, exposure, outcomes, and covariates. However, under the metrics used to indicate higher quality, many studies did not measure the exposure across more than one time period, instead using a binary definition to assess gentrification. This creates the potential to misclassify neighborhoods that gentrified, as neighborhoods not meeting thresholds for gentrification may still be changing enough to impact health. Despite not meeting this one metric, included studies met most of others, thus carrying an overall low risk of bias.

Conceptualizing gentrification
The reviewed articles used multiple phenotypes to define gentrification. Among studies using a qualitative approach, gentrification was defined through investigator-determined changes in demographics, rent prices, or household income [17,31,33,34,40,42,43,46,47]. Studies that used a quantitative approach to define gentrification generally applied two major criteria [5,  Self-reported health, depression/anxiety symptoms and Moderators/Group Differences Economically vulnerable older adults in gentrifying neighborhoods had higher SRH than those in lowincome neighborhoods and more depression and anxiety symptoms than those in affluent areas. Higher-income older adults in gentrifying neighborhoods had poorer mental health than those in low-income neighborhoods and more depression and anxiety symptoms than those in affluent areas

United States
Matching design and linear regression Gentrified: Increase in MHI, percent collegeeducated residents, median owner-occupied housing value, median rent.
(Continued ) Higher SES neighborhoods had less walkability. This was relationship was moderated by gentrification.

Gentrified:
Neighborhoods as those in top 95% of rank change.
(Continued )   N/A-This article uses qualitative focus groups to determine neighborhood stressors in the five boroughs of New York City. Although gentrification is mentioned as a common stressor, it is not discretely measured. 15 focus groups in the five boroughs of New Yok City, three in each one. The median size of the focus group was 10, though the size ranged from 6-17.

Mechanisms of Health Implications (Psychosocial Stress)
The most common perceived stressors in the boroughs of New York City were perceived neglect and physical disorder, safety and harassment by police, and gentrification and racism. With gentrification, a common theme was the effect of gentrification on people of color, and the perceived racial preferences to the gentrifiers over those living in the original communities Gentrification is defined as a phenomenon involving the "'invasion' of previously working-class neighborhoods by middle or upper-income groups and the subsequent displacement of many of the original residents" 30 Qualitative interviews of 30 residents (all older age) of the two neighborhoods studied.

Mechanisms for Health Implications (Social Capital) and Moderators/Group Differences
Gentrification may be associated with an increase in social exclusion, insecurity, and connectedness with the neighborhood. Gentrification was viewed as positive by some, though others described the process as leading to a loss in familiarity with their neighborhood.

Montreal, Canada
Inductive and deductive approaches to identify themes. Codes were generated using a grounded theory approach (Continued ) Most participants felt a strong level of identification with their neighborhood and its people, one that was being taken away as newcomers moved into the area. This changing attitude was described by tensions participants had with "outsiders" and concerns about church tourism. Participants also reported financial pressures as a concern to aging in place and aging near family.  16, 18-27, 29, 30, 32, 36-39, 41, 44, 45, 48-50]. First, neighborhoods had to be eligible to be gentrified. Eligibility was defined by comparing a neighborhood (e.g., census tract, block group) with a larger geographic area (e.g., city). Values for these variables generally had to be at or below the median of the larger geographic area or have a negative z-score compared to the larger geographic area. Second, there had to be a change in z-score value, percent, or numerical change from a baseline period to follow-up for a variable(s) of interest. Although most studies defined gentrification as a binary concept occurring over 2 time points, two studies described gentrification as a staged process [19,23], using quintiles of z-scores or percentiles as indicators of different levels of gentrification. Within these two parameters, the variables and number of variables used to define gentrification differed, as presented in Table 2. Only two studies used a single variable, one based on changes in educational attainment [45] and other based on changes to the white population [39]. Most studies used between 2 to 6 variables to define gentrification.
Across these studies, common variables incorporated in defining gentrification included neighborhood socioeconomic factors (e.g., median income, proportion with bachelor's degree), measures of housing value and stock (e.g., median home value, proportion of housing greater than 20 years old, median rent price), and changing demographics (e.g., proportion of the population Black or White, proportion aged 30-44 years, or proportion aged 65+). Additional variables included urbanization of neighborhoods, change in immigrant population, and change in type of professions. Three studies used a composite index to define gentrifying areas. The indices, the Socioeconomic Index for Areas of Relative Advantage and Disadvantage (SEIFA) in Australia, the Index of Multiple Deprivation (IMD) in the United Kingdom, and the Neighborhood Change and Gentrification Scale (NCGS) in the United States, included a variety of variables that measure household income, employment, housing, occupation, education, material objects, and health.

Health-related outcomes
Self-reported health. Six studies in this review examined the relationship between gentrification and self-reported health [21,22,25,32,36,49]. Multiple reported that gentrification was not associated with or only marginally associated with higher self-reported health [21,22,25,36,49]. However, negative associations were found in analyses that examined specific subgroups, such as African Americans [22,25] or higher-income older adults [32], with gentrification being associated with worse self-reported health. Physical and mental health diagnoses. Six studies quantified the association between gentrification and biomarkers or disease symptoms/diagnoses [5,19,29,32,37,38]. In one study, greater neighborhood affluence and gentrification was associated with a lower risk of hypertension [29]. Another study examining diabetes incidence found that compared to those living in "stable" areas, those living in "declining SES," "new housing" and "improving SES" areas have a decrease in diabetes incidence [5]. One study considered a range of health measures in evaluating the relationship between gentrification and the health of low-income children, including BMI, asthma, attention-deficit/hyperactivity disorder (ADHD), conduct disorder, and anxiety or depression prevalence, emergency department visits and hospitalizations, and proportion of children with at least one well-child visit (or routine exam). Children in rapidly gentrifying areas experienced higher rates of anxiety and depression, though no relationship between gentrification and health was found for the other outcomes [19].
The impact of gentrification on mental health was also assessed by three other studies [32,37,38]. One found that compared to their counterparts in high income areas, both economically vulnerable and higher income older adults in gentrifying areas experienced more depression and anxiety symptoms [32]. The other found that low-income and long-term adults of gentrifying neighborhoods had a higher likelihood of serious psychological distress than those in low-income and not gentrified neighborhoods [38]. However, gentrification was not associated with the likelihood of having psychotic experiences [37].
Health related behaviors. Four studies examined the impact of gentrification on healthrelated behaviors and healthcare utilization [24,26,47,48]. Two examined gentrification in relation to emergency department encounters. One found that displaced residents were more likely to have an emergency department encounter, experience hospitalizations, and make mental health visits as compared to residents who remained in gentrifying neighborhoods or non-gentrifying neighborhoods [26]. The other study found that decreases in the rate of emergency room admissions were not associated with gentrification [48].
Studies that examined the association between gentrification and alcohol use presented conflicting results. In one study, gentrification was not associated with increased drinking (i.e., binge drinking) overall but did increase the odds of binge drinking in residents who had lived in the neighborhood for less than 5 years [24]. Another study suggested that gentrifying areas are more likely to criminalize those who participated in public drinking, with several study participants reporting greater levels of isolation and increased drinking as a result [47].
Environmental changes that have implications for health or health behaviors. Seven studies considered the relationship between gentrification and environmental changes that impact health [16,18,30,34,39,41,45]. Four of these studies suggested that gentrification is related to changes in access to green space, air pollution, and walkability [16,18,41,45]. Greater exposure to active green space was associated with lower odds of reporting fair or poor health [18]. However, within gentrifying neighborhoods, researchers found that the health benefits of increased green space only held for those with high incomes or high levels of education [18]. In another study, gentrification was found to occur in closer proximity to green spaces, though researchers suggested that this "green gentrification" may result in the displacement of socially vulnerable residents [41]. In regard to its impact on environmental hazards, one study found that air pollution levels are not equally distributed among neighborhoods that did and did not gentrify; neighborhoods that did not gentrify had a greater proportion of cumulative toxic air pollution compared to other neighborhoods [16].
Three studies examined the relationship between gentrification and food accessibility [30,34,39]. One study found that gentrification increases the prevalence of both healthy and (particularly) unhealthy food options [30] while one suggested that gentrification may increase food insecurity and force residents to spend a greater proportion of their income on housing, thereby leaving less to spend on nutritious food [34]. Another study examined the relationship between food mirages (areas with physical access to food but high food prices). It found that food mirages were more likely to be found in areas experiencing gentrification, suggesting that low-income individuals in these areas may not be able to benefit from increased access to food [39].
Moderators/Group differences. Several studies we reviewed emphasized that the impact of gentrification on a neighborhood is shaped by the demographics of the neighborhood; that is, gentrification may impact one population and not another [16-20, 22-25, 27, 32-34, 38, 43, 46, 47]. For example, Huynh and Maroko reported that although gentrification overall was not associated with preterm birth, very high gentrification-defined as neighborhoods with a substantial increase in college educated and high median household income-was adversely associated with preterm birth for non-Hispanic Blacks but protective of preterm birth in non-Hispanic Whites [23].
In another study, elevated self-reported stress levels, which have been shown to negatively impact health, were higher among non-White residents within neighborhoods that experienced White gentrification as compared to neighborhoods that had experienced Black gentrification [20]. Other studies also suggested that race may moderate the impact of gentrification on health, with Black individuals reporting worse self-reported health [22,25].
Gentrification may also impact recovery efforts for drug users. Studies showed that injection drug users were more likely to be homeless in neighborhoods that gentrified [27] and were unable to access overdose prevention sites because of policing practices in gentrifying areas [43]. This can increase risks for multiple health conditions [51].
Studies suggested differential results on the impact of gentrification on health by age. Anxiety or depression diagnoses have been shown to be significantly greater among low-income children in rapidly gentrifying areas than those in persistently low SES areas, though gentrification was not associated with other health measures in this population [19]. In another study, gentrification was associated with increases in both healthy and unhealthy food options but was not associated with rates of childhood obesity [30]. Among older adults, gentrification inversely affected the self-reported health of those who were higher-income, while positively affecting that of those who were low-income, when compared to their counterparts in lowincome neighborhoods [32]. Among older adults, gentrification was also attributed to a loss of social capital [33,40,42]

Potential mechanisms
Our review also included studies that examined potential mechanisms mediating the association between gentrification and health. The current literature suggests that gentrification may impact health through social capital and psychosocial stress [17,28,31,33,35,40,42,44,46,49].
Social capital. Social capital has been described as the collective value of all social networks and the inclinations that arise from these networks to do things for each other (norms of reciprocity) [52]. In the face of change, social capital can be critical in maintaining one's health. One study examined this, finding that higher levels of identification or connectedness with one's neighborhood can act as a buffer to protect individuals from mental health strains related gentrification [44].
In our literature search, five articles linked gentrification with a decline in necessary components of social capital: neighborhood trust, social cohesion, and/or social networks [17,33,40,42,46]. All these articles examined the impact of gentrification on vulnerable communities (e.g., the elderly and/or marginalized racial and sexual backgrounds). Versey and colleagues describe common themes among older Black residents of gentrifying neighborhoods. Participants reported a decrease in neighborhood trust, collective efficacy, and a disruption of social networks, outcomes related to poor health [33]. Another common narrative was perceived conflict between gentrifiers and long-term residents; the former of whom were seen as dismantling long standing community structures. Gentrification, though viewed positively by some, was generally attributed to a loss of social spaces for long-term residents [40,42].
Only one study suggested that gentrification increases collective efficacy, another form of social capital [49]. Researchers hypothesized that incoming higher income residents, who may carry increased political influence and agency, could demand higher quality resources and promote community initiatives, benefitting long-term residents.
Psychosocial stress. Gentrification was also cited as a source of psychosocial stress [31]. Participants in several studies also noted shifting racial demographics accompanying gentrification, leading to perceived increases in preferences towards the White residents. Among non-white residents, this was seen as a source of elevated stress and decreased community cohesion [17,31,33,40]. Gentrification was also cited as a cause of decreasing comfort for certain groups such sexual/gender minority identified sex workers [46] and people who use drugs [27,43].
Research has attributed increased psychosocial stress to poor health outcomes [53]. Within the context of gentrification and social capital, one review suggested separation from a familiar environment and difficulty establishing new social networks may result in elevated plasma cortisol levels, more hospitalizations, and decreased overall wellbeing [28].

Discussion
Our systematic review of gentrification and health yields several insights that are valuable for future research in this area. Based on our review, we propose a conceptual model (Fig 2) that describes the process of gentrification as involving changes to both the physical and social environment, which in turn can indirectly affect health outcomes through several mediators. Based on accumulated evidence, we also suggest that characteristics like race and age may modify the association between gentrification and health, such that Black residents and the elderly may be more likely to have their health negatively impacted by gentrification than White and younger residents.
Our findings on the relationship between gentrification and health are consistent with prior research on the health impact of physical and social environments. In this context, the physical environment refers to the availability and quality of resources within a neighborhood, including transportation, access to healthy food, noise level, air pollution, and recreational resources. These factors impact behaviors such as physical activity, diet, and sleep. These behaviors, in turn, can affect proximal biological factors such as blood pressure, diabetes, and stress levels, which can increase risk for distal health outcomes such as cardiovascular disease [54,55]. The social environment includes safety and violence, social support and cohesion, and social norms. These factors can impact psychosocial stress levels, which in turn affect proximal biological factors and thereby impact the risk of distal health outcomes [56,57]. Our review finds that gentrification, specifically, can impact neighborhoods and the health of residents through environmental factors, stress, and social capital.
However, research on gentrification and health is limited by inconsistencies in defining gentrification. The qualitative studies we reviewed all defined gentrification differently, often using distinct sources to support their definitions. Among the quantitative studies, we identified common approaches in developing a definition. These approaches include 1) identifying neighborhoods that were eligible to be gentrified and 2) measuring a change in z-score value, percent, or numerical change from a baseline period to follow-up for a variable (s) of interest. However, within these two parameters, the number of variables used to define gentrification was inconsistent, varying from one to six. Based on the most central variables that have been used in the gentrification literature, we propose defining and quantifying gentrification based on changes to neighborhood socioeconomic status (median income, median home value/rent, educational attainment) and changing demographics (race and age).
In addition, most phenotypes of gentrification compared socioeconomic variables at the census tract or block group to a city-wide average. Implicitly, this suggests that gentrification is possible in any city, including places like San Francisco, Palo Alto, or even Beverly Hills. However, it is clear that these cities are not technically gentrifying, even if certain areas within them are below the city-wide mean, as they are already starting with strong labor markets, high employment, and other indicators of a high socioeconomic level. As such, novel phenotypes of gentrification are needed that do not require a comparison to the surrounding area. Lastly, gentrification is not a binary exposure, as it was defined by most studies; it is a process that can occur over a period of years and even decades. As such, the lack of granular socioeconomic data from prior decades at the same geographical level to define gentrification further impedes study of this phenomenon.
The limited number of health outcomes examined in gentrification literature presents another limitation to current literature. Linking neighborhood level data that defines gentrification with other sources of health data-such as the electronic health record (EHR)-may allow for a more robust examination of the association between gentrification and health [58]. Such approaches can also enable researchers to examine the health outcomes of those who may have been displaced by gentrification, another challenge faced by most of the reviewed studies.
By linking multiple data sources and collecting primary data-such as biomarkers-we might also be able to address a third gap in knowledge: the paucity of research on the mechanism(s) by which gentrification may impact health. One possible mechanism by which gentrification impacts health is through social capital. Prior work has shown that strong levels of social capital and its related constructs, such as collective efficacy, social networks, community cohesion, and social fabric are associated with multiple positive health outcomes, including: lower obesity prevalence [59], fewer hospitalizations from coronary heart disease [60], lower prevalence of sexually transmitted diseases [61], higher levels of self-reported health [62], and decreased mortality [63].
However, the way in which gentrification affects social capital remains contested. Some have proposed that gentrification results in greater social capital, providing disadvantaged communities with increased opportunities to access individuals and institutions associated with mainstream society [64]. Research also suggests that gentrification erodes social capital, particularly for disadvantaged groups [8][9][10]. Policies that promote gentrification and a decrease in social cohesion may lead to the displacement [65,66] and socio-spatial segregation [11] of cities, exacerbating social capital. Further research on the mechanisms by which gentrification may impact health is critical and necessary in answering debates like this one.

Limitations
All effort was made to ensure that a comprehensive literature search was conducted. It is possible that articles were missed for various reasons, including for not being published in English or gentrification being described using different terminology. Additionally, there is the possibility of publication bias, but we did identify articles showing no association between gentrification and health in our review.
A novel conceptualization of the association between gentrification and health is needed to inform this future research, including mechanistic studies and interventions that can mitigate any negative impact of gentrification. Fig 2 depicts our conceptualization, based on current literature, of how gentrification is associated with health. Much like prior models that depict how neighborhoods impact health, we propose gentrification changes the physical and social environment, which in turn can result in biological responses (e.g., increased levels of stress), changes in health-related behaviors (e.g., physical activity, management of pre-existing conditions), and/or changes in healthcare utilization (e.g., emergency department visits, preventative care). These proximal impacts may increase or decrease risk for long term mental and physical health outcomes. All of these changes occur over time and can be moderated by public policy such as economic incentives and subsidized housing.

Conclusion
Gaps in knowledge inform future directions for this line of research. Developing a consensus definition of gentrification requires novel phenotypes of gentrification, specifically ones that do not rely on relative comparisons to surrounding geographies. Based on the current literature, a core set of health outcomes that should be included in studies that examine the association between gentrification and health include measures of mental and physical health (e.g., depression, anxiety, hypertension, diabetes) and healthcare utilization (e.g., primary care encounters). There is a need for more proximal measures of health-such as biomarkerssince gentrification may not impact incidence of hard outcomes (e.g., myocardial infarction, stroke) in the short term. All of this should be done using analytic approaches that incorporate other data sources, including information on the pollutants in the environment and the physical resources available to residents. The results of these ongoing questions may inform interventions to address this social determinant of health. Supporting information S1