Skip to main content
NCBI home page
EPA Author Manuscripts logoLink to EPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Sep 1.
Published in final edited form as: Soc Sci Med. 2020 Jul 15;261:113211. doi: 10.1016/j.socscimed.2020.113211

Active travel and social justice: Addressing disparities and promoting health equity through a novel approach to Regional Transportation Planning

Nicole Iroz-Elardo a,b,*, Jessica Schoner a,c, Eric H Fox a, Allen Brookes d, Lawrence D Frank a
PMCID: PMC7939113  NIHMSID: NIHMS1665755  PMID: 32745821

Abstract

Public health impacts of transportation policies and infrastructure investment are becoming better understood, particularly for those associated with physical activity. Yet health impacts are not routinely evaluated within the context of the development of a Regional Transportation Plan (RTP) and subsequent programming and investment processes. This is particularly concerning because the spatial distribution of planned transportation infrastructure potentially has significant health equity implications for vulnerable populations at greater risk of chronic disease. This study discusses the application of the National Public Health Assessment Model (NPHAM) – a new approach that expands several scenario planning tools to include health – for the San Joaquin Council of Governments 2018 RTP. It demonstrates how quantifying health impacts at a finer spatial scale (census block groups) helps assess the extent to which RTP strategies are likely to benefit or harm health. It further enables a spatial form of health equity analysis that can help planners understand where infrastructure is most needed to meet social equity goals. To the knowledge of the authors, this is the first example of a quantified, health equity analysis of transport physical activity and a health outcome – body mass index - associated with an RTP; it demonstrates significant advancement in transportation planning practice and policy.

Keywords: Health equity, Scenario planning, Built environment, Physical activity, Obesity, Environmental justice

1. Introduction

Physical activity is a key health behavior that impacts risk factors, such as obesity, and a wide range of health outcomes including metabolic diseases such as diabetes (Aune et al., 2015; Jeon et al., 2007), cardiovascular disease (Carlsson et al., 2016; Murtagh et al., 2015; Williams, 2013), cancer (Li et al., 2016; Mahmood et al., 2017; Moore et al., 2016), and all-cause mortality (Williams, 2013). For two decades, researchers have also demonstrated that active travel – walking, biking, and walking to and from transit – is a viable way to modestly increase physical activity and health across a wide swath of the population (Laine et al., 2014; Wanner et al., 2012). Research is solidifying linking the built environment, physical activity and health outcomes like obesity, cardiovascular disease, and diabetes (Creatore et al., 2014; den Braver et al., 2018; Grasser et al., 2013; Malambo et al., 2016; Mayne et al., 2015). In turn, “walkability” research has focused on the environments most supportive of active travel including land use patterns that promote density, connectivity, and mixed use (Durand et al., 2011; Ewing and Cervero, 2010; Frank et al., 2006; Grasser et al., 2013; McCormack and Shiell, 2011; McGrath et al., 2015).

Health is considered an important co-benefit of active transportation, whereas sedentary time spent in cars (Frank et al., 2004) generates harmful emissions and is an adverse impact of land use and transportation planning and investment. However, long range planning processes such as a Regional Transportation Plan (RTP) do not routinely incorporate quantitative physical activity estimates as performance measures. For example, Singleton and Clifton (2014, 2017) have reviewed the state of the practice of health inclusion within RTPs twice. In their first analysis, only four of 18 RTPs included physical activity as a performance measure; only one of those investigated physical activity specifically from active travel (Singleton and Clifton, 2014). In their follow-up study, only eight of the 25 plans reviewed included physical activity performance measures; most “measure the number of walk, bicycle, and walk-to-transit trips or mode share as a proxy for physical activity” (Singleton and Clifton, 2017). Only San Francisco and Los Angeles – very large and well-resourced metropolitan planning organizations (MPOs) – reported transport physical activity measures. Not a single RTP through 2014 included a health endpoint – for example, prevalence of obesity, high blood pressure, diabetes, cardiovascular disease – as a performance measure (Singleton and Clifton, 2014, 2017). As representative national datasets, surveillance data such as the National Transportation Household Travel Survey (NHTS) and Behavioral Risk Factors Surveillance System (BRFSS) are not generally available at the fine spatial scale needed to support transportation decisions. With very few models to translate surveillance data into small area estimates of transport physical activity and health, MPOs are unable to incorporate health indicators for assumed baseline and future conditions.

Policy makers and planners are beginning to incorporate the cobenefits of health to internalize health into decision-making and address equity challenges. This is most explicit in California, where, by legislative mandate, the environmental justice element of a land use plan must “identify objectives and policies to reduce the unique or compounded health risks in disadvantaged communities by means that include… [promotion of] physical activity” (California Governor’s Office of Planning and Research, 2017; SB-1000, 2016). California may be leading the incorporation of health into planning, but other examples across the U.S. suggest widespread interest. The Nashville Area MPO has invested in customizing the Integrated Transport and Health Impact Tool (ITHIM) to translate changes in active transportation minutes into physical-activity and air-quality health outcomes (Whitfield et al., 2017). In other efforts, the authors have augmented RTPs for the MPOs of both Los Angeles, California and Madison, Wisconsin using similar land-use based modeling as presented here. Despite the advent of these advanced modeling techniques, applications, and legislative mandates, most MPOs continue to list health as a policy goal within RTPs with weak links to indicators and associated modeling practices (Singleton and Clifton, 2017).

As MPOs have struggled to move from conceptual to concrete inclusion of health in long-range planning over the last decade, a similar conversation about equity objectives and performance measures in transportation planning has been occurring (Bills et al., 2012; Federal Highway Administration, 2011; Golub and Martens, 2014; Karner and Niemeier, 2013; Martens et al., 2012). There have been calls to identify and operationalize meaningful equity objectives and performance measures of multiple dimensions of social equity (Manaugh et al., 2015) and to link those measures to policy guidance and funding (Amekudzi et al., 2012; Singleton and Clifton, 2014). Some methods are informed by spatial units of analysis using a threshold approach to identify areas of concentrated populations (Klein, 2007); others suggest a householdbased analysis to capture all individuals within a demographic group. The general consensus is that both spatially defined thresholds of concentrated equity populations (composition of neighborhoods) and individual equity individuals or households, regardless of the neighborhood composition, should be tracked for both baseline and future scenarios; however, nearly all MPOs default to the threshold approach (Brodie and Amekudzi-Kennedy, 2017; Karner; Niemeier, 2013; Manaugh et al., 2015).

While the incorporation of physical activity measures into long-range planning exercises is rare, the application of health equity analytical methods is arguably non-existent. Developing health equity methods for long range planning should be prioritized. Promoting health equity by remedying health disparities is the primary impetus for public health professionals’ interest in transportation planning as a form of population level intervention (Bambra et al., 2010). The common agreement is that low socio-economic status (SES) groups have reduced levels of physical activity and worse health than those with higher SES (Gaskin et al., 2014; Lovasi et al., 2009; Wen and Kowaleski-Jones, 2012). There is also concern that transportation infrastructure improvements disproportionately help and are used by higher SES groups (Smith et al., 2017). Research has shown that low SES groups are increasingly more likely to live in less accessible, less walkable neighborhoods (Manaugh and El-Geneidy, 2012; Smith et al., 2017). Others have suggested that different associations between the built environment and walking rates of lower SES groups may reflect different walking preferences and needs, some of which may be socially or culturally driven (Adkins et al., 2017; Casagrande et al., 2009). Still, low SES groups are more likely to walk for transport – a phenomenon akin to non-choice riders of transit – even if this does not always translate into better health. For example, an Atlanta-based study found that Black and lower income participants of a regional household travel survey walked twice as much but were twice as likely to be obese as their white counterparts (Frank et al., 2004). A methodology is needed that addresses the spatial distribution of physical activity and sedentary impacts from both utilitarian and recreational travel. Moreover, an approach is required to “hard wire” land use and transportation decisions with chronic disease endpoints to allow for the evaluation of health equity impacts of contrasting investments within RTP processes and thus inform policy and investments.

2. Methods

Methods for quantifying physical activity-based health impacts of proposed transportation and land use plans generally follow a comparative risk assessment approach to compare baseline and projected levels of physical activity and resulting health impacts. Two primary approaches to comparative risk assessment have dominated modeling efforts. The first is to use relative risks of disease for specified ages and gender culled from the literature and multiply them by changes in exposure to physical activity; this is the approach used by tools such as the Integrated Transport and Health Impact Model (Whitfield et al., 2017) and Health Economic Assessment Tool (World Health Organization, 2018). The second option is to use a land-use regression to build predictive physical activity and chronic diseases models; these types of tools have the advantage of being sensitive to differing land-use and transportation contexts as well as incorporating demographic characteristics as controls. These latter methods are able to scale down to smaller geographies with varying demographics to evaluate the spatial distribution of projected change in disease – a primary implied research question underlying many transportation agencies’ interest in health. Thus, the land-use regression approach was used for this study.

2.1. Using the National Public Health Assessment Model (NPHAM) to estimate health impacts for equity areas

This equity-centered analysis is an extension of the National Public Health Assessment Model (NPHAM) funded by the U.S. Environmental Protection Agency (EPA) and developed by Urban Design 4 Health, Inc. NPHAM is a compilation of regression-based equations that predict health behaviors such as walking minutes and health indicators such as body mass index (BMI) and poor health at the census block group (CBG) level. The portion of NPHAM used for this project was developed with address-level surveillance data from the California Health Interview Survey (CHIS) which has some similarities to the spatially aggregate federal Behavioral Risk Factor Surveillance Survey (BRFSS). The BRFSS is not available at the CBG level. Therefore, we have used the large sample size of CHIS to create a California-based model that can then predict health behaviors and outcomes at a finer scale across the U.S. for “baseline” and “future” year of an MPO’s RTP. To create these models, CHIS participants were spatially matched with detailed built environment and census demographic data at the CBG level; both census and built environment data were then used to model and predict CBG activity and health estimates. Much of the built environment data used in NPHAM comes from the National Environmental Database (NED) of built, natural, and social measures developed by Blinded with the support of the Robert Wood Johnson Foundation. The NED offers complete coverage of the 50 U.S. states, supporting seamless NPHAM applications anywhere in the US.. A more detailed description of the NPHAM coded equations and the methodology by which they were developed has been published elsewhere (Schoner et al., 2018). The following NPHAM outcomes were included in the health equity analysis: weekly transport walking minutes per person, weekly leisure walking minutes per person, and average BMI.

NPHAM is “software agnostic,” developed to work with multiple scenario planning tools (SPT), such as Calthorpe Analytics’ UrbanFootprint, Fregonese Associates’ Envision Tomorrow, and CityExplained’s CommunityViz, by including routines that transform built environment and demographic data within the SPT into variables consistent with internal equations. When the paired SPT does not have the underlying data needed for a particular variable, data is assumed to hold constant at “baseline” levels from the NED. This allows planners to pair NPHAM with the SPT of their choice that is likely already utilized in their planning efforts using only already defined MPO assumptions about scenarios for consistency across models. For the analysis undertaken in San Joaquin County, California, NPHAM was paired with Envision Tomorrow, the SPT already being used by the San Joaquin Council of Governments (SJCOG) for their planning exercises. Fig. 1 shows an example of NPHAM being used with Envision Tomorrow operating within the ESRI ArcGIS Desktop environment.

Fig. 1.

Fig. 1.

Open in a new tab

Example of National Public Health Assessment Model (NPHAM) within Envision Tomorrow’s ESRI ArcGIS software environment.

2.2. San Joaquin Council of Government’s 2018 Regional Transportation Plan and assumptions

SJCOG – as is typical of many MPOs – initiated a draft 2018 RTP with three separate scenarios under consideration for the 2035 build-out horizon. This in-depth post-processing equity analysis was performed on Scenario 2 which most closely matched the adopted final plan. The first scenario was the least aggressive in terms of density with only 20% of new housing units; the second and third scenarios assumed 39% and 49% of new housing units would be multi-family, respectively. Scenarios 2 and 3 also assumed higher levels of expanded transit (bus rapid transit and commuter rail) and active transportation facilities. These assumed changes were not concentrated at a single location, rather, changes to the built environments were spread across the region. Table 1 provides the translation of these general strategies into Envision Tomorrow variables including total population, households, employees, and workers; retail worker density; a 5-tier employment entropy (mixed use) indicator; intersection density; and percent low income households. While population growth is relatively stable across the scenarios, the spatial variation of the additional people and employment changes depended on the land use assumptions in Envision Tomorrow. Population density and employment density are more concentrated in Scenarios 2 and 3, although the specific flavor of this concentration varies. Differences between the scenarios were relatively minor for most areas of the study region, suggesting that expected differences in health outcomes by scenario should also be modest.

Table 1.

Overview of changing (baseline and % change from baseline) demographic and built environment variables for the average census block group by scenario.

NPHAM Input Variable Baseline San Joaquin County
% Change from Baseline
Scenario 1 Scenario 2 Scenario 3
Average total population in CBG 3194 +35% +44% +38%
Average total households in CBG 938 +41% +56% +49%
Average total employees (jobs) in CBG 557 +82% +74% +90%
Average percent low income households in CBG 42% −21.42% −21.37% −21.42%
Average gross population density on unprotected land (people per acre) in CBG 9.10 +11% +16% +16%
Average gross employment density on unprotected land (employees/jobs per acre) in CBG 1.73 +25% +34% +38%
Average total retail employment/jobs in CBG 95 +83% +84% +96%
Average 5-tier employment mix (entropy) in CBG 0.49 −0.6% −0.3% −0.3%
Average intersection density in CBG 71 +1.4% +4.4% +4.0%
Open in a new tab

2.3. Post-processing health predictions for an equity analysis

NPHAM, connected to Envision Tomorrow, was run on all three scenarios to estimate health outcomes. Post-processing routines were then applied to calculate different population-weighted averages for the entire region and for populations living in “equity” areas. For consistency, definitions being used by SJCOG for their environmental justice (EJ) analysis were also applied to the predicted health outcomes. CBGs were flagged as one or more “equity areas” if households living below the federal poverty level in 2015 exceeded 30 percent (“poverty areas”); if the non-White alone (non-Hispanic or Latinx) population exceeded 75 percent (“minority areas”); and/or if the CBG was in the top quartile of the California EnviroScreen (“ES areas”). California EnviroScreen is an index created and maintained by the State of California to locate disadvantaged communities and direct resources and interventions accordingly. San Joaquin County includes Stockton – a relatively modest cost of living area with lower wages compared to the larger California metropolitan areas. Therefore, 207 (52.4%) of 395 CBGs in the county qualified as the top quartile of ES areas. Of the ES areas, 136 (34.4%) and 89 (22.5%) CBGs were denoted as minority and poverty areas respectively. Post-processing routines that calculated equity and non-equity area population-weighted averages using predicted CBG values for both 2015 Baseline and Scenario 2 in 2035 (“future”) were completed using equity area definitions based on SJCOG’s recommendation. Independent 2-sample Welch’s t-test (unequal variance) was calculated for each pair of equity to non-equity areas for Baseline and future Scenario 2. A change from Baseline to Scenario 2 was also calculated based on the absolute difference between the population weighted averages for each scenario; this isolates the change in health impacts attributable to the scenario under consideration.

3. Results

Walking for transport and leisure and BMI results for the entire county overall indicate that all three scenarios are expected to modestly improve health (Table 2). When comparing across scenarios, Scenario 3 produces the most health benefits, reflecting the most aggressive increase in density and mixed land use environments that support walking to destinations. Fig. 2 illustrates the estimated Baseline conditions at the CBG for BMI and walking for transportation; it highlights poverty areas largely concentrated in south Stockton and east Lodi. Higher BMI tends to track with poverty areas as well as more outlying rural areas of the county, while higher levels of walking for transportation are focused in central Stockton and the most urbanized areas of Tracy, Manteca and Lodi with a few exceptions.

Table 2.

Overview of estimated behavior and health outcomes by scenario (baseline and % change from baseline.

NPHAM Outcome Variable Estimated Baseline % Change from Baseline
Scenario 1 Scenario 2 Scenario 3
Weekly transport walking (minutes per person) 23.31 (5.65) +0.2% +0.4% +0.5%
Weekly leisure walking (minutes per person) 50.9 (9.24) +0.2% +0.2% +0.5%
Average body mass index (BMI) 28.34 (0.61) −0.3% −0.3% −0.4%
Open in a new tab

Fig. 2.

Fig. 2.

Open in a new tab

Choropleth maps showing the Census block group level geographic distribution of body mass index and weekly minutes of walking for transportation highlighting poverty areas across San Joaquin County.

Estimated levels of walking and BMI were also compared across equity areas. Interpretation of the equity analysis requires familiarization with differences in baseline conditions and assumptions for Scenario 2 by equity area (Table 3). For some variables, both the baseline conditions and the projected changes are different depending on the equity definition. For example, high-poverty CBGs have the highest population density (12.80 gross population per acre versus 9.10 in the region) but the fewest retail jobs (55 versus 95 in the region). High-poverty CBGs are also anticipated to see the lowest growth in retail jobs in the region (45% versus 84% regionally) under Scenario 2.

Table 3.

Assumed baseline conditions and percent change under scenario 2 for the average census block group by different equity definitions.

NPHAM Input Variable Minority Poverty EnviroScreen (ES)
Base % Change Base % Change Base % Change
Average total population in CBG 3290 +27% 1653 +6% 3381 +68%
Average total households in CBG 900 +37% 516 +3% 941 +92%
Average total employees (jobs) in CBG 597 +86% 566 +16% 641 +96%
Average percent low income households in CBG 58.60% −25.47% 70.11% −17.18% 50.05% −23.65%
Average gross population density on unprotected land (people per acre) in CBG 11.71 +13% 12.80 +10% 9.34 +20%
Average gross employment density on unprotected land (employees/jobs per acre) in CBG 2.31 +43% 3.50 +20% 1.95 +44%
Average total retail employment/jobs in CBG 140 +75% 55 +45% 127 +91%
Average 5-tier employment mix (entropy) in CBG 0.47 −1.5% 0.49 −0.9% 0.49 −0.9%
Average intersection density in CBG 78 +3.7% 91 +0.1% 71 +5.4%
Open in a new tab

Table 4 shows population weighted means (and population weighted standard deviation) for the predicted walking variables and BMI by different areas of concern (“equity areas”) under Scenario 2. Table 4 also shows the population weighted average for the balance of the region and the difference between an equity area and the rest of the region; when that difference is statistically significant using Welch’s two sample t-test of unequal variance, letter a is shown. The first three rows show estimated levels at Baseline (2015). Eight out of nine comparisons at Baseline between equity areas and the rest of the county showed statistically significant differences. Current residents in areas of high poverty are estimated to walk for transport 23.9% more (about five additional minutes per week) than the balance of the region. However, those living in poverty areas in San Joaquin County are predicted to currently walk eight fewer minutes for leisure. This is consistent with low-income populations having less discretionary time for leisure travel and fewer transportation choices. Similar statistically significant differences in 2015 Baseline walking minutes are predicted for minority versus non-minority areas. With the broader ES definition of areas of concern, transport walking minutes are no longer statistically different when compared to the balance of the county; however, leisure walking minutes are statistically lower (6.6 min) for ES areas. Finally, average predicted BMI levels range from 1.7 to 2.2 percent higher for areas of concern in 2015 – a statistically significant difference.

Table 4.

Estimated health indicators for Baseline (2015) and Scenario 2 (2035) comparing poverty, minority and California EnviroScreen areas.

NPHAM Health Indicator Poverty Comparison Minority Comparison EnviroScreen (ES)Comparison
Wt. Mean (Wt. Std) Absolute Diff (%) Wt. Mean (Wt. Std) Absolute Diff (%) Wt. Mean (Wt. Std) Absolute Diff (%)
Area of Concern Balance of County Area of Concern Balance of County Area of Concern Balance of County
Estimated Baseline (2015)
Transport walking (minutes, weekly) 27.69 (5.53) 22.35 (5.21) 5.34 (23.9%)a 25.00 (5.62) 22.35 (5.46) 2.65 (11.9%)a 23.64 (5.95) 22.94 (5.31) 0.70 (3.1%)
Leisure walking (minutes, weekly) 44.86 (8.03) 52.21 (8.97) −7.35 (−14.1%)a 45.31 (6.66) 54.07 (9.00) −8.76 (16.2%)a 47.75 (8.30) 54.35 (9.01) −6.60 (−12.2)a
Body mass index (BMI) 28.85 (0.56) 28.23 (0.56) 0.62 (2.20%)a 28.71 (0.53) 28.13 (0.55) 0.58 (2.1%)a 28.57 (0.58) 28.09 (0.53) 0.48 (1.7)a
Estimated Future Scenario 2 (2035)
Transport walking (minutes, weekly) 27.88 (5.72) 22.41 (5.21) 5.47 (24.4%)a 25.16 (5.69) 22.39 (5.47) 2.77 (12.4%)a 23.72 (6.05) 23.02 (5.29) 0.70 (3.0%)
Leisure walking (minutes, weekly) 44.88 (8.00) 52.31 (8.97) −7.43 (−14.2%)a 45.17 (6.38) 54.29 (9.00) −9.12 (−16.8%)a 47.86 (8.38) 54.41 (8.96) −6.55 (−12.0%)a
Body mass index (BMI) 28.80 (0.57) 28.13 (0.53) 0.67 (2.4%)a 28.63 (0.55) 28.03 (0.52) 0.60 (2.1%)a 28.45 (0.61) 28.02 (0.51) 0.43 (1.5%)a
Open in a new tab
a

Statistically significant difference in weighted mean, p < 0.01, using Welch’s unequal variance t-test.

The last three rows of Table 4 display similar estimates for transport walking, leisure walking, and BMI in 2035, assuming that Scenario 2 is fully implemented. Again, eight out of nine comparisons between areas of concern and the rest of the county showed statistically significant differences when using a Welch’s two sample t-test of unequal variance.

Finally, the differences between Baseline (2015) and Scenario 2 (2035) were analyzed for systematic differences across equity areas as shown in Table 5. This tracks impacts attributable to the plan, accounting for population growth and household income changes. Mapping these changes shows where the plan’s impacts are most concentrated. An equity area performing better than the balance of the region suggests equity gains from the planned investments of the RTP. However, it is important to realize this usually indicates a narrowing the equity gap rather than complete parity in 2035.

Table 5.

Difference between estimated health indicators for Baseline (2015) and Scenario 2 (2035) comparing the three areas of concern with the balance of the region.

NPHAM Health Indicator Poverty Minority EnviroScreen (ES)
Area of Concern Balance of Region Ratio Area of Concern Balance of Region Ratio Area of Concern Balance of Region Ratio
Transport walking (minutes, weekly) 0.190 0.061 3.115 0.158 0.042 3.762 0.083 0.085 0.976
Leisure walking (minutes, weekly) 0.014 0.102 0.137 −0.140 0.215 −0.651 0.115 0.055 2.091
Body mass index (BMI) −0.046 −0.107 0.430 −0.085 −0.102 0.833 −0.124 −0.065 1.908
Open in a new tab

Bold indicates equity area shows larger predicted health gains than balance of the region (i.e., net equity gains).

Table 5 also shows a ratio calculated as the change in areas of concern divided by the change in the balance of the region. Positive values of this ratio greater than one indicate gains for the equity area; cells where this is true are bolded in Table 5. Values greater than zero but less than one indicate improvements in equity areas, but at a slower pace than the balance of the region. Negative values indicate that one area is improving while another is experiencing worse outcomes. Tests of statistical significance were not applied to these differences between equity areas and balance of the region due to the challenges inherent in having different population weights at the two time points.

These results estimate that implementation of Scenario 2 will result in modest gains in health behaviors and outcomes from current conditions in most areas. Further, interesting differences emerged across the three equity definitions that were considered. For example, while all areas, regardless of equity definition, gained transportation walking minutes, only minority and poverty areas showed gains (0.158 and 0.190 min per week, respectively) larger than the balance of the region; the gains in ES areas were approximately equal to the balance of the region. To put these modest gains into perspective, the transport walking gains for minority and poverty areas were more than three times larger than the gains in the balance of the region. This suggests that implementing Scenario 2 will certainly help those living in the most critical equity areas increase walking for transport.

In San Joaquin County, every CBG that qualifies as a concentrated poverty or minority area also qualify as an ES area; however, the reverse is not true. These definitions can be leveraged to quickly assess the performance of the Scenario for the lowest SES areas versus “moderate equity” areas (ES CBGs that are not also poverty or minority areas) as an informal sensitivity analysis. For example, non-white, minority areas had the dubious distinction of reduced leisure walking – the only negative health result in the Baseline to Scenario 2 – even as the ES areas showed leisure walking increases that were double those of the balance of the region. Knowing that ES areas are predicted to show positive gains above and beyond the balance of the region while areas of concentrated minorities lose leisure walking minutes suggests that “moderate equity” areas are expected to see the most gains in leisure minutes. A similar analysis can be performed for predicted changes in BMI which is expected to modestly drop for every group. However, only ES areas showed a larger reduction in BMI than the balance of the region (−0.124 versus −0.65). Because the deeply impoverished and high concentrated areas of minority households are not doing as well as the region, we can again assume that “moderate equity” areas gain the most from Scenario 2. Without further intervention to guide resources, the most deeply impoverished areas and/or ethnic enclaves – which typically also are the areas with the worst health outcomes – could be left behind in active travel and public health gains.

4. Discussion

The analysis clearly shows that residents living in areas of concern – regardless of the definition – currently walk less overall due to lower rates of leisure walking. Less walking, along with a host of other negative health risk factors and behaviors associated with poverty, results in higher body weight (BMI) and increased risk of chronic disease. The analysis also verifies the extent of these differences in both 2015 and 2035; the statistical significance in San Joaquin County by minority areas; and the predicted change that is attributable to Scenario 2. In almost every case, Scenario 2 will likely result in better health outcomes for both equity areas and the balance of the region with the exception of leisure walking for areas of concentrated poverty.

Analyzing health impacts at the CBG level facilitates mapping the predicted levels of utilitarian and leisure active travel, as well as BMI, which can help decision-makers determine where to spatially target interventions and investments. RTPs generally do not allocate infrastructure dollars at the project level. Discretion by the MPO could target leisure walking interventions in minority areas to assure that leisure walking does not decline through Scenario 2’s implementation. This could include creating additional safe and inviting public spaces, improving local parks to have more appealing walking paths, and adding culturally relevant programming. Since leisure physical activity domains are more closely associated with mental health than non-leisure physical activity (White et al., 2017), such a strategy might help address stress typically experienced by minorities due to systemic racism and thus help restore mental health. While equity areas already have higher rates of utilitarian walking, promoting even more utilitarian walking would also be beneficial to improve health outcomes. The MPO could promote municipal level incentives in equity areas to foster retail destinations through tax abatement, and target transit improvements to foster walking and reduce car dependence. Several studies show significant increases in walking and reduced BMI in association with transit facilities (Besser and Dannenberg, 2005; Brown et al., 2015; Lachapelle, 2016; MacDonald et al., 2010).

The analysis also highlights methodological implications for health, equity and EJ analyses within long-range transportation planning. This health modeling explicitly incorporates projected changes in land use while controlling for changes in demographics. This analysis is subject to limitations of predicting future health outcomes based on assumptions about long-range changes in built environment and future demographic patterns. RTP models typically rely heavily on housing types as signals of shifting household income and do not address race or ethnicity; this is also true in this analysis. Yet it is important to recognize that the modeling underpinning the NPHAM demonstrates that even though demographic changes will drive much of the shifts in healthy walking behaviors and chronic disease risk factors, built environment considerations are still crucially important.

Another challenge is that this health modeling uses built environment variables such as population density and intersection density similar to walkability indices. Notably, it does not include micro-scale measurements such as sidewalk presence or conditions known to make walking safe, viable, and enjoyable. These modeling shortcomings are discussed at length in Schoner et al. (2018); however, the benefits of explicitly incorporating health into MPO decision-making should hopefully outweigh reservations about health modeling assumptions and limitations. All else being equal, greater harm to the most under-served is likely to occur from the absence of adding health modeling to RTP’s.

In terms of equity and EJ analyses, SJCOG, like most other MPOs in the U.S., holds the relative spatial distribution of minority populations constant. Poverty is only slightly changed via predicted household income changes – based primarily on housing-type mixes within a block group – into the horizon year. However, models currently in use do not forecast how traditionally disadvantaged populations move around the region over time. The differences between how minority and poverty areas perform relative to other parts of the region demonstrate the importance of resolving this seemingly intractable problem. Failure to forecast where future equity groups will most likely be located in the horizon year constrains the planning organization’s capacity to influence the equitable distribution of a plan’s benefits. This is a particularly vexing problem for health analysis because demographic factors such as age, race, household structure, and educational attainment are seldom forecasted within the RTP context. Yet, collectively, these SES factors can have a more significant influence on risk and distribution of disease than changes to the built environment. Assuming age and race remain unchanged over time and allowing income to vary according to housing type does somewhat isolate the effect of transportation investments and related land use patterns from SES factors for future scenarios. Further, it allows this to happen in a manner that is consistent with other RTP analyses. That said, more research is needed to predict where equity populations will be located in the future.

There are additional definitional challenges most RTP equity analyses fail to address. Equity areas are typically defined by using a single absolute or relative threshold to label a specific geography as an area of concern (e.g., geographies where 50% of residents are people of color). This method is computationally simple, but it is vulnerable to the arbitrariness of threshold choice. The results of this analysis illustrate that these definitional parameters matter: the broader ES equity definition performed differently and typically better than the more restrictive definitions of minority or poverty equity areas, at times even producing effects in the opposite direction. Indeed, the analysis of difference of Baseline versus Scenario 2 suggests that “moderate equity” areas are faring better than the lowest SES areas – a result that was not explicitly sought by the MPO. The ad hoc sensitivity analysis presented in this study helped this finding surface. This implies that RTPs should consider incorporating detailed sensitivity analyses to more accurately reflect the intra-regional heterogeneity of demographic factors used to define vulnerability.

The ability of NPHAM to estimate physical activity related health behaviors at small spatial scales also supports a far more in-depth and nuanced comparison across equity definitions, and discussion about trade-offs within an RTP process. This exercise shows that modeling public health indicators in an RTP context can support nuanced policy conversations about equity in transportation planning by both (1) broadening the range of indicators typically considered to include health-related metrics such as active travel and BMI, and (2) indicating spatially where these changes are most likely to occur and with what magnitude. An updated version of NPHAM has been developed that also estimates diabetes, cardiovascular disease, and hypertension, and will further expand the scholarship in this field. With this information, policy makers can better assess health equity of baseline conditions and impacts of contrasting scenarios, incorporating traditionally externalized costs and benefits of health into the decision making process. This will enhance the ability to prioritize investments in target areas with the greatest health disparities; estimate potential health improvements for equity populations under the future scenario; and assess whether health improvements for equity areas are as great as or greater than those for non-equity areas.

Acknowledgements

Funding for the project on which this manuscript is based was provided by a Regional Transportation Plan/Sustainable Communities Strategy Tech Services agreement executed by the San Joaquin Council of Governments with Urban Design 4 Health, Inc. The authors would like to acknowledge Kim Anderson and Christine Corrales from the San Joaquin Council of Governments and Jim Chapman of Urban Design 4 Health, Inc. for their support on this project. Development of National Public Health Assessment Tool (NPHAM) was funded by the US Environmental Protection Agency Office of Sustainable Communities (Task Order 050: Contract EP-W-11-011). The content is solely the responsibility of the authors and does not necessarily represent the official view of the San Joaquin Council of Governments or the US Environmental Protection Agency. The authors would also like to thank the Robert Wood Johnson Foundation for funding the compilation of the National Environmental Database (NED) (Grant ID 72859) used in this project.

References

  1. Adkins A, Makarewicz C, Scanze M, Ingram M, Luhr G, 2017. Contextualizing walkability: do relationships between built environments and walking vary by socioeconomic context? J. Am. Plann. Assoc 83, 296–314. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Amekudzi AA, Smith MK, Brodie SR, Fischer JM, Ross CL, 2012. Impact of environmental justice on transportation. Transportation research record. J. Transport Res. Board 2320, 1–9. [Google Scholar]
  3. Aune D, Norat T, Leitzmann M, Tonstad S, Vatten LJ, 2015. Physical activity and the risk of type 2 diabetes: a systematic review and dose-response meta-analysis. Eur. J. Epidemiol 30, 529–542. [DOI] [PubMed] [Google Scholar]
  4. Bambra C, Gibson M, Sowden A, Wright K, Whitehead M, Petticrew M, 2010. Tackling the wider social determinants of health and health inequalities: evidence from systematic reviews. J. Epidemiol. Community Health 64, 284–291. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Besser LM, Dannenberg AL, 2005. Walking to public transit: steps to help meet physical activity recommendations. Am. J. Prev. Med 29, 273–280. [DOI] [PubMed] [Google Scholar]
  6. Bills TS, Sall EA, Walker JL, 2012. Activity-based travel models and transportation equity analysis. Transport. Res. Rec.: J. Transport Res. Board 2320, 18–27. [Google Scholar]
  7. Brodie S, Amekudzi-Kennedy A, 2017. Performance-based methodology for evaluating equity for transportation System users. Transport. Res. Rec.: J. Transport Res. Board 2605, 83–91. [Google Scholar]
  8. Brown BB, Werner CM, Tribby CP, Miller HJ, Smith KR, 2015. Transit use, physical activity, and body mass index changes: objective measures associated with complete street light-rail construction. Am. J. Publ. Health 105, 1468–1474. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. California Governor’s Office of Planning and Research, 2017. Chapter 4: rquired elements - designing healthy, equitable, reslient, and economically vibrant places. In: General Plan Guidelines: 2017 Update. Governor’s Office of Planning & Research, Sacramento, California. [Google Scholar]
  10. Carlsson AC, Arnlov J, Sundstrom J, Michaelsson K, Byberg L, Lind L, 2016. Physical activity, obesity and risk of cardiovascular disease in middle-aged men during a median of 30 years of follow-up. Eur. J. Prev. Cardiol 23, 359–365. [DOI] [PubMed] [Google Scholar]
  11. Casagrande SS, Whitt-Glover MC, Lancaster KJ, Odoms-Young AM, Gary TL, 2009. Built environment and health behaviors among African Americans: a systematic review. Am. J. Prev. Med 36, 174–181. [DOI] [PubMed] [Google Scholar]
  12. Creatore MI, Glazier R, Moineddin R, Fazli GS, Gozdyra P, Matheson F, et al. , 2014. The impact of neighbourhood walkability on the rise in overweight, obesity, and diabetes in Ontario, Canada, 2001–2010. Diabetes 63 A21–A21. [Google Scholar]
  13. den Braver NR, Lakerveld J, Rutters F, Schoonmade LJ, Brug J, Beulens JWJ, 2018. Built environmental characteristics and diabetes: a systematic review and meta-analysis. BMC Med. 16, 12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Durand CP, Andalib M, Dunton GF, Wolch J, Pentz MA, 2011. A systematic review of built environment factors related to physical activity and obesity risk: implications for smart growth urban planning. Obes. Rev 12, e173–182. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Ewing R, Cervero R, 2010. Travel and the built environment: a meta-analysis. J. Am. Plann. Assoc 76, 265–294. [Google Scholar]
  16. Frank LD, Andresen MA, Schmid TL, 2004. Obesity relationships with community design, physical activity, and time spent in cars. Am. J. Prev. Med 27, 87–96. [DOI] [PubMed] [Google Scholar]
  17. Frank LD, Sallis JF, Conway TL, Chapman JE, Saelens BE, Bachman W, 2006. Many pathways from land use to health: associations between neighborhood walkability and active transportation, body mass index, and air quality. J. Am. Plann. Assoc 72, 75–87. [Google Scholar]
  18. Gaskin DJ, Thorpe RJ Jr., McGinty EE, Bower K, Rohde C, Young JH, et al. , 2014. Disparities in diabetes: the nexus of race, poverty, and place. Am. J. Publ. Health 104, 2147–2155. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Golub A, Martens K, 2014. Using principles of justice to assess the modal equity of regional transportation plans. J. Transport Geogr 41, 10–20. [Google Scholar]
  20. Grasser G, Van Dyck D, Titze S, Stronegger W, 2013. Objectively measured walkability and active transport and weight-related outcomes in adults: a systematic review. Int. J. Publ. Health 58, 615–625. [DOI] [PubMed] [Google Scholar]
  21. Highway Administration, Federal, 2011. Environmental justice emerging trends and best practices guidebook. In: International, I. (Ed.), Department of Transportation, Federal Highway Administration, Washington DC. [Google Scholar]
  22. Jeon CY, Lokken RP, Hu FB, van Dam RM, 2007. Physical activity of moderate intensity and risk of type 2 diabetes: a systematic review. Diabetes Care 30, 744–752. [DOI] [PubMed] [Google Scholar]
  23. Karner A, Niemeier D, 2013. Civil rights guidance and equity analysis methods for regional transportation plans: a critical review of literature and practice. J. Transport Geogr 33, 126–134. [Google Scholar]
  24. Klein N, 2007. Spatial methodology for assessing distribution of transportation project impacts with environmental justice framework. Transport. Res. Rec.: J. Transport Res. Board 46–53 2013. [Google Scholar]
  25. Lachapelle U, 2016. Examining the relationship between transit use and active transportation in Canada: a time use approach of the active lifestyle of transit users. In: Transportation Research Board 95th Annual Meeting. Transportation Research Board 95th Annual Meeting, Washington, DC. [Google Scholar]
  26. Laine J, Kuvaja-Kollner V, Pietila E, Koivuneva M, Valtonen H, Kankaanpaa E, 2014. Cost-effectiveness of population-level physical activity interventions: a systematic review. Am. J. Health Promot 29, 71–80. [DOI] [PubMed] [Google Scholar]
  27. Li T, Wei S, Shi Y, Pang S, Qin Q, Yin J, et al. , 2016. The dose-response effect of physical activity on cancer mortality: findings from 71 prospective cohort studies. Br. J. Sports Med 50, 339–345. [DOI] [PubMed] [Google Scholar]
  28. Lovasi GS, Hutson MA, Guerra M, Neckerman KM, 2009. Built environments and obesity in disadvantaged populations. Epidemiol. Rev 31, 7–20. [DOI] [PubMed] [Google Scholar]
  29. MacDonald JM, Stokes RJ, Cohen DA, Kofner A, Ridgeway GK, 2010. The effect of light rail transit on body mass index and physical activity. Am. J. Prev. Med 39, 105–112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Mahmood S, MacInnis RJ, English DR, Karahalios A, Lynch BM, 2017. Domain-specific physical activity and sedentary behaviour in relation to colon and rectal cancer risk: a systematic review and meta-analysis. Int. J. Epidemiol 46, 1797–1813. [DOI] [PubMed] [Google Scholar]
  31. Malambo P, Kengne AP, De Villiers A, Lambert EV, Puoane T, 2016. Built environment, selected risk factors and major cardiovascular disease outcomes: a systematic review. PloS One 11, e0166846. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Manaugh K, El-Geneidy AM, 2012. Who benefits from new transportation infrastructure? Using accessibility measures to evaluate social equity in transit provision. In: Geurs K, Krizek K, Reggiani A (Eds.), Accessibility and Transport Planning: Challenges for Europe and North America. Edward Elgar, London, UK, pp. 211–227. [Google Scholar]
  33. Manaugh K, Badami MG, El-Geneidy AM, 2015. Integrating social equity into urban transportation planning: a critical evaluation of equity objectives and measures in transportation plans in North America. Transport Pol. 37, 167–176. [Google Scholar]
  34. Martens K, Golub A, Robinson G, 2012. A justice-theoretic approach to the distribution of transportation benefits: implications for transportation planning practice in the United States. Transport. Res. Pol. Pract 46, 684–695. [Google Scholar]
  35. Mayne SL, Auchincloss AH, Michael YL, 2015. Impact of policy and built environment changes on obesity-related outcomes: a systematic review of naturally occurring experiments. Obes. Rev 16, 362–375. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. McCormack GR, Shiell A, 2011. In search of causality: a systematic review of the relationship between the built environment and physical activity among adults. Int. J. Behav. Nutr. Phys. Activ 8, 125. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. McGrath LJ, Hopkins WG, Hinckson EA, 2015. Associations of objectively measured built-environment attributes with youth moderate-vigorous physical activity: a systematic review and meta-analysis. Sports Med. 45, 841–865. [DOI] [PubMed] [Google Scholar]
  38. Moore SC, Lee IM, Weiderpass E, Campbell PT, Sampson JN, Kitahara CM,et al. , 2016. Association of leisure-time physical activity with risk of 26 types of cancer in 1.44 million adults. JAMA Int. Med 176, 816–825. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Murtagh EM, Nichols L, Mohammed MA, Holder R, Nevill AM, Murphy MH, 2015. The effect of walking on risk factors for cardiovascular disease: an updated systematic review and meta-analysis of randomised control trials. Prev. Med 72, 34–43. [DOI] [PubMed] [Google Scholar]
  40. SB-1000, 2016. Land Use: General Plans: Safety and Environmental Justice. (California: ). [Google Scholar]
  41. Schoner J, Chapman J, Brookes A, MacLeod KE, Fox EH, Iroz-Elardo N, Frank LD, 2018. September 1. Bringing health into transportation and land use scenario planning: creating a National Public Health Assessment Model (N-PHAM). J. Transport Health 10, 401–418. 10.1016/j.jth.2018.04.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Singleton PA, Clifton KJ, 2014. Incorporating public health in US long-range metropolitan transportation planning: a review of guidance statements and performance measures. In: TRB 94th Annual Meeting Compendium of Papers. [Google Scholar]
  43. Singleton PA, Clifton KJ, 2017. Considering health in US metropolitan long-range transportation plans: a review of guidance statements and performance measures. Transport Pol. 57, 79–89. [Google Scholar]
  44. Smith M, Hosking J, Woodward A, Witten K, MacMillan A, Field A, et al. , 2017. Systematic literature review of built environment effects on physical activity and active transport - an update and new findings on health equity. Int. J. Behav. Nutr. Phys. Activ 14, 158. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Wanner M, Gotschi T, Martin-Diener E, Kahlmeier S, Martin BW, 2012. Active transport, physical activity, and body weight in adults: a systematic review. Am. J. Prev. Med 42, 493–502. [DOI] [PubMed] [Google Scholar]
  46. Wen, Kowaleski-Jones, 2012. The built environment and risk of obesity in the United States: racial–ethnic disparities. Health Place 18. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. White RL, Babic MJ, Parker PD, Lubans DR, Astell-Burt T, Lonsdale C, 2017. Domain-specific physical activity and mental health: a meta-analysis. Am. J. Prev. Med 52, 653–666. [DOI] [PubMed] [Google Scholar]
  48. Whitfield GP, Meehan LA, Maizlish N, Wendel AM, 2017. The integrated transport and health impact modeling tool in Nashville, Tennessee, USA: implementation steps and lessons learned. J. Transp. Health 5, 172–181. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Williams PT, 2013. Dose-response relationship of physical activity to premature and total all-cause and cardiovascular disease mortality in walkers. PloS One 8, e78777. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. World Health Organization, 2018. Examples of Applications of the Health Economic Assessment Tool (HEAT) for Cycling. World Health Organization, Regional Office for Europe, Copenhagen. [Google Scholar]

RESOURCES