Modeling health and well-being measures using ZIP code spatial neighborhood patterns.

Individual-level assessment of health and well-being permits analysis of community well-being and health risk evaluations across several dimensions of health. It also enables comparison and rankings of reported health and well-being for large geographical areas such as states, metropolitan areas, and counties. However, there is large variation in reported well-being within such large spatial units underscoring the importance of analyzing well-being at more granular levels, such as ZIP codes. In this paper, we address this problem by modeling well-being data to generate ZIP code tabulation area (ZCTA)-level rankings through spatially informed statistical modeling. We build regression models for individual-level overall well-being index and scores from five subscales (Physical, Financial, Social, Community, Purpose) using individual-level demographic characteristics as predictors while including a ZCTA-level spatial effect. The ZCTA neighborhood information is incorporated by using a graph Laplacian matrix; this enables estimation of the effect of a ZCTA on well-being using individual-level data from that ZCTA as well as by borrowing information from neighboring ZCTAs. We deploy our model on well-being data for the U.S. states of Massachusetts and Georgia. We find that our model can capture the effects of demographic features while also offering spatial effect estimates for all ZCTAs, including ones with no observations, under certain conditions. These spatial effect estimates provide community health and well-being rankings of ZCTAs, and our method can be deployed more generally to model other outcomes that are spatially dependent as well as data from other states or groups of states.

Nationwide estimation of daily ambient PM2.5 from 2008 to 2020 at 1 km2 in India using an ensemble approach.

High-resolution assessment of historical levels is essential for assessing the health effects of ambient air pollution in the large Indian population. The diversity of geography, weather patterns, and progressive urbanization, combined with a sparse ground monitoring network makes it challenging to accurately capture the spatiotemporal patterns of ambient fine particulate matter (PM2.5) pollution in India. We developed a model for daily average ambient PM2.5 between 2008 and 2020 based on monitoring data, meteorology, land use, satellite observations, and emissions inventories. Daily average predictions at each 1 km × 1 km grid from each learner were ensembled using a Gaussian process regression with anisotropic smoothing over spatial coordinates, and regression calibration was used to account for exposure error. Cross-validating by leaving monitors out, the ensemble model had an R2 of 0.86 at the daily level in the validation data and outperformed each component learner (by 5-18%). Annual average levels in different zones ranged between 39.7 µg/m3 (interquartile range: 29.8-46.8) in 2008 and 30.4 µg/m3 (interquartile range: 22.7-37.2) in 2020, with a cross-validated (CV)-R2 of 0.94 at the annual level. Overall mean absolute daily errors (MAE) across the 13 years were between 14.4 and 25.4 µg/m3. We obtained high spatial accuracy with spatial R2 greater than 90% and spatial MAE ranging between 7.3-16.5 µg/m3 with relatively better performance in urban areas at low and moderate elevation. We have developed an important validated resource for studying PM2.5 at a very fine spatiotemporal resolution, which allows us to study the health effects of PM2.5 across India and to identify areas with exceedingly high levels.

The Exposure Peaks of Traffic-Related Ultrafine Particles Associated with Inflammatory Biomarkers and Blood Lipid Profiles.

In this article, we explored the effects of ultrafine particle (UFP) peak exposure on inflammatory biomarkers and blood lipids using two novel metrics-the intensity of peaks and the frequency of peaks. We used data previously collected by the Community Assessment of Freeway Exposure and Health project from participants in the Greater Boston Area. The UFP exposure data were time-activity-adjusted hourly average concentration, estimated using land use regression models based on mobile-monitored ambient concentrations. The outcome data included C-reactive protein, interleukin-6 (IL-6), tumor necrosis factor-alpha receptor 2 (TNF-RII), low-density lipoprotein (LDL), high-density lipoprotein (HDL), triglycerides and total cholesterol. For each health indicator, multivariate regression models were used to assess their associations with UFP peaks (N = 364-411). After adjusting for age, sex, body mass index, smoking status and education level, an increase in UFP peak exposure was significantly (p < 0.05) associated with an increase in TNF-RII and a decrease in HDL and triglycerides. Increases in UFP peaks were also significantly associated with increased IL-6 and decreased total cholesterol, while the same associations were not significant when annual average exposure was used. Our work suggests that analysis using peak exposure metrics could reveal more details about the effect of environmental exposures than the annual average metric.

Impact of heatwaves on all-cause mortality in India: A comprehensive multi-city study.

BACKGROUND: Heatwaves are expected to increase with climate change, posing a significant threat to population health. In India, with the world's largest population, heatwaves occur annually but have not been comprehensively studied. Accordingly, we evaluated the association between heatwaves and all-cause mortality and quantifying the attributable mortality fraction in India. METHODS: We obtained all-cause mortality counts for ten cities in India (2008-2019) and estimated daily mean temperatures from satellite data. Our main extreme heatwave was defined as two-consecutive days with an intensity above the 97th annual percentile. We estimated city-specific heatwave associations through generalised additive Poisson regression models, and meta-analysed the associations. We reported effects as the percentage change in daily mortality, with 95% confidence intervals (CI), comparing heatwave vs non-heatwave days. We further evaluated heatwaves using different percentiles (95th, 97th, 99th) for one, two, three and five-consecutive days. We also evaluated the influence of heatwave duration, intensity and timing in the summer season on heatwave mortality, and estimated the number of heatwave-related deaths. FINDINGS: Among âˆ¼ 3.6 million deaths, we observed that temperatures above 97th percentile for 2-consecutive days was associated with a 14.7 % (95 %CI, 10.3; 19.3) increase in daily mortality. Alternative heatwave definitions with higher percentiles and longer duration resulted in stronger relative risks. Furthermore, we observed stronger associations between heatwaves and mortality with higher heatwave intensity. We estimated that around 1116 deaths annually (95 %CI, 861; 1361) were attributed to heatwaves. Shorter and less intense definitions of heatwaves resulted in a higher estimated burden of heatwave-related deaths. CONCLUSIONS: We found strong evidence of heatwave impacts on daily mortality. Longer and more intense heatwaves were linked to an increased mortality risk, however, resulted in a lower burden of heatwave-related deaths. Both definitions and the burden associated with each heatwave definition should be incorporated into planning and decision-making processes for policymakers.

Climate stress and anxiety, environmental context, and civic engagement: A nationally representative study.

There is increasing recognition that people are experiencing stress and anxiety around climate change, and that this climate stress/anxiety may be associated with more pro-environmental behavior. However, less is known about whether people's own environmental exposures affect climate stress/anxiety or the relationship between climate stress/anxiety and civic engagement. Using three waves of survey data (2020-2022) from the nationally representative Tufts Equity in Health, Wealth, and Civic Engagement Study of US adults (n = 1071), we assessed relationships among environmental exposures (county-level air pollution, greenness, number of toxic release inventory sites, and heatwaves), self-reported climate stress/anxiety, and civic engagement measures (canvasing behavior, collaborating to solve community problems, personal efficacy to solve community problems, group efficacy to solve community problems, voting behavior). Most participants reported experiencing climate stress/anxiety (61%). In general, the environmental exposures we assessed were not significantly associated with climate stress/anxiety or civic engagement metrics, but climate stress/anxiety was positively associated with most of the civic engagement outcomes (canvassing, personal efficacy, group efficacy, voter preference). Our results support the growing literature that climate stress/anxiety may spur constructive civic action, though do not suggest a consistent relationship between adverse environmental exposures and either climate stress/anxiety or civic engagement. Future research and action addressing the climate crisis should promote climate justice by ensuring mental health support for those who experience climate stress anxiety and by promoting pro-environmental civic engagement efforts.

Evaluating COVID-19 Risk to Essential Workers by Occupational Group: A Case Study in Massachusetts.

Occupational exposure to SARS-CoV-2 varies by profession, but "essential workers" are often considered in aggregate in COVID-19 models. This aggregation complicates efforts to understand risks to specific types of workers or industries and target interventions, specifically towards non-healthcare workers. We used census tract-resolution American Community Survey data to develop novel essential worker categories among the occupations designated as COVID-19 Essential Services in Massachusetts. Census tract-resolution COVID-19 cases and deaths were provided by the Massachusetts Department of Public Health. We evaluated the association between essential worker categories and cases and deaths over two phases of the pandemic from March 2020 to February 2021 using adjusted mixed-effects negative binomial regression, controlling for other sociodemographic risk factors. We observed elevated COVID-19 case incidence in census tracts in the highest tertile of workers in construction/transportation/buildings maintenance (Phase 1: IRR 1.32 [95% CI 1.22, 1.42]; Phase 2: IRR: 1.19 [1.13, 1.25]), production (Phase 1: IRR: 1.23 [1.15, 1.33]; Phase 2: 1.18 [1.12, 1.24]), and public-facing sales and services occupations (Phase 1: IRR: 1.14 [1.07, 1.21]; Phase 2: IRR: 1.10 [1.06, 1.15]). We found reduced case incidence associated with greater percentage of essential workers able to work from home (Phase 1: IRR: 0.85 [0.78, 0.94]; Phase 2: IRR: 0.83 [0.77, 0.88]). Similar trends exist in the associations between essential worker categories and deaths, though attenuated. Estimating industry-specific risk for essential workers is important in targeting interventions for COVID-19 and other diseases and our categories provide a reproducible and straightforward way to support such efforts.

Calculating access to parks and other polygonal resources: A description of open-source methodologies.

Public health studies routinely use simplistic methods to calculate proximity-based "access" to greenspace, such as by measuring distances to the geographic centroids of parks or, less frequently, to the perimeter of the park area. Although computationally efficient, these approaches oversimplify exposure measurement because parks often have specific entrance points. In this tutorial paper, we describe how researchers can instead calculate more-accurate access measures using freely available open-source methods. Specifically, we demonstrate processes for calculating "service areas" representing street-network-based buffers of access to parks within set distances and mode of transportation (e.g., 1-km walk or 20-minute drive) using OpenRouteService and QGIS software. We also introduce an advanced method involving the identification of trailheads or parking lots with OpenStreetMap data and show how large parks particularly benefit from this approach. These methods can be used globally and are applicable to analyses of a wide range of studies investigating proximity access to resources.

Characterizing temporal trends in populations exposed to aircraft noise around U.S. airports: 1995-2015.

BACKGROUND: Aircraft noise is a key concern for communities surrounding airports, with increasing evidence for health effects and inequitable distributions of exposure. However, there have been limited national-scale assessments of aircraft noise exposure over time and across noise metrics, limiting evaluation of population exposure patterns. OBJECTIVE: We evaluated national-scale temporal trends in aviation noise exposure by airport characteristics and across racial/ethnic populations in the U.S. METHODS: Noise contours were modeled for 90 U.S. airports in 5-year intervals between 1995 and 2015 using the Federal Aviation Administration's Aviation Environmental Design Tool. We utilized linear fixed effects models to estimate changes in noise exposure areas for day-night average sound levels (DNL) of 45, 65, and a nighttime equivalent sound level (Lnight) of 45 A-weighted decibels (dB[A]). We used group-based trajectory modeling to identify distinct groups of airports sharing underlying characteristics. We overlaid noise contours and Census tract data from the U.S. Census Bureau and American Community Surveys for 2000 to 2015 to estimate exposure changes overall and by race/ethnicity. RESULTS: National-scale analyses showed non-monotonic trends in mean exposed areas that peaked in 2000, followed by a 37% decrease from 2005 to 2010 and a subsequent increase in 2015. We identified four distinct trajectory groups of airports sharing latent characteristics related to size and activity patterns. Those populations identifying as minority (e.g., Hispanic/Latino, Black/African American, Asian) experienced higher proportions of exposure relative to their subgroup populations compared to non-Hispanic or White populations across all years, indicating ethnic and racial disparities in airport noise exposure that persist over time. SIGNIFICANCE: Overall, these data identified differential exposure trends across airports and subpopulations, helping to identify vulnerable communities for aviation noise in the U.S. IMPACT STATEMENT: We conducted a descriptive analysis of temporal trends in aviation noise exposure in the U.S. at a national level. Using data from 90 U.S. airports over a span of two decades, we characterized the noise exposure trends overall and by airport characteristics, while estimating the numbers of exposed by population demographics to help identify the impact on vulnerable communities who may bear the burden of aircraft noise exposure.

School Greenness and Student-Level Academic Performance: Evidence From the Global South.

Greenspace in schools might enhance students' academic performance. However, the literature-dominated by ecological studies at the school level in countries from the Northern Hemisphere-presents mixed evidence of a beneficial association. We evaluated the association between school greenness and student-level academic performance in Santiago, Chile, a capital city of the Global South. This cross-sectional study included 281,695 fourth-grade students attending 1,498 public, charter, and private schools in Santiago city between 2014 and 2018. Student-level academic performance was assessed using standardized test scores and indicators of attainment of learning standards in mathematics and reading. School greenness was estimated using Normalized Difference Vegetation Index (NDVI). Linear and generalized linear mixed-effects models were fit to evaluate associations, adjusting for individual- and school-level sociodemographic factors. Analyses were stratified by school type. In fully adjusted models, a 0.1 increase in school greenness was associated with higher test scores in mathematics (36.9 points, 95% CI: 2.49; 4.88) and in reading (1.84 points, 95% CI: 0.73; 2.95); as well as with higher odds of attaining learning standards in mathematics (OR: 1.20, 95% CI: 1.12; 1.28) and reading (OR: 1.07, 95% CI: 1.02; 1.13). Stratified analysis showed differences by school type, with associations of greater magnitude and strength for students attending public schools. No significant associations were detected for students in private schools. Higher school greenness was associated with improved individual-level academic outcomes among elementary-aged students in a capital city in South America. Our results highlight the potential of greenness in the school environment to moderate educational and environmental inequalities in urban areas.

Environmental equity and COVID-19 experiences in the United States: Results from three survey waves of a nationally representative study conducted between 2020-2022.

Certain environmental exposures, such as air pollution, are associated with COVID-19 incidence and mortality. To determine whether environmental context is associated with other COVID-19 experiences, we used data from the nationally representative Tufts Equity in Health, Wealth, and Civic Engagement Study data (n=1785; three survey waves 2020-2022). Environmental context was assessed using self-reported climate stress and county-level air pollution, greenness, toxic release inventory site, and heatwave data. Self-reported COVID-19 experiences included willingness to vaccinate against COVID-19, health impacts from COVID-19, receiving assistance for COVID-19, and provisioning assistance for COVID-19. Self-reported climate stress in 2020 or 2021 was associated with increased COVID-19 vaccination willingness by 2022 (odds ratio [OR] = 2.35; 95% confidence interval [CI] = 1.47, 3.76), even after adjusting for political affiliation (OR = 1.79; 95% CI = 1.09, 2.93). Self-reported climate stress in 2020 was also associated with increased likelihood of receiving COVID-19 assistance by 2021 (OR = 1.89; 95% CI = 1.29, 2.78). County-level exposures (i.e., less greenness, more toxic release inventory sites, more heatwaves) were associated with increased vaccination willingness. Air pollution exposure in 2020 was positively associated with likelihood of provisioning COVID-19 assistance in 2020 (OR = 1.16 per µg/m3; 95% CI = 1.02, 1.32). Associations between certain environmental exposures and certain COVID-19 outcomes were stronger among those who identify as a race/ethnicity other than non-Hispanic White and among those who reported experiencing discrimination; however, these trends were not consistent. A latent variable representing a summary construct for environmental context was associated with COVID-19 vaccination willingness. Our results add to the growing body of literature suggesting that intersectional equity issues affecting likelihood of exposure to adverse environmental conditions are also associated with health-related outcomes.

Assessing the impact of aircraft arrival on ambient ultrafine particle number concentrations in near-airport communities in Boston, Massachusetts.

Aircraft emissions contribute to overall ambient air pollution, including ultrafine particle (UFP) concentrations. However, accurately ascertaining aviation contributions to UFP is challenging due to high spatiotemporal variability along with intermittent aviation emissions. The objective of this study was to evaluate the impact of arrival aircraft on particle number concentration (PNC), a proxy for UFP, across six study sites 3-17 km from a major arrival aircraft flight path into Boston Logan International Airport by utilizing real-time aircraft activity and meteorological data. Ambient PNC at all monitoring sites was similar at the median but had greater variation at the 95th and 99th percentiles with more than two-fold increases in PNC observed at sites closer to the airport. PNC was elevated during the hours with high aircraft activity with sites closest to the airport exhibiting stronger signals when downwind from the airport. Regression models indicated that the number of arrival aircraft per hour was associated with measured PNC at all six sites, with a maximum contribution of 50% of total PNC at a monitor 3 km from the airport during hours with arrival activity on the flight path of interest (26% across all hours). Our findings suggest strong but intermittent contributions from arrival aircraft to ambient PNC in communities near airports.

Examining the Optimal Placement of Cooling Centers to Serve Populations at High Risk of Extreme Heat Exposure in 81 US Cities.

OBJECTIVE: Although extreme heat can impact the health of anyone, certain groups are disproportionately affected. In urban settings, cooling centers are intended to reduce heat exposure by providing air-conditioned spaces to the public. We examined the characteristics of populations living near cooling centers and how well they serve areas with high social vulnerability. METHODS: We identified 1402 cooling centers in 81 US cities from publicly available sources and analyzed markers of urban heat and social vulnerability in relation to their locations. Within each city, we developed cooling center access areas, defined as the geographic area within a 0.5-mile walk from a center, and compared sociodemographic characteristics of populations living within versus outside the access areas. We analyzed results by city and geographic region to evaluate climate-relevant regional differences. RESULTS: Access to cooling centers differed among cities, ranging from 0.01% (Atlanta, Georgia) to 63.2% (Washington, DC) of the population living within an access area. On average, cooling centers were in areas that had higher levels of social vulnerability, as measured by the number of people living in urban heat islands, annual household income below poverty, racial and ethnic minority status, low educational attainment, and high unemployment rate. However, access areas were less inclusive of adult populations aged ≥65 years than among populations aged <65 years. CONCLUSION: Given the large percentage of individuals without access to cooling centers and the anticipated increase in frequency and severity of extreme heat events, the current distribution of centers in the urban areas that we examined may be insufficient to protect individuals from the adverse health effects of extreme heat, particularly in the absence of additional measures to reduce risk.

Influence of geospatial resolution on sociodemographic predictors of COVID-19 in Massachusetts.

PURPOSE: When studying health risks across a large geographic region such as a state or province, researchers often assume that finer-resolution data on health outcomes and risk factors will improve inferences by avoiding ecological bias and other issues associated with geographic aggregation. However, coarser-resolution data (e.g., at the town or county-level) are more commonly publicly available and packaged for easier access, allowing for rapid analyses. The advantages and limitations of using finer-resolution data, which may improve precision at the cost of time spent gaining access and processing data, have not been considered in detail to date. METHODS: We systematically examine the implications of conducting town-level mixed-effect regression analyses versus census-tract-level analyses to study sociodemographic predictors of COVID-19 in Massachusetts. In a series of negative binomial regressions, we vary the spatial resolution of the outcome, the resolution of variable selection, and the resolution of the random effect to allow for more direct comparison across models. RESULTS: We find stability in some estimates across scenarios, changes in magnitude, direction, and significance in others, and tighter confidence intervals on the census-tract level. Conclusions regarding sociodemographic predictors are robust when regions of high concentration remain consistent across town and census-tract resolutions. CONCLUSIONS: Inferences about high-risk populations may be misleading if derived from town- or county-resolution data, especially for covariates that capture small subgroups (e.g., small racial minority populations) or are geographically concentrated or skewed (e.g., % college students). Our analysis can help inform more rapid and efficient use of public health data by identifying when finer-resolution data are truly most informative, or when coarser-resolution data may be misleading.

Inequities in COVID-19 vaccine and booster coverage across Massachusetts ZIP codes after the emergence of Omicron: A population-based cross-sectional study.

BACKGROUND: Inequities in Coronavirus Disease 2019 (COVID-19) vaccine and booster coverage may contribute to future disparities in morbidity and mortality within and between Massachusetts (MA) communities. METHODS AND FINDINGS: We conducted a population-based cross-sectional study of primary series vaccination and booster coverage 18 months into the general population vaccine rollout. We obtained public-use data on residents vaccinated and boosted by ZIP code (and by age group: 5 to 19, 20 to 39, 40 to 64, 65+) from MA Department of Public Health, as of October 10, 2022. We constructed population denominators for postal ZIP codes by aggregating census tract population estimates from the 2015-2019 American Community Survey. We excluded nonresidential ZIP codes and the smallest ZIP codes containing 1% of the state's population. We mapped variation in ZIP code-level primary series vaccine and booster coverage and used regression models to evaluate the association of these measures with ZIP code-level socioeconomic and demographic characteristics. Because age is strongly associated with COVID-19 severity and vaccine access/uptake, we assessed whether observed socioeconomic and racial/ethnic inequities persisted after adjusting for age composition and plotted age-specific vaccine and booster coverage by deciles of ZIP code characteristics. We analyzed data on 418 ZIP codes. We observed wide geographic variation in primary series vaccination and booster rates, with marked inequities by ZIP code-level education, median household income, essential worker share, and racial/ethnic composition. In age-stratified analyses, primary series vaccine coverage was very high among the elderly. However, we found large inequities in vaccination rates among younger adults and children, and very large inequities in booster rates for all age groups. In multivariable regression models, each 10 percentage point increase in "percent college educated" was associated with a 5.1 (95% confidence interval (CI) 3.9 to 6.3, p < 0.001) percentage point increase in primary series vaccine coverage and a 5.4 (95% CI 4.5 to 6.4, p < 0.001) percentage point increase in booster coverage. Although ZIP codes with higher "percent Black/Latino/Indigenous" and higher "percent essential workers" had lower vaccine coverage (-0.8, 95% CI -1.3 to -0.3, p < 0.01; -5.5, 95% CI -7.3 to -3.8, p < 0.001), these associations became strongly positive after adjusting for age and education (1.9, 95% CI 1.0 to 2.8, p < 0.001; 4.8, 95% CI 2.6 to 7.1, p < 0.001), consistent with high demand for vaccines among Black/Latino/Indigenous and essential worker populations within age and education groups. Strong positive associations between "median household income" and vaccination were attenuated after adjusting for age. Limitations of the study include imprecision of the estimated population denominators, lack of individual-level sociodemographic data, and potential for residential ZIP code misreporting in vaccination data. CONCLUSIONS: Eighteen months into MA's general population vaccine rollout, there remained large inequities in COVID-19 primary series vaccine and booster coverage across MA ZIP codes, particularly among younger age groups. Disparities in vaccination coverage by racial/ethnic composition were statistically explained by differences in age and education levels, which may mediate the effects of structural racism on vaccine uptake. Efforts to increase booster coverage are needed to limit future socioeconomic and racial/ethnic disparities in COVID-19 morbidity and mortality.

Relationship between traffic-related air pollution and inflammation biomarkers using structural equation modeling.

BACKGROUND: Evidence suggests that exposure to traffic-related air pollution (TRAP) and social stressors can increase inflammation. Given that there are many different markers of TRAP exposure, socio-economic status (SES), and inflammation, analytical approaches can leverage multiple markers to better elucidate associations. In this study, we applied structural equation modeling (SEM) to assess the association between a TRAP construct and a SES construct with an inflammation construct. METHODS: This analysis was conducted as part of the Community Assessment of Freeway Exposure and Health (CAFEH; N = 408) study. Air pollution was characterized using a spatiotemporal model of particle number concentration (PNC) combined with individual participant time-activity adjustment (TAA). TAA-PNC and proximity to highways were considered for a construct of TRAP exposure. Participant demographics on education and income for an SES construct were assessed via questionnaires. Blood samples were analyzed for high sensitivity C-reactive protein (hsCRP), interleukin-6 (IL-6), and tumor necrosis factor-α receptor II (TNFRII), which were considered for the construct for inflammation. We conducted SEM and compared our findings with those obtained using generalized linear models (GLM). RESULTS: Using GLM, TAA-PNC was associated with multiple inflammation biomarkers. An IQR (10,000 particles/cm3) increase of TAA-PNC was associated with a 14 % increase in hsCRP in the GLM. Using SEM, the association between the TRAP construct and the inflammation construct was twice as large as the associations with any individual inflammation biomarker. SES had an inverse association with inflammation in all models. Using SEM to estimate the indirect effects of SES on inflammation through the TRAP construct strengthened confidence in the association of TRAP with inflammation. CONCLUSION: Our TRAP construct resulted in stronger associations with a combined construct for inflammation than with individual biomarkers, reinforcing the value of statistical approaches that combine multiple, related exposures or outcomes. Our findings are consistent with inflammatory risk from TRAP exposure.

Impacts of the choice of distance measurement method on estimates of access to point-based resources.

BACKGROUND/OBJECTIVE: Lack of access to resources such as medical facilities and grocery stores is related to poor health outcomes and inequities, particularly in an environmental justice framework. There can be substantial differences in quantifying "access" to such resources, depending on the geospatial method used to generate distance estimates. METHODS: We compared three methods for calculating distance to the nearest grocery store to illustrate differential access at the census block-group level in the Atlanta metropolitan area, including: Euclidean distance estimation, service areas incorporating roadways and other factors, and cost distance for every point on the map. RESULTS: We found notable differences in access across the three estimation techniques, implying a high potential for exposure misclassification by estimation method. There was a lack of nuanced exposure in the highest- and lowest-access areas using the Euclidean distance method. We found an Intraclass Correlation Coefficient (ICC) of 0.69 (0.65, 0.73), indicating moderate agreement between estimation methods. SIGNIFICANCE: As compared with Euclidean distance, service areas and cost distance may represent a more meaningful characterization of "access" to resources. Each method has tradeoffs in computational resources required versus potential improvement in exposure classification. Careful consideration of the method used for determining "access" will reduce subsequent misclassifications.

Critical windows of susceptibility for the effects of prenatal exposure to heat and heat variability on gestational growth.

BACKGROUND: Studies have shown that prenatal heat exposure may impact fetal growth, but few studies have examined the critical windows of susceptibility. As extreme heat events and within season temperature variability is expected to increase in frequency, it is important to understand how this may impact gestational growth. OBJECTIVES: We investigated associations between various measures of weekly prenatal heat exposure (mean and standard deviation (SD) of temperature and heat index (HI), derived using temperature in °C and dew point) and term birthweight or odds of being born small for gestational age (SGA) to identify critical windows of susceptibility. METHODS: We analyzed data from mother-child dyads (n = 4442) in the Boston-based Children's HealthWatch cohort. Birthweights were collected from survey data and electronic health records. Daily temperature and HI values were obtained from 800 m gridded spatial climate datasets aggregated by the PRISM Climate Group. Distributed lag-nonlinear models were used to assess the effect of the four weekly heat metrics on measures of gestational growth (birthweight, SGA, and birthweight z-scores). Analyses were stratified by child sex and maternal homelessness status during pregnancy. RESULTS: HI variability was significantly associated with decreased term birthweight during gestational weeks 10-29 and with SGA for weeks 9-26. Cumulative effects for these time periods were -287.4 g (95% CI: -474.1 g, -100.8 g for birthweight and 4.7 (95% CI: 1.6, 14.1) for SGA. Temperature variability was also significantly associated with decreased birthweight between weeks 15 and 26. The effects for mean heat measures on term birthweight and SGA were not significant for any gestational week. Stratification by sex revealed a significant effect on term birthweight in females between weeks 23-28 and in males between weeks 9-26. Strongest effects of HI variability on term birthweight were found in children of mothers who experienced homelessness during pregnancy. Weekly HI variability was the heat metric most strongly associated with measures of gestational growth. The effects observed were largest in males and those who experienced homelessness during pregnancy. DISCUSSION: Given the impact of heat variability on birthweight and risk of SGA, it is important for future heat warnings to incorporate measure of heat index and temperature variability.

Validation of a neighborhood sentiment and safety index derived from existing data repositories.

BACKGROUND: The communities we live in are central to our health. Neighborhood disadvantage is associated with worse physical and mental health and even early mortality, while resident sense of safety and positive neighborhood sentiment has been repeatedly linked to better physical and mental health outcomes. Therefore, understanding where negative neighborhood sentiment and safety are salient concerns can help inform public health interventions and as a result, improve health outcomes. To date, fear of crime and neighborhood sentiment data or indices have largely been based on the administration of time consuming and costly standardized surveys. OBJECTIVE: The current study aims to develop a Neighborhood Sentiment and Safety Index (NSSI) at the census tract level, building on publicly available data repositories, including the US Census and ACS surveys, Data Axle, and ESRI repositories. METHODS: The NSSI was created using Principal Component Analysis. Mineigen and minimum loading values were 1 and 0.3, respectively. Throughout the step-wise PCA process, variables were excluded if their loading value was below 0.3 or if variables loaded into multiple components. RESULTS: The novel index was validated against standardized survey items from a longitudinal cohort study in the Northeastern United States characterizing experiences of (1) Neighborhood Characteristics with a Pearson correlation of -0.34 (p < 0.001) and, (2) Neighborhood Behavior Impact with a Pearson correlation of -0.33 (p < 0.001). It also accurately predicted the Share Care Community Well Being Index (Spearman correlation = 0.46) and the neighborhood deprivation index (NDI) (Spearman correlation = -0.75). SIGNIFICANCE: Our NSSI can serve as a predictor of neighborhood experience where data is either unavailable or too resource consuming to practically implement in planned studies. IMPACT STATEMENT: To date, fear of crime and neighborhood sentiment data or indices have largely been based on the administration of time consuming and costly standardized surveys. The current study aims to develop a Neighborhood Sentiment and Safety Index (NSSI) at the census tract level, building on publicly available data repositories, including the US Census and ACS surveys, Data Axle, and ESRI repositories. The NSSI was validated against four separate measures and can serve as a predictor of neighborhood experience where data is either unavailable or too resource consuming to practically implement in planned studies.

Missing Race and Ethnicity Data among COVID-19 Cases in Massachusetts.

Infectious disease surveillance frequently lacks complete information on race and ethnicity, making it difficult to identify health inequities. Greater awareness of this issue has occurred due to the COVID-19 pandemic, during which inequities in cases, hospitalizations, and deaths were reported but with evidence of substantial missing demographic details. Although the problem of missing race and ethnicity data in COVID-19 cases has been well documented, neither its spatiotemporal variation nor its particular drivers have been characterized. Using individual-level data on confirmed COVID-19 cases in Massachusetts from March 2020 to February 2021, we show how missing race and ethnicity data: (1) varied over time, appearing to increase sharply during two different periods of rapid case growth; (2) differed substantially between towns, indicating a nonrandom distribution; and (3) was associated significantly with several individual- and town-level characteristics in a mixed-effects regression model, suggesting a combination of personal and infrastructural drivers of missing data that persisted despite state and federal data-collection mandates. We discuss how a variety of factors may contribute to persistent missing data but could potentially be mitigated in future contexts.

Deep Ensemble Learning for Retinal Image Classification.

Purpose: Vision impairment affects 2.2 billion people worldwide, half of which is preventable with early detection and treatment. Currently, automatic screening of ocular pathologies using convolutional neural networks (CNNs) on retinal fundus photographs is limited to a few pathologies. Simultaneous detection of multiple ophthalmic pathologies would increase clinical usability and uptake. Methods: Two thousand five hundred sixty images were used from the Retinal Fundus Multi-Disease Image Dataset (RFMiD). Models were trained (n = 1920) and validated (n = 640). Five selected CNN architectures were trained to predict the presence of any pathology and categorize the 28 pathologies. All models were trained to minimize asymmetric loss, a modified form of binary cross-entropy. Individual model predictions were averaged to obtain a final ensembled model and assessed for mean area under the receiver-operator characteristic curve (AUROC) for disease screening (healthy versus pathologic image) and classification (AUROC for each class). Results: The ensemble network achieved a disease screening (healthy versus pathologic) AUROC score of 0.9613. The highest single network score was 0.9586 using the SE-ResNeXt architecture. For individual disease classification, the average AUROC score for each class was 0.9295. Conclusions: Retinal fundus images analyzed by an ensemble of CNNs trained to minimize asymmetric loss were effective in detection and classification of ocular pathologies than individual models. External validation is needed to translate machine learning models to diverse clinical contexts. Translational Relevance: This study demonstrates the potential benefit of ensemble-based deep learning methods on improving automatic screening and diagnosis of multiple ocular pathologies from fundoscopy imaging.