A Driver in Health Outcomes: Developing Discrete Categories of Transportation Insecurity.

Research suggests that transportation is an important social determinant of health, because the ability to get around is consequential for accessing health care and nutritious food and for making social connections. We used an inductive mixed-methods approach and a quantitative k-means clustering approach to identify 5 categories of transportation insecurity using the validated 16-item Transportation Security Index. The resulting 5-category measure distinguished among respondents with qualitatively different experiences of transportation insecurity. Analyzing data from 2018 that were representative of the US adult population aged 25 years or older, we demonstrated a nonparametric association between transportation insecurity and 2 different health measures (self-rated health and depressive symptoms). There was a threshold relationship between self-rated health and any level of transportation insecurity. High transportation insecurity had a very strong relationship with depressive symptoms. The categorical Transportation Security Index will be useful for clinicians who wish to screen for transportation-related barriers to health care. It will also facilitate research investigating the influence of transportation insecurity on health outcomes and provide the basis for interventions designed to address health disparities.

Street Audits to Measure Neighborhood Disorder: Virtual or In-Person?

Neighborhood conditions may influence a broad range of health indicators, including obesity, injury, and psychopathology. In particular, neighborhood physical disorder-a measure of urban deterioration-is thought to encourage crime and high-risk behaviors, leading to poor mental and physical health. In studies to assess neighborhood physical disorder, investigators typically rely on time-consuming and expensive in-person systematic neighborhood audits. We compared 2 audit-based measures of neighborhood physical disorder in the city of Detroit, Michigan: One used Google Street View imagery from 2009 and the other used an in-person survey conducted in 2008. Each measure used spatial interpolation to estimate disorder at unobserved locations. In total, the virtual audit required approximately 3% of the time required by the in-person audit. However, the final physical disorder measures were significantly positively correlated at census block centroids (r = 0.52), identified the same regions as highly disordered, and displayed comparable leave-one-out cross-validation accuracy. The measures resulted in very similar convergent validity characteristics (correlation coefficients within 0.03 of each other). The virtual audit-based physical disorder measure could substitute for the in-person one with little to no loss of precision. Virtual audits appear to be a viable and much less expensive alternative to in-person audits for assessing neighborhood conditions.

Use of Google Street View to Assess Environmental Contributions to Pedestrian Injury.

OBJECTIVES: To demonstrate an information technology-based approach to assess characteristics of streets and intersections associated with injuries that is less costly and time-consuming than location-based studies of pedestrian injury. METHODS: We used imagery captured by Google Street View from 2007 to 2011 to assess 9 characteristics of 532 intersections within New York City. We controlled for estimated pedestrian count and estimated the relation between intersections' characteristics and frequency of injurious collisions. RESULTS: The count of pedestrian injuries at intersections was associated with the presence of marked crosswalks (80% increase; 95% confidence interval [CI] = 2%, 218%), pedestrian signals (156% increase; 95% CI = 69%, 259%), nearby billboards (42% increase; 95% CI = 7%, 90%), and bus stops (120% increase; 95% CI = 51%, 220%). Injury incidence per pedestrian was lower at intersections with higher estimated pedestrian volumes. CONCLUSIONS: Consistent with in-person study observations, the information-technology approach found traffic islands, visual advertising, bus stops, and crosswalk infrastructures to be associated with elevated counts of pedestrian injury in New York City. Virtual site visits for pedestrian injury control studies are a viable and informative methodology.

Validity of an ecometric neighborhood physical disorder measure constructed by virtual street audit.

Neighborhood physical disorder is thought to affect mental and physical health, but it has been difficult to measure objectively and reliably across large geographical areas or multiple locales. Virtual street audits are a novel method for assessing neighborhood characteristics. We evaluated the ecometric properties of a neighborhood physical disorder measure constructed from virtual street audit data. Eleven trained auditors assessed 9 previously validated items developed to capture physical disorder (e.g., litter, graffiti, and abandoned buildings) on 1,826 block faces using Google Street View imagery (Google, Inc., Mountain View, California) dating from 2007-2011 in 4 US cities (San Jose, California; Detroit, Michigan; New York, New York; and Philadelphia, Pennsylvania). We constructed a 2-parameter item response theory scale to estimate latent levels of disorder on each block face and defined a function using kriging to estimate physical disorder levels, with confidence estimates, for any point in each city. The internal consistency reliability of the resulting scale was 0.93. The final measure of disorder was positively correlated with US Census data on unemployment and housing vacancy and negatively correlated with data on owner-occupied housing. These results suggest that neighborhood physical disorder can be measured reliably and validly using virtual audits, facilitating research on possible associations between physical disorder and health.

The impact of neighborhood park access and quality on body mass index among adults in New York City.

OBJECTIVE: To evaluate the association between adult individuals' body mass index (BMI) and characteristics of parks (size and cleanliness) in an urban environment taking into account the physical and social environments of the neighborhood. METHODS: Cross-sectional, hierarchical linear models were used to determine whether park effects were associated with BMI using self-reported height and weight data obtained from the Community Health Survey in New York City (2002-2006). RESULTS: Both the proportion of the residential zip code that was large park space and the proportion that was small park space had significant inverse associations with BMI after controlling for individual socio-demographic and zip code built environment characteristics (-0.20 BMI units across the inter-quartile range (IQR) for large parks, 95% CI -0.32, -0.08; -0.21 BMI units across the IQR for small parks, 95% CI -0.31, -0.10, respectively). Poorer scores on the park cleanliness index were associated with higher BMI, 0.18 BMI units across the IQR of the park cleanliness index (95% CI 0.05, 0.30). CONCLUSIONS: This study demonstrated that proportion of neighborhoods that was large or small park space and park cleanliness were associated with lower BMI among NYC adults after adjusting for other neighborhood features such as homicides and walkability, characteristics that could influence park usage.

Neighborhood walkability and active travel (walking and cycling) in New York City.

Urban planners have suggested that built environment characteristics can support active travel (walking and cycling) and reduce sedentary behavior. This study assessed whether engagement in active travel is associated with neighborhood walkability measured for zip codes in New York City. Data were analyzed on engagement in active travel and the frequency of walking or biking ten blocks or more in the past month, from 8,064 respondents to the New York City 2003 Community Health Survey (CHS). A neighborhood walkability scale that measures: residential, intersection, and subway stop density; land use mix; and the ratio of retail building floor area to retail land area was calculated for each zip code. Data were analyzed using zero-inflated negative binomial regression incorporating survey sample weights and adjusting for respondents' sociodemographic characteristics. Overall, 44 % of respondents reported no episodes of active travel and among those who reported any episode, the mean number was 43.2 episodes per month. Comparing the 75th to the 25th percentile of zip code walkability, the odds ratio for reporting zero episodes of active travel was 0.71 (95 % CI 0.61, 0.83) and the exponentiated beta coefficient for the count of episodes of active travel was 1.13 (95 % CI 1.06, 1.21). Associations between lower walkability and reporting zero episodes of active travel were significantly stronger for non-Hispanic Whites as compared to non-Hispanic Blacks and to Hispanics and for those living in higher income zip codes. The results suggest that neighborhood walkability is associated with higher engagement in active travel.

Exploring potential for selection bias in using survey data to estimate the association between institutional trust and depression.

PURPOSE: We tested the hypothesis that low institutional trust would be associated with depressive symptom elevation, with attention to potential selection bias. METHODS: The District of Columbia Area Survey (DCAS) was conducted by mail in 2018. Invitations sent to 8800 households resulted in a sample of 1061 adults. Institutional trust questions referenced nonprofit organizations, businesses, and government. Depressive symptom elevation was assessed using PHQ-9. Logistic regression model estimates were compared with and without adjustment for sociodemographic characteristics and neighborhood satisfaction; among complete cases and following multiple imputation of missing covariate data; and with and without survey weights or correction for collider selection bias. RESULTS: Of 968 participants without missing depressive symptom or trust data, 24% reported low institutional trust. Low institutional trust was associated with elevated depressive symptoms (adjusted OR following multiple imputation: 2.0; 95% CI: 1.1, 3.4), although the association was attenuated with use of survey weights (adjusted OR incorporating multiple imputation and survey weights: 1.6; 95% CI: 0.7, 3.2). CONCLUSIONS: Under contrasting scenarios where low institutional trust and depressive symptoms jointly increase nonresponse, selection bias could lead to under- or overestimation of this association. Future research could explore posited selection bias scenarios that differ in direction of bias.

Individual- and school-level sociodemographic predictors of obesity among New York City public school children.

To identify student- and school-level sociodemographic characteristics associated with overweight and obesity, the authors conducted cross-sectional analyses of data from 624,204 public school children (kindergarten through 12th grade) who took part in the 2007-2008 New York City Fitnessgram Program. The overall prevalence of obesity was 20.3%, and the prevalence of overweight was 17.6%. In multivariate models, the odds of being obese as compared with normal weight were higher for boys versus girls (odds ratio (OR) = 1.39, 95% confidence interval (CI): 1.36, 1.42), for black (OR = 1.11, 95% CI: 1.07, 1.15) and Hispanic (OR = 1.48, 95% CI: 1.43, 1.53) children as compared with white children, for children receiving reduced-price (OR = 1.17, 95% CI: 1.13, 1.21) or free (OR = 1.12, 95% CI: 1.09, 1.15) school lunches as compared with those paying full price, and for US-born students (OR = 1.54, 95% CI: 1.50, 1.58) as compared with foreign-born students. After adjustment for individual-level factors, obesity was associated with the percentage of students who were US-born (across interquartile range (75th percentile vs. 25th), OR = 1.10, 95% CI: 1.07, 1.14) and the percentage of students who received free or reduced-price lunches (across interquartile range, OR = 1.13, 95% CI: 1.10, 1.18). The authors conclude that individual sociodemographic characteristics and school-level sociodemographic composition are associated with obesity among New York City public school students.

Machine Learning Approaches for Measuring Neighborhood Environments in Epidemiologic Studies.

Purpose of review: Innovations in information technology, initiatives by local governments to share administrative data, and growing inventories of data available from commercial data aggregators have immensely expanded the information available to describe neighborhood environments, supporting an approach to research we call Urban Health Informatics. This review evaluates the application of machine learning to this new wealth of data for studies of the effects of neighborhood environments on health. Recent findings: Prominent machine learning applications in this field include automated image analysis of archived imagery such as Google Street View images, variable selection methods to identify neighborhood environment factors that predict health outcomes from large pools of exposure variables, and spatial interpolation methods to estimate neighborhood conditions across large geographic areas. Summary: In each domain, we highlight successes and cautions in the application of machine learning, particularly highlighting legal issues in applying machine learning approaches to Google's geo-spatial data.

Causal Inference with Case-Only Studies in Injury Epidemiology Research.

Purpose of Review: We review the application and limitations of two implementations of the "case-only design" in injury epidemiology with example analyses of Fatality Analysis Reporting System data. Recent Findings: The term "case-only design" covers a variety of epidemiologic designs; here, two implementations of the design are reviewed: (1) studies to uncover etiological heterogeneity and (2) studies to measure exposure effect modification. These two designs produce results that require different interpretations and rely upon different assumptions. The key assumption of case-only designs for exposure effect modification, the more commonly used of the two designs, does not commonly hold for injuries and so results from studies using this design cannot be interpreted. Case-only designs to identify etiological heterogeneity in injury risk are interpretable but only when the case-series is conceptualized as arising from an underlying cohort. Summary: The results of studies using case-only designs are commonly misinterpreted in the injury literature.

Racial residential segregation and low birth weight in Michigan's metropolitan areas.

OBJECTIVES: We examined the influence of racial residential segregation, independent of neighborhood economic factors, on the overall and specific etiological risks of low birth weight. METHODS: We geocoded all singleton births in Michigan metropolitan areas during 2000 to census tracts. We used hierarchical generalized linear models to investigate the association between low birth weight (< 2500 g) and neighborhood-level economic and racial segregation, controlling for individual and neighborhood characteristics. We analyzed competing risks of the 2 etiologies of low birth weight: intrauterine growth restriction and preterm birth. RESULTS: Living in a Black segregated area was associated with increased odds (odds ratio [OR] = 1.15; 95% confidence interval [CI] = 1.03, 1.29; P < .05) of low birth weight after adjusting for individual- and tract-level measures. The analysis suggested that the association between low birth weight and racial segregation was attributable primarily to increased risk of intrauterine growth restriction (OR = 1.19; 95% CI = 1.03, 1.37; P < .05). CONCLUSIONS: Odds of low birth weight are higher in racially segregated Black neighborhoods in Michigan's metropolitan areas, independent of economic factors. The association appears to operate through intrauterine growth restriction rather than preterm birth.

Measurement of the local food environment: a comparison of existing data sources.

Studying the relation between the residential environment and health requires valid, reliable, and cost-effective methods to collect data on residential environments. This 2002 study compared the level of agreement between measures of the presence of neighborhood businesses drawn from 2 common sources of data used for research on the built environment and health: listings of businesses from commercial databases and direct observations of city blocks by raters. Kappa statistics were calculated for 6 types of businesses-drugstores, liquor stores, bars, convenience stores, restaurants, and grocers-located on 1,663 city blocks in Chicago, Illinois. Logistic regressions estimated whether disagreement between measurement methods was systematically correlated with the socioeconomic and demographic characteristics of neighborhoods. Levels of agreement between the 2 sources were relatively high, with significant (P < 0.001) kappa statistics for each business type ranging from 0.32 to 0.70. Most business types were more likely to be reported by direct observations than in the commercial database listings. Disagreement between the 2 sources was not significantly correlated with the socioeconomic and demographic characteristics of neighborhoods. Results suggest that researchers should have reasonable confidence using whichever method (or combination of methods) is most cost-effective and theoretically appropriate for their research design.

Disparities in neighborhood food environments: implications of measurement strategies.

Public health researchers have begun to map the neighborhood "food environment" and examine its association with the risk of overweight and obesity. Some argue that "food deserts"­areas with little or no provision of fresh produce and other healthy food­may contribute to disparities in obesity, diabetes, and related health problems. While research on neighborhood food environments has taken advantage of more technically sophisticated ways to assess distance and density, in general, it has not considered how individual or neighborhood conditions might modify physical distance and thereby affect patterns of spatial accessibility. This study carried out a series of sensitivity analyses to illustrate the effects on the measurement of disparities in food environments of adjusting for cross-neighborhood variation in vehicle ownership rates, public transit access, and impediments to pedestrian travel, such as crime and poor traffic safety. The analysis used geographic information systems data for New York City supermarkets, fruit and vegetable markets, and farmers' markets and employed both kernel density and distance measures. We found that adjusting for vehicle ownership and crime tended to increase measured disparities in access to supermarkets by neighborhood race/ethnicity and income, while adjusting for public transit and traffic safety tended to narrow these disparities. Further, considering fruit and vegetable markets and farmers' markets, as well as supermarkets, increased the density of healthy food outlets, especially in neighborhoods with high concentrations of Hispanics, Asians, and foreign-born residents and in high-poverty neighborhoods.

Disparities in trajectories of changes in the unhealthy food environment in New York City: A latent class growth analysis, 1990-2010.

Disparities in availability of food retailers in the residential environment may help explain racial/ethnic and socio-economic differences in obesity risk. Research is needed that describes whether food environment dynamics may contribute to equalizing conditions across neighborhoods or to amplifying existing inequalities over time. This study improves the understanding of how the BMI-unhealthy food environment has evolved over time in New York City. We use longitudinal census tract-level data from the National Establishment Time-Series (NETS) for New York City in the period 1990-2010 and implement latent class growth analysis (LCGA) to (1) examine trajectories of change in the number of unhealthy food outlets (characterized as selling calorie-dense foods such as pizza and pastries) at the census tract-level, and (2) examine how trajectories are related to socio-demographic characteristics of the census tract. Overall, the number of BMI-unhealthy food outlets increased between 1990 and 2010. We summarized trajectories of evolutions with a 5-class model that indicates a pattern of fanning out, such that census tracts with a higher initial number of BMI-unhealthy food outlets in 1990 experienced a more rapid increase over time. Finally, fully adjusted logistic regression models reveal a greater increase in BMI-unhealthy food outlets in census tracts with: higher baseline population size, lower baseline income, and lower proportion of Black residents. Greater BMI-unhealthy food outlet increases were also noted in the context of census tracts change suggestive of urbanization (increasing population density) or increasing purchasing power (increasing income).

Neighborhood Physical Disorder in New York City.

Neighborhood physical disorder, or the deterioration of urban environments, is associated with negative mental and physical health outcomes. Eleven trained raters used CANVAS, a web-based system for conducting reliable virtual street audits, to collect data on nine indicators of physical disorder using Google Street View imagery of 532 block faces in New York City, New York, USA. We combined the block face indicator data into a disorder scale using item response theory; indicators ranged in severity from presence of litter, a weak indicator of disorder, to abandoned cars, a strong indicator. Using this scale, we estimated disorder at the center point of each sampled block. We then used ordinary kriging to interpolate estimates of disorder levels throughout the city. The resulting map condenses a complex estimation process into an interpretable visualization of the spatial distribution of physical disorder in New York City.

Measuring health-relevant businesses over 21 years: refining the National Establishment Time-Series (NETS), a dynamic longitudinal data set.

BACKGROUND: The densities of food retailers, alcohol outlets, physical activity facilities, and medical facilities have been associated with diet, physical activity, and management of medical conditions. Most of the research, however, has relied on cross-sectional studies. In this paper, we assess methodological issues raised by a data source that is increasingly used to characterize change in the local business environment: the National Establishment Time Series (NETS) dataset. DISCUSSION: Longitudinal data, such as NETS, offer opportunities to assess how differential access to resources impacts population health, to consider correlations among multiple environmental influences across the life course, and to gain a better understanding of their interactions and cumulative health effects. Longitudinal data also introduce new data management, geoprocessing, and business categorization challenges. Examining geocoding accuracy and categorization over 21 years of data in 23 counties surrounding New York City (NY, USA), we find that health-related business environments change considerably over time. We note that re-geocoding data may improve spatial precision, particularly in early years. Our intent with this paper is to make future public health applications of NETS data more efficient, since the size and complexity of the data can be difficult to exploit fully within its 2-year data-licensing period. Further, standardized approaches to NETS and other "big data" will facilitate the veracity and comparability of results across studies.

Development and deployment of the Computer Assisted Neighborhood Visual Assessment System (CANVAS) to measure health-related neighborhood conditions.

Public health research has shown that neighborhood conditions are associated with health behaviors and outcomes. Systematic neighborhood audits have helped researchers measure neighborhood conditions that they deem theoretically relevant but not available in existing administrative data. Systematic audits, however, are expensive to conduct and rarely comparable across geographic regions. We describe the development of an online application, the Computer Assisted Neighborhood Visual Assessment System (CANVAS), that uses Google Street View to conduct virtual audits of neighborhood environments. We use this system to assess the inter-rater reliability of 187 items related to walkability and physical disorder on a national sample of 150 street segments in the United States. We find that many items are reliably measured across auditors using CANVAS and that agreement between auditors appears to be uncorrelated with neighborhood demographic characteristics. Based on our results we conclude that Google Street View and CANVAS offer opportunities to develop greater comparability across neighborhood audit studies.

CREATING MEASURES OF THEORETICALLY RELEVANT NEIGHBORHOOD ATTRIBUTES AT MULTIPLE SPATIAL SCALES.

Accurately measuring attributes in neighborhood environments allows researchers to study the influence of neighborhoods on individual-level outcomes. Researchers working to improve the measurement of neighborhood attributes generally advocate doing so in one of two ways: improving the theoretical relevance of measures and correctly defining the appropriate spatial scale. The data required by the first, "ecometric" neighborhood assessments on a sample of neighborhoods, are generally incompatible with the methods of the second, which tend to rely on population data. In this article, the authors describe how ecometric measures of theoretically relevant attributes observed on a sample of city blocks can be combined with a geostatistical method known as kriging to develop city block-level estimates across a city that can be configured to multiple neighborhood definitions. Using a cross-validation study with data from a 2002 systematic social observation of physical disorder on 1,663 city blocks in Chicago, the authors show that this method creates valid results. They then demonstrate, using neighborhood measures aggregated to three different spatial scales, that residents' perceptions of both fear and neighborhood disorder vary substantially across different spatial scales.