Identifying multilevel predictors of behavioral outcomes like park use: A comparison of conditional and marginal modeling approaches.

This study compared marginal and conditional modeling approaches for identifying individual, park and neighborhood park use predictors. Data were derived from the ParkIndex study, which occurred in 128 block groups in Brooklyn (New York), Seattle (Washington), Raleigh (North Carolina), and Greenville (South Carolina). Survey respondents (n = 320) indicated parks within one half-mile of their block group used within the past month. Parks (n = 263) were audited using the Community Park Audit Tool. Measures were collected at the individual (park visitation, physical activity, sociodemographic characteristics), park (distance, quality, size), and block group (park count, population density, age structure, racial composition, walkability) levels. Generalized linear mixed models and generalized estimating equations were used. Ten-fold cross validation compared predictive performance of models. Conditional and marginal models identified common park use predictors: participant race, participant education, distance to parks, park quality, and population >65yrs. Additionally, the conditional mode identified park size as a park use predictor. The conditional model exhibited superior predictive value compared to the marginal model, and they exhibited similar generalizability. Future research should consider conditional and marginal approaches for analyzing health behavior data and employ cross-validation techniques to identify instances where marginal models display superior or comparable performance.

Unearthing Inequities in the Relationship between Multiple Sociodemographic Factors and Diverse Elements of Park Availability and Quality in a Major Southern Metropolitan Region.

Parks are critical components of healthy communities. This study explored neighborhood socioeconomic and racial/ethnic inequalities in park access and quality in a large U.S. southeastern metropolitan region. A total of 241 block groups were examined, including 77 parks. For each block group, we obtained multiple sociodemographic indicators, including unemployment rate, education level, renter-occupied housing, poverty rate, and racial/ethnic minority composition. All parks were mapped using geographical information systems and audited via the Community Park Audit Tool to evaluate their features and quality. We analyzed seven diverse elements of park quality (transportation access, facility availability, facility quality, amenity availability, park aesthetics, park quality concerns, and neighborhood quality concerns), as well as an overall park quality score by calculating the mean for all parks within each block group. The mean percent of residents below 125% of the poverty level and the percentage of renter-occupied housing units were significantly higher among block groups with any parks in comparison to block groups with no parks. In addition, there were significant positive associations between park transportation access scores and both the percentage of residents with less than high school education and the percent identifying as non-Hispanic white. Moreover, there was a significant negative association between park amenity availability and the block group's unemployed population. Further, a significant negative association between park aesthetics and the population with a lower than high school education percentage was observed. Revealed differences in park availability, park acreage, and park quality dimensions emphasized the need for targeted policy, programmatic, and infrastructure interventions to improve park access and quality and address health disparities.

Exploring Disparities in Youth Physical Activity Environments by Income and Non-Hispanic White Population Across the United States.

This study investigated relationships between youth physical activity (PA) environments and income and non-Hispanic White population across the United States, stratified by US Census region and urban-rural designation. For all counties (n = 3142), publicly accessible data were used for sociodemographic indicators (ie, median household income and percent non-Hispanic White population) and a composite PA environment index (including exercise opportunities, violent crime incidence, walkability, and access to public schools). One-way analysis of variance was used to examine differences in PA environment index values according to income and non-Hispanic White population tertiles. There were significant differences in PA environments according to tertiles of income (F = 493.5, P < .001) and non-Hispanic White population (F = 58.6, P < .001), including variations by region and urban-rural designation. Public health practice and policy initiatives, such as joint use agreements, Safe Routes to School programs, and targeted funding allocations, should be used to address more pronounced income-based disparities in Southern and metropolitan counties and race-based disparities in rural and micropolitan counties.

Examining Disparities in Food Access Between Historically Black Colleges and Universities and Non-Historically Black Colleges and Universities.

PURPOSE: Examine if Historically Black Colleges and Universities (HBCUs) are more likely to be located in low food access area (LFA) census tracts compared to public non-HBCUs. DESIGN: ArcGIS Pro was utilized to capture food environments and census tract sociodemographic data. SETTING: The sample included 98 HBCUs and 777 public non-HBCUs within the United States. 28.9% of study census tracts were classified as LFA tracts. MEASURES: University data were gathered from the National Center for Education Statistics. Census tract-level LFA classification was informed by the United States Department of Agriculture's Food Access Research Atlas. Covariates included population density and neighborhood socioeconomic status of census tracts containing subject universities. ANALYSIS: Multilevel logistic regression was employed to examine the relationship between university type and LFA classification. RESULTS: A higher percentage of HBCUs (46.9%) than public non-HBCUs (26.6%) were located in LFAs. After adjusting for population density and neighborhood socioeconomic status, university type was significantly associated with food access classification (B=0.71;p=.0036). The odds of an HBCU being located in LFA tracts were 104% greater than for a public non-HBCU (OR=2.04;95% CI=1.26,3.29). CONCLUSION: Findings underscore the need for policy interventions tailored to HBCU students to promote food security, environmental justice, and public health.

Incongruency of youth food and physical activity environments in the United States: Variations by region, rurality, and income.

Local environments are increasingly the focus of health behavior research and practice to reduce gaps between fruit/vegetable intake, physical activity (PA), and related guidelines. This study examined the congruency between youth food and PA environments and differences by region, rurality, and income across the United States. Food and PA environment data were obtained for all U.S. counties (N = 3142) using publicly available, secondary sources. Relationships between the food and PA environment tertiles was represented using five categories: 1) congruent-low (county falls in both the low food and PA tertiles), 2) congruent-high (county falls in both the high food and PA tertiles), 3) incongruent-food high/PA low (county falls in high food and low PA tertiles), 4) incongruent-food low/PA high (county falls in low food and high PA tertiles), and 5) intermediate food or PA (county falls in the intermediate tertile for food and/or PA). Results showed disparities in food and PA environment congruency according to region, rurality, and income (p < .0001 for each). Nearly 25% of counties had incongruent food and PA environments, with food high/PA low counties mostly in rural and low-income areas, and food low/PA high counties mostly in metropolitan and high-income areas. Approximately 8.7% of counties were considered congruent-high and were mostly located in the Northeast, metropolitan, and high-income areas. Congruent-low counties made up 10.0% of counties and were mostly in the South, rural, and low-income areas. National and regional disparities in environmental obesity determinants were identified that can inform targeted public health interventions.

Healthy and unhealthy food environments are linked with neighbourhood socio-economic disadvantage: an innovative geospatial approach to understanding food access inequities.

OBJECTIVE: This study examined the separate relationships between socio-economic disadvantage and the density of multiple types of food outlets, and relationships between socio-economic disadvantage and composite food environment indices. DESIGN: Cross-sectional data were analysed using geospatial kernel density techniques. Food outlet data included convenience stores, discount stores, fast-food and fast casual restaurants, and grocery stores. Controlling for urbanicity and race/ethnicity, multivariate linear regression was used to examine the relationships between socio-economic disadvantage and density of food outlets. SETTING: This study occurred in a large Southeastern US county containing 255 census block groups with a total population of 474 266, of which 77·1 % was Non-Hispanic White, the median household income was $48 886 and 15·0 % of residents lived below 125 % of the federal poverty line. PARTICIPANTS: The unit of analysis was block groups; all data about neighbourhood socio-economic disadvantage and food outlets were publicly available. RESULTS: As block group socio-economic disadvantage increased, so too did access to all types of food outlets. The total food environment index, calculated as the ratio of unhealthy food outlets to all food outlets, decreased as block group disadvantage increased. CONCLUSIONS: Those who reside in more disadvantaged block groups have greater access to both healthy and unhealthy food outlets. The density of unhealthy establishments was greater in more disadvantaged areas; however, because of having greater access to grocery stores, disadvantaged populations have less obesogenic total food environments. Structural changes are needed to reduce access to unhealthy food outlets to ensure environmental injustice and reduce obesity risk.

Associations between Walkability and Youth Obesity: Differences by Urbanicity.

Background: Attributes of the built environment, such as neighborhood walkability, have been linked to increased physical activity and reduced obesity risk. This relationship, however, has primarily been documented in adults; less is known about neighborhood walkability and youth obesity, as limited prior research has produced mixed findings. The purpose of this study was to examine the association between neighborhood walkability and youth obesity, including differences by urbanicity. Methods: Data were collected in 2013 from youth aged 7-14 years (n = 13,469) in a Southeastern county school district. Height and weight were objectively measured and utilized to calculate body mass index (BMI) z-scores. Youth demographic characteristics and addresses were obtained, and a Walk Score® was gathered for each youth's home address. Multilevel linear regression analysis, accounting for nesting within census block groups, was conducted to examine the association between Walk Score and BMI z-score and to test for the moderating effect of urbanicity. Separate multilevel analyses examined Walk Score and BMI z-score among urban, urban-rural mixed, and rural youth subsamples. Results: Overall, as Walk Score increased, youth BMI z-score decreased. Walk Score was positively associated with BMI z-score among urban youth and negatively associated with BMI z-score among rural youth; no relationship was observed between Walk Score and youth in urban-rural mixed areas. Conclusions: Neighborhood walkability may impact youth differently across geographic areas. Further study is warranted about how youth utilize a walkable environment and mechanisms through which walkability influences youth physical activity and obesity risk.