Validation of food store environment secondary data source and the role of neighborhood deprivation in Appalachia, Kentucky

Predicting access to healthful food retailers with machine learning

Many U.S. households lack access to healthful food and rely on inexpensive, processed food with low nutritional value. Surveying access to healthful food is costly and finding the factors that affect access remains convoluted owing to the multidimensional nature of socioeconomic variables. We utilize machine learning with census tract data to predict the modified Retail Food Environment Index (mRFEI), which refers to the percentage of healthful food retailers in a tract and agnostically extract the features of no access-corresponding to a "food desert" and low access-corresponding to a "food swamp." Our model detects food deserts and food swamps with a prediction accuracy of 72% out of the sample. We find that food deserts and food swamps are intrinsically different and require separate policy attention. Food deserts are lightly populated rural tracts with low ethnic diversity, whereas swamps are predominantly small, densely populated, urban tracts, with more non-white residents who lack vehicle access. Overall access to healthful food retailers is mainly explained by population density, presence of black population, property value, and income. We also show that our model can be used to obtain sensible predictions of access to healthful food retailers for any U.S. census tract.

Validation of secondary data sources for enumerating marijuana dispensaries in a state commercializing marijuana

OBJECTIVES

To assess 1) the validity of online crowdsourcing platforms in enumerating licensed brick-and-mortar marijuana dispensaries and 2) the validity of state licensing directory and online crowdsourcing platforms in enumerating active brick-and-mortar marijuana dispensaries in California.

METHODS

We obtained business lists from California Bureau of Cannabis Control (BCC) licensing directory and three online crowdsourcing platforms (Weedmaps, Leafly, and Yelp) in May 2019. Calls were made to verify street address, operation status, dispensary category (recreational-only, medical-only, recreational & medical), and presence of storefronts in May-July 2019. Validity measures, including sensitivity, specificity, positive predictive value, and negative predictive value, were calculated when applicable.

RESULTS

In identifying licensed dispensaries in BCC, Leafly had the highest sensitivity (.66) and Yelp had the highest specificity (.87). The dispensary category posted on online crowdsourcing platforms in over 25 % licensed dispensaries and the dispensary category claimed in call verification in over 10 % licensed dispensaries disagreed with the approved category in BCC. There were 2121 businesses combined from BCC and online crowdsourcing platforms, among which 826 were verified to be active brick-and-mortar dispensaries. Weedmaps had the highest sensitivity (.80) and Yelp had the highest negative predictive value (.74) in identifying verified dispensaries. Weedmaps overall had the highest sensitivity in all three dispensary categories. Weedmaps had the highest sensitivity in more populated counties whereas BCC had the highest sensitivity in less populated counties.

CONCLUSIONS

Each secondary data source has strengths and limitations. The findings inform surveillance and research regarding how to best strategize data use when resources are limited.

The use and misuse of ratio and proportion exposure measures in food environment research

Background

The food stores within residential environments are increasingly investigated as a possible mechanism driving food behaviours and health outcomes. Whilst increased emphasis is being placed on the type of study designs used and how we measure the outcomes, surprisingly little attention gets diverted to the measures of the food environment beyond calls for standardised approaches for food store coding and geographic scales of exposure. Food environments are a challenging concept to measure and model and the use of ratio and proportion measures are becoming more common in food environment research. Whilst these are seemingly an advance on single store type indicators, such as simply counting the number of supermarkets or fast food restaurants present, they have several limitations that do not appear to have been fully considered.

Main body

In this article we report on five issues related to the use of ratio and proportion food environment measures: 1) binary categorisation of food stores; 2) whether they truly reflect a more or less healthy food environment; 3) issues with these measures not reflecting the quantity of food stores; 4) difficulties when no stores are present; and 5) complications in statistical treatment and interpretation of ratio and proportion measures. Each of these issues are underappreciated in the literature to date and highlight that ratio and proportion measures need to be treated with caution.

Conclusion

Calls for the broader adoption of relative food environment measures may be misguided. Whilst we should continue to search for better ways to represent the complexity of food environments, ratio and proportion measures are unlikely to be the answer.

Examining the relationship between food environment and gestational diabetes

BACKGROUND

Studies have shown an association between the incidence of gestational diabetes and living in neighborhoods oversaturated with unhealthy foods.

OBJECTIVE

This study sought to determine if the food environment also affects the management of gestational diabetes. We hypothesized that living in areas with a higher quality of food decreased the risk of requiring medication to treat gestational diabetes.

STUDY DESIGN

This was a retrospective cohort study of singleton births at the Christiana Care Health System between 2015 and 2018. Patients with gestational diabetes who live in Delaware (N=1327) were geocoded and classified according to their census tract food environment. The food environment was assessed using the modified Retail Food Environment Index, which measures the percentage of healthy food retailers among all food retailers within a half-mile radius of the census tract boundaries. The modified Retail Food Environment Index scores were divided into 3 categories: poor (modified Retail Food Environment Index score, 0-3), average (modified Retail Food Environment Index score, 4-10), and good or above average (modified Retail Food Environment Index score, ≥11) food environments. The primary outcome was the prevalence of A2 gestational diabetes mellitus. Several neonatal and obstetrical outcomes were also examined including type II diabetes mellitus (defined as a 2-hour glucose tolerance test with at least 1 value above the threshold), cesarean delivery, shoulder dystocia, admission to the neonatal intensive care unit, neonatal hypoglycemia, neonatal hyperbilirubinemia, neonatal respiratory distress syndrome, and macrosomia.

RESULTS

A total of 689 (52%) women were diagnosed as having A2 gestational diabetes mellitus. Women in the average or good or above average food environment groups had a lower prevalence of A2 gestational diabetes mellitus than women in the poor food environment group (modified Retail Food Environment Index score, 4-10 [adjusted odds ratio, 0.58; 95% confidence interval, 0.37-0.92] and modified Retail Food Environment Index score, ≥11 [adjusted odds ratio, 0.56; 95% confidence interval, 0.40-0.82]). They also had a lower prevalence of type II diabetes mellitus (modified Retail Food Environment Index score, 4-10 [adjusted odds ratio, 0.25; 95% confidence interval, 0.09-0.72] and modified Retail Food Environment Index score, ≥11 [adjusted odds ratio, 0.48; 95% confidence interval, 0.27-0.86]). There were no differences in the other secondary outcomes of interest.

CONCLUSION

The food environment affects the requirement for medication to obtain glycemic levels that are within the target range for those with gestational diabetes.

Field Validation of Commercially Available Food Retailer Data in the Netherlands

The aim of this study was to validate a Dutch commercial dataset containing information on the types and locations of food retailers against field audit data. Field validation of a commercial dataset ("Locatus") was conducted in February 2019. Data on the location and classification of food retailers were collected through field audits in 152 streets from four urban and four rural neighborhoods in the Netherlands. The classification of food retailers included eight types of grocery stores (e.g., supermarkets, bakeries) and four types of food outlets (e.g., cafés, take away restaurants). The commercial dataset in the studied area listed 322 food retailers, whereas the field audit counted 315 food retailers. Overall, the commercially available data showed "good" to "excellent" agreement statistics (>0.71) with field audit data for all three levels of analysis (i.e., location, classification and both combined) and across urban as well as rural areas. The commercial dataset under study provided an accurate description of the measured food environment. Therefore, policymakers and researchers should feel confident in using this commercial dataset as a source of secondary data.

Government data v. ground observation for food-environment assessment: businesses missed and misreported by city and state inspection records

OBJECTIVE

To assess the accuracy of government inspection records, relative to ground observation, for identifying businesses offering foods/drinks.

DESIGN

Agreement between city and state inspection records v. ground observations at two levels: businesses and street segments. Agreement could be 'strict' (by business name, e.g. 'Rizzo's') or 'lenient' (by business type, e.g. 'pizzeria'); using sensitivity and positive predictive value (PPV) for businesses and using sensitivity, PPV, specificity and negative predictive value (NPV) for street segments.

SETTING

The Bronx and the Upper East Side (UES), New York City, USA.

PARTICIPANTS

All food/drink-offering businesses on sampled street segments (n 154 in the Bronx, n 51 in the UES).

RESULTS

By 'strict' criteria, sensitivity and PPV of government records for food/drink-offering businesses were 0·37 and 0·57 in the Bronx; 0·58 and 0·60 in the UES. 'Lenient' values were 0·40 and 0·62 in the Bronx; 0·60 and 0·62 in the UES. Sensitivity, PPV, specificity and NPV of government records for street segments having food/drink-offering businesses were 0·66, 0·73, 0·84 and 0·79 in the Bronx; 0·79, 0·92, 0·67, and 0·40 in the UES. In both areas, agreement varied by business category: restaurants; 'food stores'; and government-recognized other storefront businesses ('gov. OSB', i.e. dollar stores, gas stations, pharmacies). Additional business categories - 'other OSB' (barbers, laundromats, newsstands, etc.) and street vendors - were absent from government records; together, they represented 28·4 % of all food/drink-offering businesses in the Bronx, 22·2 % in the UES ('other OSB' and street vendors were sources of both healthful and less-healthful foods/drinks in both areas).

CONCLUSIONS

Government records frequently miss or misrepresent businesses offering foods/drinks, suggesting caveats for food-environment assessments using such records.

More than Fast Food: Development of a Story Map to Compare Adolescent Perceptions and Observations of Their Food Environments and Related Food Behaviors

The purpose of this convergent, multiphase, mixed methods study was to better understand the perceptions of adolescents' food environments and related food behaviors using grounded visualization and story mapping. Adolescents from one high school (13⁻16 years) in the southeastern United States were evaluated via data from health behavior surveys (n = 75), school environment maps, focus groups (n = 5 groups), and Photovoice (n = 6) from October 2016 to April 2017. Data from each phase were integrated using grounded visualization and new themes were identified (n = 7). A story map using ArcGIS Online was developed from data integration, depicting the newly identified themes. Participants failed to meet national recommendations for fruit and vegetable intake (2.71 cups). Focus group and Photovoice findings indicated the need for convenience food items in all environments. The story map is an online, interactive dissemination of information, with five maps, embedded quotes from focus groups, narrative passages with data interpretation, pictures to highlight themes, and a comparison of the participants' food environments. Story mapping and qualitative geographic information systems (GIS) approaches may be useful when depicting adolescent food environments and related food behaviors. Further research is needed when evaluating story maps and how individuals can be trained to create their own maps.

Assessments of residential and global positioning system activity space for food environments, body mass index and blood pressure among low-income housing residents in New York City

Research has examined how the food environment affects the risk of cardiovascular disease (CVD). Many studies have focused on residential neighbourhoods, neglecting the activity spaces of individuals. The objective of this study was to investigate whether food environments in both residential and global positioning system (GPS)-defined activity space buffers are associated with body mass index (BMI) and blood pressure (BP) among low-income adults. Data came from the New York City Low Income Housing, Neighborhoods and Health Study, including BMI and BP data (n=102, age=39.3±14.1 years), and one week of GPS data. Five food environment variables around residential and GPS buffers included: fast-food restaurants, wait-service restaurants, corner stores, grocery stores, and supermarkets. We examined associations between food environments and BMI, systolic and diastolic BP, controlling for individual- and neighbourhood-level sociodemographics and population density. Within residential buffers, a higher grocery store density was associated with lower BMI (β=- 0.20 kg/m2, P<0.05), and systolic and diastolic BP (β =-1.16 mm Hg; and β=-1.02 mm Hg, P<0.01, respectively). In contrast, a higher supermarket density was associated with higher systolic and diastolic BP (β=1.74 mm Hg, P<0.05; and β=1.68, P<0.01, respectively) within residential buffers. In GPS neighbourhoods, no associations were documented. Examining how food environments are associated with CVD risk and how differences in relationships vary by buffer types have the potential to shed light on determinants of CVD risk. Further research is needed to investigate these relationships, including refined measures of spatial accessibility/exposure, considering individual's mobility.

Factors associated with supermarket and convenience store closure: a discrete time spatial survival modelling approach

While there is a literature on the distribution of food stores across geographic and social space, much of this research uses cross-sectional data. Analyses attempting to understand whether the availability of stores across neighborhoods is associated with diet and/or health outcomes are limited by a lack of understanding of factors that shape the emergence of new stores and the closure of others. We used quarterly data on supermarket and convenience store locations spanning seven years (2006-2012) and tract-level census data in four US cities: Birmingham, Alabama; Chicago, Illinois; Minneapolis, Minnesota; San Francisco, California. A spatial discrete-time survival model was used to identify factors associated with an earlier and/or later closure time of a store. Sales volume was typically the strongest indicator of store survival. We identified heterogeneity in the association between tract-level poverty and racial composition with respect to store survival. Stores in high poverty, non-White tracts were often at a disadvantage in terms of survival length. The observed patterns of store survival varied by some of the same neighborhood sociodemographic factors associated with lifestyle and health outcomes, which could lead to confusion in interpretation in studies of the estimated effects of introduction of food stores into neighborhoods on health.

How Segregation Makes Us Fat: Food Behaviors and Food Environment as Mediators of the Relationship Between Residential Segregation and Individual Body Mass Index

OBJECTIVES

Racial residential segregation affects food landscapes that dictate residents' food environments and is associated with obesity risk factors, including individual dietary patterns and behaviors. We examine if food behaviors and environments mediate the association between segregation and body mass index (BMI).

METHODS

Non-Hispanic Whites and Blacks living in the St. Louis and Kansas City metro regions from 2012 to 2013 were surveyed on dietary behaviors, food environment, and BMI (n = 1,412). These data were combined with the CDC's modified retail food environment index and 2012 American Community Survey data to calculate racial segregation using various evenness and exposure indices. Multi-level mediation analyses were conducted to determine if dietary behavior and food environment mediate the association between racial residential segregation and individual BMI.

RESULTS

The positive association between racial segregation and individual BMI is partially mediated by dietary behaviors and fully mediated by food environments.

CONCLUSION

Racial segregation (evenness and exposure) is associated with BMI, mediated by dietary behaviors and food environment. Elements of the food environment, which form the context for dietary behaviors, are potential targets for interventions to reduce obesity in residentially segregated areas.

Validation of commercial business lists as a proxy for licensed alcohol outlets

Background

Studies of retail alcohol outlets are restricted to regions due to lack of U.S. national data. Commercial business lists (BL) offer a possible solution, but no data exists to determine if BLs could serve as an adequate proxy for license data. This paper compares geospatial measures of alcohol outlets derived from a commercial BL with license data for a large US state.

Methods

We validated BL data as a measure of off-premise alcohol outlet density and proximity compared to license data for 5528 randomly selected California residential addresses. We calculated three proximity measures (Euclidean distance, road network travel time and distance) and two density measures (kernel density estimation and the count within a 2-mile radius) for each dataset. The data was acquired in 2015 and processed and analyzed in 2015 and 2016.

Results

Correlations and reliabilities between density (correlation 0.98; Cronbach’s α 0.97–0.99) and proximity (correlations 0.77–0.86; α 0.87–0.92) measures were high. For proximity, BL data matched license in 55–57% of addresses, overstated distance in 19%, and understated in 24–26%.

Conclusions

BL data can serve as a reliable proxy for licensed alcohol outlets, thus extending the work that can be performed in studies on associations between retail alcohol outlets and drinking outcomes.

Quantifying the foodscape: A systematic review and meta-analysis of the validity of commercially available business data

This paper reviews studies of the validity of commercially available business (CAB) data on food establishments ("the foodscape"), offering a meta-analysis of characteristics associated with CAB quality and a case study evaluating the performance of commonly-used validity indicators describing the foodscape. Existing validation studies report a broad range in CAB data quality, although most studies conclude that CAB quality is "moderate" to "substantial". We conclude that current studies may underestimate the quality of CAB data. We recommend that future validation studies use density-adjusted and exposure measures to offer a more meaningful characterization of the relationship of data error with spatial exposure.

Measures of the food environment: A systematic review of the field, 2007-2015

BACKGROUND

Many studies have examined the relationship between the food environment and health-related outcomes, but fewer consider the integrity of measures used to assess the food environment. The present review builds on and makes comparisons with a previous review examining food environment measures and expands the previous review to include a more in depth examination of reliability and validity of measures and study designs employed.

METHODS

We conducted a systematic review of studies measuring the food environment published between 2007 and 2015. We identified these articles through: PubMed, Embase, Web of Science, PsycINFO, and Global Health databases; tables of contents of relevant journals; and the National Cancer Institute's Measures of the Food Environment website. This search yielded 11,928 citations. We retained and abstracted data from 432 studies.

RESULTS

The most common methodology used to study the food environment was geographic analysis (65% of articles) and the domination of this methodology has persisted since the last review. Only 25.9% of studies in this review reported the reliability of measures and 28.2% reported validity, but this was an improvement as compared to the earlier review. Very few of the studies reported construct validity. Studies reporting measures of the school or worksite environment have decreased since the previous review. Only 13.9% of the studies used a longitudinal design.

CONCLUSIONS

To strengthen research examining the relationship between the food environment and population health, there is a need for robust and psychometrically-sound measures and more sophisticated study designs.

A step-by-step approach to improve data quality when using commercial business lists to characterize retail food environments

BACKGROUND

Food environment characterization in health studies often requires data on the location of food stores and restaurants. While commercial business lists are commonly used as data sources for such studies, current literature provides little guidance on how to use validation study results to make decisions on which commercial business list to use and how to maximize the accuracy of those lists. Using data from a retrospective cohort study [Weight And Veterans' Environments Study (WAVES)], we (a) explain how validity and bias information from existing validation studies (count accuracy, classification accuracy, locational accuracy, as well as potential bias by neighborhood racial/ethnic composition, economic characteristics, and urbanicity) were used to determine which commercial business listing to purchase for retail food outlet data and (b) describe the methods used to maximize the quality of the data and results of this approach.

METHODS

We developed data improvement methods based on existing validation studies. These methods included purchasing records from commercial business lists (InfoUSA and Dun and Bradstreet) based on store/restaurant names as well as standard industrial classification (SIC) codes, reclassifying records by store type, improving geographic accuracy of records, and deduplicating records. We examined the impact of these procedures on food outlet counts in US census tracts.

RESULTS

After cleaning and deduplicating, our strategy resulted in a 17.5% reduction in the count of food stores that were valid from those purchased from InfoUSA and 5.6% reduction in valid counts of restaurants purchased from Dun and Bradstreet. Locational accuracy was improved for 7.5% of records by applying street addresses of subsequent years to records with post-office (PO) box addresses. In total, up to 83% of US census tracts annually experienced a change (either positive or negative) in the count of retail food outlets between the initial purchase and the final dataset.

DISCUSSION

Our study provides a step-by-step approach to purchase and process business list data obtained from commercial vendors. The approach can be followed by studies of any size, including those with datasets too large to process each record by hand and will promote consistency in characterization of the retail food environment across studies.

Short-Form Audit Instrument for Assessing Corner Store Healthfulness

PURPOSE

To develop a valid and feasible short-form corner store audit tool (SCAT) that could be used in-store or over the phone to capture the healthfulness of corner stores.

DESIGN

Nonexperimental.

SETTING

Four New Jersey cities.

SUBJECTS

Random selection of 229 and 96 corner stores in rounds 1 and 2, respectively.

MEASURES

An adapted version of the Nutrition Environment Measures Survey for Corner Stores (NEMS-CS) was used to conduct in-store audits. The 7-item SCAT was developed and used for round 2 phone audits.

ANALYSIS

Exploratory factor analysis and item response theory were used to develop the SCAT.

RESULTS

The SCAT was highly correlated with the adapted NEMS-CS ( r = .79). Short-form corner store audit tool scores placed stores in the same healthfulness categories as did the adapted NEMS-CS in 88% of the cases. Phone response matches indicated that store owners did not distinguish between 2% and low-fat milk and tended to round up the fruit and vegetable count to 5 if they had fewer varieties.

CONCLUSION

The SCAT discriminates between higher versus lower healthfulness scores of corner stores and is feasible for use as a phone audit tool.

Identification of Vape Shops in Two North Carolina Counties: An Approach for States without Retailer Licensing

Stores that sell electronic nicotine delivery systems (ENDS) as their primary product are a new phenomenon and often termed “vape shops”. While vape shops are now regulated by state and federal agencies, not all states maintain lists of vape shops in operation. Standard ways of identifying tobacco retailers through off-premise alcohol permits and business listing services may not identify vape shops. We used four online business listing services (i.e., Google Maps, ReferenceUSA, YellowPages.com, Yelp) to identify vape shops in two counties in North Carolina (NC). In one county, we also assessed four vaping web sites. We drove primary and secondary roads to physically validate the identified stores and attempt to identify stores not listed online. To assess the accuracy of the online searches, we calculated sensitivity and positive predictive values (PPVs). This research was conducted in spring and summer 2016 and identified 28 vape shops online. We confirmed 16 vape shops (seven in Pitt County, NC, USA, and nine in Durham County, NC, USA). Online searches ranged in sensitivity, 62.5%–81.3%, and PPVs ranged from 73.3% to 92.3%. Because of the range of sensitivity found among the business listing services, state policymakers should consider uniform licensing requirements for vape and tobacco retailers to more easily track retailers and ensure compliance with regulations.

Modified ground-truthing: an accurate and cost-effective food environment validation method for town and rural areas

Background

A major concern in food environment research is the lack of accuracy in commercial business listings of food stores, which are convenient and commonly used. Accuracy concerns may be particularly pronounced in rural areas. Ground-truthing or on-site verification has been deemed the necessary standard to validate business listings, but researchers perceive this process to be costly and time-consuming. This study calculated the accuracy and cost of ground-truthing three town/rural areas in Minnesota, USA (an area of 564 miles, or 908 km), and simulated a modified validation process to increase efficiency without comprising accuracy. For traditional ground-truthing, all streets in the study area were driven, while the route and geographic coordinates of food stores were recorded.

Results

The process required 1510 miles (2430 km) of driving and 114 staff hours. The ground-truthed list of stores was compared with commercial business listings, which had an average positive predictive value (PPV) of 0.57 and sensitivity of 0.62 across the three sites. Using observations from the field, a modified process was proposed in which only the streets located within central commercial clusters (the 1/8 mile or 200 m buffer around any cluster of 2 stores) would be validated. Modified ground-truthing would have yielded an estimated PPV of 1.00 and sensitivity of 0.95, and would have resulted in a reduction in approximately 88 % of the mileage costs.

Conclusions

We conclude that ground-truthing is necessary in town/rural settings. The modified ground-truthing process, with excellent accuracy at a fraction of the costs, suggests a new standard and warrants further evaluation.

Field validation of food outlet databases: the Latino food environment in North Carolina, USA

OBJECTIVE

Obtaining valid, reliable measures of food environments that serve Latino communities is important for understanding barriers to healthy eating in this at-risk population.

DESIGN

The primary aim of the study was to examine agreement between retail food outlet data from two commercial databases, Nielsen TDLinx (TDLinx) for food stores and Dun & Bradstreet (D&B) for food stores and restaurants, relative to field observations of food stores and restaurants in thirty-one census tracts in Durham County, NC, USA. We also examined differences by proportion of Hispanic population (

SETTING

One hundred and seventy-four food stores and 337 restaurants in Durham County, NC, USA.

RESULTS

We found that overall sensitivity of food store listings in TDLinx was higher (64 %) than listings in D&B (55 %). Twenty-five food stores were characterized by auditors as Latino food stores, with 20 % identified in TDLinx, 52 % in D&B and 56 % in both sources. Overall sensitivity of restaurants (68 %) was higher than sensitivity of Latino restaurants (38 %) listed in D&B. Sensitivity did not differ substantially by Hispanic composition of neighbourhoods.

CONCLUSIONS

Our findings suggest that while TDLinx and D&B commercial data sources perform well for total food stores, they perform less well in identifying small and independent food outlets, including many Latino food stores and restaurants.

Collapse

Key Words

• Latino

Collapse

MESH Headings

• Databases, Factual
• Diet/economics
• Diet/ethnology
• Food Supply/economics
• Hispanic or Latino
• Humans
• Language
• North Carolina
• Residence Characteristics
• Restaurants

Collapse

Grants

• R01-HL 114091 NHLBI NIH HHS
• T32 HD07168-33 NICHD NIH HHS
• N01-HC-48050 NHLBI NIH HHS
• DK56350 NIDDK NIH HHS
• HHSN268201300027C NHLBI NIH HHS
• N01HC95095 NHLBI NIH HHS
• N01-HC-48049 NHLBI NIH HHS
• R01 HL114091 NHLBI NIH HHS
• N01HC48048 NHLBI NIH HHS
• R01 HL104580 NHLBI NIH HHS
• HHSN268201300026C NHLBI NIH HHS
• P30 ES010126 NIEHS NIH HHS
• N01-HC-48047 NHLBI NIH HHS
• N01HC48049 NHLBI NIH HHS
• N01-HC-95095 NHLBI NIH HHS
• HHSN268201300025C NHLBI NIH HHS
• HHSN268200900041C NHLBI NIH HHS
• HHSN268201300028C NHLBI NIH HHS
• T32 DK007686 NIDDK NIH HHS
• N01HC48050 NHLBI NIH HHS
• N01HC48047 NHLBI NIH HHS
• T32 HD007168 NICHD NIH HHS
• P30 DK056350 NIDDK NIH HHS
• N01-HC-48048 NHLBI NIH HHS
• R24 HD050924 NICHD NIH HHS
• HHSN268201300029C NHLBI NIH HHS
• R01-HL104580 NHLBI NIH HHS
• P2C HD050924 NICHD NIH HHS

Collapse

Affiliation(s)

Pasquale E Rummo University of North Carolina at Chapel Hill, Carolina Population Center, CB# 8120, University Square, 123 West Franklin Street, Chapel Hill, NC 27516-2524, USA
Penny Gordon-Larsen University of North Carolina at Chapel Hill, Carolina Population Center, CB# 8120, University Square, 123 West Franklin Street, Chapel Hill, NC 27516-2524, USA Department of Nutrition, Gillings School of Global Public Health, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA
Sandra S Albrecht University of North Carolina at Chapel Hill, Carolina Population Center, CB# 8120, University Square, 123 West Franklin Street, Chapel Hill, NC 27516-2524, USA Department of Nutrition, Gillings School of Global Public Health, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA

Collapse

Rural and Remote Food Environments and Obesity

Within most developed countries, rural residents are more likely to be obese and overweight compared to their urban counterparts. Studies of specific rural communities have found that the limited availability of healthy foods in the community and home as well as individual characteristics and preferences contribute to poor diet and overweight. The rural food environment is varied and may be affected by climate, regional and cultural preferences, transportation access, and remoteness among other factors. Given this diversity and the vulnerabilities of rural residents, who are more likely to have low-income, substandard housing or low educational attainment compared to their urban counterparts, policy and programmatic interventions should target specific needs and communities. This review will describe the rural community, home, and individual food environments and what is known about their roles in healthy eating.

Spatial-temporal modeling of neighborhood sociodemographic characteristics and food stores

The literature on food stores, neighborhood poverty, and race/ethnicity is mixed and lacks methods of accounting for complex spatial and temporal clustering of food resources. We used quarterly data on supermarket and convenience store locations from Nielsen TDLinx (Nielsen Holdings N.V., New York, New York) spanning 7 years (2006-2012) and census tract-based neighborhood sociodemographic data from the American Community Survey (2006-2010) to assess associations between neighborhood sociodemographic characteristics and food store distributions in the Metropolitan Statistical Areas (MSAs) of 4 US cities (Birmingham, Alabama; Chicago, Illinois; Minneapolis, Minnesota; and San Francisco, California). We fitted a space-time Poisson regression model that accounted for the complex spatial-temporal correlation structure of store locations by introducing space-time random effects in an intrinsic conditionally autoregressive model within a Bayesian framework. After accounting for census tract-level area, population, their interaction, and spatial and temporal variability, census tract poverty was significantly and positively associated with increasing expected numbers of supermarkets among tracts in all 4 MSAs. A similar positive association was observed for convenience stores in Birmingham, Minneapolis, and San Francisco; in Chicago, a positive association was observed only for predominantly white and predominantly black tracts. Our findings suggest a positive association between greater numbers of food stores and higher neighborhood poverty, with implications for policy approaches related to food store access by neighborhood poverty.

Geographic scale matters in detecting the relationship between neighbourhood food environments and obesity risk: an analysis of driver license records in Salt Lake County, Utah

OBJECTIVES

Empirical studies of the association between neighbourhood food environments and individual obesity risk have found mixed results. One possible cause of these mixed findings is the variation in neighbourhood geographic scale used. The purpose of this paper was to examine how various neighbourhood geographic scales affected the estimated relationship between food environments and obesity risk.

DESIGN

Cross-sectional secondary data analysis.

SETTING

Salt Lake County, Utah, USA.

PARTICIPANTS

403,305 Salt Lake County adults 25-64 in the Utah driver license database between 1995 and 2008.

ANALYSIS

Utah driver license data were geo-linked to 2000 US Census data and Dun & Bradstreet business data. Food outlets were classified into the categories of large grocery stores, convenience stores, limited-service restaurants and full-service restaurants, and measured at four neighbourhood geographic scales: Census block group, Census tract, ZIP code and a 1 km buffer around the resident's house. These measures were regressed on individual obesity status using multilevel random intercept regressions.

OUTCOME

Obesity.

RESULTS

Food environment was important for obesity but the scale of the relevant neighbourhood differs for different type of outlets: large grocery stores were not significant at all four geographic scales, limited-service restaurants at the medium-to-large scale (Census tract or larger) and convenience stores and full-service restaurants at the smallest scale (Census tract or smaller).

CONCLUSIONS

The choice of neighbourhood geographic scale can affect the estimated significance of the association between neighbourhood food environments and individual obesity risk. However, variations in geographic scale alone do not explain the mixed findings in the literature. If researchers are constrained to use one geographic scale with multiple categories of food outlets, using Census tract or 1 km buffer as the neighbourhood geographic unit is likely to allow researchers to detect most significant relationships.

Do minority and poor neighborhoods have higher access to fast-food restaurants in the United States?

BACKGROUND

Disproportionate access to unhealthy foods in poor or minority neighborhoods may be a primary determinant of obesity disparities. We investigated whether fast-food access varies by Census block group (CBG) percent black and poverty.

METHODS

We measured the average driving distance from each CBG population-weighted centroid to the five closest top ten fast-food chains and CBG percent black and percent below poverty.

RESULTS

Among 209,091 CBGs analyzed (95.1% of all US CBGs), CBG percent black was positively associated with fast-food access controlling for population density and percent poverty (average distance to fast-food was 3.56 miles closer (95% CI: -3.64, -3.48) in CBGs with the highest versus lowest quartile of percentage of black residents). Poverty was not independently associated with fast-food access. The relationship between fast-food access and race was stronger in CBGs with higher levels of poverty (p for interaction <0.0001).

CONCLUSIONS

Predominantly black neighborhoods had higher access to fast-food while poverty was not an independent predictor of fast-food access.

Field validation of secondary data sources for enumerating retail tobacco outlets in a state without tobacco outlet licensing

Identifying tobacco retail outlets for U.S. FDA compliance checks or calculating tobacco outlet density is difficult in the 13 States without tobacco retail licensing or where licensing lists are unavailable for research. This study uses primary data collection to identify tobacco outlets in three counties in a non-licensing state and validate two commercial secondary data sources. We calculated sensitivity and positive predictive values (PPV) to examine the evidence of validity for two secondary data sources, and conducted a geospatial analysis to determine correct allocation to census tract. ReferenceUSA had almost perfect sensitivity (0.82) while Dun & Bradstreet (D&B) had substantial sensitivity (0.69) for identifying tobacco outlets; combined, sensitivity improved to 0.89. D&B identified fewer "false positives" with a PPV of 0.82 compared to 0.71 for ReferenceUSA. More than 90% of the outlets identified by ReferenceUSA were geocoded to the correct census tract. Combining two commercial data sources resulted in enumeration of nearly 90% of tobacco outlets in a three county area. Commercial databases appear to provide a reasonably accurate way to identify tobacco outlets for enforcement operations and density estimation.

The intersection of neighborhood racial segregation, poverty, and urbanicity and its impact on food store availability in the United States

BACKGROUND

Food store availability may determine the quality of food consumed by residents. Neighborhood racial residential segregation, poverty, and urbanicity independently affect food store availability, but the interactions among them have not been studied.

PURPOSE

To examine availability of supermarkets, grocery stores, and convenience stores in US census tracts according to neighborhood racial/ethnic composition, poverty, and urbanicity.

METHODS

Data from 2000 US Census and 2001 InfoUSA food store data were combined and multivariate negative binomial regression models employed.

RESULTS

As neighborhood poverty increased, supermarket availability decreased and grocery and convenience stores increased, regardless of race/ethnicity. At equal levels of poverty, Black census tracts had the fewest supermarkets, White tracts had the most, and integrated tracts were intermediate. Hispanic census tracts had the most grocery stores at all levels of poverty. In rural census tracts, neither racial composition nor level of poverty predicted supermarket availability.

CONCLUSIONS

Neighborhood racial composition and neighborhood poverty are independently associated with food store availability. Poor predominantly Black neighborhoods face a double jeopardy with the most limited access to quality food and should be prioritized for interventions. These associations are not seen in rural areas which suggest that interventions should not be universal but developed locally.

Variation in low food access areas due to data source inaccuracies

Several spatial measures of community food access identifying so called "food deserts" have been developed based on geospatial information and commercially-available, secondary data listings of food retail outlets. It is not known how data inaccuracies influence the designation of Census tracts as areas of low access. This study replicated the U.S. Department of Agriculture Economic Research Service (USDA ERS) food desert measure and the Centers for Disease Control and Prevention (CDC) non-healthier food retail tract measure in two secondary data sources (InfoUSA and Dun & Bradstreet) and reference data from an eight-county field census covering169 Census tracts in South Carolina. For the USDA ERS food deserts measure accuracy statistics for secondary data sources were 94% concordance, 50-65% sensitivity, and 60-64% positive predictive value (PPV). Based on the CDC non-healthier food retail tracts both secondary data demonstrated 88-91% concordance, 80-86% sensitivity and 78-82% PPV. While inaccuracies in secondary data sources used to identify low food access areas may be acceptable for large-scale surveillance, verification with field work is advisable for local community efforts aimed at identifying and improving food access.

Validity of secondary retail food outlet data: a systematic review

CONTEXT

Improving access to healthy foods is a promising strategy to prevent nutrition-related chronic diseases. To characterize retail food environments and identify areas with limited retail access, researchers, government programs, and community advocates have primarily used secondary retail food outlet data sources (e.g., InfoUSA or government food registries). To advance the state of the science on measuring retail food environments, this systematic review examined the evidence for validity reported for secondary retail food outlet data sources for characterizing retail food environments.

EVIDENCE ACQUISITION

A literature search was conducted through December 31, 2012, to identify peer-reviewed published literature that compared secondary retail food outlet data sources to primary data sources (i.e., field observations) for accuracy of identifying the type and location of retail food outlets. Data were analyzed in 2013.

EVIDENCE SYNTHESIS

Nineteen studies met the inclusion criteria. The evidence for validity reported varied by secondary data sources examined, primary data-gathering approaches, retail food outlets examined, and geographic and sociodemographic characteristics. More than half of the studies (53%) did not report evidence for validity by type of food outlet examined and by a particular secondary data source.

CONCLUSIONS

Researchers should strive to gather primary data but if relying on secondary data sources, InfoUSA and government food registries had higher levels of agreement than reported by other secondary data sources and may provide sufficient accuracy for exploring these associations in large study areas.

Business list vs ground observation for measuring a food environment: saving time or waste of time (or worse)?

In food-environment research, an alternative to resource-intensive direct observation on the ground has been the use of commercial business lists. We sought to determine how well a frequently used commercial business list measures a dense urban food environment like the Bronx, NY. On 155 Bronx street segments, investigators compared two different levels for matches between the business list and direct ground observation: lenient (by business type) and strict (by business name). For each level of matching, researchers calculated sensitivities and positive predictive values (PPVs) for the business list overall and by broad business categories: General Grocers (eg, supermarkets), Specialty Food Stores (eg, produce markets), Restaurants, and Businesses Not Primarily Selling Food (eg, newsstands). Even after cleaning the business list (eg, for cases of multiple listings at a single location), and allowing for inexactness in listed street addresses and spellings of business names, the overall performance of the business list was poor. For strict matches, the business list had an overall sensitivity of 39.3% and PPV of 45.5%. Sensitivities and PPVs by broad business categories were not meaningfully different from overall values, although sensitivity for General Grocers and PPV for Specialty Food Stores were particularly low: 26.2% and 32%, respectively. For lenient matches, sensitivities and PPVs were somewhat higher but still poor: 52.4% to 60% and 60% to 75%, respectively. The business list is inadequate to measure the actual food environment in the Bronx. If results represent performance in other settings, findings from prior studies linking food environments to diet and diet-related health outcomes using such business lists are in question, and future studies of this type should avoid relying solely on such business lists.