Field Validation of Commercially Available Food Retailer Data in the Netherlands

Socio-Spatial Segregation of Unhealthy Food Environments across Public Schools in Santiago, Chile

Santiago, Chile is a very segregated city, with higher childhood obesity rates observed in vulnerable areas. We compared the counts and proximity of unhealthy food outlets (UFOs) around a 400 m buffer of 443 public schools (municipal and subsidized) located in socioeconomically diverse neighborhoods in 14 municipalities of Santiago. This was a cross-sectional study in which the socioeconomic status (SES) of the population living inside the buffer was classified as middle-high, middle, and low. We used the Kruskal-Wallis test for comparisons of density and proximity between type of school, SES, and population density. We used a negative binomial model (unadjusted and adjusted by population density) to determine the expected change in counts of UFOs by SES, which was compared to the reference (middle-high). Low SES neighborhoods had significantly more counts of UFOs, and these were located much closer to schools. Low and middle SES neighborhoods had an 88% and 48% higher relative risk of having UFOs compared to middle-high SES areas; (IRR = 1.88; 95% CI 1.59-2.23) and (IRR = 1.48; 95% CI 1.20-1.82), respectively. A socio-spatial segregation of UFOs associated with childhood obesity across public schools was observed in Santiago.

The interaction between the community food environment and cooking skills in association with diet-related outcomes in Dutch adults

OBJECTIVE

We examined whether associations between the food environment, frequency of home cooking, diet quality and BMI were modified by the level of cooking skills.

DESIGN

Cross-sectional study using linear and modified Poisson regression models adjusted for age, sex, energy intake, education, income, household size and urbanisation. The frequency of home cooking was categorised into <6 and 6-7 d. Diet quality was based on a validated Dutch healthy diet index (0-150 points). Count of restaurants and food stores were determined by their count in a 1000m buffer around home and work. Cooking skills (score 1-5) were assessed using a validated questionnaire and added as interaction term.

SETTING

The Netherlands.

PARTICIPANTS

1461 adults aged 18-65 years.

RESULTS

Count of restaurants and food stores were not associated with the frequency of home cooking. A 10-unit higher count of food stores was associated with a higher diet quality (β: 0·58 (95 % CI (0·04, 1·12)), and a 10-unit higher count of restaurants was associated with a lower BMI kg/m2 (β: -0·02 (95 % CI (-0·04, -0·004)). Better cooking skills were associated with a higher likelihood of cooking 6-7 d compared with <6 d (risk ratio: 1·24 (95 % CI (1·16, 1·31)) and a higher diet quality (β: 4·45 (95 % CI (3·27, 5·63)) but not with BMI. We observed no interaction between the food environment and cooking skills (P-for-interaction > 0·1).

CONCLUSIONS

Exposure to food stores was associated with a higher diet quality and exposure to restaurants with a lower BMI. Better cooking skills were associated with a higher frequency of home cooking and better diet quality but did not modify associations with the food environment. Future studies should explore different approaches to understand how individuals interact with their food environment.

A detour for snacks and beverages? A cross-sectional assessment of selective daily mobility bias in food outlet exposure along the commuting route and dietary intakes

The evidence of selective daily mobility bias distorting exposure-health associations is limited. Using 7-day smartphone-based global positioning system (GPS) tracking data for 67 Dutch adults aged 25-45, we conducted paired Wilcoxon tests to compare the absolute and relative exposure to food outlets along actual and modelled commuting routes. We fitted Tobit regressions to examine their associations with three daily snack and soft drink intake outcomes. We found significant differences in absolute food outlet exposure between two types of routes. Adjusted regression analyses yielded unexpected associations between dietary intakes and food outlet exposures. Our results suggested no evidence of a selective daily mobility bias in the association between the food environment along commuting routes and adults' snacks and soft drink consumption in this sample.

The association between the food environment and adherence to healthy diet quality: the Maastricht Study

OBJECTIVE

The purpose of this study is to determine if healthier neighbourhood food environments are associated with healthier diet quality.

DESIGN

This was a cross-sectional study using linear regression models to analyse data from the Maastricht Study. Diet quality was assessed using data collected with a FFQ to calculate the Dutch Healthy Diet (DHD). A buffer zone encompassing a 1000 m radius was created around each participant home address. The Food Environment Healthiness Index (FEHI) was calculated using a Kernel density analysis within the buffers of available food outlets. The association between the FEHI and the DHD score was analysed and adjusted for socio-economic variables.

SETTING

The region of Maastricht including the surrounding food retailers in the Netherlands.

PARTICIPANTS

7367 subjects aged 40-75 years in the south of the Netherlands.

RESULTS

No relationship was identified between either the FEHI (B = 0·62; 95 % CI = -2·54, 3·78) or individual food outlets, such as fast food (B = -0·07; 95 % CI = -0·20, 0·07) and diet quality. Similar null findings using the FEHI were identified at the 500 m (B = 0·95; 95 % CI = -0·85, 2·75) and 1500 m (B = 1·57; 95 % CI = -3·30, 6·44) buffer. There was also no association between the food environment and individual items of the DHD including fruits, vegetables and sugar-sweetened beverages.

CONCLUSION

The food environment in the Maastricht area appeared marginally unhealthy, but the differences in the food environment were not related to the quality of food that participants reported as intake.

The quality of OpenStreetMap food-related point-of-interest data for use in epidemiological research

We assessed the quality of food-related OpenStreetMap (OSM) data in urban areas of five European countries. We calculated agreement statistics between point-of-interests (POIs) from OSM and from Google Street View (GSV) in five European regions. We furthermore assessed correlations between exposure measures (distance and counts) from OSM data and administrative data from local data sources on food environment data in three European countries. Agreement between POI data in OSM compared to GSV was poor, but correlations were moderate to high between exposures from OSM and local data sources. OSM data downloaded in 2020 seems to be an acceptable source of data for generating count-based food exposure measures for research in selected European regions.

Can mindful eating buffer against the influence of neighbourhood fast food exposure on unhealthy food intake?

A food-abundant environment is associated with unhealthy food intake, but not everyone is affected to the same degree. Mindful eating, which is eating with attention and awareness, has been associated with less external eating and less food cravings, and could act as a protective factor against influences from the food environment. The current study aimed to investigate whether the association between exposure to fast-food around the home and unhealthy food intake was moderated by mindful eating. The study was conducted in 1086 Dutch adults of 55 years and older of the Longitudinal Aging Study Amsterdam study. The mindful eating domains (Mindful Eating Behavior Scale) were tested as moderating variables in the linear regression models with absolute and relative density of fast-food outlets in the neighbourhood (400, 800 and 1600m) as independent variables and unhealthy food intake (snacks (g/d)) and saturated fat as a percentage of total energy intake (en%)) as dependent variable. Bootstrapping with 5000 samples using the pick-a-point approach showed that after adjustments, only two out of 48 interactions terms were statistically significant: Eating with Awareness (EwA) and Eating without Distraction (EwD) moderated the positive association between the relative density of fast-food outlets and saturated fat (en%) respectively in a buffer of 800m (interaction EwA: B = -0.84, 95% CI [-1.46; -0.22]) and in a buffer of 1600m (interaction EwD: B = -0.82, 95% CI [-1.61; -0.04]). The results of the current study indicate that mindful eating cannot buffer against the influence of the fast-food abundant environment on unhealthy food intake. Future research is needed to confirm these findings, for example in younger populations.

Survey-derived activity space-based exposures to fast food outlets and their cross-sectional associations with use of fast food outlets, diet quality and BMI

There is a need for conceptual and methodological innovation in food environment-health research. We compared different operationalizations of survey-derived activity space exposures to fast food outlets (FFOs) in associations with use of FFO, diet quality and body mass index (BMI). FFO exposure was determined for home, work and a maximum of sixteen other locations reported by 1728 Dutch adults. Considerable differences in count of FFO between locations were found. Adjusted linear regression analyses resulted in small, unexpected associations with use of FFO, diet quality and BMI, whereby the strength of associations differed between exposure measures. Using home and work areas may be a cost-efficient compromise to capture large parts of the exposure to FFOs.

The Changing Landscape of Food Deserts and Swamps over More than a Decade in Flanders, Belgium

Food deserts and swamps have previously been mostly studied in Anglo-Saxon countries such as the USA and Great Britain. This research is one of the first studies to map food deserts and swamps in a mainland European, densely populated but heavily fragmented region such as Flanders. The evolution of food deserts and swamps between 2008 and 2020 was assessed. Special focus was given to areas where high numbers of elderly, young people and/or families with low income live. Food deserts were calculated based on supermarket access within 1000 m and bus stop availability, while food swamps were calculated using the Modified Food Environment Retail Index. The main cause behind the formation of food deserts in Flanders is its rapidly aging population. Food deserts with a higher number of older people increased from 2.5% to 3.1% of the residential area between 2008 and 2020, housing 2.2% and 2.8% of the population, respectively. Although the area that could become a food desert in the future due to these sociospatial and demographic evolutions is large, food deserts are currently a relatively small problem in Flanders in comparison to the widespread existence of food swamps. Unhealthy retailers outnumbered healthy retailers in 74% of residential areas in 2020, housing 88.2% of the population. These food swamps create an environment where unhealthy food choices predominate. Residential areas with a higher number of elderly people, young people and families with low incomes had healthier food environments than Flanders as a whole, because these areas are mostly found in dense urban centers where the ratio of healthy food retailers to all retailers is higher. This research showed that food deserts and swamps could be a growing problem in European regions with a high population density that experience the high pressures of competing land uses.

The CIVISANO protocol: a mixed-method study about the role of objective and perceived environmental factors on physical activity and eating behavior among socioeconomically disadvantaged adults

BACKGROUND

Overweight and obesity have a strong socioeconomic profile. Unhealthy behaviors like insufficient physical activity and an unbalanced diet, which are causal factors of overweight and obesity, tend to be more pronounced in socioeconomically disadvantaged groups in high income countries. The CIVISANO project aims to identify objective and perceived environmental factors among different socioeconomic population groups that impede or facilitate physical activity and healthy eating behavior in the local context of two peri-urban Flemish municipalities in Belgium. We also aim to identify and discuss possible local interventions and evaluate the participatory processes of the project.

METHODS

This study (2020-2023) will use community-based participatory tools, involving collaborative partnerships with civic and stakeholder members of the community and regular exchanges among all partners to bridge knowledge development and health promotion for socioeconomically disadvantaged citizens. Furthermore, a mixed-methods approach will be used. A population survey and geographic analysis will explore potential associations between the physical activity and eating behaviors of socioeconomically disadvantaged adults (25-65 years old) and both their perceived and objective physical, food and social environments. Profound perceptive context information will be gathered from socioeconomically disadvantaged adults by using participatory methods like photovoice, walk-along, individual map creation and group model building. An evaluation of the participatory process will be conducted simultaneously.

DISCUSSION

The CIVISANO project will identify factors in the local environment that might provoke inequities in adopting a healthy lifestyle. The combination of perceived and objective measures using validated strategies will provide a robust assessment of the municipality environment. Through this analysis, the project will investigate to what extent community engagement can be a useful strategy to reduce health inequities. The strong knowledge exchange and capacity-building in a local setting is expected to contribute to our understanding of how to maximize research impact in this field and generate evidence about potential linkages between a health enhancing lifestyle among socioeconomically disadvantaged groups and their physical, food and social environments.

Validation of retail food outlet data from a Danish government inspection database

BACKGROUND

Globally, unhealthy diet is one of the leading global risks to health, thus it is central to consider aspects of the food environment that are modifiable and may enable healthy eating. Food retail data can be used to present and facilitate analyses of food environments that in turn may direct strategies towards improving dietary patterns among populations. Though food retail data are available in many countries, their completeness and accuracy differ.

METHODS

We applied a systematically name-based procedure combined with a manual procedure on Danish administrative food retailer data (i.e. the Smiley register) to identify, locate and classify food outlets. Food outlets were classified into the most commonly used classifications (i.e. fast food, restaurants, convenience stores, supermarkets, fruit and vegetable stores and miscellaneous) each divided into three commonly used definitions; narrow, moderate and broad. Classifications were based on branch code, name, and/or information on the internal and external appearance of the food outlet. From ground-truthing we validated the information in the register for its sensitivity and positive predictive value.

RESULTS

In 361 randomly selected areas of the Capital region of Denmark we identified a total of 1887 food outlets compared with 1861 identified in the register. We obtained a sensitivity of 0.75 and a positive predictive value of 0.76. Across classifications, the positive predictive values varied with highest values for the moderate and broad definitions of fast food, convenience stores and supermarkets (ranging from 0.89 to 0.97).

CONCLUSION

Information from the Smiley Register is considered to be representative to the Danish food environment and may be used for future research.

The Food Environment Around Primary Schools in a Diverse Urban Area in the Netherlands: Linking Fast-Food Density and Proximity to Neighbourhood Disadvantage and Childhood Overweight Prevalence

In the Netherlands, the neighbourhood food environment has received little attention in initiatives to combat overweight/obesity. This study maps the food environment around primary schools in The Hague, The Netherlands, and examines associations between neighbourhood disadvantage, the school food environment and childhood overweight using Geographic Information Systems (GIS). Linear regression analyses were performed to test the association between schools' disadvantage scores (proxy for neighbourhood disadvantage) and relative fast-food density within 400 m and 1000 m and fast-food proximity. Univariable and multivariable linear regression analyses were used to test the association between the school food environment and overweight prevalence among children in the respective sub-district in which the schools is found. Multivariable analyses were adjusted for the schools' disadvantage scores. Results show that fast-food outlets were available around most primary schools. Schools in disadvantaged neighbourhoods were closer to and surrounded by a higher number of fast-food restaurants, grillrooms and kebab shops. On the sub-district level, the density of such fast-food outlets was associated with overweight prevalence among children. These findings highlight the importance of national and local policies to improve the food environment, particularly in disadvantaged neighbourhoods.

The role of material and psychosocial resources in explaining socioeconomic inequalities in diet: A structural equation modelling approach

We examined whether material and psychosocial resources may explain socioeconomic differences in diet quality. Cross-sectional survey data from 1461 Dutch adults (42.5 (SD 13.7) years on average and 64% female) on socio-demographics, diet quality, psychosocial factors and perceptions of and objective healthiness of the food environment were used in a structural equation model to examine mediating pathways. Indicators for socioeconomic position (SEP) were income, educational, and occupational level and the 2015 Dutch Healthy Diet (DHD15) index assessed diet quality. Material resources included food expenditure, perceptions of healthy food accessibility and healthfulness of the food retail environment. Psychosocial resources were cooking skills, resilience to unhealthy food environments, insensitivity to food cues and healthy eating habits. Higher SEP was associated with better diet quality; Beducation 8.5 (95%CI 6.7; 10.3), Bincome 5.8 (95%CI 3.7; 7.8) and Boccupation 7.5 (95%CI 5.5; 9.4). Material resources did not mediate the association between SEP and diet quality and neither did the psychosocial resources insensitivity to food cues and eating habits. Cooking skills mediated between 13.3% and 19.0% and resilience to unhealthy food environments mediated between 5.9% and 8.6% of the relation between SEP and the DHD15-index. Individual-level factors such as cooking skills can only explain a small proportion of the SEP differences in diet quality. On top of other psychosocial and material resources not included in this study, it is likely that structural factors outside the individual, such as financial, work and living circumstances also play an important role.

A Proposed Research Agenda for Promoting Healthy Retail Food Environments in the East Asia-Pacific Region

PURPOSE OF REVIEW

This paper aimed to summarise existing literature on strategies to improve the healthiness of retail food environments in the East Asia and Pacific (EAP) region, and propose a prioritised research agenda on this topic.

RECENT FINDINGS

Little research on retail food environments has been conducted in the EAP region. Several approaches for measuring retail food environments were identified, although none have been tailored to the EAP context. A small number of policies and initiatives to promote healthy retail food environments have been implemented in EAP. Lessons learnt from successful implementation of initiatives in other regions could be applied in EAP. Retail food environments have a strong influence on food choices and health outcomes. Research can contribute to efforts to improve the healthiness of retail food environments in EAP by (1) describing the current state of retail food environments to highlight areas of good practice and concern and (2) identifying policies and initiatives that are likely to be effective, and mechanisms for their successful implementation.

Ultra-processed food consumption patterns among older adults in the Netherlands and the role of the food environment

PURPOSE

To describe the patterns of ultra-processed foods (UPFs) consumption in the Netherlands; to test if exposure to the food environment is associated with UPFs consumption; and if this association differed across educational levels and neighbourhood urbanisation.

METHODS

Cross-sectional study using 2015-data of 8104 older adults from the Dutch EPIC cohort. Proportion of UPFs consumption was calculated from a validated food-frequency questionnaire. Exposure to the food environment was defined as proximity and availability of supermarkets, fast-food restaurants, full-service restaurants, convenience stores, candy stores and cafés. Consumption of UPFs was expressed as both percentage of total grams and total kilocalories.

RESULTS

The study population was aged 70(± 10 SD) years and 80.5% was female. Average UPFs consumption was 17.8% of total food intake in grams and 37% of total energy intake. Those who consumed greater amounts of UPFs had a poorer overall diet quality. Adjusted linear regression models showed that closer proximity and larger availability to any type of food retailer was associated with lower UPFs consumption (both in grams and kilocalories). Somewhat stronger significant associations were found for proximity to restaurants (β = - 1.6%, 95% confidence interval (CI) = - 2.6; - 0.6), and supermarkets (β = - 2.2%, 95%CI = - 3.3; - 1.1); i.e., Individuals living within 500 m from the closest supermarket, as compared to 1500 m, had 2.6% less calories from UPFs. No differences were found on analyses stratified for urbanisation and education.

CONCLUSIONS

Using various measures of exposure to the food environment, we found that exposure to restaurants and supermarkets was associated with somewhat lower consumption of UPFs.

Residential exposure to fast-food restaurants and its association with diet quality, overweight and obesity in the Netherlands: a cross-sectional analysis in the EPIC-NL cohort

Background

Unhealthy food environments may contribute to unhealthy diets and risk of overweight and obesity through increased consumption of fast-food. Therefore, we aimed to study the association of relative exposure to fast-food restaurants (FFR) with overall diet quality and risk of overweight and obesity in a sample of older adults.

Methods

We analyzed cross-sectional data of the EPIC-NL cohort (n = 8,231). Data on relative FFR exposure was obtained through linkage of home address in 2015 with a retail outlet database. We calculated relative exposure to FFR by dividing the densities of FFR in street-network buffers of 400, 1000, and 1500 m around the home of residence by the density of all food retailers in the corresponding buffer. We calculated scores on the Dutch Healthy Diet 2015 (DHD15) index using data from a validated food-frequency questionnaire. BMI was categorized into normal weight (BMI < 25), overweight (25 ≤ BMI < 30), and obesity (BMI ≥ 30). We used multivariable linear regression (DHD15-index) and multinomial logistic regression (weight status), using quartiles of relative FFR exposure as independent variable, adjusting for lifestyle and environmental characteristics.

Results

Relative FFR exposure was not significantly associated with DHD15-index scores in the 400, 1000, and 1500 m buffers (β_Q4vsQ1= -0.21 [95 %CI: -1.12; 0.70]; β_Q4vsQ1= -0.12 [95 %CI: -1.10; 0.87]; β_Q4vsQ1 = 0.37 [95 %CI: -0.67; 1.42], respectively). Relative FFR exposure was also not related to overweight in consecutive buffers (OR_Q4vsQ1=1.10 [95 %CI: 0.97; 1.25]; OR_Q4vsQ1=0.97 [95 %CI: 0.84; 1.11]; OR_Q4vsQ1= 1.04 [95 %CI: 0.90–1.20]); estimates for obesity were similar to those of overweight.

Conclusions

A high proportion of FFR around the home of residence was not associated with diet quality or overweight and obesity in this large Dutch cohort of older adults. We conclude that although the food environment may be a determinant of food choice, this may not directly translate into effects on diet quality and weight status. Methodological improvements are warranted to provide more conclusive evidence.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12937-021-00713-5.

Socioeconomic inequalities in the food environment and body composition among school-aged children: a fixed-effects analysis

BACKGROUND

There is limited evidence regarding socioeconomic inequalities of exposure to the food environment and its contribution to childhood obesity.

METHODS

We used data from 4235 children from the Generation R Study, a large birth-cohort conducted in the city of Rotterdam, The Netherlands. We included 11,277 person-observations of body mass index (BMI) and 6240 person-observations of DXA-derived fat mass index (FMI) and fat-free mass index (FFMI) when children were between 4 and 14 years. We applied linear regression models to evaluate changes in the relative and absolute exposure of fast-food outlets, and the healthiness of the food environment within 400 m from home by maternal education. Furthermore, we used individual-level fixed-effects models to study changes in the food environment to changes in BMI, FMI and FFMI.

RESULTS

Children from lower educated mothers were exposed to more fast-food outlets at any time-point between the age of 4 and 14 years. Over a median period of 7.1 years, the absolute (0.6 fast-food outlet (95% CI: 0.4-0.8)) and relative (2.0%-point (95% CI: 0.7-3.4)) amount of fast-food outlets increased more for children from lower as compared to higher educated mothers. The food environment became more unhealthy over time, but no differences in trends were seen by maternal education level. Changes in the food environment were not associated with subsequent changes in BMI, FMI and FFMI. For children from lower educated mothers not exposed to fast-food at first, we found some evidence that the introduction of fast-food was associated with small increases in BMI.

CONCLUSIONS

Our findings provide evidence of widening inequalities in exposure to fast-food in an already poor food environment. Access to more fast-food outlets does not seem to have an additional impact on BMI in contemporary contexts with ubiquitous fast-food outlets.

Deep phenotyping meets big data: the Geoscience and hEalth Cohort COnsortium (GECCO) data to enable exposome studies in The Netherlands

Environmental exposures are increasingly investigated as possible drivers of health behaviours and disease outcomes. So-called exposome studies that aim to identify and better understand the effects of exposures on behaviours and disease risk across the life course require high-quality environmental exposure data. The Netherlands has a great variety of environmental data available, including high spatial and often temporal resolution information on urban infrastructure, physico-chemical exposures, presence and availability of community services, and others. Until recently, these environmental data were scattered and measured at varying spatial scales, impeding linkage to individual-level (cohort) data as they were not operationalised as personal exposures, that is, the exposure to a certain environmental characteristic specific for a person. Within the Geoscience and hEalth Cohort COnsortium (GECCO) and with support of the Global Geo Health Data Center (GGHDC), a platform has been set up in The Netherlands where environmental variables are centralised, operationalised as personal exposures, and used to enrich 23 cohort studies and provided to researchers upon request. We here present and detail a series of personal exposure data sets that are available within GECCO to date, covering personal exposures of all residents of The Netherlands (currently about 17 M) over the full land surface of the country, and discuss challenges and opportunities for its use now and in the near future.

The interplay between fast-food outlet exposure, household food insecurity and diet quality in disadvantaged districts

OBJECTIVE

The current study aimed to explore the interplay between food insecurity, fast-food outlet exposure and dietary quality in disadvantaged neighbourhoods.

DESIGN

In this cross-sectional study, main associations between fast-food outlet density and proximity, food insecurity status and dietary quality were assessed using Generalized Estimating Equation analyses. We assessed potential moderation by fast-food outlet exposure in the association between food insecurity status and dietary quality by testing for effect modification between food insecurity status and fast-food outlet density and proximity.

SETTING

A deprived urban area in the Netherlands.

PARTICIPANTS

We included 226 adult participants with at least one child below the age of 18 years living at home.

RESULTS

Fast-food outlet exposure was not associated with experiencing food insecurity (fast-food outlet density: b = -0·026, 95 % CI = -0·076; 0·024; fast-food outlet proximity: b = -0·003, 95 % CI = -0·033; 0·026). Experiencing food insecurity was associated with lower dietary quality (b = -0·48 per unit increase, 95 % CI = -0·94; -0·012). This association was moderated by fast-food outlet proximity (Pinteraction = 0·008), and stratified results revealed that the adverse effect of food insecurity on dietary quality was more pronounced for those with the nearest fast-food outlet located closer to the home.

CONCLUSIONS

Food insecurity but not fast-food outlet density is associated with dietary quality. However, the association between food insecurity and dietary quality may be modified by the food environment. These findings could inform policymakers to promote a healthier food environment including less fast-food outlets, with particular emphasis on areas with high percentages of food insecure households.