Tensorial Principal Component Analysis in Detecting Temporal Trajectories of Purchase Patterns in Loyalty Card Data: Retrospective Cohort Study.

BACKGROUND: Loyalty card data automatically collected by retailers provide an excellent source for evaluating health-related purchase behavior of customers. The data comprise information on every grocery purchase, including expenditures on product groups and the time of purchase for each customer. Such data where customers have an expenditure value for every product group for each time can be formulated as 3D tensorial data. OBJECTIVE: This study aimed to use the modern tensorial principal component analysis (PCA) method to uncover the characteristics of health-related purchase patterns from loyalty card data. Another aim was to identify card holders with distinct purchase patterns. We also considered the interpretation, advantages, and challenges of tensorial PCA compared with standard PCA. METHODS: Loyalty card program members from the largest retailer in Finland were invited to participate in this study. Our LoCard data consist of the purchases of 7251 card holders who consented to the use of their data from the year 2016. The purchases were reclassified into 55 product groups and aggregated across 52 weeks. The data were then analyzed using tensorial PCA, allowing us to effectively reduce the time and product group-wise dimensions simultaneously. The augmentation method was used for selecting the suitable number of principal components for the analysis. RESULTS: Using tensorial PCA, we were able to systematically search for typical food purchasing patterns across time and product groups as well as detect different purchasing behaviors across groups of card holders. For example, we identified customers who purchased large amounts of meat products and separated them further into groups based on time profiles, that is, customers whose purchases of meat remained stable, increased, or decreased throughout the year or varied between seasons of the year. CONCLUSIONS: Using tensorial PCA, we can effectively examine customers' purchasing behavior in more detail than with traditional methods because it can handle time and product group dimensions simultaneously. When interpreting the results, both time and product dimensions must be considered. In further analyses, these time and product groups can be directly associated with additional consumer characteristics such as socioeconomic and demographic predictors of dietary patterns. In addition, they can be linked to external factors that impact grocery purchases such as inflation and unexpected pandemics. This enables us to identify what types of people have specific purchasing patterns, which can help in the development of ways in which consumers can be steered toward making healthier food choices.

Purchase of ultra-processed foods in Norway: a repeated cross-sectional analysis of food sales in 2013 and 2019.

OBJECTIVE: A dietary pattern dominated by ultra-processed foods has been associated with non-communicable diseases in several studies. A previous study from 2013 found a high share of ultra-processed foods in Norwegian food sales. This study aimed to investigate the current share of ultra-processed foods in Norway and the development in expenditure on ultra-processed foods from 2013. DESIGN: A repeated cross-sectional analysis of scanner data from the Consumer Price Index from September 2013 and 2019 and an investigation of the processing degree according to the NOVA classification system. SETTING: Food sales in Norway. PARTICIPANTS: Norwegian grocery stores (n 180, for both time periods). RESULTS: The share of expenditure in 2019 was highest for ultra-processed foods (46·5 %) and minimally or unprocessed foods (36·3 %), followed by processed foods (8·5 %) and processed culinary ingredients (1·3 %). An increasing degree of processing was found for several of the food groups between 2013 and 2019; however, most effect sizes were weak. In 2019, soft drinks became the most frequently purchased food item, surpassing milk and cheese, with the highest expenditure in Norwegian grocery stores. Increases in expenditure on ultra-processed foods were mainly due to increased expenditures on soft drinks, sweets and potato products. CONCLUSIONS: A high share of expenditure on ultra-processed food was found in Norway, which may imply a high consumption of these foods. The change in expenditure of NOVA groups between 2013 and 2019 was small. Carbonated and non-carbonated soft drinks were the most frequently purchased products in Norwegian grocery stores and contributed to most of the expenditures.

Household Food Insecurity and Demographic Factors, Low Birth Weight and Stunting in Early Childhood: Findings from a Longitudinal Study in South Africa.

BACKGROUND: Low birthweight (LBW) as well as early childhood stunting are risk factors for increased childhood morbidity in low-and middle-income countries (LMIC). The Covid 19 pandemic has exacerbated food insecurity and unemployment globally, prompting concerns for maternal and child health. OBJECTIVES: We used data from the great recession of 2008 to examine the relationship between household food security and other risk factors with LBW and stunting using a longitudinal sample of South African women and their offspring. METHODS: Food security indicators, alcohol use, blood pressure and other characteristics were examined in relation to LBW (≤ 2500 g), stunting (height for age ≤ 2SD) and severe stunting (height for age ≤ 3SD). Regression modelling with clustering at maternal ID level were employed to adjust for maternal characteristics and women who gave birth more than once during the reference period. RESULTS: Birthweight data were available for 1173 children and height for age 1216 children. The prevalence of LBW was 14.7% while stunting and severe stunting was 17.8% and 14.5%. Child hunger in the household, maternal hypertension and alcohol use were associated with low birthweight. Food expenditure below the Stats SA poverty line and low dietary diversity was associated with stunting and severe stunting respectively. Maternal height and low birthweight were associated with both stunting and severe stunting. CONCLUSIONS FOR PRACTICE: Interventions that can improve household food security and nutritional status during the periconceptional and antenatal period may reduce the prevalence of low birthweight and subsequent stunting in low- and middle-income countries.

Dietary diversity, food insecurity and the double burden of malnutrition among children, adolescents and adults in South Africa: Findings from a national survey.

Childhood stunting remains a global public health problem. Many stunted children live in the same household as overweight or obese adults (the so-called double burden of malnutrition), evidence that quality as well as quantity of food is important. In recent years, food security measurement has shifted away from anthropometry (e.g., stunting) to experiential measures (e.g., self-reported hunger). However, given the continued problem of stunting, it is important that national surveys identify malnutrition. Objectives: To examine the associations between a variety of food security indicators, including dietary diversity, with adult, child (0-4 years) (5-9 years) and adolescent (10-17 years) anthropometry. To estimate the prevalence of double burden households. Methods: The study utilized cross-sectional data from the South African National Income Dynamics Survey NIDS (2008). We examined the associations between five food security indicators and anthropometry outcomes. The indicators were adult and child hunger in the household, self-reported household food sufficiency, food expenditure>60% of monthly expenditure and household dietary diversity. Multinomial and logistic regression models were employed to examine the associations with adult BMI categories and children's stunting and BMI. Results: The prevalence of stunting was 18.4% and the prevalence of wasting and overweight was 6.8 and 10.4%, respectively. Children <5 and adolescents with medium dietary diversity were significantly more likely to be stunted than children with high dietary diversity. Among children <5, child hunger and medium dietary diversity were significantly associated with wasting. None of the food security indicators were associated with stunting in children aged 5-9. Among stunted children, 70.2% lived with an overweight or obese adult. Among adults, increased dietary diversity increased the risk of overweight and obesity. Conclusion: Dietary diversity can be used as a proxy for poor nutritional status among children <5 years and adolescents but the relationship between dietary diversity and adult obesity is more complex. Given the double burden of malnutrition in many low- and middle-income countries, indicators of dietary quality remain important. These tools can be further refined to include an extra category for processed foods. Given the relative simplicity to collect this data, national surveys would be improved by its inclusion.

The cost of healthier and more sustainable food choices: Do plant-based consumers spend more on food?

Plant-based diets are often promoted as healthier and more sustainable and thus as a mechanism to achieve the targets proposed to mitigate climate change and noncommunicable diseases. However, plant-based diets can be perceived as more expensive than the common omnivorous diets, when considering the expensive novel meat substitutes and also the higher costs of fruits and vegetables, whose consumption is perceived to increase. Therefore, the present study assesses the question: Do plant-based consumers spend more on food compared to omnivorous consumers? Based on primary data (n = 1040) collected through an online survey, representative of the Portuguese population, through logistic regressions, it was possible to conclude that plant-based consumers, particularly vegan, are associated with lower food expenditures compared to omnivorous consumers. In fact, plant-based consumers are shown to spend less than all other consumers assessed. Food policies aligning healthiness and sustainability with affordability can deliver a major boost for the promotion of plant-based diets and help achieve the mitigation targets proposed.

Covid-19 pandemic and food security in India: Can authorities alleviate the disproportionate burden on the disadvantaged?

The present study examines covariates of food security and the impact of Covid-19 induced shocks, among households in India using a nationally representative survey. Using a 2SLS panel regression model, we find an important role of incomes, relative food prices, household characteristics, as well as mobility restrictions in response to the rising number of infections in a given region in explaining varying food expenditure shares prior to and during the Covid-19 pandemic. The disproportionate burden of the pandemic induced lockdowns on the disadvantaged and minorities calls for effective action on the part of policymakers to boost aggregate demand, fix supply chains and reduce food price volatility.

Positive effects of COVID-19 on food preparation and expenditure habits: a comparative study across three countries.

OBJECTIVE: This study seeks to empirically investigate how the changing eating habits affect health habits within three countries with entirely different cultures and diets to understand to what extent the pandemic may be responsible for these changes. DESIGN: Specifically, a questionnaire was conducted in China, Portugal and Turkey in early 2021. A series of statistical analyses were performed to identify how changes in individuals' eating habits have influenced their diets, considering the pandemic context and the varying cultural contexts where this research was performed. SETTING: A structured questionnaire form was developed and uploaded to an online platform with unique links for automatic distribution to respondents in each country. Data for the main survey were gathered between 3 January and 1 February 2021. PARTICIPANTS: Using snowball sampling, the authors leveraged their social networks by asking friends and colleagues to distribute the survey to potentially interested individuals. This distribution was stratified accordingly to the distribution of the population. The authors ultimately collected 319 useable surveys from China, 351 from Portugal and 449 from Turkey. RESULTS: The pandemic inspired healthier food habits, mostly because people have additional time to cook, shop differently for food and spend more money on groceries. CONCLUSIONS: The study suggests that aside from cultural values and dietary habits, the available time and the fear of the pandemic most explained the new eating habits. Several implications are provided for researchers and overall society in these three countries.

Impact of self-rationing on academic performance of college students.

OBJECTIVE: Self-rationing or one's ability to regulate allocation of resources for consumption can be an inefficient process. Despite the negative impact of self-rationing inefficiencies (eg unhealthy food choices), there is scarce research to understand this phenomenon. This study investigated college students' self-rationing efficiency of eating behavior and spending choices, and their relationship to academic performance. PARTICIPANTS AND METHODS: This study used a combination of a real-time experiment, self-reported surveys (n = 101), and interviews with college students (n = 12). Data was collected using two 24-hour recall periods, online surveys, and semi-structured interviews. RESULTS: College students demonstrated self-rationing inefficiency of eating behavior, but not of expenditure on food. Furthermore, there was a significant relationship between self-rationing inefficiencies in eating behavior and students' academic performance. CONCLUSIONS: This study directly investigated self-rationing efficiency amongst college students' food choices and food expenditure, and enhances our understanding of its impact on college students' academic performance.

Prevalence and determinants of anemia among resident female university students from Southern Punjab, Pakistan.

Health and nutritional status of the women of reproductive age (WRA) is expected to be influenced with women literacy status alongside other sociodemographic indicators. However, data are scant to validate if literate young women have lower incidence of anemia prevalence. Nexus to the problem stated, a cross-sectional study was conducted on 1,541 female resident university students (FRUS) aged 17-35 years from a public sector university located in Southern Punjab, Pakistan. Hemoglobin (Hb) led screening for anemia was performed followed by nutritional assessment and structured questionnaire-based sociodemographic and dietary assessment. The data generated were analyzed using independent t-test, Chi-square, and response surface regression models. Response rate for the prevalence of anemia in FRUS was 38 percent with mean Hb levels 10.5 g/dL. With a significant effect (p = .001) of participants' weight on anemia prevalence, 22.45 percent of the sample population was recorded as underweight. Sociodemographic and dietary parameters analyzed suggested low daily food expenditure (x2 = 20.59; p = .000) and reduced intake of meat (x2 = 12.14; p = .01), beans & pulses (x2 = 18.56; p = .001) to significantly influence rate of anemia prevalence in FRUS. The study concludes high prevalence rate of anemia among FRUS to strongly relate with students' low monthly stipend, little daily food expenditure, and substandard dietary quality.

Secular trends in Javanese adult height: the roles of environment and educational attainment.

BACKGROUND: Since its independence in 1945, Indonesia has undergone rapid socioeconomic development. The fastest changes occurred in Java, which is the main island where more than half of the Indonesian population lives. OBJECTIVE: This study aimed to analyze the secular trend in the height of adults living in Java born between 1953 and 1995 by comparing their residence (rural, small city, or large city) and considering factors that contribute to adult height. METHODS: The analysis used the following data: birth year, body height, weight, body mass index, sex, educational attainment, share of household food expenditures, and residence of 20- to 40-year-old men and women collected by the Indonesia Family Life Survey (IFLS) waves 1 to 5. Multiple linear regression was conducted to analyze several factors that influence adult height. Significance was set at p < 0.05 with a 95% confidence interval (CI). RESULTS: The study included 30,656 measurements of individuals born between 1953 and 1995 (54.9% female). Positive secular trends (95% CI) were observed for men and women: 1.3 (1.1-1.4) cm and 0.9 (0.8-1.0) cm per decade, respectively. Multiple linear regression analyses showed that, in addition to the year of birth, the adult height of both males and females was independently associated with level of education and share of household food expenditure. Stratifying the data into residence in rural areas, small cities, and large cities showed that education levels influenced the adult height of men and women living in all regions, whereas the influence of birth year and share of household food expenditure differed between areas and genders. CONCLUSIONS: We observed positive secular trends in the height of adults living in Java who were born between 1953 and 1995. The birth year, educational attainment, and share of household food expenditure significantly influenced adult height. A higher education level was consistently associated with taller adult height in both men and women living in rural areas, small cities, and large cities.

Disentangling Drivers of Food Waste in Households: Evidence from Nigeria.

Food waste is a burden on society in terms of the money wasted. There is limited information on the determinants of food waste and the amount lost to food waste by households as most previous studies were on post-harvest losses. Hence, determinants of food waste among households in Kogi West Senatorial District, Kogi State Nigeria were investigated. A three-stage sampling technique was used to select the respondents, while a structured questionnaire was used for data collection. Data were analyzed using Tobit regression and an equality test. The study revealed that food waste was higher in male headed households. The average monthly food waste proportion among urban households was significantly higher than that of rural households. The estimated amounts lost to food waste per month were ₦2103 and ₦5530 for the rural and urban households, respectively. These represented 7.2% and 13.1% of the total expenditure on food per month for rural and urban households, respectively. Among rural households, leftovers of food and lack of proper storage were the main reasons for food waste, while leftovers of food and preparation of food more than needed were the reasons for food waste among urban households. The sex of respondents, work experience, and monthly income influenced the proportion of food waste among rural households, while the dependency ratio, monthly income, and monthly food expenditure were the determinants of proportion of food waste among the urban households. Non-Governmental Organization efforts through sensitization campaigns focused on the need to reduce food waste, especially among urban households, would help to reduce the financial burden of food waste on households.

Assessment of the Cost of the Mediterranean Diet in a Low-Income Region: Adherence and Relationship with Available Incomes.

BACKGROUND: The Mediterranean Diet (MD) is recognized as heart-healthy, but the economic cost associated with this type of diet has scarcely been studied. The objective of the present study is to explore the cost and adherence of a low-income region population to the MD and its relationship with income. METHODS: A population-based study was carried out on 2,833 subjects between 25 and 79 years of age, 54% women, selected at random from the municipalities of Vegas Altas, La Siberia and La Serena in the province of Badajoz, Extremadura (Spain). Average monthly cost of each product included in the MD was computed and related to adherence to the MD using the Panagiotakos Index and average disposable income. RESULTS: The monthly median cost was 203.6€ (IQR: 154.04-265.37). Food-related expenditure was higher for men (p<0.001), age cohort between 45 and 54 years (p<0.013) and those living in urban areas (p<0.001). A positive correlation between food-related expenditure and the MD adherence was found. Monthly median cost represents 15% of average disposable income, ranging between 11% for the group with low MD adherence and 17% for the group with high MD adherence. CONCLUSIONS: The monthly cost of the MD was positively correlated with the degree of adherence to this dietary pattern. Given that the estimated monthly cost is similar to that of other Spanish regions with a higher income level, the economic effort required to be able to afford the Mediterranean diet is higher. This may represent a barrier to access, which should be analyzed in detail by public decision-makers.

Changing Food Consumption Patterns and Land Requirements for Food in the Six Geopolitical Zones in Nigeria.

Research on food consumption in Nigeria has mainly focused on food intake, household diversity, and purchasing power. We investigated a knowledge gap for food consumed by households and the land requirements for food resulting from household consumption patterns. The food consumed and the household size determine the land requirement for food. Therefore, a quantity-based analysis and a land demand methodology were applied to derive household food quantity and land requirements for food respectively. The results show that a greater percentage of household income is spent on cereals and starchy roots as the main source of calories and that cowpea is a secondary food option for households. In addition, households are changing their dietary intake from rice to maize and rice to cassava and yams as a cheaper alternative and experts' measurements of food security at the household level indicates that households in our study are moderately food insecure. Other findings show that the country's specific and per capita land requirements for food have gradually increased between 2000 and 2018. Across the six geopolitical zones, Northern regions with higher populations have high land requirements for food, especially for rice and maize (cereals), while Southern regions have high land requirements for cassava and yams (starchy roots) due to their respective consumption and household sizes. In addition, from our study, the land requirements for food show the actual cropland area of South South fed 5000 households. Consequently, a scenario analysis shows that the land requirements for food in our study exceeds the entire geographical area of Nigeria. Therefore, continued population growth without improved living standards and adequate food production output per hectare will further exacerbate food insecurity and land shortage in Nigeria.

The effect of COVID-19 on food sales.

Both COVID-19 pandemic-related restrictions and recessionary employment loss severely impacted US food sales during 2020. This article estimates the historical relationship between food expenditures and employment at the county level. Using these estimates, we simulate the impact of the loss of employment on food sales and find that, on average, employment loss increased food-at-home (FAH) sales by 1.3% and decreased food-away-from-home (FAFH) sales by 0.5% in 2020. We argue differences to the actual 4.8% increase in FAH sales and 19.5% decrease in FAFH sales in 2020 likely stem from the more drastic COVID-19 pandemic-related restrictions and behavioral changes.

Balancing a sustained pursuit of nutrition, health, affordability and climate goals: exploring the case of Indonesia.

BACKGROUND: To guide the transformation of food systems to provide for healthy and sustainable diets, countries need to assess their current diet and food supply in comparison to nutrition, health, affordability, and environmental goals. OBJECTIVES: We sought to compare Indonesia's food utilization to diets optimized for nutritional value and cost and to diets that are increasingly plant-based in order to meet further health and environmental goals, including the EAT-Lancet planetary health diet, to explore whether multiple goals could be achieved simultaneously. METHODS: We compared 13 dietary scenarios (2 current, 7 optimized, 3 increasingly plant-based, 1 EAT-Lancet) for nutrient content, cost, greenhouse gas emissions (GHGe), and water footprints, using the FAO food balance sheet, Indonesia Household Income and Expenditure Survey household food expenditure, food composition, life cycle assessment, food losses, and trade data. RESULTS: The diversity of modeled scenarios was higher than that of current consumption, reflecting nutritional deficiencies underlying Indonesia's burden of different forms of malnutrition. Nutrient intake targets were met best by nutrient- and cost-optimized diets, followed by the EAT-Lancet diet. Those diets also had high GHGe, although less than 40% of a scenario in which Indonesia would adopt a typical high-income country's diet. Only the low food chain diet had a GHGe below the 2050 target set by the EAT-Lancet commission. Its nutrient content was comparable to that of a no-dairy diet, slightly above those of fish-and-poultry and current diets, and somewhat below those of the EAT-Lancet diets. To meet nutrient needs, some animal-source foods had to be included. Costs of all except the optimized diets were above the current national average food expenditure. No scenario met all goals simultaneously. CONCLUSIONS: Indonesia's consumption of rice and unhealthy foods should decrease; food production, trade, and processing should prioritize diversification, (bio)fortification, and limiting environmental impacts; and consumer and institutional demands for healthy, nutritious, and sustainable foods should be stimulated. More granular data and tools are required to develop and assess more detailed scenarios to achieve multiple goals simultaneously.

Association of food expenditure with life expectancy in the United States, 2001-2014.

OBJECTIVES: A healthy diet is associated with reduced risk for premature death; however, data on the association between food expenditure and life expectancy are scarce. The aim of this study was to determine the association of expenditure on food groups with life expectancy in men and women. METHODS: This study used data from 1548 counties, representing >90% of the population of the United States from 2001 through the end of 2014. Multivariable adjusted Bayesian generalized linear models were used to assess the association of county-level expenditure on nine groups of individual food and combined healthy and unhealthy foods, and a constructed healthy diet score with life expectancy at 40 y of age by sex. RESULTS: A 1-unit increase in the healthy diet score based on food expenditure was associated with a 0.07 y (95% confidence interval [CI], 0.05-0.10), 0.04 y (95% CI, 0.02-0.07), and 0.06 y (95% CI, 0.04-0.08) increase in county-level life expectancy among men alone, women alone, and men and women combined, respectively. Increasing expenditure on whole grains (estimate of per 1% increase 0.07; 95% CI, 0.03-0.11), fresh fruit and vegetables (0.06; 95% CI, 0.02-0.09), and dairy products (0.05; 95% CI, 0.03-0.07), as well as reducing expenditure on sugar-sweetened beverages (-0.02; 95% CI, -0.04 to -0.01) and processed red meat (-0.05; 95% CI, -0.08 to -0.02) showed a positive association with increased county-level life expectancy. CONCLUSIONS: These findings may inform a nutritional measure against premature death and stagnation of increase in life expectancy.

Risk perception and resource scarcity in food procurement during the early outbreak of COVID-19.

OBJECTIVES: The retail food industry, a major essential business, is among the very few thriving sectors during the COVID-19 pandemic. However, such prosperity on the store side does not guarantee a sufficient food supply for all populations. This study aims to understand if people's risk perception and food security status shaped their food procurement behaviors during the early outbreak of the pandemic. STUDY DESIGN: Extended from the theory of risk perception, food consumers may behave differently during a disastrous event in terms of store patronization. The study evaluates how food procurement behaviors are affected by perceived risk aversion, resource scarcity, and consumers' food security status. METHODS: The study examines how people with different food security statuses made grocery shopping decisions at the risk of epidemic exposure based on a nationwide survey of 2590 participants in the U.S. during the early break of the pandemic in April 2020. The study uses a moderated mediation analysis on in-store shopping frequency and food expenditure. RESULTS: People having a food-secure status before the pandemic spent significantly more as a result of the reduced shopping frequency (i.e., the secure-insecure subgroup ß = -0.18, P < .01; the secure-secure subgroup ß = -0.35, P < .01). The increase in food expenditure was insignificant for people who were food-insecure before the pandemic (i.e., the insecure-insecure subgroup, ß = -0.01, P > .05; the insecure-secure subgroup, ß = -0.11, P > .05). CONCLUSIONS: The study reports that in general people reduced the frequency of grocery shopping trips to avoid epidemic exposure while increasing the food expenditure per trip. The increase in food expenditure was not statistically significant among the food-insecure populations likely due to their budget constraints.

Effects of cigarette price increase on fresh food expenditures of low-income South Korean households that spend relatively more on cigarettes.

On January 1, 2015, the cigarette tax in South Korea increased sharply. In this study, we analyzed how the fresh food expenditures of low-income smoking households that spend relatively more on cigarettes changed between 2014 and 2015 compared to their demographically comparable non-smoking counterparts. We conducted the analysis using a difference-in-differences analysis from which we derived expenditure differences between smoking households and non-smoking households and then examined whether the differences increased from 2014 to 2015. In deriving the differences, we utilized the nearest-matching method to ensure that, besides smoking status, the socio-demographics of the two groups were matched. We used data from the Household Income and Expenditure Survey of South Korea. The analysis showed that smoking households spent less on fruits and vegetables than non-smoking households and that this difference increased after the tax increase. The change was more remarkable in the poorest households that spent relatively more on cigarettes. This suggests that cigarette expenditures have a crowding out effect on fresh food expenditures for smoking households, with a significantly larger effect for households in the poorest group that spend more on cigarettes.

The role of dwelling type on food expenditure: a cross-sectional analysis of the 2015-2016 Australian Household Expenditure Survey.

OBJECTIVE: To explore differences in proportion of food budget and total food expenditure by dwelling type. DESIGN: A cross-sectional study using data from the Australian Bureau of Statistics 2015-2016 Household Expenditure Survey. Food expenditure was examined on multiple categories: fresh fruits, fresh vegetables, pre-prepared meals, meals in restaurants, hotels and clubs, and fast food and takeaway meals, using two-part models and zero-one inflated beta regression models. Dwelling types were categorised as separate house, semi-detached house, low-rise apartment and high-rise apartment. SETTING: Australia, 2015-2016. PARTICIPANTS: Seven thousand three hundred and fifty-eight households from greater capital city areas. RESULTS: Households living in high-rise apartments were estimated to allocate a greater proportion of their food budget to meals in restaurants, hotels and clubs, and to spend more (actual dollars) on that category, compared with other dwelling types. No substantial differences were estimated in the proportion of food budget allocated to the other food categories across dwelling types. CONCLUSIONS: The dwelling type households live in may play a role in their food budget. Households living in a high-rise apartment may potentially spend more on meals in restaurants, hotels and clubs than those living in other dwelling types. Given the growth in urban population and the changes in living arrangements, findings point to the critical need for a better understanding of the influence of dwelling types on food expenditure and call for research investigating the relationship between the two.

Food purchasing behaviour of shoppers from different South African socio-economic communities: results from grocery receipts, intercept surveys and in-supermarkets audits.

OBJECTIVE: To identify factors associated with food purchasing decisions and expenditure of South African supermarket shoppers across income levels. DESIGN: Intercept surveys were conducted, grocery receipts collated and expenditure coded into categories, with each category calculated as percentage of the total expenditure. In-supermarket food quality audit and shelf space measurements of foods such as fruits and vegetables (F&V) (healthy foods), snacks and sugar-sweetened beverages (SSB) (unhealthy foods) were also assessed. Shoppers and supermarkets were classified by high-, middle- and low-income socio-economic areas (SEA) of residential area and location, respectively. Shoppers were also classified as "out-shoppers" (persons shopping outside their residential SEA) and "in-shoppers" (persons shopping in their residential SEA). Data were analysed using descriptive analysis and ANOVA. SETTING: Supermarkets located in different SEA in urban Cape Town. PARTICIPANTS: Three hundred ninety-five shoppers from eleven purposively selected supermarkets. RESULTS: Shelf space ratio of total healthy foods v. unhealthy foods in all the supermarkets was low, with supermarkets located in high SEA having the lowest ratio but better quality of fresh F&V. The share expenditure on SSB and snacks was higher than F&V in all SEA. Food secure shoppers spent more on food, but food items purchased frequently did not differ from the food insecure shoppers. Socio-economic status and food security were associated with greater expenditure on food items in supermarkets but not with overall healthier food purchases. CONCLUSION: Urban supermarket shoppers in South Africa spent substantially more on unhealthy food items, which were also allocated greater shelf space, compared with healthier foods.