Consumption of foods away from home in Brazil

OBJECTIVE: To describe foods consumed away from home and associated factors in Brazil. METHODS: The study was based on the National Dietary Survey which was conducted among residents aged over 10 years old in 24% of households participating in the Household Budget Survey in 2008-2009 (n = 34,003). The consumption of food and beverages was collected through records of foods consumed, type of preparation, quantity, time and food source (inside or outside home). The frequency with which individuals consumed food away from home was calculated according to age, gender, income, household area location, family size, presence of children at home and age of head of household in Brazil and in each Brazilian region. Specifi c sampling weight and effect of the sampling design were considered in the analyses. RESULTS: Consumption of food away from home in Brazil was reported by 40% of respondents, varying from 13% among the elderly in the Midwest Region to 51% among adolescents in the Southeast. This percentage decreased with age and increased with income in all regions of Brazil and was higher among men and in urban areas. Foods with the highest percentage of consumption outside home were alcoholic beverages, baked and fried snacks, pizza, soft drinks and sandwiches. CONCLUSIONS: Foods consumed away from home showed a predominance of high energy content and poor nutritional content, indicating that the consumption of foods away from home should be considered in public health campaigns aimed at improving Brazilians’ diet. DESCRIPTORS: Food Habits. Restaurants. Food Services, utilization. Collective Feeding. Diet Surveys, utilization. Consumption of foods away from home Bezerra IN et al The proportion of food consumed away from home has increased in Brazil. Data on food spending indicated that 31% was spent on eating out in 2008-2009, compared with 24% in 2002-2003.a In spite of this evident increase in spending on food outside of the home, data on food consumption away from home in Brazil are limited in the literature, especially in relation to the type and quantity of food consumed. The only publication in Brazil on the topic, which analyses data from the 2002-2003 Household Budget Survey, contains no detailed analysis of the type or quantity of food consumed.3 In developed countries, the study of food consumed away from home has received a fair amount of attention in recent years, due to the high content of calories, fat, saturated fat, sugar and salt of food sold in restaurants and snack bars.10,12,13,15 Social, demographic and economic factors, such as the greater participation of women in the labor market,b increased household income,17,23 urbanization,21 and lack of time in modern society11 and decreasing prices of convenience foods6 which have infl uenced the increase in spending on food outside of the home, will continue to drive spending on this type of food. Thus, the increase in popularity of eating out may have a signifi cant infl uence on Brazilians’ quality of diet. The majority of studies on the impact of eating out on the population’s diet have been carried out in developed countries, principally the U.S and Europe. This study aims to analyze characteristics of food consumed outside of the home.

In the 2008-2009 Household Budget Survey, the households were selected using two-stage cluster sampling.Census tracts are the main units of the sample and are initially stratifi ed according to geographical location and mean income.Tracts were randomly selected with probability proportional to the number of households they contained.In the second stage, the households were selected for simple random sampling and a subsample of 24.5% of these households was selected to participate in the National Dietary Survey.Individuals aged 10 and over in these households were included in the fi nal National Dietary Survey sample (n = 34,003).Details on the sampling plan and study design are available in a previous publication.d Food and beverage consumption was collected using food records from two non-consecutive days.Participants were told to record everything they ate and drank, except water, in the 24 hour period.In addition to the food consumed, the method of preparation, quantity, time and source (inside or outside home) was registered.
Eating out was defi ned based on the source of the food and the place it was consumed and included all food prepared and consumed outside of the home.
In order to guide correct recording, each participant received instructional material with guidance on fi lling in the food record and photographs of utensils commonly used to serve the food and drink.
All of the records were checked by research agents, probing for foods frequently omitted (small snacks, sweets, confectionary, drinks), refi ning the information on the type of preparation used for specifi c foods such as meat and vegetables, and estimations of quantity consumed, verifying the number of items consumed and the interval between meals.All of this information was entered into a laptop computer in the household itself, using a program specifi cally designed for data entry on food consumption.
In this article, analysis of the fi rst day's food record is considered.Demographic and socio-economic characteristics were collected using a questionnaire specifi cally designed for the Household Budget Survey.Participants' age was stratifi ed into three groups: adolescents (ten to 19 years old), adults (20 to 59 years old) and the elderly (60 and over).To calculate per capita income, the sum of all the household members' income was divided by the number of residents in the household.Income was stratifi ed into four groups up to $ 124.30; between $ 124.30 and $ 239.93; between $ 239.93 and $ 457.58 and over $ 457.58, respectively.
The size of the family and age of the head of the household were stratifi ed based on median values for each variable: families with up to three and with more than three residents in the household, with head of the household younger than 45 and 45 and over.Children in the household was defi ned as the existence of at least one individual age under ten in the household.
In the data analysis, the frequency with which individuals ate out was calculated according to age group, gender, income group, area in which the household was located, family size, presence of children in the household and age of the head of the household for Brazil and for each region.Association between the frequency of eating out and age and income were examined using linear trend.To evaluate differences between frequencies of consumption outside of the home and the other variables, the Chi-squared test was used.To analyze the time of consumption of food outside of the home, the variable was stratifi ed into six periods: 6-8 AM, 9-11 AM, 12-2 PM, 3-5 PM, 6-8 PM and between 9 PM and 5 AM.
All of the food and drink reported in the fi rst day's food record of the individual food consumption survey were organized into 33 food groups: rice, corn, beans and other pulses, vegetables, legumes, potato, cassava and yam, fruit, oilseeds, other cereals, pasta, soup, bread,

RESULTS
The expanded percentages were 52% women, 65% adults, 22% adolescents and 13% elderly.Eating out was reported by 40% of the interviewees, varying between 13% for the elderly in the South of the country to 51% among adolescents in the Southeast.
The percentage of individuals who ate out decreased with age and increased with income in all regions of Brazil.More men reported eating out than women, as did residents in urban areas compared to those who lived in rural areas.The presence of children in the household did not affect the frequency of eating out, nor did family size, with the exception of the North and Northeast regions, where larger families ate out less often (Table 2).
The most common time for eating out was between 12 and 2 PM (40%); 8% of consumption was between 6 and 8 AM; 19% between 9 and 11 AM; 17% between 3 and 5 PM, 11% between 6 and 8 PM and the remaining 5% between the hours of 9PM and 5 AM.
The food groups which were, on average, most commonly consumed in Brazil were beans and other pulses and rice and the groups most frequently consumed outside of the home were alcoholic drinks and fried and roast snacks.
The contribution of eating out according to percentage of food consumed away from home in the urban area was greater than that in the rural area for the majority of groups, with the exception of oilseeds, other cereals, soups, bread, salted meat and fi sh, sausages, milk and other dairy products, cheese, oil, sauces and condiments, alcoholic drinks, soft drinks, pizza, fried and roast snacks and sandwiches.The food groups with the highest percentages of consumption outside of the home were alcoholic drinks, fried and roast snacks, pizza, soft drinks and sandwiches, in both the urban and rural areas (Table 3).
In all regions, the food group most commonly consumed outside of the home was alcoholic drinks, with the exception of the Midwest region, where pizza was the food group with the highest percentage of consumption outside of the home.Fried and roast snacks were the second most consumed group, with the exception of the Midwest region, where alcoholic drinks had the second largest frequency of consumption outside of the home.
In the North, the food groups most commonly consumed outside the home were cakes and cookies, milk and other dairy products and alcoholic drinks, in comparison to the Midwest, Southeast and South.However, total consumption of these groups was, on average, higher in these regions than in the North, with the exception of alcoholic drinks, which had similar total consumption in all of the regions.The percentage of soup consumed outside of the home was higher in the North than in the South, in spite of total consumption being similar in both regions.The same was true of pasta, when comparing the North with the Midwest region.The North also had higher percentages of consumption of cheese outside of the home than the Northeast and Southeast regions, in spite of mean consumption being half of that in the Northeast and a quarter of that in the Southeast.The contribution of fruit juices to foods consumed outside the home was greater in the North than in the Northeast, Midwest and Southeast regions, whereas the total quantity of juice consumed was similar in all of the regions, with the exception of the Northeast, where consumption was lower than in the North.The percentage of fried and roasted snacks consumed outside the home in the Southeast was lower than in the North and Northeast regions, but the mean total consumed was greater in the Southeast.The percentage of consumption of sandwiches outside of the home was similar in all regions, despite means of total consumption being higher in the Southeast, South and Midwest regions compared with the North and Northeast (Table 4).
Table 5 shows mean daily consumption of the food groups and percentages consumed outside of the home according to sex.Men had higher means of total consumption than women for the majority of groups, excepting vegetables, fruit and soup.The mean consumption of potatoes, oilseeds, other cereals, cakes and cookies, processed snacks and crackers, milk and other dairy products, cheese and sugar and sweets was similar between the sexes.The contribution of eating out to the consumption of these food groups was greater among men than women for all of the groups except sugar and sweets.The percentage of corn, oilseeds, other cereals, cakes and cookies, processed snacks and crackers, pork, fi sh and seafood, salted meat and fi sh, milk and other dairy products, cheese, oil, sauces and condiments, alcoholic drinks, juices and pizzas consumed outside the home was similar for both sexes.The groups most often consumed outside of the home were alcoholic drink and fried and roasted snacks, followed by pizza, soft drinks and sandwiches, for both sexes.
Different consumption profiles can be observed between the age groups.Adolescents had higher mean consumption of cakes and cookies, processed snacks, milk and other dairy products, cheese, sugar and sweets, juice, soft drinks, fried and roasted snacks and sandwiches.Those aged 60 and over consumed, on average, larger quantities of legumes, fruit, soup, cheese and tea and coffee.The contribution of eating out to consumption was more frequent among adults than in the other age groups for most of the food groups.Adolescents had higher percentages of eating out for soup, cakes and cookies, processed snacks and crackers, sugar and sweets and alcoholic drinks.The percentage of soft drinks, pizza, fried and roasted snacks and sandwiches was the same for adults and adolescents, although adolescents had higher mean consumption of these groups, excepting pizza.The food groups most commonly consumed outside of the home were the same for adults and adolescents: alcoholic drinks, fried and roasted snacks and sandwiches.Among the elderly, consumption of pizza was not so important; only 10% of pizza consumed by the elderly was consumed outside of the home (Table 5).

DISCUSSION
For all the regions, eating out is inversely related to age, is greater among men than among women, more frequent in urban than in rural areas and increases with income.Other studies have also identifi ed higher frequency of eating out among the young and among men, 8,12,13,23 as well as the association between demand for eating out and income or indicators of income such as schooling and employment situation. 2,14,23In Brazil, a 10% increase in income contributed to a 3% rise in food bought to be consumed outside of the home. 7 contrast to what has been observed in the United States, 2,23 the presence of children and family size were not associated with frequency of eating out in Brazil, except in the North and Northeast regions.Family size is a variable used as an indicator of a household's socio-economic level and it is estimated that smaller families have higher levels of income and, therefore, eat out more frequently.In this study, it was only in the North and Northeast regions that frequency of eating out was inversely associated with the number of people in the household, possible as these regions are those with the poorest socio-economic levels in the country, where larger families are also less well off.e In general, the food groups consumed in the largest quantities, irrespective of where they were consumed, were beans and other pulses and rice, showing the predominance of traditional eating habits in Brazil.However, data on trends suggest negative changes in the availability of these foods in the home, with a 20% decrease in mean acquisition of cereals and pulses in Forty percent of those interviewed reported consuming something outside of the home on the fi rst day's food record.In studies carried out in other countries, eating out is less frequent than reported in Brazil.Among British adults, 29% reported eating out. 13In Australia, 36% of individuals aged over 18 had consumed food prepared outside of the home on the day preceding the interview. 5In Canada, only consumption of food from 'fast food' type restaurants was considered, and on the day preceding the interview, 25% of individuals aged 19 and over reported eating out.b It should be born in mind that the defi nition of eating out is a variable which does not have a standardized defi nition in the scientifi c literature and may explain the large variations observed. 4 Brazil, the food groups most commonly consumed outside of the home in all regions, in urban and rural areas, among both sexes and for all age groups were alcoholic drinks, fried and roasted snacks, pizza, soft drinks and sandwiches, consistent with data based on spending on eating out in 2002-2003. 3Consumption of food with high calorie and low nutrient content outside of the home is consistent with studies which describe the signifi cant contribution of eating out to total calorie, fat and sugar consumption, and its low contribution to micronutrient consumption. 5,10,13,20Soft drinks, alcoholic drinks and savory snacks were those food groups most frequently consumed outside of the home in other studies. 14,18,19,24Fried and roasted snacks characteristic to Brazil are cheaper than meals and fast food. 3e percentage of high calorie, low nutrient food consumed outside of the home was greater among adolescents.These results are compatible with those seen in another study in the UK, which showed that food consumed by adolescents outside of the home had low levels of nutrients (protein, calcium, iron and vitamin A), but high levels of fat and sugar compared to food eaten at home. 1 Direct comparisons with international studies are limited due to different defi nitions of "eating out", different foci of investigation on specifi c places such as fast food restaurants, school and workplace and other methodological limitations, for example, related to sampling and methods of assessing consumption.In this study, the defi nition included all food prepared and consumed outside the home, with the aim of excluding food available to eat at home, i.e., those foods that made up the stock of food in the home.As concerns regional difference, the highest percentages of eating out were found in the Southeast, South and midwest regions.The total consumption means of the food groups were, in general, higher in the more developed regions of the country ( Southeast and South).Few food groups had higher total consumption means in the North and Northeast, with the exception of cassava and yam and fi sh and seafood.Thus, although the South and Southeast regions have lower means for eating out for some food groups, this percentage might be more important due to the higher mean consumption of these groups.An example of this is in the lower percentage of fried and roasted snacks consumed outside the home in the Southeast compared to the North and Northeast, whereas the mean total consumption of this group was higher in the Southeast.The percentage of sandwich consumption outside the home was also observed to be similar in all regions, although mean consumption was twice as high in the South and Southeast compared with the North and Northeast.
As regards the households' locations, those in rural areas tended to eat more traditional Brazilian foods, with higher means for total consumption of beans and other pulses, rice, corn and cassava and yam.The rural areas also showed lower percentages of eating out for the majority of the food groups.However, the food groups most frequently consumed outside of the house: soft drinks, pizza, fried and roasted snacks, sandwiches and alcoholic drinks, the percentages consumed outside of the home were similar for both areas, although total mean consumption is much higher in urban areas than in rural.
The advantages of this analysis are using a representative sample of the Brazilian population aged ten and over according to different demographic and socioeconomic strata and analysis of effective consumption of these food groups, this being the fi rst study to describe eating out in Brazil and its correlations, based on individual food consumption.
The analysis was based on one day's food records, and this was a limitation as it does not refl ect usual consumption of these food outside of the home.However, the analyses developed in this study are based on population means, suffi cient to provide adequate estimates of consumption for the population, especially when a suffi ciently large number of individuals is used. 25With regards to the comparative analyses of the variables used in this study, the use of one day's f Instituto Brasileiro de Geografi a e Estatística.Pesquisa de Orçamentos Familiares 2008-2009: avaliação nutricional da disponibilidade domiciliar de alimentos no Brasil.Rio de Janeiro: IBGE; 2010.

REFERENCES
food records generated conservative estimates.With correction for intra-individual variability, other differences may also be considered statistically signifi cant for a value of p < 0.05.
Other studies have also been based on one day's food records, and were also not corrected for intra-individual variability. 5,15,19,20This correction is only possible with at least two days' records.
Another limitation of the study was the impossibility of evaluating different places where food was consumed outside of the home, such as the school of workplace, as meals consumed on these premises can have an important infl uence on the quality of an individual's diet. 9,14,16n Brazil, these meals are generally planned out based on specifi c governmental programs and should, therefore, follow specifi c nutritional recommendations and even serve as tools for promoting health eating.In a study in the city of São Paulo, Southeastern Brazil, lunch was the meal most frequently eaten outside of the home. 22This analysis does not allow consumption to be analyzed according to type of meal, although the majority of food eaten outside the home was consumed between 12 and 2 PM.
In conclusion, young individuals living in urban areas and in more economically developed regions such as the South and Southeast are those who most frequently eat out.In general, food consumed outside of the home is high in calories and has poor nutritional content, such as soft drinks, pizzas and snacks.These fi ndings indicate that eating out should be taken into consideration in public health campaigns aimed at improving the Brazilian diet.Studies focusing on factors which infl uence where food is consumed, types of food consumed according to place and according to meal are important in order to better describe eating out in Brazil.

Table 1 .
Sample size (n), frequency (%) of individuals who ate out on the fi rst day's food record for the National Dietary Survey, according to socio-economic and demographic characteristics.Brazil, 2008-2009.

Table 2 .
Mean daily consumption of food groups (g/day) and percentage consumed outside of the home in the National Dietary Survey, according to location of household.Brazil, 2008-2009.
a Statistically signifi cant differences (p < 0.05) between urban and rural area

Table 3 .
Mean daily consumption per capita of food groups (g/day) and percentage consumed outside of the home in the National Dietary Survey, according to region ofBrazil.Brazil, 2008-2009.
c p < 0.05 between North and Southeast d p < 0.05 between North and South e p < 0.05 between Northeast and Midwest f p < 0.05 between Northeast and Southeast g p < 0.05 between Northeast and South h p < 0.05 between Midwest Southeast i p < 0.05 between Midwest and South j p < 0.05 between Southeast and South

Table 4 .
Mean per capita daily consumption of food groups (g/day) and percentage consumed outside of the home in the National Dietary Survey, according to sex.Brazil, 2008-2009.

Table 5 .
Mean per capita daily consumption of food groups (g/day) and percentage consumed outside of the home in the National Dietary Survey, according to age.Brazil, 2008-2009.TD_38.pdf the last 40 years.Although eating out has increased in Brazil, this does not seem to explain the lower availability of rice and beans in the home, as around 12% of these foods are consumed outside of the home.