%20-%20https://sketch.com%20--%3e%3ctitle%3e3ECBBF50-FD08-412A-943D-2CFC5E025A16%3c/title%3e%3cdesc%3eCreated%20with%20sketchtool.%3c/desc%3e%3cg%20id='Stylesheet'%20stroke='none'%20stroke-width='1'%20fill='none'%20fill-rule='evenodd'%3e%3cg%20transform='translate(-926.000000,%20-668.000000)'%20id='BUTTONS'%3e%3cg%20transform='translate(100.000000,%20597.000000)'%3e%3cg%20id='BTN-/-RIGHTS'%20transform='translate(816.000000,%2065.000000)'%3e%3cg%20id='ICON-/-Rights'%20transform='translate(10.000000,%206.000000)'%3e%3cpath%20d='M13.5752841,0%20L1.34943182,0%20C0.59375,0%200,0.586666667%200,1.33333333%20L0,22.6666667%20C0,23.4133333%200.620738636,24%201.34943182,24%20L17.6505682,24%20C18.40625,24%2019,23.3866667%2019,22.6666667%20L19,5.36%20L13.5752841,0%20Z%20M13.5752841,1.89333333%20L17.0838068,5.36%20L13.5752841,5.36%20L13.5752841,1.89333333%20Z%20M1.34943182,22.6666667%20L1.34943182,1.33333333%20L12.1988636,1.33333333%20L12.1988636,5.36%20C12.1988636,6.10666667%2012.7926136,6.69333333%2013.5482955,6.69333333%20L17.6235795,6.69333333%20L17.6505682,22.6666667%20L1.34943182,22.6666667%20Z'%20id='Shape'%20fill='%23595959'%20fill-rule='nonzero'%3e%3c/path%3e%3ccircle%20id='Oval'%20stroke='%23595959'%20stroke-width='1.5'%20fill='%23595959'%20cx='9.5'%20cy='13.5355339'%20r='5.5'%3e%3c/circle%3e%3cpolyline%20id='Path-2'%20stroke='%23FFFFFF'%20stroke-width='1.5'%20transform='translate(9.500000,%2013.035534)%20rotate(45.000000)%20translate(-9.500000,%20-13.035534)%20'%20points='8%2016.0355339%2011%2016.0355339%2011%2010.0355339'%3e%3c/polyline%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/svg%3e)
Dietary patterns play a pivotal role in health status throughout life, and consumed foods and beverages have a significant environmental influence on ecosystems and financial resources(Reference Duchin1). The environmental effects of food manufacturing are intimidating; for instance, agriculture is responsible for approximately 25 % of greenhouse gas emissions (GHGE)(Reference Vermeulen, Campbell and Ingram2). Overall, the GHGE of various foods change extensively, and meat and dairy are considered to be the major causes of dietary GHGE(Reference Millward and Garnett3). Furthermore, agricultural activities occupy 40 % of the Earth’s surface, using 70 % of the total water resources. The production of animal food also has the largest share in water consumption(Reference Ramankutty, Evan and Monfreda4).
A sustainable diet refers to the integrated concept of inventing a food system that provides healthy regimens for a growing population while declining its environmental impact and staying within the planetary boundaries(Reference Burlingame and Dernini5,Reference Gussow6) . The concept of a sustainable diet was defined in 2010 from two entirely different perspectives, which were referred to as person-centred nutrition and global sustainability in social, economic and environmental dimensions. Based on the definition of the FAO, a sustainable diet has a low environmental impact and contributes to food and nutrition security and a healthy lifestyle for the present and future generations. Furthermore, Meybeck et al. defined a sustainable diet as a diet that contributes to a favourable nutritional status and long-term well-being of the individual or the community, which contributes to and is enabled by sustainable food systems; this will result in long-term food security and nutrition.
Sustainable diets are protective, observant of biodiversity and ecosystems, culturally acceptable, accessible, economically fair and affordable, nutritionally adequate, safe and healthy. In addition, such a diet optimises natural and human resources(Reference Burlingame and Dernini5). On the other hand, the main limitations of sustainable diets are associated with prevalent food systems such as climate crises, environmental destructions, extinction of biodiversity and pollution. All nationalities must strive for a more sustainable diet to promote economic sustainability through low-impact, affordable and accessible foods while also supporting community health through adequate nutrition(Reference Macdiarmid, Kyle and Horgan7,Reference Reisch, Eberle and Lorek8) .
Sustainable diets may be affected by several agricultural, health-related, socio-cultural, environmental and economic factors(Reference Johnston, Fanzo and Cogill9). The predicted population growth and climate change in the next 10–50 years highlight the need to improve the quality and environmental sustainability of the food system(Reference Godfray, Beddington and Crute10). The quality of diets is largely influenced by the crises in agriculture, poverty, urbanisation, lifestyle changes and food production consumption(Reference Beddington, Furse and Bond11). Dietary patterns could be affected by the socio-economic diversities within the population(Reference Hulshof, Brussaard and Kruizinga12). Furthermore, income and its distribution and agricultural conditions in an area significantly affect sustainability. High-income countries can buy more diverse foods, which enhances the dietary quality that indicates significant lifestyle changes(Reference Johnston, Fanzo and Cogill9).
Socio-economic status (SES) is a term widely and accepted expression frequently used in dietary studies as the descriptive variables of education, occupation and income status(Reference Galobardes, Morabia and Bernstein13). Education affects food selection by promoting the ability to recognise the information that is representative of the nutrition science or food labels(Reference Macario, Emmons and Sorensen14,Reference Busselman and Holcomb15) . Income and occupation also demonstrate the accessibility of monetary and food resources, explicitly defining dietary quality(Reference Robinson, Caraher and Lang16,Reference Barratt17) .
Some studies have shown that higher SES improves dietary quality(Reference Galobardes, Morabia and Bernstein13,Reference Galobardes, Morabia and Bernstein18) . Nevertheless, a higher socio-economic status has been reported to increase the consumption of unhealthy food groups in some cases(Reference Benedetti, Biggeri and Laureti19). Since dietary behaviours are considerably associated with socio-economic indices(Reference Mead, Gittelsohn and Roache20), the sustainability of a diet may also be affected by the SES of the general population(Reference Wrieden, Halligan and Goffe21). Given the novelty of the concept of diet sustainability, few studies have been conducted in this regard, and it is essential to determine the most nutritionally sustainable options in various socio-economic groups.
Data are scarce regarding SES and sustainable diet. Therefore, the present study aimed to evaluate the association between the SES of households and having a more sustainable diet.
Materials and methods
Data
In the current study, the food basket and socio-economic data of 102 303 Iranian households were obtained from the Households Income, and Expenditure Survey (HIES) performed during 2016–2018. The HIES is performed annually by the Statistical Center of Iran (SCI) through three-stage cluster sampling at private and collective settled households in the urban and rural areas of Iran. The household expenditure and income database provide relevant data in four major dimensions; the first dimension is the social characteristics of the household, which includes data on the number of the household members, as well as the age, gender, education level, activity status and marital status of each household member. The second dimension consists of data on residential and home facilities (i.e. details of accommodation and home facilities), and the third dimension measures household expenses, with food and non-food expenses considered as the household investments in various sectors. The fourth dimension evaluates the household income by collecting data on the income obtained from salaried jobs and employees, income obtained from self-employment and the miscellaneous income of all the working household members. In addition, the food costs in this dimension are measured by the data on the household foods and their consumed amounts and the details of their preparation to examine the changes in the household food basket. All income and expenses were converted from Rials (currency in Iran) to US dollars.
Since data on the household food basket were collected for the households as the sampling units, they were converted into individual amounts. Rather than calculating the per capita amount, adult male equivalent units (AME) were estimated for each household member(Reference Weisell and Dop22). AME is the ratio of the energy requirement of a household member of a particular age and gender to the energy requirement of an adult male aged 18–30 years with moderate physical activity as recommended by the FAO and WHO(23). Unlike per capita measurements, this tool helps identify the share of each household member in the household food basket(Reference Claro, Levy and Bandoni24). In the present study, the total AME of the households was calculated based on the age and gender of the household members. In addition, the AME of each food item was estimated by dividing the amount of each food item by the total AME of the household.
Since purchased food is partly wasted, we calculated the real amount of the consumed foods based on the FAO estimate of the waste percentage for each food group in the consumption step of ‘from supply to consumption chain’(Reference Gustafsson, Cederberg and Sonesson25). The waste percentage is subtracted from the reported food purchased. In the present study, the AME of the energy, macronutrient and micronutrient intake was calculated using the Nutritionist IV software with adaptations to the Iranian food composition table. The households reporting the energy intake of less or more than three standard deviations from the mean energy intake (n 2729) were excluded from the study.
The diet quality index-international (DQI-I) was calculated to evaluate dietary quality using the method proposed by Kim et al.(Reference Kim, Haines and Siega-Riz26) The DQI-I was composed of four major categories, including variety (overall and protein sources; scores 0–20), adequacy (food/nutrient intake required to prevent undernutrition; scores 0–40), moderation (food/nutrient intake quality relative to chronic diseases; scores 0–30) and overall balance (macronutrient ratio and fatty acid composition; scores 0–10). The scores of each component were summarised in four main categories, and the scores of the categories were summed up to obtain the total DQI-I score within the range of 0–100 (zero showing the poorest and 100 showing the highest possible score). Higher DQI-I scores represented the higher quality of the diet with better variety, adequacy, moderation and overall balance.
Socio-economic status
The available socio-economic variables in the HIES included the characteristics of the head of the household (age, gender, education level and occupation status) and household variables (rural/urban, annual income, house area and number of household members). Based on the international standard classification of occupation (ISCO), the SCI occupation status category of the household head includes managers, professionals, technicians/associate professionals, clerical support workers, service/sales workers, skilled agricultural experts, forestry/fishery workers, crafts and related trades workers, plant and machine operators/assemblers, elementary occupations and armed forces. In the present study, the household education level was considered as the mean education level of the adult family members.
Principle component analysis was used to in our study extract the SES variable. Principle component analysis is a simple and potent method of dimensionality reduction, which employs the dependencies between variables for their representation in the lower dimensional form without a significant loss of data(Reference Jolliffe and Cadima27). In the current research, the occupation status of the household head, household income, education level and house area were analysed using the principle component analysis to obtain one variable for the SES. In addition, the imputed variables were converted into Z scores to produce comparable metrics before performing the principle component analysis. In this regard, one factor with the higher eigenvalue was selected as the SES variable, and the imputed items were loaded at 0·40 or higher on this factor.
Environmental footprint
In the present study, water and carbon footprints were used to assess the environmental dimension of a sustainable diet. A water footprint is defined as the total volume of freshwater that is used to produce the goods and services consumed by an individual or the community. Footprint data for food items are often reported as the water volume in cubic meters per ton (m3/ton); the water footprint data of Iran are available(Reference Hoekstra, Chapagain and Martinez-Aldaya28–Reference Mekonnen and Hoekstra30). In the current research, the water footprint data were converted into the water volume in cubic metres per gram (m3/g).
To calculate the level of carbon dioxide emission during food production, we applied the carbon footprint method, which is a measure of the exclusive total amount of the carbon dioxide emission that is directly and indirectly caused by an activity or is accumulated over the life stages of a product(Reference Wiedmann and Minx31). The required data on the carbon dioxide emissions of each food item were also obtained from the BCFN double pyramid database(32). For each household, the AME of each food item was multiplied by its water and carbon footprint, and the sum of these values showed the total dietary water and carbon footprint of the households (AME).
Statistical analysis
Similar to the procedure used by Masset et al. in France(Reference Masset, Vieux and Verger33) and Wrieden et al. in the UK(Reference Wrieden, Halligan and Goffe34), the median of the dietary carbon footprint, dietary water footprint, DQI (all standardised per an adult man) and dietary costs of the households were calculated for each year of the data in our study (2016, 2017 and 2018). In addition, the households were considered to have a more sustainable diet based on the dietary carbon footprint, dietary water footprint, dietary costs lower than the median and DQI higher than the median; the other households were assumed to have a less sustainable diet. The households were also classified based on the SES quartiles, and the share of the food groups in the total energy intake and socio-economic and demographic characteristics was compared in terms of the lower/higher median of the dietary carbon footprint, dietary water footprint, dietary costs, DQI and diet sustainability. For this purpose, a t-test was applied for the quantitative variables, and χ 2 was used for the qualitative variables.
Logistic regression was used to calculate the OR of having a more sustainable diet as the dependent variable based on various SES quartiles as the independent variable. In the second model, we adjusted the changing effects of the place of residence (rural/urban), the age of the household head and the number of the household members. In the third model, we adjusted the place of residence (rural/urban), age of the household head, the number of the household members and energy intake. In the fourth model, the place of residence (rural/urban), age of the household head, the number of the household members, total energy intake and the share of the food groups in the total energy intake were adjusted as well.
By considering the first SES quartile as the reference, the OR of having a more sustainable diet was determined at a 95 % CI. The mean data of 2016–2018 were also considered in the regression models. Figure 1 shows the share of various food groups in the total energy intake among the SES quartiles to determine the differences in the food group intakes that may lead to differences in the diet’s sustainability by SES. Data analysis was performed in SPSS version 22.0 (SPSS Statistics, IBM).
The share of food groups in total energy intake among different quartiles of socio-economic status (SES). 
, First quartile; 
, second quartile; 
, third quartile; 
, fourth quartile
Results
After excluding the households with lower or higher energy intakes than the three standard deviations from the mean energy intake, 100 500 households were selected for the current research. Table 1 shows the mean and median of the dietary carbon footprint, dietary water footprint, dietary costs, DQI, total energy intake, the share of the food groups in the total energy intake of the samples and the households with less and more sustainable diets. Among the studied households, 7·5 % had more sustainable diets. On the other hand, the total energy intake and share of meat, poultry, fish and eggs, legumes and nuts, fats, oils, sugars and sweets in the total energy intake were significantly higher in the households with less sustainable diets compared with those with more sustainable diets. The contribution of bread, cereal, rice, pasta, vegetables, fruits and dairies was also more significant in the households with more sustainable diets.
Characteristics of the whole sample, less sustainable diets and more sustainable diets
Table 2 shows the comparison of the socio-economic and demographic variables between the groups with lower and higher dietary carbon footprint, dietary water footprint, dietary costs, DQI and diet sustainability. Although the households with a female head had significantly higher dietary carbon footprint and dietary water footprint and lower dietary costs and DQI, no significant differences were observed between the two groups in terms of diet sustainability.
Comparison socio-economic and demographic variables between lower and higher groups of carbon footprint, water footprint, cost of diet, diet quality index and sustainability of diet
* Difference between two groups is significant (P < 0·001).
In terms of the place of residence, the rural households had significantly lower dietary carbon footprint, dietary water footprint, dietary costs and DQI and higher diet sustainability. In addition, the heads of the households with higher dietary carbon footprint, dietary water footprint and DQI and lower dietary costs and diet sustainability were significantly older. The number of the household members was significantly lower in the households with higher dietary carbon footprint, dietary water footprint, DQI, diet sustainability and dietary costs. Also, the households with higher dietary carbon footprint, dietary water footprint, dietary costs and DQI and lower diet sustainability had significantly higher income and house area.
According to the obtained results, the score of adult education level was significantly higher in the households with higher dietary carbon footprint and dietary costs and lower dietary water footprint, DQI and diet sustainability. On the other hand, the households with higher dietary carbon footprint, dietary water footprint, dietary costs and DQI, and lower diet sustainability had a significantly better occupation status and higher SES quartiles.
Our findings demonstrated that energy intake increased with the higher SES. As is shown in Fig. 1, higher SES quartiles compared with the lower quartiles were associated with a more significant share of dairies, meat, poultry, fish, eggs, legumes, nuts and fruits and a smaller share of bread, cereal, rice and pasta in the total energy intake.
Table 3 shows the OR of having a more sustainable diet based on different SES quartiles as the independent variable in the four models. In the first model and without controlling the other variables, the households in the third and fourth SES quartiles were significantly less likely to have a more sustainable diet compared with the first quartile. In the second model, the place of residence, age of the household head and the number of the household members were added to the first model and demonstrated significant, diminishing effects on having a more sustainable diet. Notably, the effects of the SES quartiles remained significant in this model.
Binary logistic regression between having a ‘more sustainable diet’ as a dependent variable and socio-economic status (SES) quartiles of households as an independent variable
In the third model, total energy intake was added to the previous model and exerted a significant, diminishing effect on having a more sustainable diet. In this model, the households in the second quartile (OR = 1·15; (95 % CI 1·07, 1·22)) were more likely to have a more sustainable diet compared with the first quartile, while the OR was less than one in the fourth quartile (OR = 0·66; (95 % CI 0·61, 0·72)). In the fourth model, the share of the food groups in total energy intake was adjusted as well. The share of vegetables and fruits in total energy intake had a significant, increasing effect on having a more sustainable diet. On the other hand, the contribution of dairies, meat, poultry, fish, eggs, fats/oils, sugars and sweets in total energy intake had a significant, diminishing effect on having a more sustainable diet.
Although the OR of having a more sustainable diet was significantly more than one in the second SES quartile, the households in the fourth quartile were less likely to have a sustainable diet compared with those in the first quartile (OR = 0·7; (95 % CI 0·66, 0·74)).
Discussion
According to the results of the present study, the higher SES of the households decreased the possibility of having a more sustainable diet. On the other hand, higher SES quartiles compared with the group with a lower SES were associated with the higher energy intake and consumption of dairy, meat, poultry, fish, eggs, legumes, nuts and fruits and lower consumption of bread, cereal, rice and pasta.
As mentioned earlier, the households with a higher SES had a less sustainable diet in the present study. To the best of our knowledge, the association of the SES and a sustainable diet has not been described in previous studies, and our findings are a novel contribution to the literature in this regard. Only Wrieden et al. have reported that families with lower income tend to have a more sustainable diet(Reference Wrieden, Halligan and Goffe21). Moreover, Vázquez-Rowe et al. observed significant correlations between higher income, higher GHGE and a balanced diet in the Lima metropolitan area(Reference Vázquez-Rowe, Larrea-Gallegos and Villanueva-Rey35).
In another research, Vieux et al. demonstrated positive, significant associations between dietary GHGE and the excessive consumption of energy (regardless of food source), fats (saturated), animal protein, alcohol and less vegetable protein and carbohydrates in a typical sample population of French adults(Reference Vieux, Darmon and Touazi36). In India, a higher socio-demographic status has been significantly associated with dietary blue water footprint considering the type of the consumed portion and food(Reference Harris, Green and Joy37). According to the results of the present study, higher SES was associated with the higher consumption rate of energy and animal-based foods, which would, in turn, lead to a less sustainable diet.
According to the literature, decreased energy intake and energy density are important indicators of a more sustainable diet(Reference Vieux, Darmon and Touazi36). Changes in dietary patterns towards higher total energy intake and animal-based food consumption increase the dietary water footprint and GHGEs. In addition, higher SES and urban areas have been significantly associated with the higher consumption of protein and animal-based foods(Reference Robinson and Pozzi38). More sustainable diets are accommodated with the highest amount of plant-based foods, especially starchy foods(Reference Masset, Vieux and Verger33,Reference Sobhani, Rezazadeh and Omidvar39) .
Due to the impact of globalisation, socio-economic changes and urbanisation, significant lifestyle changes are observed in middle-income countries, which is indicative of nutrition transition; the Middle East region is also affected by this transition. Some of the changes in this process include dietary shifts from traditional to Western diets (high-fat, high-sugar, industrial foods), which may reduce the overall quality of the diet(Reference Abassi, Sassi and El Ati40). In recent decades, the transitions in global dietary patterns within food sources, processing methods and allotment have led to the dramatic growth of processed food and beverage consumption(Reference Popkin41).
Nutrition transition refers to a wide range of demographic and socio-economic changes that lead to rapid shifts in dietary patterns and physical activity levels in major parts of the world, especially in low-income countries(Reference Popkin42). Despite the decline in population growth, population development is constantly on the rise, and such a demographic change alters the food consumption of individuals. Urbanisation is another aspect of the global transition, which is as intense in most developing countries as the Middle East where the urban population is estimated to be over 50 %. Changes in the rural to the urban lifestyle are characterised by changes in eating habits and patterns of physical activity. Developing countries have come to substitute their old-fashioned diets (rich in grains and fibres) with diets containing significant amounts of energetic sweeteners, fats, meat and dairies(Reference Drewnowski and Popkin43). Similarly, the nutrition transition in Iran as a middle-income country has been occurring rapidly due to the changes in the per capita income and urbanisation(Reference Ghassemi, Harrison and Mohammad44). Although the per capita gross domestic product in Iran has fluctuated noticeably in recent years, it increased from 1080·5 US dollars in 1990 to 5593 US dollars in 2017(45). The high rate of rural–urban migration, along with rapid socio-economic changes, has caused the urbanisation trend to accelerate in Iran in recent decades(Reference Assari and Mahesh46). It seems that the individuals with a higher SES in Iran have experienced this dietary change faster through the higher energy intake and animal-based food consumption, as well as the subsequently less sustainable diets, compared with the populations with a lower SES.
Without proper policymaking and interventions to promote nutritional behaviours, nutrition transition in developing countries such as Iran leads to a less sustainable diet. Therefore, effective and efficient policies and interventions are needed to promote healthy and sustainable eating behaviours. These policies and interventions can be including banning unhealthy food advertisements, public information campaigns, nutritional labelling, nutritional information on menus, fiscal measures (e.g. taxes or subsidies) and regulation of meals in schools and workplaces. Also, develop and implement nutrition-related standards, healthy and sustainable eating guidelines and agricultural policy should be considered(Reference Mazzocchi, Traill and Shogren47,Reference Brambila-Macias, Shankar and Capacci48) .
The main strength of the present study was its novelty to investigate the association between the SES of the households with a more sustainable diet. Furthermore, the sample size was large and on a national level, while no specific data were available on the carbon footprint of food in Iran. However, the GHGE data were obtained from the global data, which is considered the main limitation of our study. In addition, the food expenditure data of the households were used for food consumption calculation, despite the AME units and FAO estimates of the waste percentages used to calculate the real consumption rates of the individuals, underestimation and overestimation are possible.
Conclusion
According to the results, the households with a higher SES had less sustainable diets compared with those with a lower SES. Moreover, the better economic and social status of the households with a higher SES increased the consumption of energy and animal-based food, which in turn influenced water footprint, carbon footprint and dietary costs, thereby decreasing diet sustainability. As a consequence of economic development and social changes, Iran has recently experienced a rapid nutrition transition. By considering the populations with a higher SES in Iran, the impact of nutrition transition could be determined more accurately on the groups whose diets commonly contain higher energy and animal-based foods. It also seems that nutrition transition in developing countries such as Iran may lead to dietary unsustainability as these countries have mostly undergone dietary changes without proper policymaking and interventions to promote nutritional behaviours.
Acknowledgements
Acknowledgements: None. Financial support: The current study was financially supported by the National Nutrition and Food Technology Research Institute (NNFTRI) (grant number: 501). The NNFTRI had no role in the design, analysis or preparation of this article. Conflict of interest: There are no conflicts of interest. Authorship: The authors’ responsibilities were as follows – H.E.-Z.: conceptualisation, methodology, data curation and supervision; S.R.S.: writing – original draft preparation, methodology, software and investigation; N.S., G.R. and A.N.O.: Supervision, validation, writing-reviewing and editing and all authors: read and approved the final manuscript. Ethics of human subject participation: The current study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving human subjects were approved by the Ethical Committees of the NNFTRI and the Faculty of Nutrition Sciences and Food Technology, Shahid Beheshti University of Medical Sciences. Written informed consent was obtained from all subjects.
