Food systems and diets

Diets and food systems are closely linked. However, the notion of food system is generally focused on food; we propose here to revisit the notion in relation to diets.

A diet is a selection of foods, eaten by an individual, chosen between those made available by the food system. Conversely the sum of diets creates the overall food demand that directs food systems. Diets are thus both a result and a driver of food systems. Therefore, approaching food systems by adopting the perspective of diets can bring operational insights to the issue of the evolution of food systems towards sustainability. Diet(s) can be a good entry point to see what can be done individually and collectively to improve food systems (and the margins of manoeuvre).

Hammond and Dubé⁽ Reference Hammond and Dubé ^{6 )} proposed, as part of a systems framework for food and nutrition security, a definition of agrifood systems, focused on food production and linked to two other systems: the environmental system and the health and disease system; the interactions of the three systems determining outcomes on individuals.

Sobal et al.⁽ Reference Sobal, Khanb and Bisogni ^{7 )} proposed an integrated conceptual model of the food and nutrition system emphasising a focus on nutrition and the links between food production, food consumption and nutritional health. It defines the food and nutrition system as ‘the set of operations and processes involved in transforming raw materials into foods and transforming nutrients into health outcomes, all of which functions as a system within biophysical and sociocultural contexts.’ It identifies three subsystems: the producer subsystem, the consumer subsystem and the nutrition subsystem, each flowing into the subsequent one. Focused on the linear relations between these three subsystems it is less comprehensive in its coverage of determinants of food systems, here described as part of biophysical and sociocultural contexts. It further identifies several systems that interact in many points with the food nutrition system. These systems, including the health care, economic, cultural, ecological, governmental, transportation systems, have their own specific orientations and interact with each other. This approach puts the consumer at the centre of the system, as an intermediate between food production and nutrition outcomes. It is thus particularly useful for the consideration of diets, within food systems.

More recently, building upon these and a range of other works⁽ Reference Ericksen, Stewart and Dixon ^{9 – 12 )} the High Level Panel of Experts on Food Security and Nutrition (HLPE) proposed a comprehensive, descriptive definition:

We propose to build upon the HLPE definition of food systems and to integrate the idea of linear exchanges between production and consumption, focusing on diets. In addition, we emphasise two dynamic domains interacting with food systems: environmental changes and social and economic drivers as well as two other domains interacting with individual food choices and diets: attitudes and lifestyles on one side, health on the other (see Fig. 1).

Fig. 1. (Colour online) Food systems and diets^{( 8 )}.

Notions and definitions of sustainable food systems and sustainable diets

We try here to delineate the relationships between diets and food systems, in order to be able to concretely assess the sustainability of diets, intended as their contribution to the sustainability of food systems and ultimately to food security and nutrition. The internationally agreed definition of food security dates from the 1996 World Food Summit: ‘Food security exists when all people, at all times, have physical and economic access to sufficient, safe and nutritious food that meets their dietary needs and food preferences for an active and healthy life’^{( 14 )}. This definition identifies four dimensions of food security: availability of food, accessibility (economical and physical), utilisation (the way it is used and absorbed) and stability of these three dimensions.

The HLPE, in line with the original broad approach of sustainability, has provided a definition of a SFS oriented by its capacity to ensure the positive outcomes of a food system: food security now and for future generations: ‘A sustainable food system is a food system that ensures food security and nutrition for all in such a way that the economic, social and environmental bases to generate food security and nutrition of future generations are not compromised.’^{( 13 )} The HLPE has thus formalised the link between the two concepts of FSN and of SFS: the basis is that there can be no FSN (short and long terms) without SFS. FSN for all, worldwide, and the conditions for their existence over time, could be what ultimately characterises SFS.

How do sustainable diets relate to SFS? First, a diet is a notion that is person-centred, it is the set of food, beverages and nutrients that are consumed by an individual or by a community of individuals during a certain period of time^{( 15 )}. However, when the question is which diet to choose or to have, or what is an optimal diet, one has to bring in the picture elements that go beyond the diet stricto-sensu, both as impacted by the diet and as determining the space of potential choices of each individual: economic, social and cultural conditions and constraints.

As defined in 2010, sustainable diets are those diets with low environmental impacts, which contribute to food and nutrition security and to healthy life for present and future generations. Sustainable diets are protective and respectful of biodiversity and ecosystems, culturally acceptable, accessible, economically fair and affordable; nutritionally adequate, safe and healthy; while optimising natural and human resources^{( 16 )}.

Such a definition of sustainable diets, while embracing some of the elements of the food system, does not make direct reference to food systems. By introducing the SFS definition of the HLPE and how it relates to FSN, we deduct a relation between the definition of sustainable diets (with their double dimension of individual sustainability for health and of impact on global sustainability) and SFS:

This has two important consequences to understand how sustainable diets relate to SFS: First, the two notions are very linked, which can help their common assessment. The strength of the contribution of the diet to the sustainability of the food system is what characterises the sustainability of the diet. Second, sustainable diets are both an objective and an essential mean, a key driver, to achieve the transformation of food systems, which is needed to achieve FSN.

The food system, at global level, currently, is not sustainable

From the above definition, the very function of food systems is to provide food security and nutrition. As shown by current figures of hunger and malnutrition the global food system is not fulfilling its function. There is today a triple concern, often referred to as the triple burden of malnutrition, consisting first, of deficiencies in dietary energy intake (hunger defined by FAO as chronic undernourishment affecting 795 million persons according to the latest estimate^{( 17 )}); second, of nutrient deficiencies, such as iron, iodine and vitamin A, which affect some two billion people in the world; and third, from overnutrition leading to overweight, estimated by WHO to involve 1·9 billion adults in 2014, and obesity, 600 million in 2014^{( 18 )}. In addition, 42 million preschool children were suffering from overweight or obesity in 2013^{( 18 )}. These different malnutrition challenges, which very often coexist in countries, can also overlap at individual level (for instance obesity can coexist with nutrient deficiencies), with correlated individual and public health challenges.

There are however differences in the way dietary patterns change across the world. Imamura and colleagues have characterised changes in dietary patterns nationally and regionally analysing information derived from individual-based national surveys⁽ Reference Imamura, Micha and Khatibzadeh ^{19 )}. Focusing on twenty dietary factors, they have modelled two different dietary patterns: one based on relatively high consumption of healthy items, another based on relatively low consumption of unhealthy items, as well as a third one integrating all items, and derived a score for each pattern. They note substantial heterogeneity between countries, including between neighbouring countries. Between 1990 and 2010 patterns based on more healthy items improved modestly while patterns based on unhealthy items worsened to a greater extent; these trends being weakly correlated. Trends vary significantly by national incomes, with improvements on healthy items in higher-income countries; global worsening on unhealthy items, particularly in middle-income countries, some improvements in high-income countries but they remain among the worst in the world. Importantly association between socioeconomic status and diet quality varies significantly for diet patterns based on unhealthy and healthy items.

Also, there are concerns about the scale of some major environmental impacts of food production. The production of food, including deforestation and other land-use changes linked to expansion of production, is responsible for 24 % of greenhouse gases (GHG) emissions⁽ Reference Tubiello, Salvatore and Cóndor Golec ^{20 )}. Agriculture accounts for 70 % of freshwater withdrawals globally, and in many places exceeding the renewable potential^{( 21 )}. Pollution of soil and water is due to overuse of nitrogen and phosphorus fertilisers, and improper use of pesticides. Agriculture is also one major factor of biodiversity loss⁽ Reference Smith, Smith and Colleen ^{22 ,} Reference Machovina, Feeleya and Ripplec ^{23 )}, and at the same time agricultural biodiversity is decreasing: In the past 100 years, about 75 % of the genetic diversity of agricultural crops was lost⁽ Reference Shand ^{24 )}, and today more than half global plant derived energy comes from only three crop species: wheat, rice and maize^{( 25 )}.

The food system is producing, globally, enough food but is unable to ensure food security and good nutrition in the world today. And strikingly it performs proportionally worse to feed its own participants. The majority of the poor and undernourished people in the world are living in rural areas, depending on agricultural activities for their subsistence. This is why sustainability concerns for food systems expand in fact well beyond those of their inadequate environmental and natural resource use performance, questioning substantially their economic and social features.

Sustainability challenges for the future

According to FAO estimates⁽ Reference Alexandratos and Bruinsma ^{1 )}, business as usual projections of food consumption, due to increased population and, to a more important extent, to worldwide dietary changes, is expected to grow by 60 % by 2050 compared with 2007. There are major differences between countries both in increase of food consumption and in changes in dietary patterns, but in the countries experiencing food consumption increase, patterns are generally changing towards more livestock products, vegetable oils and sugar. These three food groups together now provide 29 % of total dietary energy supply of the developing countries, a share up from 20 % three decades ago, and projected to rise to 35 % in 2030 (in industrialised countries the share has been about 48 % for several decades now).

Of particular significance is the projected increase of the global consumption of animal-sourced products. Such changes are expected to have a major environmental impact, mainly evaluated in terms of GHG emissions and land-use change⁽ Reference Tilman and Clark ^{26 )}. Some studies have described alternate scenarios. For instance Agrimonde^{( 27 )} describes a baseline scenario, Agrimonde 0, with an increase of global demand of dietary energy of 83 %, and an abated scenario Agrimonde 1, where the increase is 28 %. This last scenario supposes radical changes in consumption patterns and behaviours and in worldwide distribution of food, including a decrease of 25 % of the per capita consumption in Organisation for Economic Co-operation and Development countries between 2000 and 2050, without any income reduction, and a slowdown in the increase of per capita consumption in emerging countries. Such a scenario would require radical changes in consumers' behaviours, reduction in waste, and implementation of efficient public policies to promote more balanced and healthy diets.

Another study⁽ Reference Erb, Haberl and Krausmann ^{28 )} elaborates scenarios based on diets, differing mainly by the proportion of meat: ‘western high meat’, ‘current trend’, ‘less meat’ and ‘fair-less meat’. It concludes that under the fair-less meat scenario it would probably be possible to feed the world with organic crops and an organic livestock system, with a very equitable distribution and an average daily intake of 11 715 kJ (2 800 kcal) per capita of which 20 % of protein is from animal origin. The western high meat diet, with 44 % of protein intake of animal origin would also probably be feasible, but only with a cropland expansion of 20 %, intensive crop production and intensive livestock production.

These challenges for the future, given today's unsatisfying performance of food systems, economically, socially and environmentally, poses the question of the appropriate leverage points to improve the sustainability of food systems.

Human and food systems dimensions of sustainable diets

Characterising sustainable diets requires first to distinguish between the two dimensions of the definition of sustainable diets (Fig. 2): on the one side, the nutrition and health dimension, assessed on persons, and on the other side the impact on the food system, and its sustainability, in all dimensions and measured at various levels. Assessing each dimension requires specific methodological approaches and tools.

Fig. 2. Two dimensions of sustainable diets as a driver of changes at individual level and system level⁽ Reference Meybeck, Redfern, Paoletti and Strassner ^{29 )}.

The composition of the diet has two main categories of impacts: on the individuals consuming it and on the food system as a whole. The impacts on the individuals can be assessed by the nutritional characteristics of the foods consumed and/or, with a time lag, by the assessment of certain health characteristics known to be influenced by food consumption. The composition of the diet drives the demand for and production of specific foods, with environmental, economic and social impacts. Assessing the significance of specific dietary choices on sustainability of food system brings a series of challenges. Impacts can generally only be assessed using generic indicators and figures unless the origin of the products can be traced back and that there are means to better assess specific impacts in the area of origin and along the food value chain.

Looking at the nutrition dimension involves mainly individual-level indicators, but comes with its challenges. It impacts, and to a great extent determines, the nutrition status of the person. One main challenge for the assessment of nutrition and health impacts is how to make the link between the sum of individual diets, the global consumption of the population, and the health status of a population, at country level for example, given heterogeneity of consumption, especially in a time of changing diets.

The function of a diet is to ensure proper nutrition contributing to good health. There are numerous studies and often diverging views on optimal diets from a health perspective, and the aim of the present paper is not to review those studies. Some broad characteristics seem to be generally considered to be correlated with better health outcomes. For instance, a recent review of studies on health effects of diets, organised by types (low carbohydrates, low fat, low glycaemic, Mediterranean, mixed, Paleolithic and vegan) concluded with a broad aggregation of evidence in support of diets comprising preferentially minimally processed foods, and comprising mostly plants⁽ Reference Katz and Meller ^{30 )}.

Diets that contribute to sustainable food systems

Most studies about the sustainability of diets focus only on environmental impacts, or combine them with health effects. Economic or social impacts of diets are generally not considered. Even environmental impacts are often reduced in some of them, specifically GHG emissions, land use or land-use change and water use. Impacts on water quality are much less mentioned and biodiversity is generally not taken into account or only through the proxy of land use, which does not enable proper consideration of agrobiodiversity nor of biodiversity in the field.

Assessing sustainable diets

Combining scales

Finally, one critical issue is how to deal with the effect of spatial integration, from local to regional and global, when it comes to measuring the effect of one diet on one food system. With longer and longer food chains, expanding imports and exports, and increasing interconnection of food systems, diets are more likely to have remote impacts, which can be different from local impacts, and which will necessitate to a look at the food system at global scale. The example of meat consumption in Europe is a good one, as it drives a good part of the soyabean expansion in South America in order to provide feed. Another interesting example is the success of quinoa. Is the increase of quinoa consumption in developed countries really contributing to sustainable diets and how, given all the dimensions of sustainability of food systems we talked about, locally and remote?

The original discourse about sustainable diets was very much guided by an assumption that diets and food systems are linked, and in particular that they share the same spatial limits, with a broad equivalence between consumption and production spaces. Such an equivalence is true at global level (summing all the individual diets that compose it). It was to a great extent true in ‘traditional’ food systems very much linked to a specific ‘traditional’ diet, somehow homogenous, shared by a geographic community, and therefore sharing the same geographical limits as the food production area. To a certain extent this was true for the Mediterranean diet model, abstracted from traditional diets in the Mediterranean (see Fig. 3). This is why it can be considered as a case study of a sustainable diet (arrow 1). The work done in that respect aims to devise a methodological approach to consider actual diets in the Mediterranean in relation to the notion of a sustainable diet (arrow 3) using their relationship to the Mediterranean diet model (arrow 2) as a pathway. As shown above, to assess the environmental, economic and social dimensions would require looking specifically at the food system to which a diet is linked and to assess it from a SFS perspective (arrow 5). Finally, the main question of interest in any food system could be the relationships between the diet and the food system (arrows 4) and how these determine the sustainability of the diet (arrow 3) and of the food system as a whole (arrow 5). Understanding these relationships and their dynamics could also be key to understanding drivers of change, including potential means to improve the sustainability of diets and food systems.

Fig. 3. (Colour online) From concepts to actual diets: the case of diets and food systems in the Mediterranean⁽ Reference Meybeck, Redfern, Paoletti and Strassner ^{29 )}.

This proximity of a diet to a food system, closely associated with a geographic area, is no more to be found for most modern diets. Furthermore, a diet within a particular ‘system’ is not the same for everybody; consumption within a food system is not homogenously distributed. It is the sum of diets that creates global demand and thus determines impacts on food systems. This brings an additional difficulty to assess the impact of diets on food systems, and often a way to deal with this is to look at the impact of average diets. However, from a nutrition and health perspective, the individual diets are important, rather than the average.

What drives what? Food system constraints and enablers to sustainable diets

Many drivers of food consumption choices and to their expression are to be found within food systems. If we assume that sustainable diets are both an objective and a driver of SFS, understanding the drivers of food choices is of paramount importance to design ways to improve the sustainability of both diets and food systems. This leads to particular interest for two specific groups of parameters. The first covers economic, social and cultural parameters that, both inside and outside food systems, can drive food consumption choices. The second covers those relating to consumption choices that go beyond an interest in diet composition and take into account characteristics such as quality, origin and mode of production. Such choices can have various impacts on all dimensions of sustainability. Moreover, they can be the expression of attitudes that are also grounding some choices related to diet composition.

Sustainable diets are both constrained and enabled by the food system. The state of the food system determines the space of possible diets, and therefore determines the possibility to choose sustainable diets, given the available spectrum of consumption choices and incentives, and that these diets are available at the proper scale.

The food system is shaped by many drivers, its status being the result of different diets, environmental, economic and social issues at different levels from local to global, etc. First, it is shaped by the sum of all the diets. The existence of unsustainable diets, or their predominance and influence on the market, is in itself a constraint to the emergence of sustainable diets at scale. Currently we are rather, globally, in an un-virtuous circle, by which the evolution of diets, as shown above, are shaped by and contributing to a more and more unsustainable food system. This interrelation expands, today, spatially: in one location, say at country level, diets are diverse and one could say that several dietary patterns coexist. The food system is connected (and shaped by) a variety of diets. Together with the importance of imports, diets at one place are connected to food systems at other places. Second, the food system is also shaped by many other drivers, such as food production and distribution costs, energy and input prices, the cost of labour, agricultural policies, as well as the evolution of incomes with regard to the price of foods, etc.

Prices and affordability of diets

One of the first determinants of food consumption choices (and of their feasibility) is price. From a consumer perspective, and especially poor net food buyers, the lower food prices are, the better; it facilitates diversified and nutritious diets, and favours capacity to spend on other basic needs. Conversely, low food prices can have direct negative environmental impacts by not discouraging food waste^{( 45 )}. They reduce investment capacity and thus economic sustainability. By driving the need for low production costs they also encourage low-cost practices that can be environmentally damaging and drive low income and wages for food producers and workers, with important social impacts.

Another difficulty results from the increasing discrepancy between prices of healthy v. non-healthy foods. This is particularly important as those with low income are more sensitive to prices and price change. A systematic review and meta-analysis covering ten countries found that healthier foods and diet patterns cost more than less healthy options, with the difference between much healthier and less healthy options being significant for low income families, but comparable with the societal cost of suboptimal diet quality, concluding that there would be room for public intervention⁽ Reference Rao, Afshin and Singh ^{46 )}.

This divide between the price of healthy v. unhealthy items tends to increase: in high income countries, the cost of the first category has risen more than the other during the past 30 years, and the same may apply in emerging economies, with prices of fruit and vegetables having risen more than most foods including energy dense processed foods⁽ Reference Wiggins and Keats ^{47 )}. Price-driven substitutions from healthy to unhealthy food can be amplified in case of unhealthy foods becoming more affordable, when, as shown in the case of the USA, consumption of food away from home, soft drinks, juice and meats were more responsive to price changes than other foods⁽ Reference Andreyeva, Long and Brownell ^{48 )}. Conversely, this would also lead to possible higher effects on consumption of the taxation of such produce.

However, such findings do not necessarily mean that, systematically, healthier diets are always less accessible, such as in the case of the Mediterranean diet⁽ Reference Drewnowski and Eichelsdoerfer ^{49 )}. Also, over time, some argue that income may become a weaker determinant of diets, giving thus broader scope for public policy to have a real influence in the future⁽ Reference Johnston, Fanzo and Cogill ^{50 )}. This deserves to be discussed. On one side, it can suppress barriers to more healthy and sustainable diets if they were more expensive. Conversely, it could reduce the effectiveness of any attempt to integrate negative externalities, whether for health or sustainability reasons, in food prices.

The potentially ambivalent (and at least multiform) role of prices with regard to sustainability and food security calls for clarification, clearly separating food prices as an indicator of access at consumption level, from its use and interpretation inside the food system at large, which requires breaking down the final consumption price in various components to better envision its relationships with economic and social dimensions of sustainability. Such distinctions of level of impacts inside food systems are also particularly important to better understand, conversely, the potential impacts of changing diets, on the different stages of food systems, as it will impact prices and economic exchanges⁽ Reference Adinolfi, Capone and El Bilali ^{51 )}.

Food prices should therefore also be analysed in terms of their impacts on sustainability, with different approaches for diets and for food production, for instance. The contribution of food prices to the various dimensions of sustainability can thus be different when considering only diets or food systems as a whole, particularly when integrating a long-term perspective. Prices, however, are not the only parameter influencing food choices. Other parameters enter in line, such as education and consumer information.

Consumer information

Consumers, by their choices, in terms of type of products, quantity, quality (including production modes) direct production. Consumers are directed by the information made available to them. Producers can also anticipate consumer demand and its changes and pro-actively seek new markets.

Different parts of the food and nutrition systems have different foci, goals and units and this can make communication difficult between those involved⁽ Reference Sobal, Khanb and Bisogni ^{7 )}. Nutritionists organise the world according to nutrients, producers from a commodity perspective, and consumers from a taste perspective⁽ Reference Sobal, Khanb and Bisogni ^{7 )}. In that regard the decision to move from nutrient based to food based dietary guidelines has been key to increase their readability for consumers and thus their effectiveness.

Simple messages about healthy diets can be obscured by communication on various diets and products advocated by numerous participants⁽ Reference Katz and Meller ^{30 )}.

Trends on dietary patterns towards more healthy items and towards more unhealthy items (see earlier) are not correlated which makes one wonder if it is not because these trends are driven by different policies and incentives, tendencies for increase of unhealthy items being driven by marketing and promotion of more healthy items by government policies⁽ Reference Imamura, Micha and Khatibzadeh ^{19 )}. There could also be reluctance from policy makers to propose changes that might have a negative effect on economic sectors⁽ Reference Johnston, Fanzo and Cogill ^{50 )}.

Conversely, too simple messages can have detrimental impacts on nutrition. For instance, in the UK, meat and dairy products make a considerable contribution to the intakes of fourteen nutrients for which intake is below the lower reference level⁽ Reference Buttriss and Riley ^{52 )}. Hence, a recommendation for reducing meat consumption would need to take into account the likely impacts on nutrient intake and propose a more holistic diet change to compensate for the induced changes.