The current food system negatively impacts health, food security, natural resources and ecosystems (e.g., IFAD, 2021; FAO, IFAD, UNICEF, WFP and WHO, 2024). By consequence, there is a clear need for transforming food systems to provide sustainable healthy diets and help address malnutrition in all its forms such as undernutrition and obesity. Yet, healthy diets are unaffordable to the majority of the global population. In low-income and lower-middle-income countries (LIMC), the large majority of the population cannot afford a healthy diet; in Ethiopia, for instance, this share is 83.8% (FAO, IFAD, UNICEF, WFP and WHO, 2023, 2025). These figures signify the urgency of transforming food systems. Cognizant of this, the Government of Ethiopia has developed a ten-year food systems transformation plan (2021–2030) focusing on 22 “game-changing” solutions [Federal Democratic Republic of Ethiopia (GoE), 2021]. These game-changers are grouped around six key focus areas, encapsulating a focus on stimulating sustainable healthy diets as evidenced by several of these focus areas, such as ensuring availability and accessibility of safe and nutrient-dense foods, and promoting sustainable consumption.

A joint statement by FAO and WHO (2024, p. 2) refers to a healthy diet as one that meets four core principles, including “adequacy (providing enough essential nutrients to prevent deficiencies and promote health, without excess); balanced (in energy intake, and energy sources to promote healthy weight, growth, and disease prevention); moderate (in consumption of foods, nutrients or other compounds associated with detrimental health effects), and diverse (including a wide variety of nutritious foods within and across food groups to favour nutrient adequacy and consumption of other health promoting substances).” But, how sustainability of a healthy food production, acquisition and consumption pattern is measured, is not precisely addressed. Yet, having clearly defined measures for sustainable healthy diets helps countries, such as Ethiopia, in tracking change resulting from their policy initiatives over time.

Vegetables, fruits and plant-based protein sources are considered healthy, while animal-based foods (meat, milk, eggs) have both positive and negative associations with health. At the same time, sustainability is often framed as ‘good for the planet’ and mainly related to waste generated and environmental pollution. This interpretation broadly follows the approach of the WISH (World Index for Sustainability and Health), an index that is based on globally applicable recommendations for a healthy diet for the general healthy adult population within environmental sustainability targets [see Trijsburg et al. (2021), for an explanation and application of the WISH approach]. Also in its scenario analyses, the EAT-Lancet commission highlights the environmental effects of different dietary patterns in terms of change in land and water use, nitrogen and phosphorus application and greenhouse gas (GHG) emissions (Willett et al., 2019). However, the authors also point out that a transition to sustainable healthy diets will have major social (distribution) and economic consequences as well, implicitly indicating that policy interventions aimed at strengthening sustainable healthy diets should also be evaluated on a social and economic level.

Sustainable healthy diets, as for instance defined by FAO and WHO (2019), indeed go beyond just nutrition and the environment, and encompasses economic and sociocultural dimensions too. In essence, such definitions call for a Food Systems perspective when assessing whether produced food is sustainable and healthy. Yet, current measurements of sustainable healthy diets appear to pay little attention to sociocultural and economic sustainability. When they do, studies lack commonly applied indicators in research [see for instance Jones et al. (2016); and the scoping reviews by Harrison et al. (2022), Machado et al. (2023), and Neta et al. (2023), which show great diversity in approaches and indicators used in empirical studies on sustainable diets]. Moreover, the great majority of studies focus on high-income countries, an emphasis in the literature that overlooks the livelihoods and the production and dietary alternatives most relevant to LMICs such as Ethiopia.

This manuscript aims to apply a generic assessment framework for sustainable healthy diets that adequately captures the sociocultural, economic and environmental dimensions of a sustainable food system for application in a LMIC such as Ethiopia. We examine how recent literature links dietary patterns and sustainable food system outcomes and which indicators are most commonly used to capture the socio-cultural, economic and environmental dimensions of dietary patterns. We then argue which framework and indicators seem suitable in the context of LMICs to assess sustainable healthy diets. Ideally, indicators must be quantitatively measurable and criteria of data availability and reliability are relevant for selecting the indicators. We align with the analytical framework developed by Norde et al. (2023), which was tested in Brazil at the state level. This approach seems highly suitable for application in Ethiopia. Our study thereby advances the measurement of sustainable healthy diets to the subnational state-level, adding more detail to other recent studies at the national level (e.g., Demelash and Alemu, 2024).

The structure of the paper is as follows. Section 2 reviews recent literature that defines and measures sustainability of healthy diets. This review leads to the definition of principles of a framework for measuring sustainable healthy diets in LMICs. Section 3 presents a framework for measuring sustainable healthy diets and explains the selection of indicators that can proxy the framework dimensions. Section 4–6 applies to assess sustainable healthy diets in sub-national food systems in Ethiopia. Section 7 concludes and suggests ways to improve the analytical framework, its testing and validation as well as discusses implication for policy.

Sustainability has been a prime consideration in drafting national healthy food guidelines for a long time, but a lack of metrics and a shared approach to linking the concepts of sustainability with healthy diets hampered the credibility of nutritional guidelines that propagate sustainability. To bring more unity to the different approaches, FAO and WHO organized a consultation with international experts in 2019 on guiding principles to define and measure sustainable, healthy diets. As an outcome, sustainable healthy diets were defined as “dietary patterns that promote all dimensions of individuals’ health and wellbeing; have low environmental pressure and impact; are accessible, affordable, safe and equitable; and are culturally acceptable” (FAO and WHO, 2019, p. 9). Hence, sustainable healthy diets regard health aspect as well as environmental and sociocultural aspects. Any assessment of dietary patterns and how these relate to sustainability must entail these three dimensions, allowing to evaluate potential trade-offs involved in stimulating sustainable healthy diets, and supporting policymakers in identifying appropriate interventions to mitigate these.

The FAO/WHO definition of sustainable diets entails 16 principles. Key terms in these principles are the variety of foods that is adequate in energy and nutrients (health dimension); safely produced foods (e.g., with minimal use of chemicals and antibiotics) (health dimension); maintaining GHG emissions in line with 1.5 °C target and preserving biodiversity (environmental dimension); and accessible, desirable foods built on and respecting local culture, avoiding adverse gender-related impacts (as social dimension). While the FAO/WHO definition has set a common definition, how to assess compliance with or progress with respect to these multi-dimensional goal embodied in the definition is not yet elaborated in the guiding principles document.

Most sustainability assessments of healthy diets focus on only one or two dimensions but rarely on all three dimensions simultaneously. For instance, Eme et al. (2019) shows in a systematic review that out of fifty-two studies that considered environmental, nutritional and health, or socioeconomic dimensions of sustainable diets, thirty-three studies assessed the environmental effects of different dietary practices, twelve of the studies examined the nutrition and health indicators used for assessing sustainable food systems, and eight studies assessed the social and economic costs of diets.¹

In a more recent systematic review Webb et al. (2023) concludes that “a rising number of studies consider the cost/affordability of dietary scenarios in relation to optimized environmental and health outcomes. However, only six publications incorporate social sustainability outcomes, which represents an under-explored dimension of food system concerns” (ibid, p. 1). This review further points out that studies on the sustainability impact of dietary shifts often focus on only one actor in the food system (for example on the farmer or on the consumer) and fall short in motivating a particular choice of indicators. Webb et al. (2023) argue for a greater need to capture social and economic dimensions of sustainable healthy diets, as well as an integrated assessment along the whole value chain, and not just for consumers [see also Nicholls and Drewnowski (2021)].

Machado et al. (2023) examine the extent to which 48 commonly used metrics for measuring diet quality take into account the 16 principles captured in the FAO/WHO definition. They conclude that none of the 48 assessed dietary pattern metrics captures all dimensions of sustainable diets. For instance, none of the assessed metrics examined in Machado et al. (2023) consider the principle to preserve biodiversity or to minimize the use of antibiotics in food production, whereas only one (EAT-Lancet Diet Index, EAT-LDI) considers maintaining GHG emissions, water and land use, nitrogen and phosphorus application, and chemical pollution within set targets. Also, none of the assessed metrics considers the principle of food accessibility or to avoid adverse gender-related impacts. Moreover, the authors point out that “regarding conceptual underpinnings, very few metrics were developed with the intended purpose of reflecting simultaneously health, environmental, and sociocultural aspects of diets” (ibid, p. 156).

Other measures not covered in Machado et al. include, for example, the Reference Diet Deprivation (ReDD) index, a measure of diet quality based on a comparison of household consumption of food groups with optimal consumption levels in a selected reference diet (Pauw et al., 2023). As the method includes diet costs and consumption gaps against the reference diet some aspects of sustainability are captured implicitly. The Global Diet Quality Score (GDQS) is another diet quality metric that partially captures sustainability by measuring the quantities of healthy and unhealthy food groups consumed, yielding a better understanding of dietary risks for nutrient inadequacy or non-communicable diseases (NCDs) (Bromage et al., 2021). Yet, the environmental and sociocultural aspects of sustainable healthy diets are generally missing in these metrics.

Neta et al. (2023) assess studies that use an index or metric for assessing sustainable healthy diets based on the EAT-Lancet Commission Summary Report. Examples of such indices are WISH (as mentioned above) assessing the impact of a diet on disease (risk) and environmental indicators, and SHED (Sustainable Healthy Diets) an approach evaluating 30 items allowing for the assessment and scoring of sustainable and healthy eating patterns (e.g., Liz Martins et al., 2023; Tepper et al., 2021). Neta et al. (2023) find that the 14 analysed indices often do not, or hardly, take the social dimension into account. Yet, as embodied in the FAO/WHO definition, social aspects are critical to assess who wins and who loses as diets change and food systems transform. These findings are in line with previous reviews (Jones et al., 2016; Harrison et al., 2022), demonstrating that sociocultural factors are least included in studies evaluating sustainable healthy diets.

Based on this brief overview of the most relevant and recent literature on the definition and measurement of sustainable healthy diets, we conclude that:

A sustainable food system can provide sustainable healthy diets, and can be examined holistically as its performance is the result of connections and relationships among all actors in the system (FAO and WHO, 2019). Several authors have proposed additional or alternative indicators to better capture the four food system dimensions for assessing sustainability of healthy diets. For example, Aldaya et al. (2021) and Eme et al. (2019) propose a comprehensive set of indicators that can be applied from the national to a micro-level of (groups of) households. The proposed indicators include not only frequently used indicators such as carbon footprint or water footprint but also elements that are generally missing in other assessments such as agrobiodiversity or human toxicity (see Supplementary Tables SM1, SM2 for indicators proposed by these two sources). Other examples include Béné et al. (2019) and Hebinck et al. (2021) (see Supplementary Tables SM3, SM4). All four studies mentioned base their integrated approach on extensive analyses of existing frameworks to measure sustainable food systems and suggest indicators and metrics that can be sourced from existing (publicly available) databases. However, these studies are primarily conceptual and methodological in nature, highlighting shortcomings and underrepresented aspects in the literature reviewed, but except for Béné et al. (2019) do not empirically test their frameworks.

A comprehensive multidimensional index for sustainable food systems developed by Norde et al. (2023) builds on de Carvalho et al. (2021) to measure sustainability of food systems in 26 Brazilian states. Norde et al. (2023) expanded the metrics applied by de Carvalho et al. (2021) to 46 indicators that together form the index spanning environmental, social, economic, and nutritional dimensions. Dimensions and themes covered in this framework are summarised in Table 1 below. The full composition of the multidimensional index measuring sustainability of the food system is presented in Supplementary Table SM5.

Norde et al. (2023) implement their multidimensional index making use of official Brazilian datasets. Using normalization and aggregation techniques, scores for each dimension and an overall sustainability index score were constructed, making a performance ranking of the 26 Brazilian states possible. Analyses of correlation between dimensions reveal an inverse relation between the economic and environmental dimension, while the economic dimension positively related to the social and nutritional dimension. In addition, states could be grouped into four clusters differing in their food system sustainability profile.

The analytical framework developed by Norde et al. (2023) is more comprehensive in its elaboration of indicators compared to the ones by Eme et al. (2019), Béné et al. (2019), Aldaya et al. (2021), and Hebinck et al. (2021). In addition, to the authors’ knowledge, Norde et al. (2023) is one of the very few recent studies that covers all four dimensions of sustainable food systems and empirically tests sustainability performances at a sub-national level. With its empirical application in Brazil and the broad set of social and nutritional indicators (which are typically brief or missing altogether in other studies), the framework by Norde et al. (2023) provides a useful starting point for assessing the sustainability of current food consumption patterns in LMICs, such as Ethiopia which, like Brazil, is a federal state where states have some degree of administrative autonomy and which is being used as case study to test the approach at subnational level. While our approach allows for a relatively straightforward application in the Ethiopian context and highlights distinct and useful differences between the different Ethiopian states, we have not pursued a deeper contextualization. Such would include the possible inclusion of indicators and metrics that capture distinct elements of the Ethiopian or even subnational food systems. At the same time, we recognize a potential trade-off between generating more local contextualisation vis-à-vis maintaining a generic set of indicators, allowing for comparison across regions and even countries. We leave this for a future exercise.

In this case study, we consider food systems at sub-national level (across regional states³) in Ethiopia, based on sustainability scores of the four dimensions described above, including social equity, nutrition security, environmental protection, and economic access. Lists of indicators for each dimension with available data are compiled from various sources, including the Ethiopia Socio-economic Survey (ESS) (also known as Ethiopia LSMS-ISA 2018/19), the Demographic and Health Survey (EDHS), the Ethiopia statistical abstract, the Institute for Health Metrics and Evaluation (IHME), Global Burden of Disease Study 2021 (GBD 2021) results, and the International Food Policy Research Institute (IFPRI), 2020, AReNA’s DHS-GIS Database. It should also be noted that the data used in this study predate major shocks experienced in the country, including the COVID-19 pandemic and the conflict in northern Ethiopia. Also, not all indicators were observed in the same year. Although the use of datasets from different years (and data sources) could be a limitation of this study, its likely effect on the findings is not expected to be substantial. This is because the degree of structural transformation and economic inclusion in Ethiopia has remained persistently low over time, as documented by Rohne Till (2022) using historical data spanning 1950–2020.

The number of (sub)indicators for which data is available is eleven for the social equity dimension, eight for nutritional security dimension, seven for environmental protection dimension, and nine for economic access dimension. Tables 2–5 summarize indicators for the four dimensions along with corresponding mean values and their standard deviation (SD), and the sign of the potential relationships (RS) with sustainability. For example, the relationship between the share of household expenditures on food and sustainability is assumed to be negative (see the RS-column in Table 2). This is because a higher share of food expenditures points to lower food affordability and, consequently, lower sustainability. More detailed interpretations of how each (sub)indicator is associated with sustainability are provided by Norde et al. (2023).

The list of indicators and corresponding values for each regional state is in Table 6. For sub-indicators that capture very similar aspects within a particular dimension, the mean value of those sub-indicators is considered. The (sub)indicators with available data for the social equity dimension are organized under thematic areas of food security, equity and work conditions (Table 2). While the choice of indicators follows Norde et al. (2023), the number of indicators across themes within each dimension is not the same as in Norde et al. (2023) due to data availability. For example, in the social equity dimension, more data is available for (sub)indicators in the food security theme than for social equity and working conditions. Similarly, for the nutritional security dimension, there exists more data for (sub)indicators on nutrition-related diseases than for dietary intake (Table 3). Similarly, relatively few data are available for indicators in the environmental protection dimension, with only seven (sub)indicators to capture water availability, soil use and toxicity, and deforestation can be considered (Table 4). In contrast, for the economic access dimension data for nine (sub)indicators capturing broader themes, including profitability, productivity, job market, food cost and infrastructure are available (Table 5).

We emphasize that although we follow the approach and thus the themes and indicators per dimension defined by Norde et al. (2023) as closely as possible in this study, some comments are warranted. For example, it could be argued that undernourishment, a subtheme of food security (Table 2), would be part of nutrition security (Table 3). Unlike the other tables, the indicators for measuring performance on the environmental dimension present a mix of food system environmental drivers or external effects, as well as food system outcomes. While precipitation is an exogenous environmental driver, higher levels are generally associated with increased possibilities for arable production systems, including availability of nutritious diets. Deforestation, typically being an environmental outcome of food systems change may have an ambiguous effect on healthy diets—in the short run increases in agriculture production at the result of forests could actually increase sustainable healthy diets. In the long-run deforestation may pose a risk to the ecological foundation, including water and nutrient cycles, that shape food production.

Also, indicators of the economic access dimension could benefit from adding indicators on infrastructure and connectivity to capture regional trade. Cultural differences among regional States are implicitly captured, as the administrative organization of Ethiopia’s regional States is largely based on key cultural attributes such as language and ethnicity, as well as broad underlying differences in the culture of food consumption and production. While this argument may be weaker in the case of the SNNP region, given its greater internal diversity of ethnic groups and languages, food cultures and food system practices across these groups are largely similar based on the dominant cultivation of Enset and roots and tuber crops (e.g., Borrell et al., 2020) Nonetheless, this approach avoids making differences in sociocultural sustainability across regional states explicit, beyond the social indicators listed in Table 1. This is a pragmatic choice motivated by the observation that there is neither clear conceptual guidance on measuring the sociocultural dimension of food systems [see, e.g., Biesbroek et al. (2023) and House et al. (2024)], nor specific contextualized relevant metrics in the case of Ethiopia. This demonstrates that choosing indicators to measure a food system dimension is a difficult process; all studies aiming to measure food system sustainability struggle with this. We are also aware of the limitations and potential criticisms that can be raised regarding the choice of and proxies for indicators used in this study.

Closely examining the (sub)indicators in each dimension, the data reveal stark differences over food system outcomes across regional states (Table 6). For example, in the social equity and nutrition security dimensions, Addis Ababa tends to be among top two regions with relatively better food system outcomes, including food affordability, food accessibility, the share of underweight in the population, food security prevalence, the share of people with higher dietary diversity, and the prevalence of stunting among children. In addition, gender equity in agricultural activities is the largest in SNNP and Benishangul Gumuz; the share of child labour in agriculture is the largest in Amhara and SNNP; and poverty prevalence highest in Amhara and Gambella. On the other hand, other regions appear to have relatively better outcomes in the environmental protection dimension, for example, Benihangul Gumuz in terms of precipitation, vegetation cover, deforestation; Gambella in terms of land use and the application of chemical fertilizers; and SNNP in terms of crop diversity and precipitation. Similarly, regions with better performance in the economic dimension include Benishangul Gumuz and Oromia in agricultural income, Harari in land productivity, Afar in total farm income, SNNP in food production employment, Tigray in income distribution, Harari in access to electricity by farm households, Benishangul Gumuz in access to agricultural extension services, and Somali in inequality of land distribution. While treating individual indicators can be useful for specific policy targeting, aggregating various (sub)indicators within a dimension may capture synergies and trade-offs among them and may help to better understand the sustainability performance of each regional state across food system dimensions. Hence, in what follows, we briefly describe the method of analysis followed before presenting the relative performance of regional states in Ethiopia in terms of providing sustainable healthy diets.

Scores for each of the eleven regional states in Ethiopia are calculated involving multistep procedures, following Norde et al. (2023) and de Carvalho et al. (2021):

Note that while the aggregation of heterogeneous indicators represents a limitation of this study, this approach remains the subject of ongoing academic debate and is nevertheless widely used. Further, due to severe data constraints and the small number of States included in the analysis, sensitivity analyses—such as varying the number of indicators, altering combinations of indicators within and across food system dimensions, or applying weighting schemes in the absence of theoretical or empirical justification—was not feasible. Accordingly, the results of this study should be interpreted with caution and can at best be used to rank States according to their relative performance across food system dimension scores and the composite index. The study may therefore be viewed as a preliminary effort intended to highlight data gaps and motivate further research using more comprehensive and relevant sub-national level data for Ethiopia.

Figure 1 presents the relative performance of each regional state in Ethiopia across four food system dimensions. In the social equity dimension, the Benishangul Gumuz and the Somali regions are the top and bottom performers, respectively (Figure 1a). The social equity dimension of sustainable healthy diets are above the mean in Dire Diwa, Addis Ababa, Oromia, and Harari, including the largest urbanized regions in the country (Addis Ababa, Dire Dawa and Harari). With respect to the nutrition security dimension, the mean score for Gambella is the largest, followed by Harari, Somali and Dire Dawa that also rank above the mean (Figure 1b). In contrast, Tigray followed by Amhara, Addis Ababa, Oromia and SNNP rank relatively low on this dimension. This is notwithstanding the fact that all these regions, except for Addis Ababa, are key in domestic food production. For example, major staple cereals in the country such as teff, wheat, maize, sorghum and barley are produced mainly in Amhara, Oromia, SNNPR, and Tigray (Tafesse et al., 2011), pulses mainly in Amhara and Oromia (International Food Policy Research Institute (IFPRI), 2010); and low-land and pastoralist regions such as Afar and Somali are traditionally known for livestock production (USAID and GoE, 2010). This finding appears to suggest that the nutrition security dimension of sustainable healthy diets goes beyond food production alone.

As noted above, the data used for this analysis predates the two-years long conflict in northern Ethiopia that began in November 2020 and has had negative impacts on the economy as a whole and the disruption of food systems among regions that were directly affected by the conflict, including Tigray, Amhara, and Afar (Center for Preventive Action, 2023). Hence, the results represent best-case historical scenario rather than current conditions. Nonetheless, the aforementioned three regions score below the median in three out of four food system dimensions, including social, nutritional, and environmental. The Somali region ranks relatively lower in the environmental protection and economic access dimension scores, while Benishangul Gumuz and Oromia ranks relatively higher in these dimensions(Figures 1c,d).

As shown in Figure 1, no regional state consistently ranks at the top, or at the bottom in each of the dimensions of sustainable healthy diets; but each state ranks above the median in at least one dimension. This observation suggests that describing sustainability across the four dimensions as presented in this chapter, appears to provide better insights for federal and state governments to prioritize investments and to assess the policy and institutional support needed to improve sustainable healthy diets across all dimensions. The approach followed in this study, relying on scores for the different dimensions of sustainable healthy diets, could further be used to track and assess the impact of policies aimed at food system transformation, both spatially as well as over time. Such nationwide government policy initiatives targeting food system transformation for healthy and sustainable diets in Ethiopia, albeit with noticeable differences in state-level entry points (including focus on crops)include Yelemat Tirufat (“bounty of the basket,” an initiative aimed at improving dairy, poultry, egg, honey, fish and meat production), the Irrigated Wheat Production Movement (that expands wheat production in new frontiers of irrigated systems in both lowland and highland areas), the Green Legacy Initiative (major tree planting undertakings that include fruit trees aimed at reducing vulnerability to climate change and other environmental losses).

In addition, an aggregate measure that brings together all food system dimensions is equally useful in providing a more general snapshot. As discussed, sustainable healthy diets is the outcome of connections and relationships among all actors and components of the food system, which is captured by the multidimensional aggregate score of sustainable healthy diets (Figure 2), despite the limitations of an aggregate indicator. The Somali and Benishangul Gumuz take bottom and top places in the aggregate sustainability ranking, respectively. The Somali region is characterized by pastoral and agro-pastoral livelihoods, while livelihoods in Benishangul Gumuz are based on production of major cereals, coffee, pulses as well as mining (USAID and GoE, 2010). Again this finding drives the argument that sustainable healthy diets are the result of the broader food system and not just of food production. It is worth noting that the relative performance of four regional states including Somali, Tigray, Afar, and Amhara in dimensions scores was below the median in three out of four dimensions which is also reflected in the ranking of states with the aggregate score (Figure 2), only suggesting that sustainability healthy diets is understood better by the sustainability of its dimensions (see Figure 1).

Much has been written about sustainability albeit with variations in frameworks, dimensions, and metric propositions. What seems to be key behind the concept of sustainability is the need to increasing the quality of life of both the present and future generations while paying attention to environmental, social and economic considerations. In the context of healthy diets, attaining sustainability requires full understanding of all food systems activities and the potential synergies and trade-offs among them, and the needed change in behaviors among different actors. Based on a review of earlier studies that link sustainability with healthy diets, we observe that four dimensions of sustainability appear to be almost universally acknowledged, including ecological, economic, sociocultural, and food & nutrition security. However, the specific metrics and indicators for each of these dimensions could depend on the context, for example, the scale at which sustainability is assessed, the geographic focus, type of livelihoods, and the stage of technological advancement in the food system, among others. This is not just because of data availability, but also because each indicator may not carry the same meaning and relevance across different regions or countries.

In this review, we have identified potential indicators that are broadly used to assess healthy diets sustainability in the food system (see Supplementary material for summary tables of indicators proposed by different sources). For empirical demonstration, we considered the case of Ethiopia at sub-national level, applying an approach by Norde et al. (2023) for Brazilian federal states. We found that, as expected, being a food producer or having relatively better production potential may not guarantee better food system outcomes such as food security and nutrition outcomes, much less sustainability of the same. For example, regional states known for production of cereals in the country, including Amhara, Oromia, SNNP, and Tigray; and of pulses including Amhara and Oromia; and of livestock products including Afar and Somali did not show higher relative performance in the sustainability score across food system dimensions or in the overall aggregate score. In fact, Addis Ababa and Dire Dawa, the two most urbanized regions in the country with little to no food production, ranked fairly better in terms of food affordability and accessibility, and underweight prevalence in the population, suggesting that food system outcomes and sustainability could rather be the result of connections and relationships among all food system components.

A comprehensive analysis of sustainable healthy diets that takes into account various dimensions would require substantial data at various levels in the food system. As noted above, such analysis can facilitate better understanding of complexities in the food system, including synergies, trade-offs, and unintended consequences of potential interventions at different parts of the food system. Thus, it is necessary to broaden the view of activities in the food supply chains from those of companies in the food environment to activities focused on production, transport and processing. Doing so may allow the application of more advanced methodologies such as life cycle assessments that are used to quantify the environmental impact of products or services that may yield more accurate results that would determine healthy diets sustainability (e.g., see Broekema et al., 2020). This would require data collection from all actors in the food supply chains including farmers, transport companies and food processors. Approaching them with the aim of measuring the sustainability performance of production method and trading modalities may be a time-consuming (and hence costly) matter, but is necessary to gain a more complete and therefore better insight into the true sustainability performance of producing a healthy diet.