Where do animal-source foods fit in our concepts of ‘healthy’ and ‘unhealthy’ foods, and how do we measure them?

By Jody Harris
World Vegetable Center and Institute of Development Studies

Twitter: @justjody23 @go_vegetables @IDS_UK


This week, I attended the fabulous Agriculture, Nutrition and Health Academy Week – ANH2020 – for my tenth straight year (well, it was the LCIRAH conference before 2016, but who’s counting?). The conference was supposed to be held in Malawi this time, but due to COVID-19, it was held online instead. While I really miss seeing friends and colleagues, there has been one major plus of this format: the pre-recorded presentations, which meant it was much easier to jump between sessions. 

In my new role as lead specialist on food systems for healthy diets for the World Vegetable Center, I was particularly interested in how we understand who has a healthy diet, so I jumped to some fantastic talks on this topic. Several were looking at metrics for measuring a healthy diet but had very different interpretations of which foods are healthy or unhealthy, and that got me thinking…

Classically in international nutrition research, we have focused on increasing the nutrient content of monotonous, staple-based diets for poor people in poor countries. This has led to dietary metrics such as the WHO Individual Dietary Diversity Score for children aged 24-59 months, and the FAO Minimum Dietary Diversity Score for Women. These have been widely used and have driven forward our ability to work on diets, which are the natural link between food systems and nutrition outcomes.

The emergence of the Nutrition Transition in low- and middle-income countries has shown up the limitations of these positive nutrient-focused metrics however: there are also foods we know people shouldn’t be eating for their health, such as ultra-processed foods and sugar-sweetened beverages, which confer no nutrition benefit and are positively associated with non-communicable diseases (NCDs) such as diabetes and heart disease.

A new generation of metrics is therefore trying to incorporate both nutrient adequacy and NCD risk into measurement of diets (and sometimes other things, such as affordability or sustainability, too). But in watching the ANH Academy talks, it seems we are not quite in agreement on what constitutes a healthy or unhealthy food group. The metrics I’ve seen count fruits, vegetables, fish, legumes and nuts as healthy; and saturated fats and sugar-sweetened foods and beverages (and ultra-processed foods where included) as unhealthy. Where they differ seems to be on animal source foods.

At the conference, Laura Trijsberg (Wageningen University) presented the World Index of Sustainability and Health (WISH Index) which aims to measure both nutrition and sustainability aspects of diets. The metric contains 13 food groups (Figure 1), of which 8 are classed as healthy, 3 as unhealthy and 2 as neutral. The study has tested how the score correlates with nutrient recommendations in Vietnam, finding low correlations with nutrient values. In another presentation, Sabri Bromage (Harvard University) presented the Global Diet Quality Score (GDQS). This 25-group metric (Figure 2) classes 16 foods as healthy, 7 as unhealthy and 2 as ‘non-linear’, meaning eating a little is healthy but a lot is unhealthy. The score correlates with measures of under- and over-nutrition in multiple large datasets, and requires validation in primary studies.

Screenshot from Laura Trijsburg's presentation on the WISH index (click to watch on YouTube).
Screenshot from Sabri Bromage's presentation on the GDQS (click to watch on YouTube).

What particularly intrigued me about these metrics is how they classified animal-source foods (foods of animal origin, including meats, dairy and eggs). Classically, these have been described as necessary for malnourished children in low-income countries because of their high nutrient-to-calorie ratio, meaning they are particularly dense in nutrients. This makes it easier to supply nutrients in the small portion sizes needed by young children, so many studies and programmes have recommended increasing both supply and demand of animal source foods in countries with lots of undernutrition. More recently, however, the high consumption of some animal foods has been associated with rising prevalence of NCDs. The contested EAT-Lancet Commission Report recommended a reduction in some animal foods in countries or population groups with higher intakes.

Both metrics presented at the ANH Academy take their start from the EAT-Lancet diet, and both recognise the need for nutrient-dense diets that avoid NCDs. But both categorise animal-source foods quite differently (Table 1). Eggs are classed as healthy in GDQS but neutral in WISH; dairy is healthy in WISH but divided into low-fat (healthy) and high-fat (non-linear) in GDQS; and red meat is unhealthy in WISH but non-linear in GDQS. Other speakers classified red meat as ‘healthy’ if it was boiled, and ‘unhealthy’ if it was fried, focusing on the preparation method rather than the food itself, but always as ‘energy-dense/nutrient-rich’; and some research respondents themselves classified meat as ‘unhealthy’ if it was perceived to be unclean or unsafe.

Table 1: Animal food classifications in two new diet metrics






Eggs, low-fat dairy


Red meat

Processed meat


Eggs, chicken

High-fat dairy, red meat

So who is correct? Do we even know? I’m not sure we do – and as the intense and sometimes vitriolic debate in the wake of the EAT-Lancet report highlights, there are some powerful interests at play, not least from the meat and dairy industries and the newer plant-based processed food companies. This blog hasn’t touched on the environmental or ethical aspects of animal foods, which add further polarising dimensions in particular to debates on animal foods. 

Academic work on healthy diets (and on the best food system responses towards this goal) now needs further work on understanding and extending the evidence behind the food classifications in the metrics we use, and it will be important to do this for different contexts and populations with different profiles of nutrition outcomes. An important start will be conversations after conferences like this to start to nuance our messages on what constitutes ‘healthy’ and ‘unhealthy’, and for whom.

More information about ANH2020 can be found here, including the interactive programme.

Jody Harris is a senior scientist with a research interest in policy and social interventions for healthy diets and nutrition. She conducts research into power in societies, including work on equity and marginalization; power in politics, including political science work on policy processes; and power in food systems, including the roles of public and private sector actors. Jody completed her PhD in international development policy at the School of Oriental and African Studies (SOAS) and her MSc in international nutrition at the London School of Hygiene and Tropical Medicine (LSHTM), both at the University of London in the UK. She has previously worked for the International Food Policy Research Institute (IFPRI) and the Institute of Development Studies (IDS) as well as several universities and international NGOs, conducting research and designing programs in various contexts in Asia and Africa, particularly Vietnam and Zambia. Jody will be bringing her food systems and policy focus to WorldVeg’s work on healthy diets, helping to update the flagship strategy, advise on the Center’s diet and nutrition work globally, and undertake research to further understanding of the role of vegetables in these complex systems.