A Need for a Paradigm Shift in Healthy Nutrition Research

Abstract

Research in the field of sustainable and healthy nutrition is calling for the application of the latest advances in seemingly unrelated domains such as complex systems and network sciences on the one hand and big data and artificial intelligence on the other. This is because the confluence of these fields, whose methodologies have experienced explosive growth in the last few years, promises to solve some of the more challenging problems in sustainable and healthy nutrition, i.e., integrating food and behavioral-based dietary guidelines. Focusing here primarily on nutrition and health, we discuss what kind of methodological shift is needed to open current disciplinary borders to the methods, languages, and knowledge of the digital era and a system thinking approach. Specifically, we advocate for the adoption of interdisciplinary, complex-systems-based research to tackle the huge challenge of dealing with an evolving interdependent system in which there are multiple scales-from the metabolome to the population level-, heterogeneous and-more often than not- incomplete data, and population changes subject to many behavioral and environmental pressures. To illustrate the importance of this methodological innovation we focus on the consumption aspects of nutrition rather than production, but we recognize the importance of system-wide studies that involve both these components of nutrition. We round off the paper by outlining some specific research directions that would make it possible to find new correlations and, possibly, causal relationships across scales and to answer pressing questions in the area of sustainable and healthy nutrition.

Keywords: complex systems; data science; digitalization; health; nutrition.

Figure 1

Multilayer schematic representation of the interactions that are involved at three different complexity levels. These interactions link together food composition and socio-cultural covariates. Each layer is represented as a network of interactions among the components at that scale.

Figure 2

Percentage of countries whose food composition table quantifies each chemical compound for olive oil. We include as a quantified chemical those with an estimated quantity larger than 0, regardless of the specific amount or units shown in the database.

Figure 3

The number of chemical compounds associated with each source of vegetable oil in FooDB as of April 2020. From the total number of chemicals associated with each crop, the large majority are “expected [to be present] but not quantified”. The others can be associated with the ingredient in its raw form or in its oil. Most quantified compounds are associated with the raw product, and only a tiny fraction is explicitly associated with the oil. Note that for palm and rapeseed the amount of chemicals associated with the oil is almost the same as the quantified fraction, likely because these products are mostly consumed in their oil form.

Figure 4

A complex systems view of food consumption. (A) shows a comparison of the intake of legumes in Southern, Western and Northern Europe, which reveals the role of sociocultural factors. Looking at the ingredients that are commonly paired with these legumes in Spanish cuisine, one finds that this consumption is heavily related to the consumption of non-dairy animal products, although this is not true for the Indian cuisine sample (B). In (C), each circle represents a single ingredient used in at least one recipe containing beans, chickpeas or lentils. The size of the circle is proportional to the number of recipes containing the ingredient and the color the group they belong to. In all cases only the top 10 ingredients are labeled.