The ecological nutrition paradigm

The German zoologist Ernst Haeckel is credited with coining the term ecology (‘Oekologie’) which he conceptualised as a subdiscipline of evolutionary theory focussing on the interplay between organisms and their environment^{(Reference Watts, Hoßfeld and Levit19)}. There is no evidence that he used the term in a nutrition context. Instead, it was the US scientist Ellen Swallow Richards, who in 1892 after previously visiting Haeckel’s laboratory, pioneered thinking about nutrition within an ecological context with her studies into ‘human ecology’ that focused on how humans influence environmental conditions which in turn influence human health^{(Reference Dyball and Carlsson20)}.

It was not until the mid-1970s that nutrition science scholarship within an ecological context began to gain significant momentum for nutrition research and policy applications when Leitzmann and colleagues developed a research program in what they referred to as ‘Nutrition ecology’^{(Reference Leitzmann21)}. This program was closely followed by the publication of Gussow’s seminal text, ‘The Feeding Web: Issues in Nutritional Ecology’^{(Reference Gussow22)}, which further stimulated scholarship in this nascent field. Independently, Leitzmann and Gussow emphasised that nutrition science needed to conceptualise food and diet as being inexorably connected to nature and sustainability. In her address to the pertinently titled, ‘Ellen Swallow Richards Lecture Series’, Gussow criticised nutrition science for looking at ever smaller and smaller aspects of food, breaking it down into microscopic pieces and,

Several years after this address Gussow along with Clancy proposed that future iterations of dietary guidelines should incorporate environmental as well as health goals^{(Reference Gussow and Clancy24)}. This proposal has now been taken up by the UN’s Food and Agriculture Organization and World Health Organization⁽²⁵⁾ and over one-third of the committees responsible for preparing national dietary guidelines around the world^{(Reference James-Martin, Baird and Hendrie26)}.

The conceptual basis to ecological nutrition was advanced significantly with the publishing and launch of the highly influential ‘New Nutrition Science project’ in 2005^{(Reference Cannon and Leitzmann27)}. The project provided a new conceptual framework for nutrition science that in effect proposed a ‘paradigm shift’ for nutrition science^{(Reference Cannon and Leitzmann2)}. The project was formulated at a workshop meeting of leading nutrition scientists, including several with significant experience in ecological nutrition research and policy scholarship, and held at the Justus Liebig University Giessen. The choice of location was symbolic as it was at Giessen that Justus von Liebig developed nutrition science in its original scope as a biochemical discipline. The key output from the workshop was The Giessen Declaration, which explains that the three dimensions of the new nutrition science are biological, social and environmental and that the science is concerned with personal, population and planetary health. The Declaration begins by stating:

A transdisciplinary approach to ecological nutrition policymaking

This review highlights that proponents of the ecological nutrition paradigm challenge the conventional nutrition science discipline’s ‘one-size-fits-all’ policymaking approach for tackling unhealthy and unsustainable diets. They argue there is a need to change what counts as evidence and how it is synthesised and translated. This argument aligns with McPhee and colleagues’^{(Reference McPhee, Bliemel and van der Bijl-Brouwer35)} assessment that complex policy problems need to be tackled with a transdisciplinary policymaking approach. Their illustration comparing multidisciplinary, interdisciplinary and transdisciplinary policymaking approaches is adapted for the policy problem of unhealthy and unsustainable diets and illustrated in Figure 1 which sets out the:

1. i) Multidisciplinary approach: Draws on the disciplines of nutrition science and ecological science to provide complementary sets of evidence to consider in policymaking.

2. ii) Interdisciplinary approach: Integrates the nutrition science and ecological science disciplines by combining their knowledge and methods, drawing on ecological and evolutionary theoretical insights.

3. iii) Transdisciplinary approach: Collaborates with policymakers to apply innovative ‘fit-for-purpose’ policymaking such that the policy problem itself is conceptualised and used to inform (arrow is double-headed) the design of fit-for-purpose decision-making processes to determine how ecological nutrition evidence is synthesised and translated in policymaking.

Figure 1

Comparing multidisciplinary, interdisciplinary and transdisciplinary policymaking approaches for the policy problem of unhealthy and unsustainable diets.

Nutrition paradigm guidance for nutrition policymaking

At its core nutrition policymaking involves using decision-making processes to convert nutrition evidence into nutrition policies to tackle nutrition problems. The main decision-making processes are: evidence synthesis methods used to inform nutrient reference values, nutrition policy statements and dietary guidelines; and evidence translation metrics used to convert this nutrition science guidance into an individual food’s ‘healthiness’ rating for informing nutrition policy actions such as front-of-pack food labelling, food taxation, food procurement and restrictions on food marketing. Both processes are associated with scientifically and politically mediated ‘trade-offs’.

The design of evidence synthesis methods needs to manage a trade-off between internal and external validity^{(Reference Guba36)}. In a nutrition policymaking context, internal validity broadly refers to the degree of confidence with which a study’s findings of an association between a nutrition exposure and a health and sustainability outcome can be considered accurate. External validity broadly refers to the extent to which a study’s findings relate to the ‘real world’ within which food and diets are consumed. The trade-off being that designing a study method to strengthen one type of validity often weakens the other type of validity. For example, evidence synthesis methods designed to assess evidence quality on the ability of a study design to control for internal bias may be well suited for investigating associations between isolated nutrient supplements and short-term physiological outcomes. However, their use is limited for explaining and predicting complex and context-rich associations between real world dietary pattern exposures and long-term health and sustainability outcomes^{(Reference Hébert, Frongillo and Adams12,Reference Blumberg, Heaney and Huncharek37)} .

The design of evidence translation metrics needs to manage a trade-off between reductionism and holism^{(Reference Lawrence, Dickie and Woods38)}. There are three main types of food classification system providing metrics for assessing a food’s healthiness:

1. i) Nutrient profiling-based systems such as the Health Star Rating (HSR)⁽³⁹⁾ and Nutri-Score⁽⁴⁰⁾ which assess a food’s healthiness according to its content of certain nutrients⁽⁴¹⁾.

2. ii) Food profiling-based systems such as the Nova food processing classification system^{(Reference Monteiro, Cannon and Levy42)} which includes the ultra-processed food (UPF) concept in assessing a food’s healthiness according to its extent and purpose of industrial processing.

3. ii) Dietary pattern profiling-based systems which assess a food’s healthiness according to dietary guideline recommendations, e.g. using a binary approach to assess a food categorised into a nutritious food group as ‘healthy’ and a food categorised into the ‘discretionary’ food group as ‘unhealthy’^{(Reference Lee, Rangan and Allman-Farinelli43)}.

In a nutrition policymaking context, reductionism refers to nutrient-based food classification systems and holism refers to food- and dietary pattern-based food classification systems. The trade-off being that designing a food-level metric based on a nutrient-level exposure often misaligns the food’s healthiness assessment based on food- and dietary pattern-based exposures, and vice versa. For example, 73 % of foods displaying a ‘pass’ rating of 2·5/5 or more ‘health’ stars, according to the HSR nutrient profiling-based system were reported as being ultra-processed foods (UPFs)^{(Reference Dickie, Woods and Baker44)}. Similarly, 53 % of foods displaying a rating of 2·5/5 or more health stars were categorized as discretionary foods according to the Australian dietary guidelines^{(Reference Dickie, Woods and Baker44)}. There is a lack of consensus on which type, or combination, of food classification system(s), provides the most robust evidence translation metric for informing nutrition policy actions. A circular argument can ensue. Decision-makers may argue it’s not clear whether nutrient-based systems are resulting in unintended consequences by creating ‘health halos’ for UPFs and discretionary foods, or whether food- and dietary pattern-based systems are inadequately capturing nutrient profiles.

Nutrition paradigms can influence nutrition policymaking by shaping how the scientific and political uncertainties associated with these two trade-offs play out. The next section of this review examines the contrast in two characteristics of policymaking guidance being used by the dominant medical nutrition paradigm with that that would be consistent with the ecological nutrition paradigm. These contrasting characteristics are outlined in Table 1.

Policymaking characteristic 2: Design of decision-making processes

Medical nutrition paradigm

Evidence synthesis method

The most prominent evidence synthesis method used to inform the development of nutrition policy guidelines is Grading of Recommendations Assessment, Development and Evaluation (GRADE). It is used by more than 120 organisations worldwide^{(Reference Guyatt, Hultcrantz and Agoritsas46)}, including by the World Health Organization in its guideline development activities⁽⁴⁷⁾. GRADE emerged from the evidence-based medicine movement of the 1990s^{(Reference Guyatt, Sackett and Cook48)}, with the purpose of evaluating medical interventions to support decision-making by clinicians concerned with safe and effective care of patients^{(Reference Sackett, Rosenberg and Gray11,Reference Djulbegovic and Guyatt49)} and has extended to public health settings over the following decades^{(Reference Guyatt, Hultcrantz and Agoritsas46)}.

GRADE uses a hierarchical design to rate evidence quality in systematic reviews in terms of a primary study design’s ability to control for internal bias, with evidence from randomized controlled trials (RCTs) rated high relative to evidence from observational studies^{(Reference Lichtenstein, Petersen and Barger50)}. In this way it manages the internal validity-external validity trade-off dilemma by prioritising internal validity as an integral dimension of the evidence synthesis method design. Judgements about the strength of a recommendation are then made by considering factors such as the balance between benefits and harms^{(Reference Blake, Durão and Naude51)}.

There are scientific uncertainties with using an evidence synthesis method based on assessing evidence quality using a hierarchy of evidence approach in which evidence from a RCT study design is rated relatively highly^{(Reference Penders, Wolters and Feskens6,Reference Jukola52)} . The RCT design has limited ability to specify and measure complex dietary pattern synergistic interactions among nutrients within foods or the cocktail affect among multiple novel ingredients in foods and their cumulative exposures over time. Also, the RCT study design often controls for the presence of contextual factors, that may help explain associations and instead viewing them as potential confounders.

There can also be compliance^{(Reference Temple53)}, and cost constraints^{(Reference Magni, Bier and Pecorelli54,Reference Williams, Ashwell and Prentice55)} when trial participants need to adhere to specific dietary patterns over decades to detect health and sustainability outcomes. These challenges can skew the body of evidence available to inform nutrition policy actions. An analysis of the Cochrane Library found that of the 470 systematic reviews available to inform nutrition policy actions, more were based on evaluations of nutrient supplement exposures than food and dietary pattern exposures combined, and only 5 % of those reviews included non-RCT study designs^{(Reference Naude, Durao and Harper56)}. Unintended consequences resulting from conventional nutrition policy decision-making processes have been observed. For example, guidelines recommending the substitution of industrial trans-fatty acids for SFA in certain foods did not result in reduced CVD risk as intended, and instead inadvertently increased CVD risk^{(Reference Schleifer57)}.

Evidence translation metric

Global^(58–60), regional^{(61,Reference Turck and Bohn62)} and national^{(Reference Labonté, Poon and Gladanac63)} nutrition organisations have prioritised nutrient-profiling-based food classification systems to guide the designs of evidence translation metrics to assess a food’s healthiness to inform nutrition policy actions. This prioritization is consistent with the observation that decision-making for managing the reductionism-holism trade-off dilemma when designing evidence translation metrics is influenced by a nutrient-centric worldview towards the causes of and solutions to nutrition problems^{(Reference Fardet and Rock64,Reference Scrinis65)} .

Nutrient profiling models reduce complex associations between dietary patterns and health outcomes down to the amounts of a small number of certain nutrients in a food to ascribe a healthiness rating to that food. The reductionist worldview that underpins the design of these models, and the algorithms they use to calculate nutrient scores to inform healthiness ratings, do not capture the biological and ecological underpinnings of dietary patterns. Instead, the designs of nutrient profiling models rely on simplistic abstractions of complex nutrition concepts and technical assumptions, creating uncertainties about their construct validity in measuring a food’s healthiness^{(Reference Cooper, Pelly and Lowe66)}. For example, the models are unable to account for the complex and context-rich interactions among the thousands of other nutrients and bioactive compounds in foods. Also, they are unable to determine if the food ‘vehicles’ within which the target nutrients are located are nutritious minimally processed foods or unhealthy UPFs and discretionary foods^{(Reference Dickie, Woods and Baker44)}.

Ecological nutrition paradigm

Decision-making for the ecological nutrition paradigm is currently not overtly practiced. The intention is that it would replace the medical nutrition paradigm’s prescriptive evidence synthesis method and nutrient-centric evidence translation metrics with fit-for-purpose decision-making processes designed to be relevant to the biological, social and environmental dimensions of associations between nutrition exposures and nutrition outcomes. Identifying a relevant design for decision-making processes would involve drawing on theories to help conceptualise the causes of and solutions to nutrition problems.

Ecological and evolutionary theories provide insights into how the physical and chemical characteristics of different types and forms of nutrition exposures influence their associations with population^{(Reference Ulijaszek, Mann and Elton67–Reference Garnås71)} and planetary^{(Reference Rockström, Edenhofer and Gaertner72–Reference Elton74)} health outcomes. These insights can help contextualise epidemiological and mechanistic evidence of associations between nutrition exposures and health and sustainability outcomes. This contextualising facilitates greater consideration of external validity and holism concerns in decision-making than is evident in current decision-making processes. As Kreiger comments,

‘modern epidemiology often seems more concerned with intricately modeling complex relationships among risk factors than with understanding their origins and implications for public health. … models do not exist independently of theories. Theories attempt to explain why phenomena exist and are interrelated.’ ^{(Reference Krieger75)}, pp. 887–91.

This review considers two well established theories that are relevant to ecological nutrition: the nutrition transition theory^{(Reference Popkin76)}; and the food synergy theory^{(Reference Jacobs, Gross and Tapsell77)}. These theories help explain the context-rich nature of associations between nutrition exposures and nutrition outcomes. This explanation is important because the complexity of an association has implications for designing decision-making processes.

Nutrition transition theory

The nutrition transition theory proposes that modernization, urbanization, technological change and increased wealth associated with economic development results in changing nutrition exposures over time, thereby affecting the nutritional status of populations^{(Reference Popkin76,Reference Popkin78)} . The theory is informed by evolutionary principles that human physiology has evolved over millions of years for eating and metabolizing diets sourced from ‘ecologically stable food systems’, i.e. food systems which operate within ecological parameters to be sustainable for current and future generations. Food and ecological systems were inter-linked and ‘hard-wired’ through two immutable forces: evolutionary processes; and the planet’s physical boundaries operating within finite limits for accessing and disposing of food between organisms and the environment^{(Reference Eaton, Konner and Shostak79–Reference McMichael81)}.

Later studies added to the nutrition transition theory, proposing that the industrialization of food systems has led to, and been enabled by, an increasing use of food processing. Food processing can contribute to healthy and sustainable diets by helping increase a nutritious food’s shelf-life and storage as well as its convenience for transportation, trade and home preparation. However, a problem that has emerged with the transition to industrial food systems is its association with a global proliferation of UPFs characterized by novel physical structures and novel chemical compositions^{(Reference Monteiro, Moubarac and Cannon82,Reference Popkin83)} . This transition has occurred too rapidly for human physiology and planetary ecology to adapt to these unprecedented food exposures, resulting in adverse health and sustainability outcomes^{(Reference Ulijaszek, Mann and Elton67,Reference Pontzer and Wood84–Reference Popkin, Ng and Taillie88)} . A phenomenon Boyden refers to as ‘evodeviation’^{(Reference Boyden89)}.

Food synergy theory

In 2003 Jacobs and Steffen advocated for nutrition research to take more account of the interactions among different types of nutrition exposures, i.e. nutrients, foods and dietary patterns^{(Reference Jacobs and Steffen90)}. They proposed a ‘Food synergy’ framework (theory), where food synergy was defined as the additive or more than additive influences of foods and food constituents on health. According to the food synergy theory the activity of nutrients within physiological systems is influenced by the composition and structure of the foods within which they are consumed, and the dietary pattern within which those foods are consumed^{(Reference Jacobs, Gross and Tapsell77,Reference Jacobs and Steffen90,Reference Jacobs, Tapsell and Temple91)} . From a food compositional perspective, there are chemical interactions among nutrients within foods and among foods within dietary patterns which can enhance or mitigate their bioavailability. For example, the presence of vitamin C in a food or diet can promote the absorption of nonheme iron^{(Reference Lynch and Cook92)}. From a food structure perspective, the physical structure of the food matrix can influence nutrient bioavailability^{(Reference Miller, Ragalie-Carr and Torres-Gonzalez93)}. For example, the metabolic effect of saturated fat present in butter and cheese differs because the contrasting food matrix structures of these two dairy foods has a differential influence on the bioavailability of this nutrient^{(Reference Thorning, Bertram and Bonjour94)}.

The synergistic interactions among the different types of nutrition exposures affects their health potential. A nutrient’s health potential is determined by the presence of other nutrients in a food, a food’s health potential is determined by more than the sum of the nutrients it contains, and a dietary pattern’s health potential is determined by more than the sum of the foods (and nutrients within those foods) it contains^{(Reference Hoffmann95–Reference Ottaviani, Ensunsa and Fong97)}.

These interactions among the different types of nutrition exposures and their associations with physiological systems have ecological underpinnings. From ecological and evolutionary perspectives, consuming diverse diets with multiple inherent nutrient interactions has conferred a survival advantage. This is because the ability of synergistic interactions to increase nutrient bioavailability means that nutrient requirements can more securely be met by consuming different combinations of many foods, even moderately nutrient-dense foods, than having to rely on the constant availability of a relatively small number of nutrient-dense foods. The differential survival advantage has resulted in the natural selection of the gene combinations, and consequent physiological programming, of those humans who have consumed diverse diets.

Insights from the nutrition transition and food synergy theories highlight that interactions among nutrition exposures and their associations with health and sustainability outcomes vary with exposure form (degree and purpose of industrial processing) and type (nutrients, foods, dietary patterns). Specific aspects of these interactions and associations may be modified by contexts such as a population’s age, gender, lifestage^{(Reference Whitney, Rolfes and Crowe98)} and when foods and diets are consumed^{(Reference Manoogian, Chaix and Panda99,Reference Taylor, Allman-Farinelli and Chen100)} . The theoretical insights into the importance of distinguishing the form and type of nutrition exposures helps identify how a relevance criterion can be incorporated into the designs of evidence synthesis methods and evidence translation metrics.

Table 1

Nutrition policymaking characteristics of the medical and the ecological nutrition paradigms

Is an ecological nutrition paradigm shift warranted?

Kuhn popularised thinking about scientific paradigms with his theory that science progresses through a series of paradigm shifts^{(Reference Kuhn10)}. He argued that when an existing paradigm, characterised by shared theories, assumptions and methods, is no longer able to adequately explain a phenomenon a ‘scientific revolution’ occurs, and a new paradigm is adopted. Throughout its history, nutrition science has evolved through a series of eras marked by paradigm shifts in explaining and responding to nutrition problems^{(Reference Ridgway, Baker and Woods101–Reference Mozaffarian, Rosenberg and Uauy103)}. The current contested views and uncertainties towards nutrition policymaking to inform fit-for-purpose policies to promote healthy and sustainable diets are consistent with conditions that Kuhn set out as predictive for potential paradigm shifts.

Here the policymaking orientation and design of decision-making processes of the ecological nutrition paradigm are assessed for their fit for the purpose of informing policies to promote healthy and sustainable diets, to help determine if a paradigm shift is warranted.

Implications of the ecological nutrition paradigm shift for future decision-making processes

An ecological nutrition paradigm shift will be transformative for nutrition research and policymaking. The operationalisation of a fit-for-purpose policymaking approach will have significant implications for the design of decisions-making processes. Evidence synthesis method innovations will be needed to incorporate the relevance^{(Reference Williams, Buttriss and Whelan109,Reference Katz, Karlsen and Chung110)} or appropriateness^{(Reference Parkhurst and Abeysinghe16)} of a study design as a criterion when assessing evidence quality. For example, incorporating a relevance criterion might lead to evidence derived from non-experimental study designs, such as observational studies and anthropological investigations, receiving greater status in dietary guideline development. Also, the ecological nutrition paradigm’s focus on foods and dietary patterns being more than the sum of their nutrients, provides a strong conceptual rationale to avoid the development of nutrient-based dietary guideline recommendations. In doing so it would be undesirable to simply replace internal bias considerations with external bias considerations. The challenge will be how to resolve the intractable trade-off between internal and external validity considerations in nutrition study designs.

Similarly, evidence translation metric innovations will be needed to incorporate a food- or diet pattern-based food classification system for assessing a food’s healthiness^{(Reference Dickie, Woods and Machado111)}. The challenge will be how to resolve the trade-off between reductionism and holism in food classification systems. For example, it may be that a food processing classification system such as Nova is used to classify a food as healthy (Nova food groups 1–3), or unhealthy (Nova food group 4 (UPF)). Then within that broad classification approach, a nutrient composition cut off level is applied to processed foods (Nova food group 3) such that a processed food containing an amount added sugar or salt above the cut off receives a lower overall health rating than an equivalent processed foods with no added sugar or salt.

Epistemology and the ecological nutrition paradigm shift

A shift to an ecological nutrition paradigm challenges current nutrition research and policymaking norms. Some scientists are questioning the paradigm’s robustness by evaluating its evidence base using evidence quality methods concordant with the medical nutrition paradigms’ parameters. For example, in an editorial in the journal, The Lancet Gastroenterology & Hepatology, the author(s) wrote there is a ‘vacuum of good evidence on UPFs’ to support UPF-related nutrition policy actions and consequently they state the evidence base ‘demands epistemological humility’⁽¹¹²⁾. However, it is an epistemological contradiction to use methods characteristic of an existing paradigm to evaluate evidence quality for a nutrition action aligned with a new paradigm. This is especially pertinent when it is these very methods the new paradigm is challenging on the grounds they are incongruent for the research question being asked or the policy purpose being sought. Yet to inform urgently needed ecological nutrition policy actions the editorial’s author(s) expects the UPF concept to demonstrate ‘clinical effects’ and use an evidence quality standard based on a hierarchy of evidence scale originally designed for evaluating health care in clinical settings for unwell individuals.

Previously, Monteiro^{(Reference Monteiro113)} noted that the science needed to prove his UPF concept was not available when it was first proposed. It struggled to be accepted by nutrition scientists who used studies that grouped foods according to their nutrient composition, rather than their extent and purpose of processing. At that time Monterio recognised the need for the equivalent of an ecological nutrition paradigm shift when he sought to expand the scope of the nutrition science discipline to encompass social, economic and environmental dimensions^{(Reference Monteiro113)}.

Conclusion

This review has shown how the ecological nutrition paradigm brings a transdisciplinary approach to nutrition research and policymaking. The approach is necessary to tackle the complexity and context-rich nature of modern diet-related health and sustainability problems. Its fit-for-purpose orientation when elaborated with ecological and evolutionary theoretical insights helps conceptualise the biological, social and environmental causes of and potential solutions to these dietary problems. These insights help identify relevant nutrition research and decision-making processes to inform policies. Their attention to relevance engenders confidence in the likely effectiveness and ability to avoid unintended consequences of policies informed within an ecological nutrition paradigm. A shift to the ecological nutrition paradigm to transform nutrition research and policymaking for promoting healthy and sustainable diets is overdue.