Abstract
Research on the impacts of dietary patterns on human and planetary health is a rapidly growing field. A wide range of metrics, datasets, and analytical techniques has been used to explore the role of dietary choices/constraints in driving greenhouse gas (GHG) emissions, environmental degradation, health and disease outcomes, and the affordability of food baskets. Many argue that each domain is important, but few have tackled all simultaneously in analyzing diet-outcome relationships.
Generated Summary
This systematic review examines studies published between January 2015 and December 2021 (inclusive) that analyzed dietary patterns in relation to at least two of the following four thematic pillars: (i) planetary health, including climate change, environmental quality, and natural resource impacts, (ii) human health and disease, (iii) economic outcomes, including diet cost/affordability, and (iv) social outcomes, e.g., wages, working conditions, and culturally relevant diets. The review systematically screened 2,425 publications by title and abstract and included data from 42 eligible publications. The study aimed to determine how many studies include social and economic indicators in their analysis of dietary patterns impacts on human and planetary health, what analytical approaches, methods, and metrics are used to address social and economic concerns, and what conceptual and empirical gaps persist that represent impediments to meaningful target-setting and policy change.
Key Findings & Statistics
- The majority of included publications (N = 29 of 42, 69.0%) were published between 2019 and 2021.
- Economic, environment, and health outcomes were the most widely represented sustainability pillars measured across all publication years.
- Social outcomes were rarely examined (N = 6 of 42, 14.3%).
- A total of 226 dietary patterns were identified from the 42 papers, with the vast majority being statistically estimated or simulated diets (65.9%).
- On average, 5.4 dietary patterns (SD = 6.6, min = 1, max = 36) were identified per publication.
- Nine (21.4%) studies examined one dietary pattern; five (11.9%) studies examined two dietary patterns; four (9.5%) studies examined three dietary patterns; 11 (26.2%) studies examined four dietary patterns; four (9.5%) studies examined five dietary patterns; one (2.4%) study examined six dietary patterns; and eight (19.1%) studies examined seven or more dietary patterns (Supplementary File 2).
- Nearly half of the publications included in the review (19 of 42, 45.2%) compared an observed or approximated diet (i.e., current diet) to one or more statistically estimated or simulated diets (Supplementary File 2).
- Six papers (14.3%) compared an observed or approximated diet to one or more dietary pattern archetypes.
- The most frequently analyzed outcomes focused on the environment domain (n = 132).
- The environmental indicators that have existed the longest in the literature and have repeatedly been mainly climate-related outcomes, such as GHG emissions (n = 44, 33.3%), land use (n = 20, 15.2%), and water use (n = 18, 13.6%).
- A smaller number of studies examine toxicity outcomes (n = 9, 6.8%), energy outcomes (n = 7, 5.3%), eutrophication (n = 7, 5.3%), air pollution (n = 6, 4.5%), nitrogen or phosphorus-related outcomes (n = 6, 4.5%), acidification (n = 4, 3.0%), and biodiversity (n = 2, 1.5%).
- In the publications included for this review, a total of 95 health and disease-related outcomes were captured.
- The most frequent categories were cancer (n = 22, 23.2%); cardiovascular diseases (n = 20, 21.1%); mortality, number of deaths averted, or number of years of life saved (n = 15, 15.8%); T2D (n = 12, 12.6%); and stroke (N = 10, 10.5%).
- A total of 46 economic outcomes were analyzed by included publications: more than half (n = 26, 56.5%) were food price/cost, 15.2% (n = 7) were economy-level outcomes, 13.0% (n = 6) were healthcare costs, 6.5% (n = 3) were productivity costs, and 4.3% (n = 2) were employment-related outcomes (Table 3).
- Of the 26 papers that analyzed food price and cost, 13 compared current diet costs to the costs of an optimized diet, and 1 compared current diet costs to reference diet costs (Supplementary File 3).
- A total of six social outcomes were analyzed across the 42 included papers: two (33.3%) were acceptability outcomes, two (33.3%) were desirability outcomes (relating to taste preferences), one (16.7%) was food availability, and one (16.7%) was social risk.
- The most represented countries in this review are Spain (N = 6), France (N = 6), the United Kingdom (N = 6), and the United States (N = 4).
- Of the 42 publications included in this review, only six are explicitly focused on LMIC contexts (Figure 4).
Other Important Findings
- Most dietary patterns used were statistically estimated or simulated rather than observed.
- 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.
- Most studies that compared diets compared an observed or approximated diet to one or more statistically estimated or simulated diets.
- Current dietary patterns were typically constructed as “average” intakes derived from nationally representative or sub-population surveys.
- The lack of representation of the social pillar can be partially attributed to lack of data and metrics, as well as limited understanding of how best to define social sustainability.
- The suite of indicators used for modeling of climate and environmental dynamics is relatively standardized in relation to accepted concepts of planetary boundaries.
- The inclusion of measures on economic dimensions is relatively recent and tends to be narrowly framed.
- The social dimension of diet-environment-health relationships is also under-represented in the current literature and has been limited to some socioeconomic indicators.
Limitations Noted in the Document
- The review was restricted to including a relatively small number of publications.
- Publications were excluded if they did not pair dietary patterns with at least two pillars of interest.
- Studies were excluded if they were not in the English language.
- Methodological approaches that did not allow for direct comparison among studies were excluded.
- The variation in study descriptors and analysis limited the ability to do a meta-analysis of data or report on the bias/limitations of existing research.
- Most studies reviewed relied on assessments published in earlier literature rather than calculating their own relationships, which can present problems.
- The lack of standardized LCA methodologies and LCA databases to accurately represent marketplace foods has led to variability across studies.
- Studies often rely on assessments of foods assessed as “whole commodities” (such as meat or dairy or cereals or fruits), leading to a lack of understanding of the environmental, health and social impacts of food choices that are most common in urban settings.
Conclusion
The findings of this review suggest a need for greater transparency and clarity in methods in all future publications to better allow for comparison and replication. Researchers need to prioritize building our understanding of the social and economic dimensions of sustainability to address food security and nutrition both equitably and affordably. There is a need for greater inclusion of processed foods in assessment of diets to reflect the reality of consumer choices globally. The literature reviewed here lacks specificity when it comes to their references to social concerns, and even fewer include social metrics in their formal analyses. There is a strong need for greater attention to both metric creation for social outcomes of diets and sub-national analyses that disaggregate results by social groups. The study found that the number of metrics used to assess human and planetary impacts of dietary choices has expanded but remains centered around diet-related non-communicable disease states and GHG emissions and a range of natural resource depletion and/or pollution outcomes. The range of suggested metrics continues to grow but data availability, validation, and effective incorporation into modeling remain challenging. The study suggests that the field requires more creativity and complexity rather than standardization of metrics or approaches, but in allowing for that much greater exposition is needed on assumptions, particularly those relating to system boundaries applied to the analytical approach, data limitations, sensitivity analyses on the effects of methodological choices, and interpretability of results. This review identified no papers that included measures relevant to environmental, health, economic and social concerns simultaneously. Researchers must move toward novel, more complex frameworks that make looking across a variety of measures across all four pillars commonplace.