Simple dietary substitutions can reduce carbon footprints and improve dietary quality across diverse segments of the US population

Anna H Grummon; Cristina J Y Lee; Thomas N Robinson; Eric B Rimm; Donald Rose

doi:10.1038/s43016-023-00864-0

. Author manuscript; available in PMC: 2024 Nov 1.

Published in final edited form as: Nat Food. 2023 Oct 26;4(11):966–977. doi: 10.1038/s43016-023-00864-0

Simple dietary substitutions can reduce carbon footprints and improve dietary quality across diverse segments of the US population

Anna H Grummon ^1,², Cristina J Y Lee ¹, Thomas N Robinson ^1,³, Eric B Rimm ^4,⁵, Donald Rose ⁶

PMCID: PMC10725296 NIHMSID: NIHMS1948153 PMID: 37884673

Abstract

Changing what foods we eat could reduce environmental harms and improve human health, but sweeping dietary change is challenging. We used dietary intake data from a nationally representative sample of 7,753 US children and adults to identify simple, actionable dietary substitutions from higher- to lower-carbon foods (e.g., substituting chicken for beef in mixed dishes like burritos, but making no other changes to the diet). We simulated the potential impact of these substitutions on dietary carbon emissions and dietary quality. If all consumers who ate the high-carbon foods instead consumed a lower-carbon substitute, the total dietary carbon footprint in the US would be reduced by more than 35%. Moreover, if adopted, these substitutions would improve consumers’ overall dietary quality by 4%–10%, with benefits projected for all age, gender, and racial/ethnic groups. These results suggest a “small changes” approach could be a valuable starting point for addressing diet’s impact on climate and health.

EDITOR’S SUMMARY

Dietary shifts can be hard to implement for economic, cultural and technical reasons – and their impacts can differ across population groups. Using dietary intake data from more than 7,000 US children and adults, this study identifies relatively simple and actionable dietary substitutions from higher- to lower-carbon foods and estimates the impact of these substitutions on greenhouse gas emissions and the overall healthfulness of Americans’ diets.

The foods we eat have major implications for both personal and planetary health. Food production is a key contributor to climate change,¹ accounting for approximately a third of total human-caused greenhouse gas emissions.² Poor dietary quality is also a leading cause of morbidity and mortality, increasing risk of obesity, type 2 diabetes, heart disease, and some cancers and contributing to an estimated 11 million deaths worldwide every year.³ Experts agree that substantial changes to food systems are needed to meet the goals of the Paris Agreement on climate change and curb rising rates of diet-related diseases.^4–7

Adopting diets high in fruits, vegetables, and legumes and lower in red and processed meats (particularly meats from ruminant animals) is one strategy for individuals to reduce both their personal carbon footprints from food production and their risk of diet-related diseases,^1,4,8–16 making such diets a win-win. Changing eating habits, however, is notoriously difficult,^17,18 in part due to the volume and complexity of the food choices consumers face.^19,20 One promising strategy for overcoming these challenges is to encourage consumers to make simple, achievable substitutions in which they change just one aspect of their diet. This “small change” approach is relevant for both mitigating climate change and improving health because otherwise similar foods – e.g., a beef burrito and a chicken burrito – can have decidedly different average environmental impacts and nutritional profiles. Moreover, compared to more substantial dietary shifts (e.g., becoming vegetarian), simple substitutions may be easier for consumers to implement, reach a wider audience (including those uninterested in radical changes), and boost self-efficacy, a key ingredient to sustained behavior change.^21,22 Research is needed, however, to identify simple dietary substitutions that could yield both environmental and health benefits. One prior study found that if US adults made small changes to their consumption of specific foods (e.g., replacing cuts of beef with poultry), they would meaningfully reduce their dietary carbon footprints and improve their dietary quality.²³ What remains unknown is whether simple dietary substitutions systematically identified across a range of food groups could yield environmental and health benefits.

More importantly, it is unclear how the environmental and health impacts of adopting these dietary substitutions might vary depending on the demographic characteristics of the people who adopt them. Nonprofits and government agencies already communicate about dietary changes that could have environmental and health benefits; identifying the population subgroups for whom small dietary substitutions would yield disproportionate benefits could help prioritize these communication efforts. Moreover, given persistent disparities in diet-related diseases,^24–26 it is crucial to understand how adopting these substitutions would affect dietary quality across population subgroups.

We addressed these gaps by simulating the environmental and health impacts of simple dietary substitutions in a nationally representative, population-based sample. Our approach linked detailed dietary intake data reported by children and adults participating in the 2015–2016 National Health and Nutrition Examination Survey (NHANES) to comprehensive information on the greenhouse gas emissions associated with production of specific foods (i.e., foods’ “carbon footprints”).^15,27,28 Most prior studies on dietary changes with health and environmental co-benefits have focused on protein sources;^11,29,30 we identified dietary substitutions across a four broad food groups (protein foods, mixed dishes, milk and dairy products, and non-alcoholic beverages) that together account for 85% of the total dietary carbon footprint in the US (Table 1). Within each of these food groups, we identified high-carbon-impact foods (i.e., foods that contribute most to the population’s total dietary carbon footprint and that have a high “carbon intensity” [high emissions per gram of food produced]). We then identified foods that would offer culinary or menu item equivalence to the high-carbon-impact food, but have lower carbon intensity (i.e., “lower-carbon-impact substitute foods”). (Although we did not analyze nutritional equivalence in selecting substitute foods, we kept our substitutions to like foods within the same broad food group). Importantly, we emphasized identifying simple substitutions that would be easy to communicate and adopt – for example, replacing a beef burger with a turkey burger but making no other changes to the diet. Next, we simulated the environmental and health impacts that would be expected if all NHANES respondents who ate the high-carbon-impact food instead ate the lower-carbon-impact substitute food. We specifically examined how the substitutions affected dietary carbon footprints, assessed in CO₂-equivalents, and overall dietary quality, assessed using the Healthy Eating Index (HEI), a composite score that reflects alignment with the Dietary Guidelines for Americans³¹ and is predictive of disease outcomes.^32–34 We also examined impacts on intakes of key nutrients of concern (e.g., iron, calcium).^35–37 We simulated effects across the population overall and within demographic subgroups defined by age, gender, and race/ethnicity. Together, our analyses illuminate which types of simple dietary substitutions would be most meaningful for reducing carbon footprints and improving dietary quality and identify which population groups would benefit the most from adopting these substitutions.

Table 1. Overall dietary carbon footprint of major food groups, NHANES 2015–2016.

Table shows the total carbon emissions to produce these foods for the total amount consumed by the US population (i.e., scaled to represent the US population) expressed in kg of CO₂-eq (“overall carbon footprint”) and the food group’s contribution to the overall carbon footprint. Numbers do not sum to 100% due to rounding. The example foods (column 2) are not necessarily listed in order of their carbon intensity or their contribution to total dietary carbon footprint.

Food Group	Examples of foods included	Carbon footprint, kg CO₂ equivalents	% contribution to total US carbon footprint
Mixed Dishes	Meat, poultry, and seafood mixed dishes Bean, pea, legume, and vegetable mixed dishes Rice, pasta, and other grain-based dishes Asian mixed dishes (e.g. stir-fry, lo mein, sushi) Mexican mixed dishes (e.g., burritos, tacos, nachos) Pizza Sandwiches, burgers, and frankfurters Soups	515,242,797	39%
Protein Foods	Beef, pork, lamb, goat, and liver and organ meats Chicken, turkey, and duck and other poultry Fish and shellfish Eggs and omelets Cold cuts and cured meats, bacon, frankfurters, and sausages Beans, peas, legumes, nuts, and seeds Processed soy products	412,542,464	31%
Milk and Dairy	Unflavored and flavored milks Milk shakes and other dairy drinks Milk substitutes Cheese Yogurt	144,866,068	11%
Non-alcoholic beverages	100% citrus, apple, other fruit, and vegetable juice Diet drinks and sweetened drinks (e.g., soft drinks, fruit drinks) Coffee and tea	57,840,081	4%
Snacks and Sweets	Chips, popcorn, pretzels, snack mixes, and crackers Cereal bars and nutrition bars Cakes, pies, cookies, and brownies Doughnuts, sweet rolls, and pastries Candy Ice cream, pudding, gelatins, ices, and sorbets	55,922,780	4%
Vegetables	Tomatoes, carrots, broccoli, spinach, lettuce, string beans, cabbage, onions, and corn White potatoes, French fries, and mashed potatoes	34,736,619	3%
Grains	Cooked rice and pasta Breads, rolls, bagels, English muffins, and tortillas Biscuits, muffins, pancakes, waffles, and French toast Cereal, oatmeal, and grits	23,229,387	2%
Fruit	Apples, bananas, grapes, peaches and nectarines, strawberries, blueberries, citrus fruits, dried fruits, pears, pineapple, mango, and papaya	20,771,456	2%
Alcoholic Beverages	Beer, wine, liquor, and cocktails	19,174,896	1%
Fats and Oils	Butter and animal fats, margarine, cream, cream cheese, sour cream, whipped cream, mayonnaise, and salad dressings	14,927,864	1%
Condiments and Sauces	Tomato-based condiments, soy-based condiments, mustard, olives, pickles, pasta sauces, dips, and gravies	7,904,291	1%
Sugars	Sugars, honey, sugar substitutes, jams, syrups, and toppings	4,256,857	0%

Open in a new tab

RESULTS

Sample Characteristics

The sample included n=7,753 respondents aged 2 and older participating in NHANES 2015–2016. The weighted sample of respondents was approximately half women (51.5%) (Table 2). About 21% were younger than 18, while 22% were aged 60 or older. Consumption of a high-carbon-impact food for which we simulated a substitution was associated with age and race/ethnicity, but not gender, educational attainment, or family income. Supplementary Table 1 details characteristics of those who consumed the high-carbon-impact foods in each of the four food groups we examined (proteins, mixed dishes, milk and dairy products, and beverages).

Table 2. Sample characteristics, overall and among respondents simulated to have any dietary substitution.

The proportions shown here account for NHANES sample weights. The “Other race/ethnicity” category included respondents who reported their race as non-Hispanic American Indian or Alaska Native, Native Hawaiian or Pacific Islander, or “Other.” Income to poverty ratio is the total family income divided by the poverty threshold. P-values are from Pearson Chi-Squared tests (corrected for survey design), comparing respondents simulated to make dietary substitutions to those without simulated substitutions.

Sample characteristic	Overall sample n=7,753 %	Had substitution n=5,676 %	p-value
Gender
Female	51.5	50.3	0.074
Male	48.5	49.7	0.074
Age in years
2–5	4.4	5.6	<0.001
6–11	7.3	8.7
12–17	9.2	9.3
18–25	9.7	8.8
26–39	18.9	17.1
40–49	13.9	13.7
50–59	14.5	14.2
60+	22.1	22.6
Race/ethnicity
Non-Hispanic White	62.9	64.0	<0.001
Non-Hispanic Black	11.7	10.1
Mexican American and other Hispanic	16.8	17.5
Non-Hispanic Asian	4.7	4.5
Other race/ethnicity	4.0	3.8
Highest education attainment in household
Less than high school	11.9	11.9	0.801
High school graduate or GED	16.4	16.0
Some college	32.9	33.0
College degree or higher	38.8	39.0
Income to poverty ratio
Missing income	6.4	6.2	0.746
<1.0	14.1	13.8
1.0 to <2.0	19.7	19.8
2.0 to <5.0	37.1	36.9
5.0 or higher	22.8	23.2

Open in a new tab

Dietary Substitutions

We identified dietary substitutions in each of the four food groups, ranging from 79 substitutions in the beverage group to 180 substitutions in the mixed dish group. Supplementary Tables 2–5 detail all substitutions. Each substitution involved replacing a single high-carbon-impact food with a similar, lower-carbon-impact food. For proteins, most substitutions involved replacing cuts of beef with poultry. For mixed dishes, substitutions involved replacing beef or pork entrees with poultry or vegetarian versions of those entrees. Milk and dairy substitutions involved replacing animal-milk-based products with soy- or almond-milk-based products, and most beverage substitutions involved replacing juice with whole fruit.

Potential Impact of Simple Substitutions on Carbon Footprints

For each of the four food groups we examined, making the target dietary substitutions would substantially reduce respondents’ personal dietary carbon footprints (i.e., the carbon emissions associated with the production of foods the respondent consumed on a given day), as shown in Table 3. For example, those simulated to make protein substitutions would reduce their personal dietary carbon footprint by an average of 3.79 kg CO₂-equivalents per person per day (95% CI: −4.22, −3.37), a 50.2% reduction compared to their before-substitution diets. Similarly, respondents simulated to make a mixed dish substitution would reduce their personal carbon footprint 3.50 kg CO₂-equivalents per person per day, a 52.6% reduction. Average reductions in personal dietary carbon footprints were more modest for respondents simulated to adopt milk and dairy (8.2% reduction) and beverage (8.1% reduction) substitutions.

Table 3. Impact of protein, mixed dish, milk and dairy, and beverage substitutions on dietary carbon footprints.

Table shows the potential impact of replacing high-carbon-impact foods with lower-carbon-impact substitutes on dietary carbon footprint for those with substitutions (i.e., any individual who consumed one of the high-carbon-impact foods on the 24-hour recall day) and scaled to represent the entire US population (including those without substitutes). When the respondent consumed the high-carbon-impact food more than once, all instances were replaced with the substitute. When there were multiple feasible substitutes for a given high-carbon-impact food, primary analyses (shown here) assumed respondents replaced the high-carbon-impact food with the lowest-carbon intensity substitute. P-values are from paired two-tailed t-tests for per-person impacts and two-tailed t-tests for total impacts.

	Impact on personal dietary carbon footprint per person per day, among those with substitutions						Impact on total dietary carbon footprint, scaled to entire US population
	n (%)	Before substitution (kg CO₂-eq) Mean ± SE	After substitution (kg CO₂-eq) Mean ± SE	Absolute difference (kg CO₂-eq) Mean ± SE	Proportional difference	p	Before substitution (metric tons CO₂-eq) Total ± SE	After substitution (metric tons CO₂-eq) Total ± SE	Absolute difference (metric tons CO₂-eq) Total ± SE	Proportional difference	p
Protein	1,281 (17%)	7.56 ± 0.27	3.77 ± 0.13	−3.79 ± 0.20	−50.2%	<0.001	1,318,391 ± 80,022	1,118,877 ± 68,597	−199,514 ± 16,751	−15.1%	<0.001
Mixed dishes	1,653 (21%)	6.65 ± 0.16	3.15 ± 0.12	−3.50 ± 0.06	−52.6%	<0.001	1,318,391 ± 80,022	1,091,420 ± 67,678	−226,971 ± 19,403	−17.2%	<0.001
Milk and dairy	3,534 (46%)	4.25 ± 0.09	3.90 ± 0.08	−0.35 ± 0.01	−8.2%	<0.001	1,318,391 ± 80,022	1,270,571 ± 75,691	−47,820 ± 4,461	−3.6%	<0.001
Beverages	1,614 (21%)	4.52 ± 0.14	4.15 ± 0.15	−0.37 ± 0.01	−8.1%	<0.001	1,318,391 ± 80,022	1,298,835 ± 79,265	−9,556 ± 1,355	−1.5%	<0.001
All groups	5,676 (73%)	4.99 ± 0.09	2.74 ± 0.04	−2.25 ± 0.07	−45.0%	<0.001	1,318,391 ± 80,022	824,529 ± 49,940	−493,862 ± 32,942	−37.5%	<0.001

Open in a new tab

Although only 17–46% of respondents (depending on food group) reported consuming the high-carbon-impact foods for which we simulated a substitution, adopting the substitutions could have meaningful effects at the population level. Protein substitutions, for example, were projected to reduce the total dietary carbon footprint by 199,514 metric tons CO₂-equivalents per day when reductions were scaled to the size of the US population (95% CI: −235,218 to −163,811), a 15.1% reduction over before-substitution levels (Table 3). Similarly, mixed dish substitutions would reduce overall dietary carbon footprints by 226,971 metric tons CO₂-equivalents per day, a 17.2% reduction. Potential carbon footprint reductions from milk and dairy (−3.6%) and beverage (−1.5%) substitutions were more modest, while adopting substitutions in all four food groups simultaneously would reduce the total US population carbon footprint by 37.5%. Reductions in personal and total dietary carbon footprints were somewhat smaller in secondary analyses that used higher-carbon-intensity substitutes instead of the lowest-carbon-intensity substitute (Supplementary Table 6).

Compared to children, adults tended to have larger average personal carbon footprint reductions for all food groups except milk and dairy (though not all differences were statistically significant), while male respondents had larger personal carbon footprint reductions than females for all food groups (Figure 1, Supplementary Tables 7–8). Relative reductions for each racial/ethnic group varied across the four food categories (Supplementary Table 9). Respondents who identified as Asian, for example, had somewhat smaller carbon footprint reductions (−3.25 kg CO₂-eq per person per day) from protein substitutions than those who identified as non-Hispanic Black (−4.30) or Mexican American/Other Hispanic (−4.20), and non-Hispanic Black respondents had somewhat smaller personal carbon footprint reductions from mixed dish substitutions (−3.11 kg CO₂-eq per person per day) than both Mexican American/Other Hispanic respondents (−3.76) and Asian respondents (−3.78), though these differences were not significant after correction for multiple comparisons. Reductions in personal carbon footprints from milk and dairy substitutions were similar across racial/ethnic groups, while reductions from beverage substitutions were smallest for respondents in the “Other” race/ethnicity category.

Figure 1. — Figure shows the potential impact of replacing high-carbon-impact foods with lower-carbon-impact substitutes on personal dietary carbon footprint for those with substitutions (i.e., any individual who consumed one of the high-impact foods on the 24-hour recall day). Results are presented in kg of carbon dioxide equivalents per person per day. Bars show changes in dietary carbon footprint for 7,753 US adults and children, including the following sample sizes for each demographic group: age 2–5: n=654; age 6–11: n=1025; age 12–17: n=933; age 18–25: n=682; age 26–39: n=1178; age 40–49: n=818; age 50–59: n=802; age 60 and over: n=1661; gender: male: n=3793; female: n=3960; race/ethnicity: non-Hispanic white: n=2509; non-Hispanic Black: n=1676; Mexican American/Other Hispanic: n=2480; non-Hispanic Asian: n=718; Other: n=370). Data are presented as means ± 1 SE. When the respondent consumed the high-carbon-impact food more than once, all instances were replaced with the substitute. When there were multiple feasible substitutes for a given high-carbon-impact food, primary analyses (shown here) assumed respondents replaced the high-carbon-impact food with the lowest-carbon-intensity substitute. The “Other” race/ethnicity category included respondents who reported their race as non-Hispanic American Indian or Alaska Native, Native Hawaiian or Pacific Islander, or “Other.”

When scaling carbon footprint reductions to the population level, adults making the simulated dietary substitutions would contribute more to total reductions than children making these substitutions, while males would contribute more than females (Supplementary Figure 1). Across race/ethnic groups, non-Hispanic white populations making the simulated substitutions were projected to contribute the most to total carbon footprint reductions, followed by Mexican American/Other Hispanic populations and non-Hispanic Black populations. For example, non-Hispanic white populations making the simulated protein substitutions would contribute reductions of 119,193 metric tons CO₂-eq per day, while Mexican American/Other Hispanic and non-Hispanic Black populations would contribute 35,771 and 22,517 metric tons CO₂-eq per day, respectively.

Potential Impact of Simple Substitutions on Dietary Outcomes

The simulated dietary substitutions also improved overall dietary quality. Those simulated to make a protein substitution had an average total HEI score of 51.08 after substitution, 2.09 points (95% CI: 1.85, 2.33) higher than their before-substitution diets, a 4.3% improvement (Table 4). Benefits from mixed dish substitutions were even larger: those simulated to make these changes had a 4.78-point (10.3%) improvement in their average HEI score after substitution. Milk and dairy (2.76 points [5.2%]) and beverage (2.21 points [4.1%]) substitutions also improved HEI scores. Generally, improvements in total HEI scores were driven by improvements in seafood and plant proteins, fatty acids, and saturated fat component scores (Supplementary Table 10), benefits that were partially (but not fully) offset by reductions in component scores for sodium, refined grains, and dairy from substitutions in some food groups. Substitutions in all food groups led to improvements in fiber, folate, potassium, and added vitamin B12, but (except for beverage substitutions) led to reductions in zinc intake. The substitutions had varying effects on calcium, iron, and natural vitamin B12, and vitamin D intake, depending on food group. Improvements in total HEI scores persisted but were somewhat smaller in secondary analyses that used the higher-carbon-intensity substitutes instead of the lowest-carbon-intensity substitutes (Supplementary Table 11).

Table 4. Impact of protein, mixed dish, milk and dairy, and beverage substitutions on dietary quality among those with substitutions.

Table shows the potential impact of replacing high-carbon-impact foods with lower-carbon-impact substitutes on dietary quality for those with substitutions (i.e., any individual who consumed one of the high-impact foods on the 24-hour recall day). When the respondent consumed the high-carbon-impact food more than once, all instances were replaced with the substitute. When there were multiple feasible substitutes for a given high-carbon-impact food, primary analyses (shown here) assumed respondents replaced the high-carbon-impact food with the lowest-carbon-intensity substitute. P-values are from paired two-tailed t-tests.

	Healthy Eating Index Scores (0–100)
	n (%)	Before substitution, Mean ± SE	After substitution, Mean ± SE	Absolute difference, Mean ± SE	Proportional difference	p
Protein	1,281 (17%)	48.99 ± 0.53	51.08 ± 0.51	2.09 ± 0.11	4.3%	<0.001
Mixed dishes	1,653 (21%)	46.44 ± 0.57	51.22 ± 0.58	4.78 ± 0.30	10.3%	<0.001
Milk and dairy	3,534 (46%)	53.47 ± 0.54	56.24 ± 0.51	2.76 ± 0.10	5.2%	<0.001
Beverages	1,614 (21%)	54.16 ± 0.73	56.37 ± 0.60	2.21 ± 0.19	4.1%	<0.001
All groups	5,676 (73%)	51.02 ± 0.51	55.18 ± 0.45	4.16 ± 0.14	8.2%	<0.001

Open in a new tab

The simulated substitutions led to improvements in dietary quality for all demographic groups (Figure 2, Supplementary Tables 7–9), with generally larger improvements for younger age groups (except for mixed dishes) and for males compared to females, though most differences were not statistically significant after correction for multiple comparisons. Protein substitutions led to somewhat larger improvements in dietary quality for individuals who identified as non-Hispanic white or Black compared to the “Other” racial/ethnic groups, while mixed dish substitutions were projected to be most beneficial for respondents identifying as Mexican American or another Hispanic ethnicity. Milk and dairy substitutions led to somewhat larger improvements in dietary quality for non-Hispanic Black, Mexican American/Other Hispanic, Asian, and Other racial/ethnic groups compared to non-Hispanic white respondents, while beverage substitutions would yield the largest improvements for respondents identifying as non-Hispanic Black.

DISCUSSION

In this simulation study using data from a nationally representative population-based sample of children and adults, we found that simple, actionable dietary substitutions across a range of food groups could simultaneously reduce dietary carbon footprints and improve overall dietary quality. Although substantial changes to the way we eat are likely required to meet climate mitigation goals and curb noncommunicable diseases, these results suggest a “small changes” approach to improving diet could be a valuable starting point for encouraging eating habits that lessen the environmental impacts of food production and reduce the prevalence of diet-related diseases.

Simulated substitutions in each of the four food groups we examined reduced carbon footprints. The largest benefits were found for substitutions to protein foods and mixed dishes: if all consumers who ate the high-carbon-impact foods in these food groups instead consumed the lower-carbon substitute, their changes would result in a more than 30% reduction in the total dietary carbon footprint in the US. These substitutions typically involved replacing beef with poultry or plant-based proteins; our results therefore align with prior studies finding environmental and health benefits from poultry and plant-based proteins.^{1,4,8–14,23} The total dietary carbon footprint reductions from milk and dairy and beverage substitutions were more modest, though individuals making these changes would still see meaningful reductions in their personal dietary carbon footprints. Most prior studies on dietary changes with health and environmental co-benefits have focused on protein sources^11,29,30 or examined substitutions that may be challenging to communicate to the public (e.g., replacing a single food with a mix of other foods).³⁸ This study simulated simple, easy-to-communicate dietary changes chosen from several broad food groups.²³ Our results suggest that consumers have a variety of options for reducing their personal dietary carbon footprints, though proteins and mixed dishes remain the most important food groups for addressing total dietary carbon footprints. Mixed dish substitutions may be an especially promising target for behavior change interventions given that these substitutions typically involved replacing only one ingredient in a meal (e.g., replacing beef with chicken in a taco).

When scaling carbon footprint reductions to the US population, adults making the simulated substitutions were projected to contribute more to total carbon footprint reductions than children, males more than females, and non-Hispanic white populations more than other racial/ethnic groups. These differences reflect both the total population size simulated to make the substitution (e.g., there are more adults than children in the US) as well as baseline levels of intake of the high-carbon-impact foods (e.g., males tend to consume larger quantities of food than females). These results suggest that tailoring and targeting dietary change interventions for certain groups (e.g., adult males) could yield the highest return on investment in terms of overall carbon footprint reductions. However, this assumes that all groups examined are equally responsive to interventions, while prior work has suggested that some population groups may be more open than others to changing their diets for environmental reasons.^11,39,40

Consumers choosing to adopt the proposed substitutions would see substantial reductions in their personal dietary carbon footprint, ranging from 8% reductions from milk and dairy or beverage substitutions to >50% reductions from protein or mixed dish substitutions. Prior research suggests the promise of messaging interventions emphasizing that small dietary changes can have meaningful environmental impacts.^41,42 These messaging interventions could leverage our results, for example, by including quantitative information about how much individuals could reduce their personal carbon footprint by adopting these dietary substitutions.

The simulated dietary changes also improved overall dietary quality. For example, those making simulated substitutions had total HEI scores 2–5 points (4–10%) higher than their before-substitution diets. Improvements were seen across each of the population groups we examined, suggesting that promoting more healthful and environmentally sustainable foods could improve nutritional outcomes across demographic groups. These HEI improvements are substantial in light of the relatively minor behavior changes simulated (i.e., replacing specific foods with similar substitutes), and important given that HEI scores are predictive of incident cardiovascular disease⁴³ as well as cardiovascular, cancer, and all-cause mortality.^32–34 Analyses of the Nurses’ Health Study and Health Professionals Follow-up Study cohorts, for example, found that improving HEI scores from the first quintile (median scores ranging from 52–55) to just the second quintile (median scores ranging from 60–63) was associated with 7% lower risk of incident cardiovascular disease.⁴³ Moreover, the improvements in HEI scores from the simulated substitutions are similar in magnitude to the total improvements in average HEI scores observed in the US from 2005–2016.⁴⁴

The simulated dietary substitutions were projected to improve average dietary quality for each of the population subgroups we examined. The magnitude of the projected improvements to dietary quality, however, varied by demographic group and food group. For example, mixed dish substitutions were projected to be most beneficial for respondents identifying as Mexican American or another Hispanic ethnicity, while milk and dairy substitutions would improve dietary quality the most for younger children. Interventions focused on specific population groups could leverage these findings to identify dietary substitutions that may produce the greatest health benefits for the population of interest.

The improvements in total HEI scores were accompanied with decreases in some component scores (e.g., sodium scores worsened as a result of substitutions in the protein, mixed dish, and milk and dairy food groups), suggesting nutritional tradeoffs from these substitutions. Similarly, substitutions in all food groups would increase intake of several important nutrients (e.g., fiber, folate, potassium, and added vitamin B12), but reduce intake of others (e.g., zinc, except for beverage substitutions). The substitutions we considered might also generate other tradeoffs beyond dietary intakes; for example, replacing beef with poultry might cause greater animal suffering in some ethical frameworks.⁴⁵ These tradeoffs suggests that interventions promoting these dietary substitutions should take care to avoid or mitigate any unintended effects.

Although not all consumers are likely to make the target dietary substitutions, the “small change” dietary substitutions we identified may be appealing to individuals who want to change their diet for health or environmental reasons. Recent research suggests a growing proportion of Americans might be amenable to making these dietary substitutions. For example, 41% of US adults say they are already eating less meat for environmental reasons,⁴⁶ 46% say they are willing to use dairy alternatives,⁴⁰ and more than half say they are willing to eat more plant-based meat alternatives and less red meat.⁴⁰ For these consumers, messages delivered at the point-of-selection (e.g., when selecting a restaurant meal or checking out at the grocery store) could persuade them to adopt the targeted dietary substitutions.^41,47–51 Another approach that would rely less on individual consumer interest in these substitutions would be to change the physical environment, for example by making the lower-carbon-impact substitute the default option in restaurants or cafeterias,^52,53 increasing the availability of lower-carbon-impact foods and beverages relative to the high-carbon-impact options,^54–58 and implementing changes to institutional procurement policies.^59,60

Strengths of this study included that we used nationally representative data on dietary intake, examined a wide range of potentially achievable dietary changes, and used high-quality data on the environmental impacts of food production.²⁷ We note several limitations. First, our study simulated the effects of dietary substitutions, rather than measuring actual changes in dietary intake after consumers made the substitutions. Our approach required making assumptions, including assuming that consumers would make no other changes to their diets besides the substitution and would maintain constant calorie intake. Second, we were unable to adjust for the bioavailability of nutrients, which depends on current nutritional status,^61,62 so our estimates of changes in nutrient intakes should be interpreted with caution. Third, we examined only carbon footprints; future studies should examine outcomes such as water use, land use, biodiversity, soil health, or eutrophication. Fourth, we did not examine costs, though prior research suggests that lower-carbon diets tend to be less expensive than higher-carbon diets.^11,12 Fifth, we used estimates of the average carbon footprint of foods based on a comprehensive review of life cycle analyses,²⁷ and these averages may not apply in all contexts. For example, one study found that shifting some livestock production toward ruminants could reduce greenhouse gas emissions in some settings,⁶³ in contrast to most other studies (including the present study) finding that shifting away from ruminant consumption would lower emissions.^11,23,64,65 Sixth, we examined food groups that contributed the most to the dietary carbon footprint in the US, but this does not imply that food groups excluded from analysis (e.g. sweets and snacks, sugar-sweetened beverages) are more healthful. Seventh, we only considered foods consumed by NHANES respondents as potential substitutes, so did not consider substitutions toward newer products such as lab-grown meat, non-dairy cheeses, or oat milk, as these were not consumed by any 2015–2016 NHANES respondents.

To conclude, using a nationally representative sample, this simulation study found that simple food substitutions would improve dietary quality and meaningfully reduce dietary carbon footprints for individuals and the population. Future research should identify scalable strategies for encouraging these substitutions.

METHODS

Participants

Participants were respondents in the National Health and Nutrition Examination Surveys (NHANES), a continuous, multistage probability survey that is nationally representative of the civilian, non-institutionalized US population. We used data from the 2015–2016 NHANES cycle to match the most recently available data on the greenhouse gas emissions associated with food production (described below). The analytic sample included all respondents aged 2 years and older who completed a valid 24-hour dietary recall (n=7,753).

Data

Dietary intake data were obtained from NHANES. Trained interviewers collected 24-hour dietary recalls from respondents using the Automated Multiple Pass Method.⁶⁶ A proxy adult assisted children aged 6–11 years and completed the recall for children aged ≤5 years. We analyzed the first day of recalls only, consistent with recommendations for estimating mean usual intakes at the population level.⁶⁷ The 24-hour dietary recall has been shown to accurately assess mean total calorie and nutrient intake,^66,68,69 including for children.^70,71

Dietary recall data included each food and beverage consumed by the participant on the recall day along with the item’s weight in grams and nutritional content (e.g., calories, nutrients). We linked items consumed to their equivalent weight in food groups (e.g., fruits, dairy) using the Food Patterns Equivalent Database (FPED) to enable calculation of Healthy Eating Index scores (described below).

We linked foods consumed in NHANES to information about the environmental impacts of producing those foods. We focused on greenhouse gas emissions because they are the most thoroughly documented environmental impact of food production.⁴ Greenhouse gas emissions data came from the Database of Food Recall Impacts on the Environment for Nutrition and Dietary Studies (dataFRIENDS). DataFRIENDS provides information on the greenhouse gas emissions associated with the production of specific commodities that make up foods reported by NHANES respondents in their 24-diet recalls.^15,28 The commodity database underlying this, referred to as dataFIELD, was based on an exhaustive review of more than 300 life cycle assessments, as detailed elsewhere.^13,23,27 Emissions were quantified from cradle to farm gate (i.e., the production stage) and expressed as carbon dioxide equivalents (CO₂-eq) per 100 grams of food. We refer to these impacts as a food’s “carbon footprint.” We linked items consumed by NHANES respondents to greenhouse gas emissions using the items’ 8-digit food codes.

Procedures

First, we identified high-carbon-impact foods that could be potential targets for dietary substitutions. Our goal was to maximize the reduction in dietary carbon footprints that could be achieved through specific, actionable dietary changes that could be easily communicated and adopted. We sought to identify substitutions across a variety of food groups, given that some consumers may be unable or unwilling to make substitutions in some groups. We started by identifying the broad food groups that contribute the most to dietary carbon footprints. We estimated total dietary carbon footprints for each of 12 food groups defined by the US Department of Agriculture’s What We Eat in America food categorization system,⁷² excluding water (which has a negligible carbon footprint) and baby foods and other foods (for which carbon footprint data were not available). As shown in Table 1, four food groups – mixed dishes, protein foods, milk and dairy foods, and non-alcoholic beverages – account for approximately 85% of total dietary carbon emissions in the US. These groups also include both foods with both higher and lower carbon intensity (i.e., CO₂-eq emitted per 100 grams of food produced, see also below) that could act as substitutes for one another. We focused on identifying substitutions within these four broad food groups.

In each food group, we identified high-carbon-impact foods to target with substitutions. We ranked the foods consumed by NHANES respondents based on their relative environmental impact, similar to a prior study.²³ We ranked foods in two ways: by the total carbon emissions to produce the total amount of food consumed scaled to the US population level and expressed in metric tons of CO₂-eq (“overall carbon footprint,” estimated using survey weights) and by the relative intensity of emissions expressed in CO₂-eq per 100g of food (“carbon intensity,” as recorded in the dataFRIENDS database). Within each food group, we identified the top 100 emitting foods on each dimension. We first examined the top emitting foods separately by age, gender, and race/ethnicity; however, similar foods appeared on each list, so subsequent analyses did not stratify by demographic group. We then identified the foods that were highly ranked on both overall carbon footprint and carbon intensity, yielding 4 lists of 26–60 high-carbon-impact foods (Supplementary Tables 12–15) for which we sought potential substitutes. NHANES uses detailed definitions of foods consumed to provide accurate nutrition information for each food. For example, ‘frankfurter or hot dog sandwich, beef, plain, on white bun’ is considered a distinct food from ‘frankfurter or hot dog sandwich, beef, plain, on wheat bun.’ These foods have nearly identical carbon intensity, but production of the former contributes more to the total dietary carbon footprint (and was therefore ranked higher as a contributor to the overall carbon footprint) because it was consumed in greater amounts by NHANES respondents. Because potential interventions to encourage dietary substitutions would likely not distinguish between such similar foods, we included as targets for substitution lower-ranking foods that were very similar to high-ranking foods.

Next, we sought to identify lower-carbon-impact foods that would offer culinary or menu equivalence to the high-carbon-impact foods. For example, we identified ‘meat loaf made with chicken or turkey, with tomato-based sauce’ as a substitute for ‘meat loaf made with beef, with tomato-based sauce.’ To ensure the substitutes were feasible, we considered only foods consumed by NHANES 2015–2016 respondents as potential substitutes. Some high-carbon-impact foods lacked feasible substitutes (e.g., many cheeses); we did not implement substitutions for these foods, though they are still listed as high-impact foods in Supplementary Tables 12–15. Other high-impact foods had multiple feasible substitutes (e.g., a beef taco could be replaced with either a chicken taco or a bean taco); for these foods, we examined both the lowest-carbon-intensity and highest-carbon-intensity potential substitute, as described below. Substitutions are detailed in Supplementary Tables 2–5.

Outcome Measures

We quantified the potential impact of the dietary substitutions on two co-primary outcomes: dietary carbon footprints and overall dietary quality. First, dietary carbon footprints were calculated by summing the carbon emissions associated with producing all foods and beverages consumed by the respondent on the recall day. We focused on carbon emissions because they are the most thoroughly documented environmental impact of food production.⁴ Second, we examined overall dietary quality, assessed using Healthy Eating Index-2015 (HEI) scores.³¹ We selected HEI as our measure of overall dietary quality because healthier diets as identified by HEI are associated with lower incidence of cardiovascular disease⁴³ and lower all-cause, cancer, and cardiovascular disease mortality.^32–34 Briefly, HEI is a density-based 100-point index that reflects alignment with the Dietary Guidelines for Americans. It is based on nine “adequacy” components recommended for inclusion in a healthy diet (e.g., total vegetables) and four “moderation” components to be consumed sparingly (e.g., saturated fats).^31,32 We calculated HEI scores using the mean ratio method.⁷³ Finally, as a secondary outcome, we also assessed how the substitutions would affect consumption of 9 nutrients that have been identified as nutrients of public health concern by the Dietary Guidelines for Americans or in prior research: calcium, dietary fiber, folate (examined as dietary folate equivalents), potassium, iron, vitamin B12 (both naturally occurring and added), vitamin D, and zinc.^35–37

Statistical Analysis

To quantify the potential impact of consumers making the proposed dietary substitutions, we assumed that any respondent who consumed the high-carbon-impact food would instead consume the lower-carbon-impact food. When there were multiple feasible substitutes for a given high-carbon-impact food, we modeled two scenarios, replacing the high-carbon-impact food with: 1) its least-carbon-intensive substitute (primary analyses) and 2) its most-carbon-intensive substitute (secondary analyses). (High-carbon-impact beverages had only one substitute, so secondary analyses for beverages are identical to the primary analyses). These two scenarios provide a range of possible effects of the substitutions.

We simulated isocaloric substitutions that replaced the high-carbon-impact food with the lower-carbon-impact food on a calorie-for-calorie basis. This scenario would occur if consumers sought to maintain constant energy intake before and after substitutions, consciously or otherwise. To implement the substitutions, we identified and replaced all instances of the high-carbon-impact food codes with the lower-carbon-impact substitute food code based on the substitution lists (Supplementary Tables 2–5). We made these substitutions at the individual level, for every respondent who reported consuming any of the high-carbon-impact foods on the recall day. If a respondent consumed the high-carbon-impact food more than once, we replaced all instances with the lower-carbon-impact substitute foods. Similarly, if a respondent consumed more than one high-carbon-impact food, we replaced all high-carbon-impact foods with their respective lower-carbon-impact substitutes. We scaled the carbon footprint and nutritional outcomes for the substitute food by multiplying the calories of the high-carbon-impact food consumed on the recall day by the carbon footprint and nutritional content of the lower-carbon-impact substitute food per calorie.

For substitutions of juices and smoothies in the beverage category, we assumed that respondents would replace these beverages with a similar whole fruit on a calorie-for-calorie basis (e.g., replacing orange juice with a whole orange). Messaging interventions aiming to encourage adoption of these substitutions might encourage consumers to replace juice with fruit as well as water (e.g., to ensure adequate hydration); our analyses only modeled substitution to whole fruit (not whole fruit plus water) because water production is associated with a negligible carbon footprint (approximately 0.0000467 kg CO₂ per 100g of water) and because water does not include enough minerals to meaningfully affect HEI scores.

We calculated environmental and nutritional outcomes for each respondent before and after the hypothetical substitutions were simulated. We simulated the substitutions separately for each food group (e.g., implementing only protein substitutions, but no changes in other food groups) as well as simultaneously across all four food groups. We then examined differences in outcomes from before to after substitution. We examined the environmental impacts in two ways: 1) average changes in dietary carbon footprint for individuals with substitutions (“personal dietary carbon footprint”), which represent the average benefit an individual could expect from making the proposed substitutions, and 2) total changes in dietary carbon footprint scaled to the population level (“total dietary carbon footprint,” i.e., including those without substitutions), which represent the benefit the US as a whole could expect from the substitutions. Changes in total dietary carbon footprint were calculated using the total estimator with survey weights. For dietary impacts, we examined changes in dietary quality and nutrient intakes for individuals making the substitutions. We tested the significance of differences from before to after substitutions using t-tests. We examined the effects of the substitutions across the population overall and among demographic groups defined by age, gender, and race/ethnicity. We also used t-tests to examine whether the effects of the substitutions on personal dietary carbon footprint and dietary quality differed demographic groups; we used the Bonferroni method to correct p-values from these tests for multiple comparisons, considering 28 pairwise comparisons among age groups and 10 pairwise comparisons among race/ethnic groups.

Analyses were conducted in 2022–2023 in Stata MP version 17 and used a two-tailed critical alpha of 0.05. All analyses accounted for NHANES’s complex survey design. The Harvard Longwood Campus Institutional Review Board determined this study to be exempt (IRB #21–1624) and the Stanford University Institutional Review Board approved this study (IRB #67669).

Supplementary Material

Online Supplementary Materials

Supplementary Table 1. Sample characteristics, overall and among those simulated to have a protein, mixed dish, milk or dairy, or non-alcoholic beverage substitution

Supplementary Table 2. Descriptions of high-carbon-impact protein foods and their lower-carbon-impact substitutes

Supplementary Table 3. Descriptions of high-carbon-impact mixed dishes and their lower-carbon-impact substitutes

Supplementary Table 4. Descriptions of high-carbon-impact milk and dairy items and their substitutes

Supplementary Table 5. Descriptions of high-carbon-impact non-alcoholic beverages and their lower-carbon-impact substitutes

Supplementary Table 6. Impact of protein, mixed dish, milk and dairy, and beverage substitutions on dietary carbon footprint, secondary analyses examining replacing high-carbon-impact foods with the highest-carbon-intensity lower-carbon-impact substitutes

Supplementary Table 7. Changes in personal dietary carbon footprint and overall dietary quality after protein, mixed dish, milk and dairy, and beverage substitutions, by age group

Supplementary Table 8. Changes in personal dietary carbon footprint and overall dietary quality after protein, mixed dish, milk and dairy, and beverage substitutions, by gender

Supplementary Table 9. Changes in personal dietary carbon footprint and overall dietary quality after protein, mixed dish, milk and dairy, and beverage substitutions, by race/ethnicity

Supplementary Table 10. Impact of protein, mixed dish, milk and dairy, and beverage substitutions on Healthy Eating Index component scores and nutrient intakes among those with substitutions, primary analyses

Supplementary Table 11. Impact of protein, mixed dish, milk and dairy, and beverage substitutions on dietary quality among those with substitutions, secondary analyses examining replacing high-carbon-impact foods with the highest-carbon-intensity lower-carbon-impact substitutes

Supplementary Table 12. 47 protein foods that were highest ranked on overall dietary GHGE impact and on consumed intensity of GHGE

Supplementary Table 13. 26 mixed dishes that were highest ranked on overall dietary GHGE impact and on consumed intensity of GHGE

Supplementary Table 14. 60 milk and dairy foods that were highest ranked on overall dietary GHGE impact and on consumed intensity of GHGE

Supplementary Table 15. 55 non-alcoholic beverages that were highest ranked on overall dietary GHGE impact and on consumed intensity of GHGE

Supplementary Figure 1. Contribution of population subgroups to changes in total dietary carbon footprint (i.e., scaled to the entire US population) after protein, mixed dish, milk and dairy, and beverage substitutions

NIHMS1948153-supplement-Online_Supplementary_Materials.docx^{(211KB, docx)}

ACKNOWLEDGEMENTS

This research was supported by a grant from the National Institutes of Health, National Heart, Lung, and Blood Institute (K01 HL158608, AHG).

Footnotes

CODE AVAILABILITY

All code is available at https://github.com/clee321/NATFOOD-dietarysubs.

COMPETING INTERESTS

T.N.R. receives research grant support from the National Institutes of Health and the Stanford Maternal and Child Health Institute for nutrition- and health-related research and served as a member of the Scientific Advisory Board of WW International, Inc. until December 2022. D.R. currently receives grant support from DHHS-Health Resources and Services Administration and from USDA-National Institute of Food and Agriculture. He has received grants in the past for his diet-sustainability research from Wellcome Trust, the National Cancer Institute, and the Center for Biological Diversity. The remaining authors declare no competing interest.

DATA AVAILABILITY

This study used data from the National Health and Nutrition Examination Surveys (NHANES) and from the Database of Food Recall Impacts on the Environment for Nutrition and Dietary Studies (dataFRIENDS). NHANES data are publicly available from the Centers for Disease Control website (https://wwwn.cdc.gov/nchs/nhanes/). The dataFIELD database is available at http://css.umich.edu/page/datafield and the dataFRIENDS database is available at https://sph.tulane.edu/sbps/diet-environmental-impacts.

REFERENCES

1.Poore J, Nemecek T. Reducing food’s environmental impacts through producers and consumers. Science. 2018;360(6392):987–992. [DOI] [PubMed] [Google Scholar]
2.Crippa M, Solazzo E, Guizzardi D, Monforti-Ferrario F, Tubiello FN, Leip A. Food systems are responsible for a third of global anthropogenic GHG emissions. Nat Food. 2021;2(3):198–209. doi: 10.1038/s43016-021-00225-9 [DOI] [PubMed] [Google Scholar]
3.Afshin A, Sur PJ, Fay KA, et al. Health effects of dietary risks in 195 countries, 1990–2017: A systematic analysis for the Global Burden of Disease Study 2017. Lancet. 2019;393(10184):1958–1972. doi: 10.1016/S0140-6736(19)30041-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Willett W, Rockström J, Loken B, et al. Food in the Anthropocene: the EAT–Lancet Commission on healthy diets from sustainable food systems. The Lancet. 2019;393(10170):447–492. [DOI] [PubMed] [Google Scholar]
5.Rose D, Heller MC, Roberto CA. Position of the Society for Nutrition Education and Behavior: The Importance of Including Environmental Sustainability in Dietary Guidance. J Nutr Educ Behav. 2019;51(1):3–15.e1. doi: 10.1016/j.jneb.2018.07.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Harmon AH, Gerald BL, American Dietetic Association. Position of the American Dietetic Association: food and nutrition professionals can implement practices to conserve natural resources and support ecological sustainability. J Am Diet Assoc. 2007;107(6):1033–1043. doi: 10.1016/j.jada.2007.04.018 [DOI] [PubMed] [Google Scholar]
7.United States Department of Agriculture and Department of Health and Human Services. Scientific Report of the 2015 Dietary Guidelines Advisory Committee: Advisory Report to the Secretary of Health and Human Services and the Secretary of Agriculture.; 2015. https://health.gov/our-work/nutrition-physical-activity/dietary-guidelines/previous-dietary-guidelines/2015/advisory-report
8.Clark MA, Springmann M, Hill J, Tilman D. Multiple health and environmental impacts of foods. Proc Natl Acad Sci USA. 2019;116(46):23357. doi: 10.1073/pnas.1906908116 [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Nelson ME, Hamm MW, Hu FB, Abrams SA, Griffin TS. Alignment of Healthy Dietary Patterns and Environmental Sustainability: A Systematic Review. Advances in Nutrition. 2016;7(6):1005–1025. doi: 10.3945/an.116.012567 [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Van Horn L, Carson JAS, Appel LJ, et al. Recommended dietary pattern to achieve adherence to the American Heart Association/American College of Cardiology (AHA/ACC) guidelines: a scientific statement from the American Heart Association. Circulation. 2016;134(22):e505–e529. [DOI] [PubMed] [Google Scholar]
11.Willits-Smith A, Aranda R, Heller MC, Rose D. Addressing the carbon footprint, healthfulness, and costs of self-selected diets in the USA: a population-based cross-sectional study. The Lancet Planetary Health. 2020;4(3):e98–e106. doi: 10.1016/S2542-5196(20)30055-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Conrad Z, Drewnowski A, Belury MA, Love DC. Greenhouse gas emissions, cost, and diet quality of specific diet patterns in the United States. The American Journal of Clinical Nutrition. Published online 2023. [DOI] [PubMed] [Google Scholar]
13.Rose D, Heller MC, Willits-Smith AM, Meyer RJ. Carbon footprint of self-selected US diets: nutritional, demographic, and behavioral correlates. The American journal of clinical nutrition. 2019;109(3):526–534. [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Pollock BD, Willits-Smith AM, Heller MC, Bazzano LA, Rose D. Do diets with higher carbon footprints increase the risk of mortality? A population-based simulation study using self-selected diets from the USA. Public Health Nutrition. 2022;25(8):2322–2328. doi: 10.1017/S1368980022000830 [DOI] [PMC free article] [PubMed] [Google Scholar]
15.O’Malley K, Willits-Smith A, Rose D. Popular diets as selected by adults in the United States show wide variation in carbon footprints and diet quality. Am J Clin Nutr. Published online February 20, 2023:S0002–9165(23)00511–7. doi: 10.1016/j.ajcnut.2023.01.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Scarborough P, Clark M, Cobiac L, et al. Vegans, vegetarians, fish-eaters and meat-eaters in the UK show discrepant environmental impacts. Nat Food. 2023;4(7):565–574. doi: 10.1038/s43016-023-00795-w [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Hill JO, Wyatt HR, Reed GW, Peters JC. Obesity and the environment: Where do we go from here? Science. 2003;299(5608):853. doi: 10.1126/science.1079857 [DOI] [PubMed] [Google Scholar]
18.Roberto CA, Swinburn B, Hawkes C, et al. Patchy progress on obesity prevention: Emerging examples, entrenched barriers, and new thinking. The Lancet. 2015;385(9985):2400–2409. doi: 10.1016/S0140-6736(14)61744-X [DOI] [PubMed] [Google Scholar]
19.Leng G, Adan RAH, Belot M, et al. The determinants of food choice. Proceedings of the Nutrition Society. 2017;76(3):316–327. doi: 10.1017/S002966511600286X [DOI] [PubMed] [Google Scholar]
20.Sobal J, Bisogni C, Devine C, Jastran M. A conceptual model for the food choice process over the life course. In: Shepherd R, Raats M, eds. The Psychology of Food Choice. CABI Publishing; 2006. 10.1079/9780851990323.0001 [DOI] [Google Scholar]
21.Bandura A Health promotion from the perspective of social cognitive theory. Psychol Health. 1998;13(4):623–649. doi: 10.1080/08870449808407422 [DOI] [Google Scholar]
22.Hill JO. Can a small-changes approach help address the obesity epidemic? A report of the Joint Task Force of the American Society for Nutrition, Institute of Food Technologists, and International Food Information Council. The American Journal of Clinical Nutrition. 2009;89(2):477–484. doi: 10.3945/ajcn.2008.26566 [DOI] [PubMed] [Google Scholar]
23.Rose D, Willits-Smith AM, Heller MC. Single-item substitutions can substantially reduce the carbon and water scarcity footprints of US diets. The American Journal of Clinical Nutrition. 2022;115(2):378–387. doi: 10.1093/ajcn/nqab338 [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Centers for Disease Control and Prevention. Racial and Ethnic Disparities in Heart Disease. Centers for Disease Control and Prevention; 2019. https://www.cdc.gov/nchs/hus.htm [Google Scholar]
25.Flegal K, Kruszon-Moran D, Carroll M, Fryar C, Ogden C. Trends in obesity among adults in the United States, 2005 to 2014. JAMA. 2016;315(21):2284–2291. doi: 10.1001/jama.2016.6458 [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Centers for Disease Control. By the Numbers: Diabetes in America. Centers for Disease Control and Prevention. Published October 25, 2022. Accessed February 16, 2023. https://www.cdc.gov/diabetes/health-equity/diabetes-by-the-numbers.html [Google Scholar]
27.Heller MC, Willits-Smith A, Meyer R, Keoleian GA, Rose D. Greenhouse gas emissions and energy use associated with production of individual self-selected US diets. Environmental Research Letters. 2018;13(4):044004. [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Willits-Smith A Food Choices in the United States: Opportunities for Health and Sustainability Co-Benefits [Doctoral Dissertation, Tulane University]. ProQuest Dissertations & Theses Global; 2022. [Google Scholar]
29.Farchi S, Sario MD, Lapucci E, Davoli M, Michelozzi P. Meat consumption reduction in Italian regions: Health co-benefits and decreases in GHG emissions. PLOS ONE. 2017;12(8):e0182960. doi: 10.1371/journal.pone.0182960 [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Jarmul S, Dangour AD, Green R, Liew Z, Haines A, Scheelbeek PF. Climate change mitigation through dietary change: a systematic review of empirical and modelling studies on the environmental footprints and health effects of ‘sustainable diets.’ Environ Res Lett. 2020;15(12):123014. doi: 10.1088/1748-9326/abc2f7 [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Krebs-Smith SM, Pannucci TE, Subar AF, et al. Update of the Healthy Eating Index: HEI-2015. Journal of the Academy of Nutrition and Dietetics. 2018;118(9):1591–1602. doi: 10.1016/j.jand.2018.05.021 [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Reedy J, Lerman JL, Krebs-Smith SM, et al. Evaluation of the Healthy Eating Index-2015. Journal of the Academy of Nutrition and Dietetics. 2018;118(9):1622–1633. doi: 10.1016/j.jand.2018.05.019 [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Panizza CE, Shvetsov YB, Harmon BE, et al. Testing the Predictive Validity of the Healthy Eating Index-2015 in the Multiethnic Cohort: Is the Score Associated with a Reduced Risk of All-Cause and Cause-Specific Mortality? Nutrients. 2018;10(4):452. doi: 10.3390/nu10040452 [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Hu EA, Steffen LM, Coresh J, Appel LJ, Rebholz CM. Adherence to the Healthy Eating Index–2015 and Other Dietary Patterns May Reduce Risk of Cardiovascular Disease, Cardiovascular Mortality, and All-Cause Mortality. The Journal of Nutrition. 2019;150(2):312–321. doi: 10.1093/jn/nxz218 [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Beal T, Ortenzi F. Priority Micronutrient Density in Foods. Frontiers in Nutrition. 2022;9. Accessed May 20, 2023. https://www.frontiersin.org/articles/10.3389/fnut.2022.806566 [DOI] [PMC free article] [PubMed] [Google Scholar]
36.United States Department of Agriculture and Department of Health and Human Services. Dietary Guidelines for Americans, 2020–2025. Published online 2020. Accessed November 10, 2021. www.dietaryguidelines.gov
37.Sun H, Weaver CM. Decreased Iron Intake Parallels Rising Iron Deficiency Anemia and Related Mortality Rates in the US Population. The Journal of Nutrition. 2021;151(7):1947–1955. doi: 10.1093/jn/nxab064 [DOI] [PubMed] [Google Scholar]
38.Stylianou KS, Fulgoni VL, Jolliet O. Small targeted dietary changes can yield substantial gains for human health and the environment. Nat Food. 2021;2(8):616–627. doi: 10.1038/s43016-021-00343-4 [DOI] [PubMed] [Google Scholar]
39.Lacroix K, Goldberg M, Marlon J, et al. Understanding Differences in Americans’ Motivations for Eating Plant-Rich Foods. Yale Program on Climate Change Communication; 2022. Accessed April 4, 2023. https://climatecommunication.yale.edu/publications/understanding-differences-in-americans-motivations-for-eating-plant-rich-foods/ [Google Scholar]
40.Leiserowitz A, Ballew M, Rosenthal S, Semaan J. Climate Change and the American Diet. Yale University and EarthDay Network; 2020. [Google Scholar]
41.Blondin S, Attwood S, Vennard D, Mayneris V. Environmental Messages Promote Plant-Based Food Choices: An Online Restaurant Menu Study. World Resources Institute; 2022. Accessed January 16, 2023. https://www.wri.org/research/environmental-messages-promote-plant-based-food-choices-online-restaurant-menu-study [Google Scholar]
42.Roberto CA, Kawachi I. Introduction to behavioral economics and public health. In: Roberto CA, Kawachi I, eds. Behavioral Economics and Public Health. Oxford University Press; 2015. [Google Scholar]
43.Shan Z, Li Y, Baden MY, et al. Association Between Healthy Eating Patterns and Risk of Cardiovascular Disease. JAMA Internal Medicine. 2020;180(8):1090–1100. doi: 10.1001/jamainternmed.2020.2176 [DOI] [PMC free article] [PubMed] [Google Scholar]
44.US Department of Agriculture and US Department of Health and Human Services. Figures & Infographics | Dietary Guidelines for Americans. Published 2020. Accessed February 17, 2023. https://www.dietaryguidelines.gov/figures-infographics
45.Mathur MB. Ethical drawbacks of sustainable meat choices. Science. 2022;375(6587):1362–1362. doi: 10.1126/science.abo2535 [DOI] [PubMed] [Google Scholar]
46.Desilver D Americans Say They’re Changing Behaviors to Help the Environment – but Is It Making a Difference? Pew Research Center; 2019. Accessed May 22, 2023. https://www.pewresearch.org/short-reads/2019/12/19/americans-say-theyre-changing-behaviors-to-help-the-environment-but-is-it-making-a-difference/ [Google Scholar]
47.Forwood SE, Ahern AL, Marteau TM, Jebb SA. Offering within-category food swaps to reduce energy density of food purchases: a study using an experimental online supermarket. International Journal of Behavioral Nutrition and Physical Activity. 2015;12(1):1. [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Koutoukidis DA, Jebb SA, Ordóñez-Mena JM, et al. Prominent positioning and food swaps are effective interventions to reduce the saturated fat content of the shopping basket in an experimental online supermarket: a randomized controlled trial. International Journal of Behavioral Nutrition and Physical Activity. 2019;16(1):50. doi: 10.1186/s12966-019-0810-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Jansen L, van Kleef E, Van Loo EJ. The use of food swaps to encourage healthier online food choices: a randomized controlled trial. International Journal of Behavioral Nutrition and Physical Activity. 2021;18(1):156. doi: 10.1186/s12966-021-01222-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
50.Sparkman G, Walton GM. Dynamic norms promote sustainable behavior, even if it is counternormative. Psych Science. 2017;28(11):1663–1674. [DOI] [PubMed] [Google Scholar]
51.Wolfson JA, Musicus AA, Leung CW, Gearhardt AN, Falbe J. Effect of Climate Change Impact Menu Labels on Fast Food Ordering Choices Among US Adults: A Randomized Clinical Trial. JAMA Network Open. 2022;5(12):e2248320. doi: 10.1001/jamanetworkopen.2022.48320 [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Meier J, Andor MA, Doebbe FC, Haddaway NR, Reisch LA. Review: Do green defaults reduce meat consumption? Food Policy. 2022;110:102298. doi: 10.1016/j.foodpol.2022.102298 [DOI] [Google Scholar]
53.Reisch LA, Sunstein CR, Andor MA, Doebbe FC, Meier J, Haddaway NR. Mitigating climate change via food consumption and food waste: A systematic map of behavioral interventions. Journal of Cleaner Production. 2021;279:123717. doi: 10.1016/j.jclepro.2020.123717 [DOI] [Google Scholar]
54.Pechey R, Hollands GJ, Reynolds JP, Jebb SA, Marteau TM. Is altering the availability of healthier vs. less-healthy options effective across socioeconomic groups? A mega-analysis. International Journal of Behavioral Nutrition and Physical Activity. 2022;19(1):88. doi: 10.1186/s12966-022-01315-y [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Pechey R, Cartwright E, Pilling M, et al. Impact of increasing the proportion of healthier foods available on energy purchased in worksite cafeterias: A stepped wedge randomized controlled pilot trial. Appetite. 2019;133:286–296. doi: 10.1016/j.appet.2018.11.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Bianchi F, Garnett E, Dorsel C, Aveyard P, Jebb SA. Restructuring physical micro-environments to reduce the demand for meat: a systematic review and qualitative comparative analysis. The Lancet Planetary Health. 2018;2(9):e384–e397. doi: 10.1016/S2542-5196(18)30188-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Garnett EE, Balmford A, Sandbrook C, Pilling MA, Marteau TM. Impact of increasing vegetarian availability on meal selection and sales in cafeterias. Proceedings of the National Academy of Sciences. 2019;116(42):20923–20929. doi: 10.1073/pnas.1907207116 [DOI] [PMC free article] [PubMed] [Google Scholar]
58.Marteau TM, Hollands GJ, Pechey R, Reynolds JP, Jebb SA. Changing the assortment of available food and drink for leaner, greener diets. BMJ. 2022;377:e069848. doi: 10.1136/bmj-2021-069848 [DOI] [PubMed] [Google Scholar]
59.Reuters. New York City’s public school system goes meat-free on Fridays. Reuters. https://www.reuters.com/world/us/new-york-citys-public-school-system-goes-meat-free-fridays-2022-02-04/. Published February 4, 2022. Accessed May 20, 2023. [Google Scholar]
60.Sievert K, Lawrence M, Parker C, Baker P. Understanding the political challenge of red and processed meat reduction for healthy and sustainable food systems: a narrative review of the literature. Int J Health Policy Manag. 2020;10:793–808. [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Hurrell R, Egli I. Iron bioavailability and dietary reference values1234. The American Journal of Clinical Nutrition. 2010;91(5):1461S–1467S. doi: 10.3945/ajcn.2010.28674F [DOI] [PubMed] [Google Scholar]
62.Allen LH, Carriquiry AL, Murphy SP. Perspective: proposed harmonized nutrient reference values for populations. Advances in Nutrition. 2020;11(3):469–483. [DOI] [PMC free article] [PubMed] [Google Scholar]
63.Cheng L, Zhang X, Reis S, Ren C, Xu J, Gu B. A 12% switch from monogastric to ruminant livestock production can reduce emissions and boost crop production for 525 million people. Nat Food. 2022;3(12):1040–1051. doi: 10.1038/s43016-022-00661-1 [DOI] [PubMed] [Google Scholar]
64.Aleksandrowicz L, Green R, Joy EJ, Smith P, Haines A. The impacts of dietary change on greenhouse gas emissions, land use, water use, and health: a systematic review. PloS one. 2016;11(11):e0165797. [DOI] [PMC free article] [PubMed] [Google Scholar]
65.Godfray HCJ, Aveyard P, Garnett T, et al. Meat consumption, health, and the environment. Science. 2018;361(6399):eaam5324. doi: 10.1126/science.aam5324 [DOI] [PubMed] [Google Scholar]
66.Moshfegh AJ, Rhodes DG, Baer DJ, Murayi T, Clemens JC, Rumpler WV. The US Department of Agriculture Automated Multiple-Mass Method reduces bias in the collection of energy intakes. Am J Clin Nutr. 2008;88(2):324–332. [DOI] [PubMed] [Google Scholar]
67.National Cancer Institute. Summary Tables: Recommendations on Potential Approaches to Dietary Assessment for Different Research Objectives Requiring Group-level Estimates. National Cancer Institute Dietary Assessment Primer. Accessed December 3, 2019. https://dietassessmentprimer.cancer.gov/approach/table.html [Google Scholar]
68.Blanton CA, Moshfegh AJ, Baer DJ, Kretsch MJ. The USDA Automated Multiple-Pass Method Accurately Estimates Group Total Energy and Nutrient Intake. The Journal of Nutrition. 2006;136(10):2594–2599. doi: 10.1093/jn/136.10.2594 [DOI] [PubMed] [Google Scholar]
69.Rhodes DG, Murayi T, Clemens JC, Baer DJ, Sebastian RS, Moshfegh AJ. The USDA Automated Multiple-Pass Method accurately assesses population sodium intakes. The American Journal of Clinical Nutrition. 2013;97(5):958–964. doi: 10.3945/ajcn.112.044982 [DOI] [PubMed] [Google Scholar]
70.Zhang FF, Roberts SB, Must A, et al. Assessing Dietary Intake in Childhood Cancer Survivors: Food Frequency Questionnaire Versus 24-Hour Diet Recalls. Journal of Pediatric Gastroenterology and Nutrition. 2015;61(4):499–502. doi: 10.1097/MPG.0000000000000826 [DOI] [PMC free article] [PubMed] [Google Scholar]
71.Börnhorst C, Bel-Serrat S, Pigeot I, et al. Validity of 24-h recalls in (pre-)school aged children: Comparison of proxy-reported energy intakes with measured energy expenditure. Clinical Nutrition. 2014;33(1):79–84. doi: 10.1016/j.clnu.2013.03.018 [DOI] [PubMed] [Google Scholar]
72.Agricultural Research Service, United States Department of Agriculture. What are WWEIA Food Categories? Food Surveys Research Group: Beltsville, MD. Published August 3, 2022. Accessed August 15, 2022. https://www.ars.usda.gov/northeast-area/beltsville-md-bhnrc/beltsville-human-nutrition-research-center/food-surveys-research-group/docs/dmr-food-categories/ [Google Scholar]
73.National Cancer Institute. Basic Steps in Calculating HEI Scores. Published July 24, 2020. Accessed August 3, 2020. https://epi.grants.cancer.gov/hei/calculating-hei-scores.html

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials