Impact of taxes and warning labels on red meat purchases among US consumers: A randomized controlled trial

Choice architecture promotes sustainable choices in online food-delivery apps

Greenhouse gas emissions from the food system constitute about one-third of the global total, hence mitigation in this sphere of human activity is a vital goal for research and policy. This study empirically tests the effectiveness of different interventions to reduce the carbon footprint of food choices made on food-delivery apps, using an incentive-compatible online randomized controlled trial with 4,008 participants. The experiment utilized an interactive web platform that mimics popular online food-delivery platforms (such as Just Eat) and included three treatment conditions: a sign-posted meat tax, a carbon-footprint label, and a choice-architecture intervention that changed the order of the menu so that the lowest carbon-impact restaurants and dishes were presented first. Results show that only the choice-architecture nudge significantly reduced the average meal carbon footprint-by 0.3 kg/CO2e per order (12%), driven by a 5.6 percentage point (13%) reduction in high-carbon meal choices. Moreover, we find evidence of significant health and well-being co-benefits. Menu repositioning resulted in the average meal order having greater nutritional value and fewer calories, whilst significantly increasing self-reported satisfaction with the meal choice. Simple back-of-the-envelope calculations suggest that menu repositioning would be a highly cost-effective policy instrument if implemented at scale, with the return on investment expected to be in the range of £1.28 to £3.85 per metric ton of avoided CO2 emissions, depending on implementation costs.

Estimated effects of reductions in processed meat consumption and unprocessed red meat consumption on occurrences of type 2 diabetes, cardiovascular disease, colorectal cancer, and mortality in the USA: a microsimulation study

BACKGROUND

High consumption of processed meat and unprocessed red meat is associated with increased risk of multiple chronic diseases, although there is substantial uncertainty regarding the relationship for unprocessed red meat. We developed a microsimulation model to estimate how reductions in processed meat and unprocessed red meat consumption could affect rates of type 2 diabetes, cardiovascular disease, colorectal cancer, and mortality in the US adult population.

METHODS

We used data from two versions of the US National Health and Nutrition Examination Survey, one conducted during 2015-16 and one conducted during 2017-18, to create a simulated US population. The starting cohort was restricted to respondents aged 18 years or older who were not pregnant and had 2 days of dietary-recall data. First, we used previously developed risk models to estimate the baseline disease risk of an individual. For type 2 diabetes we used a logistic-regression model and for cardiovascular disease and colorectal cancer we used Cox proportional-hazard models. We then multiplied baseline risk by relative risk associated with individual processed meat and unprocessed red meat consumption. Prevented occurrences of type 2 diabetes, cardiovascular disease, colorectal cancer, and mortality were computed by taking the difference between the incidence in the baseline and intervention scenarios. All stages were repeated for ten iterations to correspond to a 10-year time span. Scenarios were reductions of 5%, 10%, 30%, 50%, 75%, and 100% in grams consumed of processed meat, unprocessed red meat, or both. Each scenario was repeated 50 times for uncertainty analysis.

FINDINGS

The total number of individual respondents included in the simulated population was 8665, representing 242 021 876 US adults. 4493 (51·9%) of 8665 individuals were female and 4172 (48·1%) were male; mean age was 49·54 years (SD 18·38). At baseline, weighted mean daily consumption of processed meat was 29·1 g, with a 30% reduction being 8·7 g per day, and of unprocessed red meat was 46·7 g, with a 30% reduction being 14·0 g per day. We estimated that a 30% reduction in processed meat intake alone could lead to 352 900 (95% uncertainty interval 345 500-359 900) fewer occurrences of type 2 diabetes, 92 500 (85 600-99 900) fewer occurrences of cardiovascular disease, 53 300 (51 400-55 000) fewer occurrences of colorectal cancer, and 16 700 (15 300-17 700) fewer all-cause deaths during the 10-year period. A 30% reduction in unprocessed red meat intake alone could lead to 732 600 (725 700-740 400) fewer occurrences of type 2 diabetes, 291 500 (283 900-298 800) fewer occurrences of cardiovascular disease, 32 200 (31 500-32 700) fewer occurrences of colorectal cancer, and 46 100 (45 300-47 200) fewer all-cause deaths during the 10-year period. A 30% reduction in both processed meat and unprocessed red meat intake could lead to 1 073 400 (1 060 100-1 084 700) fewer occurrences of type 2 diabetes, 382 400 (372 100-391 000) fewer occurrences of cardiovascular disease, 84 400 (82 100-86 200) fewer occurrences of colorectal cancer, and 62 200 (60 600-64 400) fewer all-cause deaths during the 10-year period.

INTERPRETATION

Reductions in processed meat consumption could reduce the burden of some chronic diseases in the USA. However, more research is needed to increase certainty in the estimated effects of reducing unprocessed red meat consumption.

FUNDING

The Wellcome Trust.

The impact of an eco-score label on US consumers' perceptions of environmental sustainability and intentions to purchase food: A randomized experiment

Front-of-package labels indicating a product's environmental footprint (i.e., eco-score labels) offer promise to shift consumers towards more sustainable food choices. This study aimed to understand whether eco-score labels impacted consumers' perceptions of environmental sustainability and intentions to purchase sustainable and unsustainable foods. US parents (n = 1,013) completed an online experiment in which they were shown 8 food products (4 sustainable and 4 unsustainable). Participants were randomized to a control (n = 503, barcode on product packaging) or eco-score label group (n = 510, eco-score label on product packaging). The eco-score label was color-coded with a grade of A-F based on the product's environmental footprint, where "A" indicates relative sustainability and "F" indicates relative unsustainability. Participants rated each product's environmental sustainability and their future likelihood of purchase. We used multilevel mixed-effects linear regression models and examined moderation by product category and sociodemographic characteristics. The eco-score label lowered perceived sustainability of unsustainable products by 13% in relative terms or -0.4 in absolute terms (95% CI -0.5, -0.3; p<0.001). The eco-score label increased perceived sustainability of sustainable products by 16% in relative terms or 0.6 in absolute terms (95% CI 0.5, 0.7, p<0.001). Effects on purchase intentions were smaller, with a 6% decrease for unsustainable products (p = 0.001) and an 8% increase for sustainable products (p<0.001). For unsustainable products, the effect of eco-score labels on sustainability perceptions was greater for older adults, men, participants with higher educational attainment, and participants with higher incomes. For sustainable products, the effect of ecolabels on sustainability perceptions was greater for those with higher educational attainment. Eco-score labels have the potential to direct consumers towards more sustainable products. Future studies should investigate eco-score label effectiveness on behavioral outcomes.

Effects of red meat taxes and warning labels on food groups selected in a randomized controlled trial

BACKGROUND

High consumption of red and processed meat contributes to both health and environmental harms. Warning labels and taxes for red meat reduce selection of red meat overall, but little is known about how these potential policies affect purchases of subcategories of red meat (e.g., processed versus unprocessed) or of non-red-meat foods (e.g., cheese, pulses) relevant to health and environmental outcomes. This study examined consumer responses to warning labels and taxes for red meat in a randomized controlled trial.

METHODS

In October 2021, we recruited 3,518 US adults to complete a shopping task in a naturalistic online grocery store. Participants were randomly assigned to one of four arms: control (no warning labels or tax), warning labels only (health and environmental warning labels appeared next to products containing red meat), tax only (prices of products containing red meat were increased 30%) or combined warning labels + tax. Participants selected items to hypothetically purchase, which we categorized into food groups based on the presence of animal- and plant-source ingredients (e.g., beef, eggs, pulses), meat processing level (e.g., processed pork versus unprocessed pork), and meat species (e.g., beef versus pork). We assessed the effects of the warning labels and tax on selections from each food group.

RESULTS

Compared to control, all three interventions led participants to select fewer items with processed meat (driven by reductions in processed pork) and (for the tax and warning labels + tax interventions only) fewer items with unprocessed meat (driven by reductions in unprocessed beef). All three interventions also led participants to select more items containing cheese, while only the combined warning labels + tax intervention led participants to select more items containing processed poultry. Except for an increase in selection of pulses in the tax arm, the interventions did not affect selections of fish or seafood (processed or unprocessed), eggs, or plant-based items (pulses, nuts & seeds, tofu, meat mimics, grains & potatoes, vegetables).

CONCLUSIONS

Policies to reduce red meat consumption are also likely to affect consumption of other types of foods that are relevant to both health and environmental outcomes.

TRIAL REGISTRATION

NCT04716010 on www.

CLINICALTRIALS

gov .

Correction: Impact of taxes and warning labels on red meat purchases among US consumers: A randomized controlled trial

[This corrects the article DOI: 10.1371/journal.pmed.1004284.].