Impacts of a Healthy Checkout Ordinance on Added Sugars, Sodium, and Other Nutrients at Checkout 1 Year After Implementation: Berkeley, California, 2021–2022

Abstract

Objectives. To assess the impact of the world’s first mandatory healthy checkout policy on added sugars, sodium, and other nutrients available at retail checkouts in Berkeley, California.

Methods. We collected data on food and beverage facings before (2021) and after (2022) policy implementation in 23 Berkeley (8786 facings) and 75 California comparison stores (40 617 facings). Our analysis used difference-in-differences models.

Results. Between 2021 and 2022, there was a 70% reduction in added sugars per serving of food and beverage facings (b = 0.30; 95% confidence interval [CI] = 0.15, 0.59) and a 53% reduction in sodium per serving of beverage facings (b = 0.47; 95% CI = 0.28, 0.81) at Berkeley checkouts relative to comparison city checkouts. There was a 41% reduction in saturated fat and a 47% increase in fiber per serving of food and beverage facings (b = 0.59; 95% CI = 0.44, 0.79 and b = 1.47; 95% CI = 1.05, 2.06, respectively).

Conclusions. One year after implementation, the policy reduced added sugars and saturated fat per serving of food and beverage facings, reduced sodium per serving of beverage facings, and increased fiber per serving of food and beverage facings.

Public Health Implications. Healthy checkout policies have potential to increase the healthfulness of checkout food environments. (Am J Public Health. 2025;115(10):1700–1712. https://doi.org/10.2105/AJPH.2025.308159)

Retail food environments influence food purchasing behaviors and, thus, diet quality of populations, partially through the placement of products in prominent store locations.^1–3 Checkout areas hold particular significance. All customers must walk through a checkout to complete their purchases, and roughly 1 in 3 customers purchase foods or beverages displayed at checkout.⁴ Because of marketing contracts between retailers and manufacturers, checkout areas encourage sales of items like candy, sugar-sweetened beverages (SSBs), and salty snacks⁵ by capitalizing on consumers’ last-minute, impulse purchasing decisions.^6–8 These items are often high in nutrients of concern like added sugars and sodium, which can increase the risk of type 2 diabetes, high blood pressure, heart disease, and stroke when consumed in excess.^9–11 Checkout areas may also contribute to nutrition and health disparities,^4,12 as stores in neighborhoods with higher poverty and higher Black and Hispanic residential compositions are more likely to stock unhealthy foods and beverages at checkout.^5,13,14

To ensure that large food retailers offer healthy options rather than unhealthy foods at checkout, Berkeley, California, implemented a healthy checkout ordinance (HCO) in 2021,¹⁵ the world’s first mandatory healthy checkout policy, to our knowledge.^16–19 This ordinance (i.e., city-level law) requires that stores display only the following at checkout: nonfood and nonbeverage products, unsweetened beverages, sugar-free gum and mints, fruits, vegetables, nuts, seeds, legumes, yogurt, cheese, and whole grains with no more than 5 grams of added sugars and no more than 200 milligrams of sodium per serving.¹⁵ The Berkeley HCO was championed by a grass-roots community coalition.¹⁹ Before implementation, only 29% of food and beverage products at checkout met the HCO requirements.⁵ One year after implementation, candy, SSBs, and other sweets were replaced with a combination of nonfood products and compliant food and beverage products.²⁰ However, no study has examined the impact of the HCO on the nutrient content of food and beverage products remaining at checkout.

To evaluate the HCO’s impact on the nutrients it limits, we assessed changes in added sugars and sodium in the food and beverage products available in store checkouts in Berkeley before and 1 year after implementation compared with other Northern California cities. We also examined the HCO’s impacts on nutrients related to the HCO food group restrictions, including saturated fat and fiber, as well as calories.

METHODS

The Berkeley HCO implementation in March 2021¹⁵ created a natural experiment in which we conducted a cohort study of products available at retail checkouts in Berkeley and comparison cities before (February 2021; ≤ 1 month before implementation) and 1 year after (February 2022) implementation.¹⁵ This study included a census of Berkeley stores initially identified as subject to the HCO (n = 23) and comparison stores, matched by store type and chain, from 3 California cities matched to Berkeley by sociodemographic characteristics: Davis, Oakland, and Sacramento (n = 75). The sampling strategy is described elsewhere.²⁰ Table A (available as a supplement to the online version of this article at https://ajph.org) shows the store sample.

All product facings⁵ in up to 3 randomly selected checkouts (i.e., designated waiting area including endcaps and any space within 3 feet of the register) were assessed in each store. A product facing refers to each product visible to customers, not including products stacked behind it. Each facing was counted separately, including side-by-side facings for identical products. Using the highly reliable Store CheckOUt Tool (SCOUT),²¹ trained data collectors photographed and assessed characteristics (e.g., brand, flavor, and size) of each facing. Using methods reported previously,⁵ nutrient and ingredient data from manufacturer and retailer websites were linked for each food and beverage facing. Of the resulting 50 061 food and beverage facings, 658 (1%) did not have nutrient information for at least 1 outcome (e.g., imported or discontinued products with no manufacturer website), leaving an analytic sample of 49 403 (8786 in Berkeley and 40 617 in comparison cities).

Of the 49 403 food and beverage facings evaluated, 8258 (17%) did not have retrievable nutrient data for the exact product or package size. These facings were linked with nutrient information using similar products with a different flavor (e.g., strawberry instead of watermelon) or size (> 1 unit [e.g., ounce, milliliter] difference in size). We conducted a sensitivity analysis without these 8258 facings, restricting to the 41 145 facings with nutrient data from the same product or package size.

Outcomes and Measures

Based on the HCO nutrient limits, the 2 primary outcomes are average grams of added sugars and milligrams of sodium per serving of food and beverage facings. Based on HCO food category limits, 3 secondary outcomes include average grams of saturated fat and fiber, and calories per serving of food and beverage facings. To assess changes in the whole product shoppers take home, we also report outcomes per package of food and beverage facings (amount per serving multiplied by servings per package).

To understand variation in HCO implementation, we report the primary outcomes stratified by store type (chain supermarket, independent supermarket, chain mass merchandiser, chain specialty food, chain dollar, independent grocery, and chain drug), defined by the North American Industry Classification System, store observations, and name recognition.

Statistical Analysis

We report summary statistics for all outcomes before and after policy implementation in Berkeley and comparison cities. We then report changes in each outcome from before to after implementation using difference-in-differences generalized linear models with cluster-robust standard errors clustered at the store level, log link, and γ distribution to account for the right skew of nutrient distributions. Specifically, we estimated the following model:

$${\text{Nutrient}}_{\text{ist}}=\text{exp}({\text{β}}_0+{\text{β}}_1{\text{Post}}_{\text{t}}+{\text{β}}_2{\text{Berkeley}}_{\text{s}}+{\text{β}}_3\left({\text{Post}}_{\text{t}}\times {\text{Berkeley}}_{\text{s}}\right)+{\text{ε}}_{\text{ist}})$$ (1)

where Nutrient_ist is the nutrient outcome for product facing i, in site s, at time t. Two indicator variables, Post and Berkeley, indicate if the product facing was at checkout before or after implementation and in Berkeley or a comparison city, respectively. With this modeling approach, exponentiated coefficients (i.e., [b−1]×100 for b < 1.0 and [1‒b]×100 for b ≥ 1.0) are interpreted as the percent change in Berkeley relative to comparison cities. We ran models for all food and beverage facings (including gum and mints) and for food facings and beverage facings separately. We also report percent change in nutrient outcomes attributable to changes in product type (e.g., SSBs, fruits) offered at checkout. We then stratified the analyses for primary outcomes by store type.

We conducted analyses in August to October 2024 with Stata/SE version 18.0 (StataCorp LLC, College Station, TX). We followed Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines.

RESULTS

The number of food and beverage facings in the analytic sample decreased by 42% (5572–3214 facings) after implementation in Berkeley, whereas there was little change (4%) in the number of food and beverage facings in comparison cities (19 914–20 703 facings; Table 1).

TABLE 1—

Nutrient Composition of Food and Beverage Products per Serving and per Package Facing at Retail Checkouts Before and After Implementation of the Berkeley Healthy Checkout Ordinance: Berkeley, California, and Comparison Cities, 2021‒2022

	Berkeley, CA		Comparison Cities		b (95% CI)	% Change,a Berkeley vs Comparison
	Preimplementation	Postimplementation	Preimplementation	Postimplementation
Added sugars (grams)
Per serving food and beverage					0.30 (0.15, 0.59)	−70
Mean (SD)	11.1 (15.6)	3.4 (10.7)	11.4 (15.8)	11.7 (16.3)
No.	5 498	3 135	19 594	20 114
Per serving beverage					0.28 (0.11, 0.73)	−72
Mean (SD)	23.0 (26.0)	6.7 (17.9)	22.8 (26.7)	24.1 (26.6)
No.	1 110	864	3 741	4 134
Per serving food					0.27 (0.16, 0.45)	−73
Mean (SD)	8.0 (9.4)	2.1 (5.4)	8.6 (10.2)	8.5 (10.1)
No.	4 388	2 271	15 853	15 980
Per package food and beverage					0.29 (0.16, 0.53)	−71
Mean (SD)	29.6 (63.5)	7.6 (29.1)	28.0 (67.4)	24.7 (50.3)
No.	5 394	3 009	19 099	19 743
Per package beverage					0.27 (0.10, 0.72)	−73
Mean (SD)	24.9 (28.7)	7.3 (18.5)	24.7 (28.0)	26.9 (33.0)
No.	1 105	803	3 580	3 986
Per package food					0.30 (0.17, 0.51)	−70
Mean (SD)	30.8 (69.7)	7.7 (32.1)	28.8 (73.5)	24.1 (53.8)
No.	4 289	2 206	15 519	15 757
Sodium (milligrams)
Per serving food and beverage					0.81 (0.57, 1.14)	NS
Mean (SD)	78.6 (119.6)	66.1 (109.0)	85.1 (170.6)	88.9 (163.7)
No.	5 543	3 142	19 608	20 506
Per serving beverage					0.47 (0.28, 0.81)	−53
Mean (SD)	97.3 (99.8)	45.0 (76.1)	90.6 (99.6)	88.9 (99.5)
No.	1 115	812	3 667	4 081
Per serving food					0.94 (0.64, 1.37)	NS
Mean (SD)	73.9 (123.7)	73.4 (117.5)	83.9 (183.0)	88.9 (176.0)
No.	4 428	2 330	15 941	16 425
Per package food and beverage					0.60 (0.35, 1.04)	NS
Mean (SD)	271.0 (1202.1)	194.5 (447.6)	243.2 (892.9)	290.7 (1391.8)
No.	5 435	3 062	19 178	20 142
Per package beverage					0.95 (0.41, 2.21)	NS
Mean (SD)	110.1 (109.5)	108.7 (277.3)	103.6 (136.2)	107.3 (237.4)
No.	1 110	801	3 576	3 994
Per package food					0.59 (0.33, 1.05)	NS
Mean (SD)	312.3 (1343.3)	224.8 (490.6)	275.2 (985.0)	336.0 (1546.6)
No.	4 325	2 261	15 602	16 148
Saturated fat (grams)
Per serving food and beverage					0.59 (0.44, 0.79)	−41
Mean (SD)	2.0 (2.7)	1.2 (2.0)	2.0 (2.8)	2.0 (2.8)
No.	5 536	3 148	19 589	20 486
Per serving beverage					1.85 (0.28, 12.00)	NS
Mean (SD)	0.2 (0.7)	0.1 (0.4)	0.2 (0.7)	< 0.1 (0.3)
No.	1 115	819	3 684	4 087
Per serving food					0.62 (0.46, 0.83)	−38
Mean (SD)	2.4 (2.8)	1.5 (2.2)	2.4 (2.9)	2.4 (2.9)
No.	4 421	2 329	15 905	16 399
Per package food and beverage					0.65 (0.43, 0.96)	−35
Mean (SD)	6.8 (16.5)	4.3 (10.0)	5.2 (12.1)	5.0 (12.4)
No.	5 427	3 061	19 155	20 128
Per package beverage					3.72 (0.33, 42.02)	NS
Mean (SD)	0.4 (3.4)	0.4 (3.2)	0.3 (2.8)	0.1 (1.7)
No.	1 110	801	3 589	4 006
Per package food					0.67 (0.45, 1.00)	NS
Mean (SD)	8.5 (18.1)	5.6 (11.1)	6.3 (13.1)	6.2 (13.6)
No.	4 317	2 260	15 566	16 122
Fiber (grams)
Per serving food and beverage					1.47 (1.05, 2.06)	47
Mean (SD)	0.8 (1.4)	1.3 (2.4)	0.7 (1.4)	0.8 (1.5)
No.	5 568	3 149	19 840	20 623
Per serving beverage					1.91 (0.38, 9.56)	NS
Mean (SD)	< 0.1 (0.2)	< 0.1 (0.3)	< 0.1 (0.4)	< 0.1 (0.3)
No.	1 115	815	3 684	4 093
Per serving food					1.56 (1.13, 2.15)	56
Mean (SD)	1.0 (1.5)	1.7 (2.6)	0.9 (1.5)	0.9 (1.6)
No.	4 453	2 334	16 156	16 530
Per package food and beverage					1.28 (0.88, 1.87)	NS
Mean (SD)	3.0 (7.3)	4.4 (9.2)	2.0 (5.7)	2.2 (6.3)
No.	5 441	3 064	19 242	20 170
Per package beverage					1.20 (0.28, 5.11)	NS
Mean (SD)	0.1 (0.8)	0.2 (1.0)	0.1 (0.8)	0.1 (1.1)
No.	1 110	803	3 590	4 012
Per package food					1.36 (0.95, 1.95)	NS
Mean (SD)	3.8 (8.0)	5.9 (10.3)	2.4 (6.2)	2.8 (6.9)
No.	4 331	2 261	15 652	16 158
Calories (kilocalories)
Per serving food and beverage					0.73 (0.60, 0.88)	−27
Mean (SD)	115.7 (94.9)	85.9 (87.9)	113.4 (96.4)	115.8 (95.0)
No.	5 550	3 135	19 657	20 525
Per serving beverage					0.42 (0.23, 0.78)	−58
Mean (SD)	101.8 (103.1)	42.7 (72.8)	103.7 (103.5)	102.6 (100.4)
No.	1 115	813	3 684	4 089
Per serving food					0.82 (0.69, 0.99)	−18
Mean (SD)	119.1 (92.4)	101.1 (87.7)	115.6 (94.6)	119.1 (93.4)
No.	4 435	2 322	15 973	16 436
Per package food and beverage					0.88 (0.65, 1.19)	NS
Mean (SD)	413.7 (622.1)	355.4 (561.1)	334.9 (537.5)	328.0 (508.7)
No.	5 441	3 063	19 227	20 161
Per package beverage					0.71 (0.34, 1.49)	NS
Mean (SD)	122.3 (147.1)	86.3 (196.8)	119.7 (136.4)	118.1 (143.2)
No.	1 110	802	3 593	4 002
Per package food					0.93 (0.69, 1.26)	NS
Mean (SD)	488.4 (673.3)	450.8 (614.8)	384.3 (581.3)	380.0 (551.5)
No.	4 331	2 261	15 634	16 159

Note. CI = confidence interval; NS =not significant. The sample size was n = 49 403. Variation in cell sizes between nutrients is attributable to missing data for some nutrients. Variation in cell sizes between the main outcome (per serving) and sensitivity analyses (per package of food-and-beverage facings) is attributable to missing data in the servings per package variable.

^a % change calculated as (b−1)×100 for b < 1.0 and (1‒b)×100 for b ≥ 1.0.

Primary Outcomes

One year after Berkeley HCO implementation, there was a 70% reduction (95% confidence interval [CI] = −85%, −41%) in grams of added sugars per serving of food and beverage facings remaining at checkouts in Berkeley relative to comparison cities (b = 0.30; 95% CI = 0.15, 0.59). When we examined beverage facings and food facings separately, there were 72% (95% CI = −89%, −27%) and 73% (95% CI = −84%, −55%) relative reductions in added sugars per serving, respectively (b = 0.28; 95% CI = 0.11, 0.73 and b = 0.27; 95% CI = 0.16, 0.45). This pattern was consistent in sensitivity analyses using added sugars per package of food and beverage facings (Figure 1; Table 1).

FIGURE 1—

Mean (a) Added Sugars and (b) Sodium Content of Food and Beverage Products at Retail Checkouts Before and After Implementation of the Healthy Checkout Ordinance in Berkeley, California, and Comparison Cities (Comp): 2021‒2022

Note. Difference-in-differences models were used to determine the significance of nutrient changes compared to comparison cities, reported as b (95% confidence interval). Dotted lines indicate the maximum content of added sugars (5 grams) and sodium (200 mg) per serving of food products permitted at checkout by the Berkeley Healthy Checkout Ordinance.

There was not a statistically significant change in milligrams of sodium per serving of food and beverage facings (b = 0.81; 95% CI = 0.57, 1.14) or per serving of food facing alone (b = 0.94; 95% CI = 0.64, 1.37), likely because before implementation, only 9% of food facings at Berkeley checkouts were high in sodium (i.e., > 200 milligrams per serving). However, there was a 53% reduction (95% CI = −72%, −19%) in sodium per serving of beverage facings at checkout in Berkeley relative to comparison cities (b = 0.47; 95% CI = 0.28, 0.81). There was no statistically significant relative reduction in sodium content per package of food and beverage facings in Berkeley (Figure 1; Table 1).

Secondary Outcomes

There were changes in saturated fat, fiber, and calorie content of food and beverage facings at Berkeley checkouts. There was a 41% reduction (95% CI = −56%, −21%) in saturated fat content per serving of food and beverage facings displayed at checkout in Berkeley relative to comparison cities (b = 0.59; 95% CI = 0.44, 0.79), with a 38% reduction (95% CI = −54%, −17%) in saturated fat content per serving of food facings (b = 0.62; 95% CI = 0.46, 0.83). There was also a 47% increase (95% CI = 5%, 106%) in fiber content per serving of food and beverage facings displayed at checkout in Berkeley relative to comparison cities (b = 1.47; 95% CI = 1.05, 2.06), again, driven by a 56% increase (95% CI = 13%, 115%) in fiber content per serving of food (b = 1.56; 95% CI = 1.13, 2.15) in Berkeley relative to comparison cities. Beverage facings contained negligible saturated fat and fiber content per serving; thus, there were no significant changes for these nutrients from these products. Finally, there was a 27% reduction (95% CI = −40%, −12%) in calories per serving of food and beverage facings displayed at checkout in Berkeley relative to comparison cities (b = 0.73; 95% CI = 0.60, 0.88), an 18% reduction (95% CI = −31%, −1%) for food and a 58% reduction (95% CI = −77%, −22%) for beverage facings (b = 0.82; 95% CI = 0.69, 0.99 and b = 0.42; 95% CI = 0.23, 0.78, respectively; Figure A [available as a supplement to the online version of this article at https://ajph.org]; Table 1).

Nutrient and calorie reductions in food and beverage facings were largely driven by the removal of candy and SSBs (e.g., soda, energy drinks) from checkouts, which accounted for 50% and 39% of the reduction in added sugars, 88% and 2% of the reduction in saturated fat, and 72% and 8% of the reduction in calories, respectively. The increase in fiber was largely driven by the addition of bars (granola and protein) and nuts, which accounted for 47% and 25% of the increase in fiber, respectively. The sodium reduction in beverage facings was driven by the removal of regular and diet energy drinks, which accounted for 44% and 26% of the reduction, respectively.

The pattern of results for saturated fat, but not fiber or calories, was similar in sensitivity analyses examining changes per package of food and beverage facings (Table 1). There was a 35% reduction (95% CI = −57%, −4%) in saturated fat content per package of food and beverage facings displayed at checkout in Berkeley relative to comparison cities (b = 0.65; 95% CI = 0.43, 0.96). However, there were no statistically significant changes in fiber or calories per package of food and beverage facings at checkout in Berkeley relative to comparison cities. For fiber, the lack of significant increase per package was attributable to the large proportion of products that were gum and mints. After excluding these products, there was a significant 57% increase (95% CI = 9%, 124%) in fiber per package in Berkeley relative to comparison cities (b = 1.57; 95% CI = 1.09, 2.24). For calories, the lack of significant reduction per package was attributable to the larger package sizes (i.e., more servings per package) of healthier compliant products, including nuts, seeds, and whole grains. In a sensitivity analysis excluding these products, there was a significant 41% reduction (95% CI = −55%, −22%) in calories per package in Berkeley relative to comparison cities (b = 0.59; 95% CI = 0.45, 0.78).

Results did not vary substantially when we restricted the sample to the 41 145 food and beverage facings with exact nutrient data (Table B, available as a supplement to the online version of this article at https://ajph.org). However, there were 2 differences for secondary outcomes: saturated fat per package and calories per serving. The reductions in saturated fat per package of food and beverage facings and calories per serving of food displayed at checkout in Berkeley relative to comparison cities were no longer statistically significant, but the magnitude of change was similar to when including the full sample. Thus, these differences may be attributable to a loss in power from restricting the sample.

Primary Outcomes by Store Type

One-year changes in added sugars and sodium per serving of food and beverage facings at checkouts varied by store type (Figure 2, Table 2). Relative to comparison checkouts, in Berkeley checkouts, there were reductions in added sugars content per serving of food and beverage facings by 97% (95% CI = −98%, −94%) in chain specialty food stores (b = 0.03; 95% CI = 0.02, 0.06), 87% (95% CI = −88%, −86%) in chain supermarkets (b = 0.13; 95% CI = 0.12, 0.14), 85% (95% CI = −96%, −39%) in independent supermarkets (b = 0.15; 95% CI = 0.04, 0.61), 80% (95% CI = −88%, −68%) in chain mass merchandisers (b = 0.20; 95% CI = 0.12, 0.32), 72% (95% CI = −88%, −32%) in chain drug stores (b = 0.28; 95% CI = 0.12, 0.68), and 59% (95% CI = −81%, −9%) in independent grocery stores (b = 0.41; 95% CI = 0.19, 0.91). There was no statistically significant change in added sugars content per serving of food and beverage facings displayed in chain dollar store checkouts in Berkeley relative to comparison cities. This was driven by 1 of 2 dollar stores in Berkeley, which reduced the number of food and beverage facings at sampled checkouts (275–148 facings) but did not reduce the added sugars content of the foods and beverage facings at checkout (24.4–27.9 grams per serving).

FIGURE 2—

Mean (a) Added Sugars and (b) Sodium Content of Food and Beverage Products at Retail Checkouts by Store Type Before and After Implementation of the Healthy Checkout Ordinance in Berkeley, California, and Comparison Cities (Comp): 2021‒2022

Note. Difference-in-differences models were used to determine the significance of nutrient changes compared to comparison cities, reported as b (95% confidence interval).

TABLE 2—

Added Sugar and Sodium per Serving of Food and Beverage Products at Retail Checkouts Before and After Implementation of the Berkeley Healthy Checkout Ordinance by Store Type: Berkeley, California, and Comparison Cities, 2021‒2022

	Berkeley, CA		Comparison Cities		b (95% CI)	% Change,a Berkeley vs Comparison
	Preimplementation	Postimplementation	Preimplementation	Postimplementation
Added sugars (grams)
Chain supermarket					0.13 (0.12, 0.14)	−87
Mean (SD)	9.0 (15.6)	1.2 (4.6)	10.5 (15.9)	10.4 (15.8)
No.	538	602	6131	6360
Independent supermarket					0.15 (0.04, 0.61)	−85
Mean (SD)	5.3 (8.3)	0.9 (1.6)	10.9 (14.3)	12.2 (14.5)
No.	319	295	1597	1602
Chain mass merchandiser					0.20 (0.12, 0.32)	−80
Mean (SD)	10.0 (16.3)	1.9 (5.4)	11.8 (18.1)	11.6 (19.4)
No.	895	464	2503	2330
Chain specialty food					0.02 (0.02, 0.06)	−97
Mean (SD)	5.7 (6.3)	0.2 (0.7)	6.0 (7.8)	5.4 (7.2)
No.	461	317	1027	1277
Chain dollar					1.13 (0.82, 1.57)	NS
Mean (SD)	16.5 (19.9)	21.5 (27.4)	17.3 (21.2)	19.9 (22.8)
No.	857	191	1702	1755
Independent grocery					0.41 (0.19, 0.91)	−59
Mean (SD)	12.0 (10.5)	4.4 (9.8)	12.3 (10.9)	11.0 (10.4)
No.	280	185	647	805
Chain drug					0.28 (0.12, 0.68)	−72
Mean (SD)	11.8 (15.3)	3.5 (9.6)	11.3 (14.2)	11.9 (14.7)
No.	2148	1081	5987	5985
Sodium (milligrams)
Chain supermarket					0.45 (0.34, 0.60)	−55
Mean (SD)	90.1 (128.9)	41.9 (90.4)	87.1 (144.9)	89.1 (138.5)
No.	543	603	6137	6491
Independent supermarket					1.05 (0.61, 1.80)	NS
Mean (SD)	22.9 (42.7)	22.5 (50.7)	129.8 (278.3)	121.4 (246.1)
No.	327	307	1650	1728
Chain mass merchandiser					0.38 (0.28, 0.51)	−62
Mean (SD)	79.9 (101.6)	34.9 (64.3)	70.6 (109.5)	81.6 (126.5)
No.	897	464	2467	2314
Chain specialty food					0.57 (0.28, 1.16)	NS
Mean (SD)	92.1 (119.3)	75.7 (119.6)	40.3 (76.7)	58.2 (94.0)
No.	477	296	1026	1283
Chain dollar					0.68 (0.29, 1.61)	NS
Mean (SD)	72.3 (102.9)	49.7 (100.6)	102.6 (267.5)	103.4 (267.1)
No.	854	188	1660	1801
Independent grocery					2.66 (0.82, 8.67)	NS
Mean (SD)	59.8 (76.0)	120.9 (172.6)	130.4 (255.3)	99.0 (215.0)
No.	294	200	672	872
Chain drug					1.00 (0.67, 1.50)	NS
Mean (SD)	85.8 (139.0)	95.4 (116.5)	74.6 (133.8)	82.8 (128.6)
No.	2151	1084	5996	6017

Note. CI = confidence interval; NS = not significant. The sample size was n = 49 403.

^a % change calculated as (b−1)×100 for b < 1.0 and (1‒b)×100 for b ≥ 1.0.

Relative to comparison checkouts, in Berkeley checkouts, there were reductions in sodium content per serving of food and beverage facings by 62% (95% CI = −72%, −49%) in chain mass merchandisers (b = =0.38; 95% CI = 0.28, 0.51) and 55% (95% CI = −66%, −40%) in chain supermarkets (b = 0.45; 95% CI = 0.34, 0.60). There were no statistically significant changes in sodium content per serving of food and beverage facings in other store types.

DISCUSSION

To our knowledge, this study is the first to quantify changes in the nutrient content of food and beverage facings available at retail checkouts before and after a healthy checkout policy. Roughly 1 year after the Berkeley HCO was implemented, there was a 42% reduction in the number of food and beverage facings available at sampled checkouts. Among the remaining food and beverage facings that were available at Berkeley checkouts, there were significant reductions in the added sugars and saturated fat content per serving of food and beverage facings, a significant reduction in sodium per serving of beverage facings, and a significant increase in fiber per serving of food and beverage facings. There was also a significant reduction in calorie content per serving. These findings are consistent with a previous evaluation of Berkeley HCO compliance, which found decreased presence of noncompliant products like SSBs, candy, and other sweets, along with increased presence of both nonfood and nonbeverage products and compliant food and beverage facings at checkout.²⁰

The reduction in candy and SSB availability at checkout was responsible for most of the reduction in added sugars. SSBs were often replaced by unsweetened beverages like water, sparkling water, and coconut water.²⁰ Though the Berkeley HCO prohibits products with nonsugar sweeteners at checkout, previous work did not find a significant reduction in the presence of these products at checkout.²⁰ Future research should assess reasons for retailer noncompliance.

Though Berkeley’s HCO intended to limit sodium at checkout, significant reductions were only seen in the sodium content per serving of beverage facings. This was because of a floor effect (only 9% of food facings in Berkeley checkout areas were high in sodium before implementation) and the lack of a significant decrease in noncompliant salty snacks in Berkeley checkouts.²⁰ Though the sodium standard set by the Berkeley HCO did not significantly decrease sodium at checkout, it did prevent the potential unintended consequence of replacing high-added-sugars products with high-sodium products.

We also examined changes in the nutrients saturated fat and fiber, which are not limited by the HCO, as well as calories. In Berkeley, relative to comparison checkouts after the HCO, there was a significant reduction in saturated fat at checkout, explained by a 38% reduction per serving of food facings, and a significant increase in fiber at checkout, explained by a 56% increase in the fiber content per serving of food facings. This increase in fiber is largely attributed to increased nuts, which are naturally high in fiber and were more frequently placed at checkout after the HCO.²⁰ There was also a significant relative reduction in calories, explained by a large (58%) reduction in calories per serving of beverage facings and, to a lesser extent, an 18% reduction in calories per serving of food facings. When we examined saturated fat, fiber, and calories per package (i.e., the whole product the shopper takes home), there were no significant changes. For fiber, this was explained by the high prevalence of gum and mints, and for calories, this was explained by package sizes of healthier products (e.g., nuts, seeds, whole grains) being particularly large.

The HCO nutritional standards used a hybrid approach, limiting added sugars and sodium and restricting to mostly health-promoting food groups at checkout.¹⁵ Though we examined additional nutrients and calories as secondary outcomes, it should be noted that calorie content may not be a good indicator of the nutritional quality of some foods, such as nuts and seeds.

The earlier evaluation of the Berkeley HCO took a whole-food approach, finding a significant increase in the presence of healthy food and beverage facings and nonfood items, and a decrease in unhealthy items available at checkout.²⁰ The current study builds on these results, focusing on changes in specific nutrients that are overconsumed by the US population. Roughly two thirds of US adults overconsume added sugars, up to 97% overconsume sodium, three quarters overconsume saturated fat, and up to 97% underconsume fiber,²² increasing risk of chronic diseases, including heart disease, type 2 diabetes, and certain cancers.^22,23 Fiber plays a key role in satiety, regulating blood sugar levels, and maintaining healthy cholesterol levels.²² Promoting nutrient-dense, fiber-rich foods via HCOs may encourage healthier purchasing and consumption patterns, potentially reducing diet-related chronic disease burdens.

HCOs may be an especially helpful intervention to counteract high consumption of added sugars, which is increasing globally, particularly in the form of SSBs.²⁴ By limiting added sugars at checkout, HCOs could reduce the purchase and consumption of high-sugar items, though these behavioral outcomes have not yet been evaluated under a mandatory HCO. An evaluation of voluntary but meaningful healthy checkout restrictions in the United Kingdom found fewer purchases of the sweets and salty snacks previously available at checkout.²⁵ In the United States, no chain has voluntarily adopted meaningful healthy checkout restrictions, and incentive-based healthy retail programs are often resource-intensive and include a small number of stores.²⁶ By contrast, mandatory healthy checkout policies like Berkeley’s HCO can have broad population reach. Future research should examine the impact of HCOs on consumer purchases and explore public perceptions of these policies. Given that public support for healthy retail initiatives varies,²⁷ potentially because of how they are framed, understanding how consumers respond to HCOs could help inform future policy design, adoption, and implementation strategies.

Regarding changes by store type, all store types except dollar stores significantly reduced added sugars content per serving of food and beverage facing, with the largest reductions observed in chain specialty food stores, supermarkets, and chain mass merchandisers. The overall number of food and beverage facings at checkout decreased after implementation,²⁰ and some remaining food and beverage products, particularly at dollar stores, remained high in added sugar. Because there were only 2 dollar stores in this study sample, further research is needed to better understand how HCO implementation differs by store type. However, previous research has documented issues in dollar stores, like understocked shelves, understaffing, and local managers with limited authority, that could pose barriers to compliance with local food policies.^28,29 In contrast to added sugars, for sodium, only chain mass merchandisers and chain supermarkets reduced sodium content per serving of food and beverage facings at checkout. Additional research³⁰ is needed on the Berkeley HCO implementation and enforcement process to better understand variation by store types.

Limitations and Strengths

Limitations of this study include having neither data before 2021 to test the parallel trends assumption for difference-in-differences models nor monthly data to examine fluctuations in policy impact. We also only assessed a sample (not a census) of checkouts in each store. Nonetheless, this study has several strengths, including the use of an objective and reliable tool to assess food and beverage characteristics and the inclusion of comparison cities to ensure that changes observed in Berkeley were not attributable to larger market trends. In addition, the large sample size of more than 49 000 food and beverage facings, linked with nutritional information, and the near census of stores subject to the policy provide robust data.

Public Health Implications

Healthy checkout policies can meaningfully improve the healthfulness of retail food environments at checkout. The overall healthfulness of food environments may further be improved by combining healthy checkout policies with other product placement efforts (e.g., healthy aisle endcaps and front-of-store displays), as has been done in the United Kingdom,²⁵ and policies like front-of-package nutrient labels and SSB taxes that have catalyzed reformulation.^16,31,32 To potentially scale healthy product placement nationally, eligibility requirements for retailers authorized to accept Supplemental Nutrition Assistance Program (SNAP) benefits could be expanded.^33,34 SNAP retailer requirements, which include stocking a variety of staple foods, could additionally require nutrition standards for prominently placed products, thus reducing promotion of and exposure to nutrients of concern in these stores.^34,35 Because the majority of food retailers are SNAP-authorized, a healthy product placement requirement could improve food environments for most households in the United States.