Abstract
Background:
Calorie labeling is mandated in restaurant chains with ≥20 locations nationwide, but the effect of labeling on the nutritional quality of purchased meals in fast food settings is unclear, especially for adolescents and children.
Objective:
To estimate the effect of calorie menu labeling on the nutrient content and composition of fast food meal purchases in McDonald’s restaurants vs. control restaurants.
Design:
From 2010 to 2014, customers were sampled during repeated visits to McDonald’s restaurants, which voluntarily labeled menus with calorie information in 2012, and five control fast food restaurant chains that had not labeled their menus over the study period.
Participants/setting:
Restaurant receipts and questionnaires were collected from 2,883 adults, 2,131 adolescents, and 433 children in four New England cities.
Main outcome measures:
Saturated fat, sugar, sodium, and fiber content of purchased meals, as well as nutrient densities (i.e. nutrient content adjusted for total calories) were calculated after linking purchases to nutrition data from restaurant websites.
Statistical analyses:
Difference-in-differences analyses and linear mixed models were used to examine whether calorie labeling in McDonald’s was associated with changes in nutrient content and nutrient density of purchased meals separately in adults, adolescents, and children.
Results:
After implementation of calorie labeling at McDonald’s, adult meals contained on average 4.0% fewer calories from sugar (95% CI: −7.5, −0.5), but 1.8% more calories from saturated fat (95% CI: 0.7, 2.9) compared to control chains. The fiber content of children’s purchased meals was lower at McDonald’s after calorie labeling (change=−1.4g, 95% CI: −2.5, −0.3). Calorie labeling was not associated with changes in nutrient quality of adolescent meals.
Conclusions:
Calorie labeling at McDonald’s was associated with a possibly small positive change in the nutritional quality of meals in adults, but not in adolescents or children. Efforts are needed to improve the nutritional quality of restaurant meals.
Keywords: calorie labeling, diet quality, natural experiment, nutrients, obesity prevention
INTRODUCTION
The Affordable Care Act’s menu labeling provision requires restaurant chains with 20 or more locations nationwide to label their menus with calorie information.1 This policy was implemented in May 2018 with the goal of encouraging customers to purchase lower-calorie items. Several studies have examined the effect of calorie labeling on the calorie content of meal purchases in fast food settings, with most showing no association2 or only a modest association.3,4 However, the nutrient content of purchased restaurant meals could also change due to labeling. This could occur either as a byproduct of changes in calorie content (e.g. if calorie content decreases then amounts of some macronutrients must also decrease) or could be an unintended consequence of labeling, even if calorie content remains constant. For example, if customers purchase more salads after labeling, this could increase the fiber content of purchased meals. Similarly, to decrease calorie content while maintaining palatability, chains might increase sodium, which does not add calories.
Despite the importance of nutritional quality for chronic disease management and prevention,5 only two studies have estimated the effect of calorie labeling on the nutritional content of purchased fast food meals,6,7 both of which found no associations. Both of these studies included only adults and were conducted in New York City after calorie labeling was implemented there in 2008. It is important to evaluate the impact of fast food menu labeling on nutritional quality in other age groups. The majority of adolescents and children have poor-quality diets8 and one-third consume fast food on a given day.9 It is also important to evaluate this policy in other cities. Because of differences in demographics, food environments, and local policies, effects may differ from what was observed in New York City.
To address these gaps, this study examined associations of calorie labeling on the nutritional content of meals purchased in adults, adolescents, and children in four New England cities from 2010 to 2014 in McDonald’s, which voluntarily implemented calorie labeling in 2012, compared to five control restaurant chains that did not label their menus over this period. A previous study using these data found that calorie labeling was not associated with large changes in calorie content of meal purchases,10 but labeling may have affected customer meal choices or McDonald’s menu offerings, which could affect nutritional content. This analysis specifically examined saturated fat, sugar, dietary fiber, and sodium, all of which are associated with chronic disease.11
METHODS
Details regarding study design and participant enrollment have been discussed previously.10,12 The nutrient content of purchased meals was examined before and after calorie labeling at McDonald’s in 2012 compared to purchases in a group of control restaurant chains: Burger King, Subway, KFC (except for adolescents), Wendy’s, and Dunkin’ Donuts (only for adolescents). These were chosen as control chains because they were similar to McDonald’s in popularity and price of meals but did not label their menus over the study period. The study area included Boston and Springfield, Massachusetts; Hartford, Connecticut; and Providence, Rhode Island. In 2010, three McDonald’s, three Burger Kings, two Subways, one Kentucky Fried Chicken, and one Wendy’s were randomly selected in each city to recruit adults (≥18 years) and children (3-15 years). To recruit adolescents (11-20 years), two McDonald’s, two Burger Kings, two Subways, two Dunkin’ Donuts, and one Wendy’s that were within one mile of a high school were selected.
The same restaurant locations were revisited every year (from April to August) until 2014 except when management refused or when very few customers were recruited at those restaurants, in which case the nearest restaurant of the same chain was selected. Forty-eight restaurants were visited to obtain the adult and child samples, 35 of which were revisited every year; the remaining 13 either had to be replaced or served as replacement study sites at some point. Fifty-three restaurants were visited to obtain the adolescent sample, 29 of which were revisited every year and 24 of which had to be replaced or served as replacement sites.
Participants were recruited upon entering the restaurant and were asked, upon exiting, to return their receipt and complete a questionnaire that clarified which items on the receipt were purchased for their (or their child’s) personal consumption and that gave details that were not clear from the receipt (e.g. added condiments, beverage choices, etc.). The questionnaire also collected information on participants’ demographic characteristics and food choice preferences.10,12 Information was gathered directly from adults and adolescents, but questions about children’s purchases were directed to their parent or guardian. To limit potential bias in responses, participants were told that the study was collecting information on “food and drink choices at fast food restaurants.” Response rates were 42% for adults, 45% for adolescents, and 44% for children, and generally similar in the pre- and post-intervention periods for all.
The calorie, saturated fat (g), sugar (g), dietary fiber (g), and sodium (mg) content of purchases were determined by linking participants’ purchased items to contemporaneous nutrition information posted on restaurant websites and summing over all items. Nutrient densities were also calculated. For saturated fat and sugar, this was done by calculating the calories contributed by these nutrients (i.e. multiplying saturated fat content by 9 calories/g and sugar content by 4 calories/g), dividing this by total calorie content, and multiplying by 100% to get percent calories from saturated fat and sugar. For fiber and sodium, which do not contribute to calories, the nutrient density was expressed as nutrient content per 100 calories.
Analyses were conducted separately in adults, adolescents, and children. Participants who purchased >5,000 calories (<1% in all samples) or who had missing data on covariates (5% of adults and adolescents; 14% of children) were excluded. The analyses estimated the effect of calorie labeling on the absolute nutrient content and nutrient densities of purchases using a difference-in-differences approach. Both outcomes were considered because some nutrients (saturated fat and sugar) are more often recommended as a percent of calorie intake, whereas others (fiber and sodium) are often recommended in absolute amounts and do not contribute calories. Mixed linear regression models were fit with terms for intervention (McDonald’s vs. control), period (pre- vs. post-labeling), and an interaction between intervention and period (βinteraction), which denoted differences in nutrient content pre vs. post-labeling between McDonald’s and control chains and was the measure of interest. Restaurant location was included as a random effect. A propensity score weighting strategy13 was used to ensure that the four groups (defined by intervention and period status) were balanced on covariates whose distributions may have changed over time differentially between intervention arms: age, sex, race/ethnicity (White, Black, Asian, Hispanic, other), body mass index (kg/m2), city, restaurant chain, and self-reported importance of price, taste, and convenience in making meal choices. We also estimated the pre-post difference within McDonald’s and control restaurants separately.
All analyses were conducted in SAS version 9.4 (Cary, NC). Two-sided 95% confidence intervals (CI) were calculated for all statistical tests. This study was approved by the institutional review board of Harvard Pilgrim Health Care. Because data collection was anonymous and queried non-sensitive information, the board approved this study without requiring written informed consent from participants.
RESULTS
The study population (Table 1) included 2,883 adults (mean age, 37.6 years [standard deviation, SD: 15.8]; 43% female), 2,131 adolescents (mean age, 16.3 years [SD: 2.7]; 48% female), and 433 children (mean age, 8.2 years [SD: 3.2]; 50% female). Most participants were Black or Hispanic. The mean calorie content of purchased meals was 804 (SD: 464) calories for adults, 741 (SD: 445) for adolescents, and 697 (SD: 356) for children.
Table 1.
Characteristicsa of adults, adolescents, and children recruited in McDonald's and other chains in the period before and after implementation of calorie labeling in McDonald's
| Adults |
Adolescents |
Children |
||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| McDonald's | Other chain | McDonald's | Other chain | McDonald's | Other chain | |||||||
| Pre (n=571) |
Post (n=338) |
Pre (n=1223) |
Post (n=751) |
Pre (n=537) |
Post (n=331) |
Pre (n=736) |
Post (n=527) |
Pre (n=125) |
Post (n=58) |
Pre (n=166) |
Post (n=84) |
|
| Age, years | 37.8 (16.6) |
38.3 (15.7) |
37.0 (15.8) |
38.0 (15.1) |
15.8 (2.9) |
15.7 (2.7) |
16.5 (2.7) |
16.7 (2.5) |
7.8 (3.0) |
8.5 (3.3) |
8.4 (3.3) |
8.2 (3.1) |
| BMI, kg/m2 | 28.1 (6.1) |
27.7 (5.8) |
28.0 (6.3) |
28.1 (6.0) |
23.1 (4.6) |
23.5 (5.0) |
23.6 (4.8) |
24.2 (5.3) |
20.0 (6.0) |
22.5 (7.1) |
22.1 (6.9) |
22.3 (7.6) |
| BMI-for-age-and-sex z-score | -- | -- | -- | -- | 0.51 (0.92) |
0.54 (1.09) |
0.50 (1.03) |
0.59 (0.98) |
0.75 (1.92) |
1.20 (1.58) |
1.24 (1.54) |
1.27 (1.69) |
| Calories purchased | 721 (475) |
636 (423) |
880 (451) |
817 (468) |
754 (480) |
709 (408) |
751 (424) |
736 (459) |
718 (370) |
589 (288) |
777 (350) |
585 (345) |
| Sex (%) | ||||||||||||
| Male | 52 | 52 | 59 | 61 | 52 | 55 | 51 | 50 | 50 | 51 | 51 | 48 |
| Female | 48 | 48 | 41 | 39 | 48 | 45 | 49 | 50 | 50 | 49 | 49 | 52 |
| Race/ethnicity (%) | ||||||||||||
| White | 34 | 37 | 40 | 39 | 17 | 17 | 22 | 20 | 15 | 17 | 23 | 15 |
| Black | 33 | 37 | 29 | 29 | 32 | 35 | 35 | 38 | 38 | 36 | 29 | 33 |
| Hispanic | 20 | 17 | 19 | 20 | 30 | 25 | 26 | 26 | 27 | 29 | 33 | 35 |
| Asian | 4 | 2 | 4 | 5 | 7 | 8 | 5 | 5 | 3 | 3 | 4 | 4 |
| Other | 9 | 8 | 9 | 7 | 14 | 14 | 12 | 11 | 17 | 15 | 12 | 13 |
| City (%) | ||||||||||||
| Boston | 28 | 40 | 24 | 37 | 55 | 70 | 44 | 54 | 30 | 44 | 35 | 52 |
| Hartford | 28 | 21 | 33 | 24 | 16 | 16 | 17 | 14 | 32 | 19 | 36 | 21 |
| Providence | 30 | 28 | 23 | 27 | 18 | 7 | 24 | 20 | 23 | 19 | 16 | 15 |
| Springfield | 15 | 12 | 20 | 11 | 11 | 7 | 15 | 12 | 15 | 19 | 14 | 12 |
| Restaurant (%) | ||||||||||||
| McDonald's | 100 | 100 | -- | -- | 100 | 100 | -- | -- | 100 | 100 | -- | -- |
| Burger King | -- | -- | 43 | 48 | -- | -- | 35 | 29 | -- | -- | 52 | 67 |
| Kentucky Fried Chicken | -- | -- | 18 | 16 | -- | -- | -- | -- | -- | -- | 21 | 12 |
| Subway | -- | -- | 24 | 22 | -- | -- | 21 | 21 | -- | -- | 14 | 14 |
| Wendy's | -- | -- | 15 | 14 | -- | -- | 15 | 17 | -- | -- | 13 | 7 |
| Dunkin' Donuts | -- | -- | -- | -- | 29 | 33 | -- | -- | -- | -- | ||
Mean (standard deviation) or %
Among adults, calorie labeling at McDonald’s was associated with lower sugar content (βinteraction=−7.6g, 95% CI: −12.6, −2.6) (Table 2) and lower percent calories from sugar (βinteraction=−4.0%, 95% CI: −7.5, −0.5) (Table 3), but higher percent calories from saturated fat (βinteraction=1.8%, 95% CI: 0.7, 2.9) than what would be expected had McDonald’s not implemented labeling. Fiber content was also lower (βinteraction=−0.5g, 95% CI: −1.0, 0.0) at McDonald’s post- vs. pre-labeling compared to control chains, though fiber density was unchanged. There were no changes in sodium content or density associated with labeling.
Table 2.
Multivariable-adjusted changes (95% CI) in the nutrient content of meals purchased by adults, adolescents, and children after calorie labeling in McDonald's restaurants compared to other fast food restaurants
| McDonald's |
Other Chains |
Difference-in- differencesb |
P diff- in-diff |
|||||
|---|---|---|---|---|---|---|---|---|
| Prea | Posta | Differenceb,c | Prea | Posta | Differenceb,c | |||
| Saturated fat (g) | ||||||||
| Adults | 9.6 (7.6) | 8.6 (6.4) | −0.9 (−1.8, 0.0) | 12.4 (8.0) | 10.8 (7.6) | −1.8 (−2.5, −1.1) | 1.0 (−0.1,2.1) | 0.07 |
| Adolescentsd,e | 9.5 (8.2) | 9.3 (6.9) | −0.2 (−1.2, 0.9) | 10.8 (7.3) | 11.1 (7.8) | 0.2 (−0.7, 1.0) | −0.4 (−1.8, 0.9) | 0.55 |
| Childrene | 7.7 (5.7) | 6.6 (4.0) | −0.8 (−2.1,0.6) | 9.7 (5.9) | 7.1 (5.1) | −2.7 (−3.9, −1.4) | 1.6 (−0.3, 3.5) | 0.09 |
| Sugar (g) | ||||||||
| Adults | 42.7 (35.0) | 35.4 (29.7) | −7.1 (−11.4, −2.9) | 36.6 (34.3) | 34.3 (34.2) | 0.3 (−2.8, 3.5) | −7.6 (−12.6, −2.6) | 0.003 |
| Adolescentsd,e | 45.9 (33.5) | 40.6 (31.1) | −5.0 (−9.5, −0.4) | 42.7 (36.3) | 40.6 (34.5) | 0.7 (−3.2, 4.6) | −5.6 (−11.6, 0.3) | 0.06 |
| Childrene | 44.8 (29.9) | 36.7 (24.6) | −10.2 (−17.8, −2.5) | 40.1 (27.9) | 31.4 (26.3) | −10.8 (−17.9, −3.7) | 0.9 (−9.4, 11.2) | 0.87 |
| Dietary fiber (g) | ||||||||
| Adults | 4.1 (3.3) | 3.6 (3.1) | −0.4 (−0.8, 0.0) | 4.9 (3.4) | 4.9 (3.7) | 0.1 (−0.2, 0.4) | −0.5 (−1.0, 0.0) | 0.04 |
| Adolescentsd,e | 4.2 (3.3) | 3.8 (2.9) | −0.4 (−0.8, 0.1) | 4.2 (3.6) | 4.0 (3.5) | 0.0 (−0.3, 0.4) | −0.4 (−0.9, 0.2) | 0.20 |
| Childrene | 4.2 (2.9) | 3.0 (2.1) | −0.9 (−1.7, −0.2) | 4.1 (2.6) | 4.2 (4.0) | 0.5 (−0.4, 1.4) | −1.4 (−2.5, −0.3) | 0.01 |
| Sodium (mg) | ||||||||
| Adults | 1074 (937) | 929 (760) | −135 (−244, −26) | 1760 (1046) | 1579 (1056) | −93 (−180, −7) | −24 (−158, 109) | 0.72 |
| Adolescentsd,e | 1116 (964) | 998 (705) | −110 (−223, 3) | 1288 (969) | 1197 (961) | 7 (−91, 106) | −109 (−261,44) | 0.16 |
| Childrene | 989 (646) | 835 (497) | −143 (−308, 21) | 1483 (793) | 1063 (723) | −334 (−514, −156) | 215 (−24, 455) | 0.08 |
Mean (SD) presented
Adjusted for age (years, continuous), sex (female, male), race/ethnicity (White, Black, Asian, Hispanic, other), BMI (kg/m2, continuous), city (Boston, Hartford, Providence, Springfield), and restaurant chain (Burger King, KFC, Subway, Wendy's), importance of price (none, a little, a lot), importance of taste (none, a little, a lot), and importance of convenience (none, a little, a lot). Restaurant location was included as a random effect.
Because pre-post differences were multivariable-adjusted, they may not equal the difference between the mean unadjusted nutrients in the pre and post periods
Control restaurants included Burger King, Wendy's, Subway, and Dunkin' Donuts
Adjusted for BMI-for-age-and-sex-z score (continuous) instead of BMI
Table 3.
Multivariable-adjusted changes (95% CI) in nutrient densities of meals purchased by adults, adolescents, and children after calorie labeling in McDonald's restaurants compared to other fast food restaurants
| McDonald's |
Other Chains |
Difference-in- differencesb |
P diff- in-diff |
|||||
|---|---|---|---|---|---|---|---|---|
| Prea | Posta | Differenceb,c | Prea | Posta | Differenceb,c | |||
| Saturated fat (%calories) | ||||||||
| Adults | 12.4 (8.2) | 13.1 (8.4) | 0.7 (−0.4, 1.8) | 12.6 (6.2) | 12.0 (5.2) | −1.1 (−1.6, −0.6) | 1.8 (0.7, 2.9) | 0.001 |
| Adolescentsd,e | 11.0 (6.4) | 11.7 (5.9) | 0.5 (−0.3, 1.4) | 13.2 (8.0) | 14.4 (8.5) | 0.1 (−0.7, 0.8) | 0.4 (−0.8, 1.5) | 0.52 |
| Childrene | 9.4 (5.9) | 10.8 (5.1) | 1.6 (0.1, 3.0) | 11.0 (4.1) | 10.9 (4.7) | 0.0 (−1.0, 1.1) | 1.3 (−0.4, 3.1) | 0.14 |
| Sugar (%calories) | ||||||||
| Adults | 29.6 (25.6) | 27.3 (24.2) | −2.9 (−6.2, 0.3) | 19.0 (21.5) | 19.2 (19.7) | 0.6 (−1.4, 2.6) | −4.0 (−7.5, −0.5) | 0.03 |
| Adolescentsd,e | 30.6 (26.8) | 26.9 (23.3) | −3.3 (−6.9, 0.3) | 27.5 (26.5) | 26.3 (23.3) | −0.2 (−2.9, 2.6) | −2.7 (−7.2, 1.8) | 0.23 |
| Childrene | 27.8 (20.9) | 27.0 (16.7) | −1.4 (−7.1, 4.3) | 22.2 (16.9) | 26.3 (22.7) | 0.8 (−3.9, 5.5) | −2.1 (−9.4, 5.2) | 0.57 |
| Dietary fiber (g/100 calories) | ||||||||
| Adults | 0.51 (0.34) | 0.51 (0.42) | 0.03 (−0.05, 0.06) | 0.56 (0.38) | 0.58 (0.37) | 0.02 (−0.01, 0.05) | −0.01 (−0.07, 0.05) | 0.71 |
| Adolescentsd,e | 0.51 (0.34) | 0.51 (0.30) | −0.01 (−0.05, 0.04) | 0.50 (0.38) | 0.48 (0.34) | −0.01 (−0.04, 0.02) | 0.00 (−0.05, 0.06) | 0.86 |
| Childrene | 0.56 (0.28) | 0.47 (0.25) | −0.04 (−0.12, 0.03) | 0.52 (0.27) | 0.66 (0.70) | 0.19 (0.04, 0.34) | −0.23 (−0.40, −0.06) | 0.008 |
| Sodium (mg/100 calories) | ||||||||
| Adults | 130.2 (77.3) | 130.8 (79.0) | 1.4 (−8.8, 11.6) | 195.9 (84.6) | 187.0 (81.3) | −3.4 (−9.8, 3.1) | 5.7 (−5.6, 17.1) | 0.32 |
| Adolescentsd,e | 130.1 (72.3) | 134.5 (72.6) | 5.2 (−4.9, 15.3) | 154.6 (79.5) | 145.7 (77.8) | −5.9 (−13.3, 1.4) | 11.3 (−1.5, 24.2) | 0.08 |
| Childrene | 131.2 (56.4) | 135.5 (56.7) | 5.6 (−10.5, 21.8) | 188.8 (68.7) | 173.7 (83.0) | −2.9 (−18.2, 12.4) | 7.9 (−14.5, 30.3) | 0.49 |
Mean (SD) presented
Adjusted for age (years, continuous), sex (female, male), race/ethnicity (White, Black, Asian, Hispanic, other), BMI (kg/m2, continuous), city (Boston, Hartford, Providence, Springfield), and restaurant chain (Burger King, KFC, Subway, Wendy's), importance of price (none, a little, a lot), importance of taste (none, a little, a lot), and importance of convenience (none, a little, a lot). Restaurant location was included as a random effect.
Because pre-post differences were multivariable-adjusted, they may not equal the difference between the mean unadjusted nutrient densities in the pre and post periods
Control restaurants included Burger King, Wendy's, Subway, and Dunkin' Donuts
Adjusted for BMI-for-age-and-sex-z score (continuous) instead of BMI
Calorie labeling was not associated with changes in the nutrient content or nutrient density of purchases in adolescents. In children, purchased meals had lower fiber content (βinteraction=−1.4g, 95% CI: −2.5, −0.3) and lower fiber density (βinteraction=−0.23g/100 calories, 95% CI: −0.40, −0.06) at McDonald’s post- vs. pre-labeling compared to other chains. Labeling was not associated with the absolute saturated fat, sugar, or sodium content of children’s purchased meals, though some reductions were observed in both McDonald’s and control restaurants separately over time. Labeling was also not associated with the density of these nutrients in children’s purchased meals; these nutrient densities did not appear to change over time within each intervention group, unlike results from analyses of absolute nutrient content.
DISCUSSION
In a previous analysis, McDonald’s calorie labeling was not associated with large changes in the calorie content of meal purchases after labeling implementation in any age group.10 The results of the current study further suggest that calorie labeling was not associated with large changes in overall nutritional quality among adolescents or children. To the authors’ knowledge, this is the first study to examine associations between labeling and the nutritional quality of fast food meals in adolescents and children, showing no positive changes after labeling. Associations between labeling and the nutritional quality of purchased adult meals were varied; purchased meals had lower percent calories from sugar, higher percent calories from saturated fat, and lower fiber content after labeling. The relatively larger reduction in sugar may indicate a slight overall benefit to health for adults dining at fast food restaurants after labeling.
Saturated fat and sugar are associated with risk of chronic diseases such as heart disease14,15 and diabetes,16,17 whereas fiber is inversely associated with these diseases.18 The lower sugar density of adult purchased fast food meals after calorie labeling suggests a beneficial change, while the lower fiber content and higher saturated fat density after labeling suggest adverse changes, making it difficult to know the net impact of labeling on the overall nutritional quality of these meals. However, when considering the recommendations released by the 2015-2020 Dietary Guidelines for Americans (DGA),19 the changes we observed in adults may point to a slight benefit overall. The DGA recommends that <10% of daily calorie intake should be from saturated fat and <10% should be from added sugars. Labeling was associated with a 4% reduction in percent calories from sugar and only a 1.8% increase in calories from saturated fat, which may point to a possible overall slight improvement in the nutritional quality of adult purchased meals after labeling (assuming all reductions in sugar came from added sugar). The total impact of labeling on health would be less than this because adults on average only obtain about 10% of calories from fast food meals,20 but such a change may be clinically significant at the population level. Confirming this would require data on whether customers compensated for the nutritional changes during meals eaten outside of the restaurant (e.g. if adults consumed a higher percent of calories from sugar at meals later in the day).
Although the absolute fiber content of purchased adult meals was lower after labeling, the fiber density did not change. This could be observed if adults ordered fewer calories but did not change the fiber composition of their meals (i.e. if total grams of fiber per 100 calories in meals was constant). There was no association of labeling with calorie content of purchased meals in a previous analysis of these data, though a very small reduction could not be ruled out based on sample size.10 To this point, previous studies using larger databases of restaurant transactions have found small reductions in calorie content of purchased meals.3,4 If a similarly small reduction occurred in McDonald’s, it could explain the differences observed between analyses of absolute fiber content and fiber density. By comparison, labeling was associated with lower absolute sugar content and sugar density, suggesting that the sugar composition of purchased meals was lower in adults after labeling independent of calories.
In children, the fiber content and density of purchased meals were lower in McDonald’s post-labeling compared to control chains, suggesting a potential adverse and unintended consequence of labeling. Potential mechanisms for this association are unclear. Given the low number of children in the study, these results should be interpreted cautiously, as the sampled children may not be representative of children who consume fast food meals. The large number of tests we conducted also increases the possibility of observing a chance finding. Although the reduction in fiber raises concerns, the mean saturated fat, sugar, and sodium content of children’s purchased meals were generally lower in both McDonald’s and control restaurants post-labeling. This was also true for calorie content in a previous analysis of these data,10 which explains why nutrient content was lower in the present analysis whereas nutrient densities were unchanged. However, children’s purchased meals in the post-labeling period still contained, on average, nearly 40% of daily recommended saturated fat and sodium intake and nearly 100% of daily recommended added sugar intake for a child of the average age in the sample (about eight years of age) according to the DGA.19 Moreover, the nutritional quality of restaurant offerings specifically marketed for children (i.e. “kid’s meals”) has not improved over time despite voluntary commitments by many restaurants,21 underscoring the need for continued intervention in this population.
Two previous studies of adults in fast food restaurants in New York City did not find associations of calorie labeling with the nutrient content of meal purchases.6,7 Both of these studies considered the same nutrients as in the present analysis except for dietary fiber. Compared with those studies, the present study included adolescents and children and sampled individuals from four different cities in New England. Two studies conducted in full-service restaurants observed improvements to the nutrient content of meal purchases associated with labeling,22,23 but the labels in those studies included specific nutrient content (e.g. fat, sodium, etc.), whereas McDonald’s labels only included calories. It is unclear whether results in those studies are due to the inclusion of this nutrient information on the labels or whether labeling is more effective in full-service settings than fast food settings.
This study has many strengths, including the use of a natural experiment, repeated sampling of restaurants over five years, a racially and ethnically diverse population, and enrollment in popular restaurants, which may increase the generalizability of the results. However, it also has several limitations. First, participation rates were <50%; these rates were comparable to those of other studies and were similar in pre- and post-labeling periods. Second, this study only considered purchased, not consumed food. It is possible that calorie labeling could affect intake through mechanisms other than purchasing (e.g. post-purchasing behavior). Third, there could be errors in the reported nutrient content of items on restaurant websites. However, this error would have to be non-random (i.e. nutrients are under- or overestimated on average) and differential with respect to time and intervention group to substantially affect the differences-in-differences. While this cannot be tested without knowing the true nutrient content of restaurant items, the general consistency of restaurant offerings over time may limit the degree to which this error affected the primary outcomes. Fourth, we may have had limited power to detect very small changes in nutrient content. Because previous studies have suggested that labeling may only have a very small effect on calorie content of purchased foods,3,4 it is possible that changes in nutrient content may also be very small and difficult to detect with our sample size. Lastly, this study considered only four nutrients that have important effects on health. Intake of other nutrients and specific food groups (e.g. vegetables, whole/refined grains, red/processed meat, etc.) are also crucial indicators of diet quality.24 Further study of nutrition substitution and changes in food group composition of purchases in response to calorie labeling should be considered.
CONCLUSIONS
In summary, this study demonstrates that calorie labeling was not associated with large improvements in nutritional quality of purchased meals for adolescents or children. Although it is difficult to determine the overall impact of labeling on health in adults due to inconsistent changes in saturated fat, sugar, and fiber density, the decrease in sugar density may suggest an overall slight improvement. Additional interventions, either in the presence or absence of calorie labeling, should be considered for improving the nutritional quality of restaurant meals, especially in adolescents and children.
RESEARCH SNAPSHOT.
Research question: Does labeling fast food menus with calorie information affect the nutritional quality of customers’ meal purchases?
Key findings: This natural experiment used customer receipt data from 2,883 adults, 2,131 adolescents, and 433 parents/guardians of children at McDonald’s (which implemented labeling in 2012) and control fast food restaurants. Calorie labeling at McDonald’s was associated with some slight positive changes in nutritional quality of adult purchases but no positive changes in nutritional quality of adolescent or children’s purchases compared to control restaurants. More interventions are needed to improve diet quality in restaurants, especially in adolescents and children.
Acknowledgements:
The authors would like to thank the participants for their valuable contributions to this study.
Funding: This study was supported by a career development award from the National Heart, Lung, and Blood Institute (grant number K23HL111211, PI: Block), a grant from the Healthy Eating Research program of the Robert Wood Johnson Foundation to JPB, and an unrestricted grant from the McLaughlin Family Foundation. JP is supported by T32HL098048. Funding sources had no role in the study design, data collection and analysis, interpretation of results, writing of the manuscript, or decision to submit the article for publication.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflicts of interest: There are no conflicts of interest to report.
REFERENCES
- 1.Food and Drug Administration. Food Labeling; Nutrition Labeling of Standard Menu Items in Restaurants and Similar Retail Food Establishments; Extension of Compliance Date. In: Food and Drug Administration, ed. Vol 82 FR 20825: Federal Register; 2017:20825–20829. [PubMed] [Google Scholar]
- 2.Bleich SN, Economos CD, Spiker ML, et al. A Systematic Review of Calorie Labeling and Modified Calorie Labeling Interventions: Impact on Consumer and Restaurant Behavior. Obesity (Silver Spring). 2017;25(12):2018–2044. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Bollinger B, Leslie P, Sorensen A. Calorie Posting in Chain Restaurants. Am Econ J: Economic Policy. 2011;3(1):91–128. [Google Scholar]
- 4.Petimar J, Zhang F, Cleveland LP, et al. Estimating the effect of calorie menu labeling on calories purchased in a large restaurant franchise in the southern United States: quasi-experimental study. BMJ. 2019;367:l5837. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Hruby A, Manson JE, Qi L, et al. Determinants and Consequences of Obesity. Am J Public Health. 2016;106(9):1656–1662. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Elbel B, Kersh R, Brescoll VL, Dixon LB. Calorie labeling and food choices: a first look at the effects on low-income people in New York City. Health Affairs. 2009;28(6):w1110–1121. [DOI] [PubMed] [Google Scholar]
- 7.Cantor J, Torres A, Abrams C, Elbel B. Five Years Later: Awareness Of New York City's Calorie Labels Declined, With No Changes In Calories Purchased. Health Affairs. 2015;34(11):1893–1900. [DOI] [PubMed] [Google Scholar]
- 8.Liu J, Rehm CD, Onopa J, Mozaffarian D. Trends in Diet Quality Among Youth in the United States, 1999–2016. JAMA. 2020;323(12):1161–1174. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Vikraman S, Fryar CD, Ogden CL. Caloric Intake From Fast Food Among Children and Adolescents in the United States, 2011–2012. NCHS Data Brief. 2015(213):1–8. [PubMed] [Google Scholar]
- 10.Petimar J, Ramirez M, Rifas-Shiman SL, et al. Evaluation of the impact of calorie labeling on McDonald's restaurant menus: a natural experiment. Int J Behav Nutr Phys Act. 2019;16(1):99. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Willett WC, Stampfer MJ. Current evidence on healthy eating. Annu Rev Public Health. 2013;34:77–95. [DOI] [PubMed] [Google Scholar]
- 12.Block JP, Condon SK, Kleinman K, et al. Consumers' estimation of calorie content at fast food restaurants: cross sectional observational study. BMJ. 2013;346:f2907. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Stuart EA, Huskamp HA, Duckworth K, et al. Using propensity scores in difference-in-differences models to estimate the effects of a policy change. Health Serv Outcomes Res Methodol. 2014;14(4):166–182. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Yang Q, Zhang Z, Gregg EW, Flanders WD, Merritt R, Hu FB. Added sugar intake and cardiovascular diseases mortality among US adults. JAMA Intern Med. 2014;174(4):516–524. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Zong G, Li Y, Wanders AJ, et al. Intake of individual saturated fatty acids and risk of coronary heart disease in US men and women: two prospective longitudinal cohort studies. BMJ. 2016;355:i5796. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Imamura F, O'Connor L, Ye Z, et al. Consumption of sugar sweetened beverages, artificially sweetened beverages, and fruit juice and incidence of type 2 diabetes: systematic review, meta-analysis, and estimation of population attributable fraction. BMJ. 2015;351:h3576. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Guasch-Ferre M, Becerra-Tomas N, Ruiz-Canela M, et al. Total and subtypes of dietary fat intake and risk of type 2 diabetes mellitus in the Prevencion con Dieta Mediterranea (PREDIMED) study. Am J Clin Nutr. 2017;105(3):723–735. [DOI] [PubMed] [Google Scholar]
- 18.Veronese N, Solmi M, Caruso MG, et al. Dietary fiber and health outcomes: an umbrella review of systematic reviews and meta-analyses. Am J Clin Nutr. 2018;107(3):436–444. [DOI] [PubMed] [Google Scholar]
- 19.U.S. Department of Health and Human Services and U.S. Department of Agriculture. 2015–2020 Dietary Guidelines for Americans. 2015.
- 20.Fryer CD, Ervin RB. Caloric intake from fast food among adults: United States, 2007–2010. NCHS Data Brief. 2013(114):1–8. [PubMed] [Google Scholar]
- 21.Moran AJ, Block JP, Goshev SG, Bleich SN, Roberto CA. Trends in Nutrient Content of Children's Menu Items in U.S. Chain Restaurants. Am J Prev Med. 2017;52(3):284–291. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Auchincloss AH, Mallya GG, Leonberg BL, Ricchezza A, Glanz K, Schwarz DF. Customer responses to mandatory menu labeling at full-service restaurants. Am J Prev Med. 2013;45(6):710–719. [DOI] [PubMed] [Google Scholar]
- 23.Pulos E, Leng K. Evaluation of a voluntary menu-labeling program in full-service restaurants. Am J Public Health. 2010;100(6):1035–1039. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Willett W Nutritional Epidemiology. 3 ed. New York: Oxford University Press; 2012. [Google Scholar]