|Home | About | Journals | Submit | Contact Us | Français|
The accuracy of stated energy contents of reduced-energy restaurant foods and frozen meals purchased from supermarkets was evaluated. Measured energy values of 29 quick-serve and sit-down restaurant foods averaged 18% more than stated values, and measured energy values of 10 frozen meals purchased from supermarkets averaged 8% more than originally stated. These differences substantially exceeded laboratory measurement error but did not achieve statistical significance due to considerable variability in the degree of underreporting. Some individual restaurant items contained up to 200% of stated values and, in addition, free side dishes increased provided energy to an average of 245% of stated values for the entrees they accompanied. These findings suggest that stated energy contents of reduced-energy meals obtained from restaurants and supermarkets are not consistently accurate, and in this study averaged more than measured values, especially when free side dishes were taken into account. If widespread, this phenomenon could hamper efforts to self-monitor energy intake to control weight, and could also reduce the potential benefit of recent policy initiatives to disseminate information on food energy content at the point of purchase.
The prevalence of obesity has risen markedly in the past 30 years (1), and energy intake has also risen substantially during this period (2,3). Reducing energy intake is therefore a cornerstone of weight management (4,5), and self-monitoring of food intake is widely recommended to facilitate success (6,7). However, achieving energy intake goals through self-monitoring depends on the accuracy of energy information available for consumed foods.
All foods, including those prepared at home, may potentially introduce error in self-monitoring of energy intake, but ready-prepared foods purchased from restaurants and supermarkets may be an area of particular concern. Consumption of these foods has increased in recent years (8,9), and meals purchased away from home are reported to contain more energy than home-prepared foods (10–14). Furthermore, information on the energy content of restaurant foods is provided without any required verification or oversight (15,16). Two recent media reports (17,18) suggested that the true energy contents of restaurant meals may be considerably more than the amounts of energy stated by the vendors, a suggestion that has important implications but is at odds with the results of a previous study (19). Inaccuracies in reported food energy content, if replicated, could contribute to inadvertent overeating, hamper efforts to use self-monitoring for weight control, and reduce the usefulness of recent government initiatives to make information on the energy content of restaurant foods available at the point of purchase (20–22).
A pilot study was therefore conducted to assess the accuracy of reported energy contents of restaurant and supermarket foods with reduced energy suitable for weight control.
This study involved measurement of the energy content of 39 commercially prepared restaurant foods and supermarket frozen convenience meals obtained in the Boston, MA, area, and comparison of measured values with nutrition information stated by the vendor or manufacturer. The restaurant chains included in the study were selected as a convenience sample of quick-serve and sit-down restaurant chains with broad distribution throughout the United States who provided information on nutrient contents (or reliable information was available from other Web sites). Because the goal of the study was to examine the accuracy of stated energy contents of foods apparently favorable for weight control, specific restaurant menu items were chosen based on three criteria: having <500 kcal, being typical American foods, and having among the lowest stated energy contents on the menu. Supermarket purchases were focused on frozen complete meals that would be alternative choices to eating out. All items were purchased between December 2007 and June 2008. This study was deemed exempt under federal regulation 45 CFR §46.101(b).
One unit of each food item was purchased and transported upright in the original container to the research center. Whole food items were weighed to determine portion size, homogenized, freeze dried (Virtis Benchmark 1000 Lyophilizer, Virtis Co, Gardiner, NY), ground to a fine powder, and then analyzed in duplicate for gross energy using an Isoperibol Bomb Calorimeter (Parr model 1261, Parr Instrument Co, Moline, IL). Weights obtained throughout the sample preparation were used to adjust for small food losses at different stages of processing. Benzoic acid standards (Benzoic Acid 1g pellets, Parr Instrument Co, Moline, IL) were analyzed and values were within 1% of the known heat of combustion for benzoic acid. The gross energy content of each food was calculated by multiplying the average kilocalories per gram of dry solids by the total weight of dry solids.
Two test “meals” were created out of basic food items with consistent composition (white wheat flour, granulated white sugar, corn oil, and nonfat dry instant milk) at three energy levels and with macronutrient balances spanning the range of test foods. The test meals were processed for energy content as described above and analyzed for energy content. Mean measured energy values were −1.9% ± 0.3% of values calculated from stated compositions of the individual ingredients, demonstrating the accuracy and precision of the methodology used for direct measurements of food energy content.
Bomb calorimetry directly measures the heat of combustion of a food and thus gives values for gross energy (23), whereas stated food values are metabolizable energy estimations (ie, gross energy corrected for assumed obligatory energy losses before energy is available to the body). Therefore, it was first confirmed that stated metabolizable energy contents in study foods could be reliably calculated from stated macronutrients, and then gross energy was calculated for each food item from stated macronutrients using the US Department of Agriculture multiplicative factors appropriate for mixed foods (24). This conversion introduced no error to the comparison and allowed measured values to be compared directly with values calculated from stated macronutrients (henceforth called stated values). One restaurant reported energy and all macronutrients except protein, which was calculated by subtraction. Two other restaurants reported only energy, fat, and fiber. For these two restaurants the carbohydrate content from www.calorieking.com was adopted and protein content was calculated by subtraction. This Web site employs registered dietitians who perform independent calculations, including using the US Department of Agriculture database (25), and thus the information was deemed acceptable. Gross energy values are henceforth called energy in the text for simplicity.
For foods that provided a portion weight on the restaurant Web sites or food package, stated values were compared to actual served portion weight as measured in the laboratory. For some analyses, measured energy values were adjusted for portion size discrepancies by multiplying the measured gross energy value by the percent difference in portion size.
Exploratory quantitative variables (eg, cost of food item, cost per kilocalorie, and restaurant size) and categorical variables (eg, whether the item was a free side dish or entrée, restaurant type was quick-serve or sit-down, or whether the meal would typically be eaten for breakfast or lunch/dinner) were assigned to food items. Cost of food was the menu price at the time and place of purchase, and restaurant size was the total number of locations as indicated in the annual report for each company. Free side dishes were defined as food items accompanying the entrée that were not specifically ordered. Menus typically listed these items as accompaniments with the entrée, but the Web site energy information for side dishes was listed separately from the entrée, or in one case, not at all. Sit-down restaurants were defined as those where table service from wait staff was available. Restaurants not meeting this criterion were considered quick-serve.
Although formal statistical methods do not apply to convenience samples, standard statistical tests were employed to summarize the data for exploratory purposes and to suggest directions for future study. Differences in energy content (measured vs stated) were therefore compared by using two-sided t tests of whether the observed mean was within sampling variability of zero. Multifactor analysis of variance was used to examine potential predictors of the percent difference between stated and measured energy for restaurant foods. Statistical significance was set at P <0.05 (two sided). Statistical analyses were performed by using SAS version 9.1 (2002, SAS Institute Inc, Cary, NC).
The energy contents of individual foods are given in the Table, and Figure 1 illustrates percent differences between measured and stated values. On average, restaurant foods contained 18% more energy than stated; however, there was substantial variability in the difference between measured and stated values, and some foods contained twice as much energy as stated. The measured energy content of supermarket-purchased meals was also greater than stated values, by 8%. There was no statistically significant difference between measured and stated values in either restaurant foods (P = 0.12) or supermarket meals (P = 0.12), and no significant difference between the mean accuracy of restaurant vs supermarket foods (P = 0.64). However, the difference between the standard deviation of the energy discrepancies in restaurant vs supermarket foods approached significance (P = 0.06) and became significant once the energy discrepancies were adjusted for discrepancies of portion size (P = 0.05). Thus, there is more uncertainty regarding stated energy values in restaurant foods than in supermarket packaged foods. Five restaurants provided side dishes at no extra cost. The mean additional energy provided by these items was 471 ± 167 kcal, which was greater than the 443 kcal mean value for the entrées they accompanied. Portion size discrepancies (measured minus stated values) were a significant predictor of percent energy difference between measured and stated energy contents values when added to the model predicting provided food energy (P <0.0001), indicating that greater portions than stated did contribute to the energy discrepancy.
There was no statistically significant association between differences in measured vs stated energy based on whether the item was a side dish or an entrée, type of restaurant (quick-serve or sit-down), meal at which the food is usually consumed (breakfast or lunch/dinner), restaurant size (number of franchises), cost per unit calculated energy content, or where the meal was purchased (restaurant or supermarket). It should also be noted that, since single samples of each food were obtained, the variability in results pertains to the general sample of foods studied, and individual results cannot be attributed to specific restaurants and brands of foods.
These results indicate that, in contrast to two recent reports in the media (17,18), restaurant meals and prepared meals purchased in US supermarkets do not typically contain substantially more energy than stated. However, measured energy values did average 18% higher in restaurant foods and 8% higher in supermarket meals than stated. The underreporting of energy by restaurants and food manufacturers notwithstanding, the majority of foods tested were not out of compliance with US Food and Drug Administration regulations because most fell within the 20% overage the Administration allows for packaged food (no ceiling of overage is specified for restaurant foods). In relation to this observation, it should be noted that, as outlined in Figure 2, the US Food and Drug Administration considers noncompliance in packaged foods to include food energy (average of 12 samples) in excess of 120% of stated energy or <99% of stated weight (average of 48 samples) (26,27), which because of greater leniency in standards on the side of overprovision compared to underprovision, may contribute to provided energy values being greater than measured. It should also be noted that three individual supermarket-purchased meals and seven restaurant foods did contain up to twice the stated energy, so some individual discrepancies were substantial.
The extra mean measured energy in this study compared to stated information may cause substantial weight gain over time. For example, positive energy balance of only 5% per day for an individual requiring 2,000 kcal/day could lead to a 10-lb weight gain in a single year. It is also notable that free side dishes on average contained more energy than the entrées alone. It is unclear whether consumers are aware of how much energy free side dishes contain, and providing more accessible information on meal energy contents both with and without side dishes could help increase attention to the potential of these casual food items to more than double meal energy intake.
Mean measured energy contents of reduced-energy restaurant and supermarket meals in this pilot study exceeded vendor-stated amounts by substantially more than could be accounted for by laboratory measurement error. Although the discrepancies were within acceptable limits based on federal regulations for most packaged and restaurant foods (which are not subject to these federal regulations) some restaurant foods did have measured energy contents that were double those stated by the restaurant, and free side dishes contained more energy on average than the entrées they accompanied. On an individual level, discrepancies of this magnitude, if widespread, are likely to substantially hamper efforts to control weight by individuals self-monitoring their energy intake. On a public scale, the emerging policy initiatives on requiring energy information at the point of purchase may not translate into improved dietary intake if foods typically contain more energy than stated. Based on these findings, registered dietitians can advise consumers about the wide variability in accuracy of stated energy contents for prepared reduced-energy foods. Approaches to improving the accuracy of stated energy information may include increased attention to quality control in food preparation. Additional measures may also be needed such as improved federal and state regulations and a monitoring system to ensure compliance.
The authors thank Paul Fuss for providing training in bomb calorimetry techniques.
FUNDING/SUPPORT: This work was supported by National Institutes of Health grant no. HL069772-06, and the US Department of Agriculture under agreement no. 58-1950-4-401. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of the US Department of Agriculture.
STATEMENT OF POTENTIAL CONFLICT OF INTEREST:
No potential conflict of interest was reported by the authors.