| dc.description.abstract | BACKGROUND Diet is an important risk factor for chronic disease and obesity. The growing proportion of dietary intake from food consumed at restaurants and fast food outlets is an important contributing factor to the overall poor diet quality in Canada. Currently, Canadians receive little or no nutrition information when they are making purchases in away-from-home settings. Providing nutrition information on menus is a population-level intervention aimed at improving the dietary choices of consumers in restaurant settings. To date, the evidence for menu labelling is considerably mixed. Although menu labelling appears to increase noticing and use of nutrition information in these settings, the impact of menu labelling on food choices is unclear.
In January, 2011, The Ottawa Hospital implemented a nutrition information program on digital menu boards at the Civic Campus cafeteria, providing information for calories, sodium, saturated fat and total fat for meals or food items on digital menu boards. The program was later implemented at the General campus cafeteria in September, 2012. The implementation provided an opportunity to conduct a quasi-experimental study of menu labelling in a naturalistic setting.
OBJECTIVES The objective of the current study is to examine the impact of displaying nutritional information on menu boards on consumer behaviour. Specifically, the study examined: 1) the impact of menu labelling on self-reported noticing and use of nutrition information in the cafeteria, 2) the impact of menu labelling on calories, sodium, saturated fat and total fat purchased, 3) the impact of menu labelling on consumers’ ability to estimate the calorie content of meals, and 4) how the impact of menu labelling may vary between population subgroups.
METHODS Data were collected from the Civic cafeteria (the “comparison” condition) and the General cafeteria (the “intervention” condition) at three time points: before and 3 months after implementation at the General cafeteria, using the Civic cafeteria as a comparison site, with a one-year follow up. Exit surveys were conducted with approximately 500 patrons at each site during data collection wave, for a total of 3,061 participants. Surveys were approximately 10 minutes in length, and examined food and drink selection and consumption, noticing and perceived influence of nutritional information, and perceived calorie content of meal items, as well as socio-demographics and nutrition-related attitudes and behaviours. Analyses were conducted using linear regression for continuous outcomes (self-reported nutrients purchased) and logistic regression for binary outcomes (self-reported noticing and use of nutrition information, correct calorie estimation within 50 kcal of objective amount), and included an interaction term between wave and site to examine the impact of the intervention over time.
RESULTS In the first wave of data collection, there was significantly greater awareness and influence of menu labelling in the Civic cafeteria where nutrition information was presented on digital menu boards (75.1% noticed, 25.4% used), compared to the General cafeteria, which provided information inconsistently on paper signs throughout the cafeteria (31.8% noticed, 9.0% used). There were significant increases in the proportion of participants that noticed and used menu labelling at the intervention cafeteria immediately (3-months) after menu labelling was implemented (51.6% noticed, 14.2% used) and at the one-year follow up (51.5% noticed, 16.0% used), compared to the comparison cafeteria (p<0.05 for all). Across all waves and sites, approximately one-third of those who noticed nutrition information reported that they used that information to inform their food choice. Hospital staff, those who visited the cafeteria more often, younger participants, those with higher levels of education and income, those who reported using nutrition labels more frequently when shopping for food, and those who knew the estimated energy requirements for calorie intake were all more likely to notice menu labelling (p<0.05 for all). The only significant correlate of self-reported use of menu labelling to inform food choice was the frequency of using nutrition labels when shopping for food, whereby those who reported using this information more frequently were more likely to use menu labelling to inform their food choice. Among the entire sample, calories were the most commonly noticed nutrient (39.4% of the entire sample), and participants were equally likely to report purchasing less sodium (5.7%) and fewer calories (5.4%) in response to noticing the nutrition information.
Nutritional information matching self-reported meal items purchased was available for 2,781 participants. At the first wave of data collection, significantly fewer calories (-131 kcal), and less sodium (-323 mg), saturated fat (-2.4 g) and total fat (-9.7 g) were purchased at the Civic cafeteria which had implemented menu labelling as one aspect of an effort to improve the nutritional quality of food items, including reformulation of some food items (p<0.001 for all). From Wave 1 to Wave 2, there were more favourable changes in the calories, sodium, saturated fat and total fat that was purchased at the General cafeteria, which had implemented the digital menu boards with nutrition labelling, compared to the comparison cafeteria (p<0.05 for all). When the analysis was conducted from Wave 1 to Wave 3, the effect of menu labelling was significant for saturated fat and total fat (p<0.05 for both), but was not significant for calories or sodium. The findings for calories were similar when examined for food purchases only, and there were no significant differences in calories for drink purchases between sites or over time.
When participants were asked to estimate the calorie content of their meal, 12.7% of the sample correctly estimated the calorie content of their meal within ±50 kcal of the objective calorie content, while 23.4% of participants did not know the content and did not provide a guess for at least one item in their meal. Approximately the same proportion of participants overestimated (32.9%) and underestimated (29.4%) the energy content of their meal. There were no differences between sites or over time in the proportion that correctly estimated the calorie content; however, those who noticed menu labelling were more likely to correctly estimate the calorie content of their meal (p=0.03). There was also no significant difference in the absolute difference between actual and estimated calorie content between sites or over time. Overall, there were high levels of support for menu labelling among cafeteria patrons for both menu labelling in the hospital cafeterias (95%), as well as in chain restaurants more broadly (91%). When asked, participants mostly commonly wanted to see calories on menu boards (72%), followed by fat (55%) and sodium (49%).
CONCLUSIONS The current study was among the first quasi-experimental studies examining menu labelling in Canada, and the first such study in a naturalistic setting using longitudinal data over a longer period. Overall, the results suggest a modest, but positive, effect of menu labelling on the nutritional quality of items purchased, with little effect on estimation of the calorie content of the meal. The intervention had similar impact between population sub-groups. The results provide evidence to inform policy development in Canada as policymakers consider menu labelling regulations at the federal, provincial and municipal level. | en |
