Eating-Out: a Study of the Nutritional Quality of Canadian Chain Restaurant Foods and Interventions to Promote Healthy Eating

Eating-Out: A Study of the Nutritional Quality of Canadian Chain Restaurant Foods and Interventions to Promote Healthy Eating

Mary J. Scourboutakos

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Department of Nutritional Sciences University of Toronto

Mary J. Scourboutakos

Doctor of Philosophy

Department of Nutritional Sciences University of Toronto

2016 Abstract Canadians are increasingly eating outside-the-home. At the outset of this thesis there were no data on the nutritional quality of Canadian chain restaurant foods, the Sodium Working Group’s plan to monitor sodium reductions in the food supply was abandoned, and despite interest and numerous bills, there was no existing menu-labelling legislation in Canada.

The specific objectives of this thesis were to 1) investigate the nutritional quality of the Canadian chain restaurant food supply; 2) explore consumers’ use of menu-labelling; and 3) test the potential of alternative forms of labelling in non-chain restaurant settings.

Objective 1 was investigated by developing and analyzing a national database of over 9000 menu-items from Canadian fast-food and sit-down chain restaurants which was created in 2010. There was wide variation in calorie levels within each restaurant and food category; furthermore, portion size, as opposed to calorie density, was the most important driver of this variation. Sodium levels in menu items often exceed daily recommendations and despite reported efforts by the restaurant sector to improve, as of 2013, reductions were minimal.

Objective 2 used an online, national consumer survey to test three menu-labelling treatments (calories; calories and sodium; and calories, sodium and serving size labelling). The effect of labelling on consumer choice varied depending on the restaurant setting, however, overall, labelling sodium in addition to calories led consumers to choose meals with significantly less sodium. There was no additional benefit from adding serving size information.

Objective 3 was examined in a quasi-experimental, population-level nutrition labelling/education intervention study in a campus cafeteria. Results showed that this intervention could modestly increase fruit and vegetable consumption, and decrease sugar- sweetened beverage consumption among University students.

Overall, this thesis provides food supply and consumer data to inform public health policy debates around issues concerning food consumed outside-the-home.

iii

Dedication

This thesis is dedicated to my Mom, who told me what a PhD was when I was twelve, and then encouraged me to do one.

Acknowledgments

In the summer following my second year of undergrad, I wrote down an idea for a research project…

―test healthy take-out food‖

…little did I know that two years later, Dr. Mary L’Abbé would propose a somewhat similar idea, that set me on a journey that ended up being more than I ever dreamed it would be.

Throughout the journey, the famous quotation attributed to Sir Isaac Newton…

―If I have seen further than others it is because I have stood on the shoulders of giants‖

…has crossed my mind on numerous occasions, because there have been so many giants in my life, without whom, none of this would have been possible. Here I will attempt to acknowledge all of the special people on whose shoulders I stand…

First and foremost, thank you to Dr. Mary L’Abbé. I can’t imagine what my life would be like if I hadn’t met you. Words cannot express how fortunate I feel for having had the privilege of being your student. Thank you for teaching and mentoring me, for being patient with me (especially in the early days), for encouraging and inspiring me, for supporting me and for giving me opportunities that exceeded my wildest dreams. So much of what I have achieved, I owe to you.

Second, one of the greatest gifts I gained as part of this experience was my lab sisters, aka, the L’Abbé Lab girls, who journeyed with me. Each of you has brought so much to my life. I truly cherish the friendships we’ve formed and I feel so privileged to have shared this experience with you.

Specifically, I must recognize my ―big sisters‖ JoAnne Arcand, Teri Emrich, Christina Wong and Alyssa Schermel. Thank you for nurturing me when I was a naive undergraduate who was way too excited and probably asked you far too many questions. Words cannot express how lucky I feel to have benefited from all of your wisdom, insight, advice and guidance. You have truly been the best role models I could have asked for. You pushed me and inspired me to achieve more than I ever could have without you.

Additionally, thank you to the second generation of L’Abbé lab girls: Chelsea Murray, Mahsa Jessri, Mavra Ahmed, Jodi Bernstein, Sheida Noorhosseini, Marie-Eve Labonté, and Beatriz Franco Arellano. Thank you for listening to my never ending stories, my research rants, and for making Fitz 87 a very special place to be.

An immense thanks is owed to my thesis committee members. First, thank you to Dr. Thomas Wolever for bringing scientific rigor to my committee meetings and for pushing my thinking and

v my research to new levels. You have been an inspiration and I immensely appreciate the multi- year commitment you made to my academic development. Thank you to Dr. Catherine Mah for being a fantastic mentor right from the beginning. I truly appreciate the time you devoted to fostering my intellectual development. Thank you for challenging and encouraging me. I will never forget that you were a key agent in connecting me with the real policy world and so much more.

Thank you to my examiners, Dr. Loren Vanderlinen and Dr. Alison Duncan who generously agreed to play this role, challenged me with their insight and helped me produce an even better finished product.

Numerous professors have greatly shaped and contributed to this intellectual endeavour. Thank you to Dr. Paul Corey, for taking such a genuine interest in my work, for contributing your expertise so generously and for all the stimulating discussions about nutrition science. Dr. Gillian Einstein for pushing my thinking into new realms, for cheering me on, and for introducing me to the world of women’s health. Thank you to the late Dr. Sharon Parker who believed in me way back in my undergrad and who wrote countless reference letters on my behalf. Taking your fourth year course in my second year was an immense joy and the first taste that got me hooked on public health nutrition.

The administrators in the Department of Nutritional Sciences are like angels that take care of all the DNS students. A special thanks is owed to Louisa Kung, Emelia D’Souza, Patrice Lee and Lucile Lo.

I am indebted to numerous volunteers who have contributed to my research, notably, Zhila Semnani-Azad and Sahar Qassem who motivated me by enthusiastically entering data that led to some of the papers I am most proud of. As well, thank you to Sarah Murphy, Frank Mazza, and Andriana Chomka, who diligently spent many evenings with me, watching people make food selections.

A special thanks is owed to the Victoria College community, especially Chef Nathan Barrett and Bill McFadden, for letting me into your cafeteria, for being so generous with your time, and for supporting my research. A huge thanks is owed to former Vic President Paul Gooch who has been my guardian angel since undergrad. Thank you for reading my column and seeing potential in me way back. You have opened so many doors for me and I am eternally grateful for your kindness and constant support.

Throughout this journey, Massey College has been my home away from home. A huge thank you is owed to Former Master John Fraser for your wisdom, support, constant encouragement and ego pumping. But most of all, thank you for your dedication to creating such an inspiring and loving environment for graduate students. A special thanks to my Quadrangler mentors Elizabeth Wilson and David Scott, to Senior Fellows Dr. Aubie Angel for running the intellectually enriching Massey Grand Rounds which I learned so much from, and especially,

Michael Valpy for being a constant source of encouragement and nutritional experimentation. Most importantly, thank you to my Junior Fellow friends who have inspired me and with whom I have shared so much of this journey.

A special thanks is owed to the team at Toronto Public Health especially, Loren Vanderlinden and Dia Mamatis, also to MPP France Gelinas, who gave me the opportunity to use my research to inform policy.

Thank you to all of the journalists who reported on our work and helped our message reach so many people.

Thank you to all of the teachers I’ve had since Kindergarten who fostered my love of learning, to say the least. Especially Dr. Gabriel Ayyavoo, my high school science teacher, whose love for the science fair inspired me to pursue my first nutrition research project.

I owe a huge thank you to the staff, students and professors in the department of Nutritional Sciences. A special thanks goes to Beatrice Boucher for being my mentor, Julie Mason, who has helped and inspired me throughout the years, and to Katie Hopperton who co-founded the DNS journal club with me, and whose insight I learned from at each of our meetings.

Thank you to the friends that have supported me. Especially Jemy Joseph who has been a friend, mentor and constant source of inspiration for the past 13 years. I owe so much to you. Also, thank you to Nan Zheng for support and encouragement and to my PhD piano teachers who gave me a break from research, especially, Midori Koga, Megumi Okamoto, Melody Chan, Matthew Li, and Selena Shen.

Thank you to my extended family, especially my Grandma who read all of my papers and even came to one of my lectures!

I owe everything to my Mom, Dad and brother Pete. Thank you for listening to me talk about my research 24/7. Thank you to my Mom for reading every proposal, paper, manuscript, blog post, etc. over the years. Thank you for encouraging my ideas, cooking for me, listening to me and every other little thing that would take innumerable pages to document. Words cannot express how grateful I am for everything you had done for me. I owe my entire success to you.

And finally, thank you to God for the life I’ve been gifted with.

Funding Acknowledgement

I have been extremely fortunate to have received numerous scholarships throughout the course of my PhD. The Canadian Institutes of Health Research provided funding through the generous Vanier Canada Graduate Scholarship that allowed me to spend more time digging deeper and learning more about nutrition. I was also the recipient of two Ontario Graduate Scholarships. My sincerest thanks goes out to Barb Riley (and all of the founders of PICDP) as well as Rob

vii

Schwartz, whose training programs in Population Intervention for Chronic Disease Prevention and Public Health Policy, respectively, not only provided financial support, but were also key elements of my training and academic development. I would like to thank you to the American Society for Nutrition for various travel awards and honorariums.

viii

Table of Contents

Acknowledgments...... iv Table of Contents ...... ix List of Tables ...... xv List of Figures ...... xvii List of Appendices ...... xix List of Abbreviations ...... xx Literature Review ...... 1 1.1. Diet-related chronic disease globally and in Canada ...... 1 1.2. Eating-Out ...... 1 1.2.1. The prevalence and trend towards eating-out ...... 1 1.2.2. Eating-out is associated with a lower diet quality and larger portion sizes ...... 2 1.2.3. Calorie and sodium levels of meals purchased from fast-food restaurants ...... 4 1.2.4. The association between eating-out and obesity ...... 4 1.3. Nutritional quality of restaurant foods ...... 5 1.3.1. Fast-food in Australia ...... 5 1.3.2. Restaurant Kids’ meals in the United States ...... 5 1.3.3. Restaurant foods in Canada ...... 6 1.4. Menu-Labelling ...... 6 1.4.1. The Rationale for Calorie-Labelling ...... 7 1.4.2. Menu-Labelling in the Canadian Context ...... 7 1.4.2.1. The history of menu-labelling in Canada and Ontario ...... 7 1.4.2.2. The CRFA voluntary nutrition disclosure program for restaurants ...... 8 1.4.2.3. The Informed Dining Program ...... 9 1.4.2.4. Menu-labelling research in Canada ...... 9 1.4.3. Menu-labelling in the United States ...... 10 1.4.3.1. The history of menu-labelling in the United States ...... 10 1.4.3.2. Natural experiments evaluating calorie labelling in New York City ...... 11 1.4.3.3. Long-term evaluation of menu-labelling in New York ...... 12

1.4.4. Experimental tests of calorie-labelling in non-restaurant settings ...... 13 1.4.5. Calorie-labelling in independent (non-chain) restaurants and cafeterias ...... 13 1.5. Alternative forms of nutrition labelling ...... 14 1.5.1. Physical Activity Calorie Equivalent (PACE) Labelling...... 14 1.5.2. Non-Numerical forms of labelling/education ...... 15 1.6. Sodium-related public health initiatives in restaurants ...... 16 1.6.1. Introduction to Sodium as a public health issue ...... 16 1.6.2. The Canadian Sodium Working Group ...... 16 1.6.3. The United States’ National Salt Reduction Initiative ...... 18 Chapter 2 ...... 20 Objectives and Hypothesis of Thesis ...... 20 2.1. Objectives ...... 20 2.2. Specific hypotheses ...... 20 2.3. Preview of chapters ...... 21 2.4. Student Contribution ...... 22 Chapter 3 ...... 23 Restaurant menus – calories, caloric density, and serving size (AJPM, 2012) ...... 23 3.1. Abstract ...... 23 3.2. Background ...... 24 3.3. Methods ...... 25 3.3.1. Data Collection ...... 25 3.3.2. Construction of the Database ...... 25 3.3.3. Inclusion Criteria ...... 26 3.3.4. Exclusion Criteria ...... 26 3.3.5. Statistical Analysis ...... 26 3.4. Results ...... 27 3.5. Discussion ...... 35 3.5.1. Study Limitations ...... 36 3.6. Conclusion ...... 37 Chapter 4 ...... 38 Sodium Levels in Canadian Sit-Down and Fast-Food Restaurants (CJPH, 2013) ...... 38

4.1. Abstract ...... 38 4.2. Introduction ...... 39 4.3. Methods ...... 40 4.3.1. Database construction ...... 40 4.3.2. Exclusion and Inclusion Criteria ...... 41 4.3.3. NSRI targets ...... 41 4.4.4. Data Analysis ...... 42 4.5. Results ...... 42 4.5.1. Sodium in single menu items compared to recommended daily intake levels ...... 42 4.5.2. Sodium in children’s meal items compared to recommended daily intake levels ...... 43 4.7. Conclusion ...... 53 Chapter 5 ...... 54 A longitudinal study of changes in sodium levels in a sample of Canadian chain restaurant foods from 2010 to 2013 (CMAJ Open, 2014) ...... 54 5.1. Abstract ...... 54 5.2. Background ...... 55 5.3. Methodology ...... 55 5.3.1. Restaurants Included in the Study ...... 56 5.3.2. Database construction ...... 60 5.3.3. Statistical Analysis ...... 60 5.4. Results ...... 61 5.4.1. Overall change in sodium levels ...... 61 5.4.2. Changes in sodium by food category ...... 61 5.4.3. Changes in sodium by restaurant ...... 61 5.4.4. Proportion of foods exceeding sodium’s recommended daily intake levels ...... 62 5.4.5. Sodium levels in newly reported, discontinued and persisting foods ...... 62 5.4.6. Changes in serving size and calories ...... 62 5.5 Interpretation ...... 70 5.5.1. Main Findings ...... 70 5.5.2. Comparison with other studies ...... 70 5.5.3. Limitations ...... 71

5.6. Conclusion ...... 72 5.7. Acknowledgements ...... 72 Chapter 6 ...... 73 Restaurant Menu-Labelling – Is it worth adding sodium to the label? (CJPH, 2014) ...... 73 6.1. Abstract ...... 73 6.2. Introduction ...... 74 6.3. Subjects and Methods...... 75 6.3.1. Participants ...... 75 6.3.2. Experimental Design and Survey Structure ...... 75 6.3.4. Treatments/different types of menu-labelling that were tested ...... 77 6.3.5. Data Analysis ...... 77 6.4. Results ...... 78 6.4.1. Participants ...... 78 6.4.2. Proportion of consumers who changed their order after seeing nutrition information 78 6.4.3. Sodium level of meals ordered before, versus after seeing labelled menus (all consumers) ...... 78 6.4.4. Sodium level of meals ordered before versus after seeing labelled menus (sub-set of consumers who changed their order) ...... 79 6.4.5. Effect of serving size information on the calorie and sodium density of consumer’s choices ...... 79 6.4.6. Consumer’s rationale for why the information influenced or did not influence their order ...... 79 6.4.7. Effect of demographics on whether or not the nutrition information influenced consumer’s decisions ...... 80 6.5. Discussion ...... 88 6.5.1. Strengths ...... 89 6.5.2. Weaknesses ...... 89 6.6. Conclusion ...... 90 Chapter 7 ...... 91 Does a Healthy Eating Intervention in a Buffet-Style University Dining Hall Change Students’ Food and Beverage Choices? ...... 91 7.1. Abstract ...... 91

xii

7.2. Introduction ...... 92 7.3. Methodology ...... 93 7.3.1. Study Setting...... 93 7.3.2. The Intervention ...... 94 7.3.4. Fruit and Vegetable Education ...... 95 7.3.5. Intervention Outcomes ...... 96 7.3.6. Data Collection ...... 96 7.3.6.1. Direct-measured Observational Data Collection ...... 96 7.3.6.2. Cafeteria Inventory Data Collection ...... 97 7.3.7. Analysis ...... 97 7.4. Results ...... 98 7.4.1. Beverages (direct observational collection) ...... 98 7.4.2. Fruits and Vegetables (direct observational collection) ...... 99 7.4.3. Inventory Data Results ...... 99 7.5.1. Limitations ...... 109 7.5.2. Strengths ...... 109 7.6. Conclusion ...... 110 Chapter 8 ...... 111 Overall Discussion ...... 111 8.1. The restaurant food supply ...... 111 8.1.2. Calories ...... 111 8.1.3. Sodium ...... 112 8.1.3.1. The Challenge of Reducing Sodium ...... 113 8.1.4. The big picture of restaurant nutrition…beyond calories and sodium ...... 114 8.2. Menu-Labelling ...... 115 8.2.1. Serving size labelling...... 115 8.2.2. Sodium labelling ...... 116 8.2.2.1. Sodium warning labels ...... 117 8.2.3. The gap in current evaluations of calorie labelling in New York City ...... 118 8.3. Labelling/education in a cafeteria setting...... 119 8.4. Future Directions ...... 121

xiii

Chapter 9 ...... 125 Summary of Key Findings ...... 125 Chapter 10 ...... 127 Conclusions ...... 127 References ...... 128 APPENDIX A ...... 142 Restaurant meals - Almost a full day’s worth of calories, fats and sodium (JAMA Intern Med, 2013)...... 142 APPENDIX B ...... 147 Sugar levels in kids’ meals from Canadian chain restaurants (Prev Med Reports, 2015) ...... 147 Supplementary tables from ―Restaurant Menus – Calories, Caloric Density and Serving Size‖ ...... 154 APPENDIX D ...... 157 Supplementary tables from ―Changes in sodium levels in chain restaurant foods‖ ...... 157 APPENDIX E ...... 163 Supplementary Files from ―Restaurant menu-labelling: Is it worth adding sodium to the label‖ ...... 163 APPENDIX F...... 170 Supplementary tables/figures from ―Does a Healthy Eating Intervention Change Students’ Food and Beverage Choices‖ ...... 170

xiv

List of Tables Chapter 3

Table 3.1. Comparison of serving size, calories per serving and caloric density in sit- down an fast-food restaurants

Chapter 4

Table 4.1. Sodium levels in Canadian sit-down restaurant menu items compared to the daily dietary reference intake (DRI) recommendations

Table 4.2. Sodium levels in Canadian fast-food restaurant menu items compared to the daily dietary reference intake (DRI) recommendations

Table 4.3. Sodium levels in Canadian fast-food and sit-down restaurant side dishes compared to the daily dietary reference intake (DRI) recommendations

Table 4.4. Sodium levels in children's meal menu items and side dishes from Canadian fast-food and sit-down restaurants compared to the daily dietary reference intake (DRI) recommendations for children aged four to eight

Table 4.5. Canadian and US restaurant sodium levels in comparison to the US National Salt Reduction Initiative (NSRI) Targets

Chapter 5

Table 5.1. Characteristics of the 61 restaurants included in the study

Table 5.2. Descriptive statistics for sodium levels in 2010 and 2013 and the average change over time in 2198 Canadian chain restaurant foods

Table 5.3. Changes in sodium levels from 2010 to 2013 within each food category (n=1580)

Table 5.4. Percentage of foods (n=1603) at each restaurant whose sodium level increased or decreased, and the average percent change

Chapter 6

Table 6.1. Demographic characteristics of the panelits in the study

Table 6.2. Comparing the calorie and sodium content of meals ordered by panelists who saw calorie or calorie and sodium labelling

Table 6.3. Comparing the calorie density ordered after seeing menu-labelling among those who opted to change their order

Table 6.4. Key themes identified in open-ended question that asked panelists to explain why the nutrition information on the menu: influenced their decision, somewhat influenced their decision, or did not influence their decision

Table 6.5. Odds ratios classified by restaurant and demographic predictor of who uses menu-labelling

Chapter 7

Table 7.1. Calorie content of beverages at the cafeteria and estimated minutes of jogging using in the Physical Activity Calorie Equivalent (PACE)* labelling

Table 7.2. Sample characteristics at baseline including visits to the beverage and fruit station and frequency of beverage and fruit consumption

Table 7.3. Inventory data illustrating the weekly average number of litres of each beverage taken before and after the intervention

xvi

List of Figures Chapter 3

Figure 3.1. Median and range of calories per serving in menu items (including main items and side dishes, excluding desserts) from sit-down restaurants

Figure 3.2a. Median and range of calories in different food categories from sit-down restaurants arranged in increasing order by median calorie intake

Figure 3.2b. Median and range of calories in different food categories from fast-food restaurants arranged in increasing order by median calorie intake

Figure 3.2c. Median and range of calories in different side-dish food categories, arranged in increasing order by median calorie intake

Figure 3.3. Mean serving size and caloric density of meal items and side dishes grouped according to 100 calorie intervals

Chapter 5

Figure 5.1. Percent change in sodium levels (per serving) in Canadian chain restaurant foods (n=2198) from 2010 to 2013

Figure 5.2. Percentage of entrees (n=1004) exceeding the daily recommended Adequate Intake level (1500 mg) and Tolerable Upper Intake level (2300 mg) in 2010 and 2013

Chapter 6

Figure 6.1. Proportion of panelists in each restaurant who changed their order after seeing menu-labelling, and proportion of panelists from each treatment who changed their order

xvii

Chapter 7

Figure 7.1. Images of the beverage, fruit and vegetable labelling/education intervention

Figure 7.2. Proportion of students who chose each beverage before and after the intervention

Figure 7.3. Proportion of students who visited the fruit station and vegetable bar before and after the invention

Figure 7.4. Inventory data trends in the cases of fruit ordered each month

Chapter 8

Figure 8.1. New York City’s sodium warning label for restaurant menus

xviii

List of Appendices

Appendix A Additional manuscript: Restaurant meals – Almost a full day’s worth of calories, fats and sodium

Appendix B Additional manuscript: Added sugars in kids’ meals from chain restaurants

Appendix C Supplementary materials accompanying chapter 3: ―Restaurant menus – calories, caloric density and serving size‖

Appendix D Supplementary materials accompanying chapter 5: ―Changes in sodium levels in Canadian chain restaurant foods‖

Appendix E Supplementary materials accompanying chapter 6: ―Restaurant menu-labelling: Is it worth adding sodium to the label‖

Appendix F Supplementary materials accompanying chapter 7: ―Does a healthy eating intervention in a buffet-style University dining hall change students’ food and beverage choices?‖

xix

List of Abbreviations

AI Adequate intake level

AMA American Medical Association

ANOVA Analysis of Variance

BMI Body Mass Index

CCHS Canadian Consumer Health Survey

CCM Canadian Consumer Monitor

CI Confidence Interval

CRFA Canadian Restaurant and Foodservice Association

DV Daily Value

FDA Food and Drug Administration

FFR Fast-food restaurant

GRAS Generally recognized as safe

HDL High density lipoprotein

IOM Institute of Medicine

KCAL Calories

NGO Non-Governmental Organization

NSRI National Sodium Reduction Initiative

NHANES National Health and Nutrition Examination Survey

OMA Ontario Medical Association

SAS Statistical Application Software

SD Standard Deviation

SDR Sit-down restaurant

SE Standard Error

SWG Sodium Working Group

UL Upper tolerable intake level

USDA United States Department of Agriculture

WHO World Health Organization

xxi

Chapter 1 Literature Review

1.1. Diet-related chronic disease globally and in Canada

Chronic, non-communicable diseases, such as obesity, diabetes, cardiovascular disease, cancer, and osteoporosis, have risen dramatically worldwide.1 Currently, four chronic diseases – cardiovascular disease, diabetes, cancer, and chronic respiratory diseases—account for 63% of global deaths annually.2

The most common factors contributing to chronic disease include tobacco use, physical inactivity and unhealthy diet.2 As a result, to decrease risk for chronic disease, the World Health Organization has recommended that dietary intakes of saturated fat, trans fat, cholesterol, sugar, sodium and total calories should be limited.1 Similarly, Health Canada, via the 2007 food guide, Eating Well with Canada’s Food Guide, recommends that Canadians should limit intakes of foods and beverages high in calories, fat, trans fat, sugar and salt.3

In Canada, the obesity rate has tripled since 1985,4, 5 and presently, one out of every four Canadian adults are obese.6 It has been estimated that 60 to 100% of weight gain from the late 1970s to the 2000s was due to dietary excess.7 Many of the leading chronic diseases in Canada, including cancer, heart disease, stroke and diabetes, have also been linked with diet.8 Furthermore, 51.6% of Canadians over the age of twenty have one chronic disease, while 14.8% have two chronic diseases.9 Treatment for chronic disease consumes 67% of all direct health care costs, and costs the Canadian economy $190 billion annually.10 Moreover, health expenditures to treat chronic diseases are rising faster than Canada’s economic growth rate.10

1.2. Eating-Out

1.2.1. The prevalence and trend towards eating-out

Alongside the increases in diet-related disease, there have been simultaneous changes in where people eat. Longitudinal studies in the United States have shown that between 1977-78 and 1994-96 the total calories consumed from foods outside-the-home almost doubled, from

18% to 32% due to the increased frequency of eating-out.11 While in 1953, Americans spent twice as much of their food dollars on food prepared at home compared to food prepared outside- the-home,12 today, nearly half (43.1%) of every food dollar in the United States is spent on food consumed outside-the-home.13

Longitudinal data on the prevalence of eating-out in Canada do not exist. However, cross-sectional data from the Canadian Consumer Health Survey (CCHS) cycle 2.2 conducted in 2004, showed that 16% of food consumed daily is from restaurants or fast-food outlets.14 Furthermore, on any given day 25% of the population ate something prepared in a fast-food outlet and another 21% ate something prepared in a sit-down restaurant, cafeteria or other food venue.15, 16 As a result, one in every ten meal occasions are sourced from restaurants and each day almost half of the population consumes something prepared outside-the-home.15, 16 According to Restaurants Canada (a Canadian not-for-profit that represents the restaurant industry), 80,800 commercial foodservice units across Canada are visited by 17.7 million people, daily.17 Furthermore, it has been estimated that 28% of every food dollar in Canada is spent on foods consumed outside-the-home.18

Despite the lack of longitudinal data on the prevalence of eating-out in Canada, research has shown that between 1976 and 2003, the per capita estimated availability of energy in Canada increased by 18% (417 calories per day) and this increase may be partially due to calories consumed outside-the-home.19 This net increase in energy availability is attributed to foods/ingredients that are commonly found in fast-food/convenience foods including: fats and oils (common in deep-fried foods and salad dressings), wheat flour, soft drinks, shortening (formerly used by the fast-food industry for frying), rice, chicken (a popular fast-food item and one of the largest fast-food subsectors) 20 and cheese (a primary ingredient in pizza, one of the largest fast-food subsectors).20

1.2.2. Eating-out is associated with a lower diet quality and larger portion sizes

Numerous dietary surveys have found an association between eating-out and poorer diet quality. For example, cross-sectional data from the United States’ Continuing Survey of Food

Intakes by Individuals which contained two consecutive 24-hour food recalls from 17,370 American adults and children between 1994 to 1998 illustrated that adults consuming foods outside-the-home had higher intakes of total calories, fat, saturated fat, sodium and carbonated beverages, while having lower intakes of vitamin A, vitamin C, milk, fruits and vegetables.21 Meanwhile, children who ate fast-food on one of the two survey days had higher intakes of total calories, total calories per gram of food, total fat, total carbohydrates, added sugars, and sugar- sweetened beverages, alongside lower fiber intakes, less milk and fewer fruits and non-starchy vegetables.22 Similarly, a longitudinal study in the United States showed that between 1977-78 and 1994-96, meals and snacks purchased away-from-home were higher in calories, total fat and saturated fat as a percentage of total calories, while being lower in fiber, calcium and iron, compared to foods prepared at home.11

The higher calorie intakes that are consistently observed at restaurants have been attributed to larger portion sizes.23, 24 Data in the United States from the National Food Consumption Survey and the Continuing Survey of Food Intake by Individuals showed that between 1977 and 1996 food portion sizes increased both inside and outside-the-home, with the largest portion sizes being found at fast-food restaurants.25

Despite this, research investigating 14-year trends in the energy and nutrient content of foods from eight of the leading fast-food chains in the United States between 1997/98 and 2009/10 found that median energy content of menu items remained stable,26 and that the healthy eating index score (a quantitative measure of diet quality in relation to US federal dietary guidance)27 improved over this time period.28 However, it should be noted that these conflicting studies were only conducted in a small number of restaurants, and their trend was not consistent across restaurants. Therefore, these studies illustrate the existing variability among restaurants, hence, studies with a small sample size, may not always be applicable to all scenarios.

It should be noted that the data discussed in this section are from the United States, due to the fact that similar data in Canada has not been evaluated to date. Some of the studies in this thesis address these gaps in the Canadian literature.

1.2.3. Calorie and sodium levels of meals purchased from fast-food restaurants

In addition to investigating the nutrient intakes of consumers who eat-out, two studies investigated the calorie and sodium content of lunchtime purchases in New York City by surveying customers who were exiting fast-food restaurants. They found that lunchtime fast-food purchases on average contained 827 calories and 1751 mg of sodium.29,30 Moreover, they found that one-third of purchases were over 1000 calories, 57% of meals contained more than 1500 mg of sodium (the daily Adequate Intake level [AI]), and 20% contained more than 2300 mg (the daily Tolerable Upper Intake level [UL]).

1.2.4. The association between eating-out and obesity

Ecological data has shown that consumption of foods prepared outside-the-home rose during the same time period that obesity increased.31 Similarly, cross sectional data has demonstrated associations between eating-out and overweight/obesity.32-36 For example, NHANES data (2005-2010) (n=10,953), demonstrated a positive association between frequency of eating-out and BMI.37 They found that those who ate-out greater than six times per day had a BMI or 28.7 while those who never ate-out had a BMI of 27.2. In addition, a community-based study among 1449 adults and adolescents in Minnesota showed a positive association between eating-out and percent body fat.38

Furthermore, many studies suggest an association between eating-out and overweight/obesity.39 For example, in a 15 year prospective study, participants consuming fast- food more than twice a week gained an additional 4.5 kilograms of bodyweight and had a two- fold greater increase in insulin resistance.23 Meanwhile, the USDA report on the impact of food away from home estimated that for the average consumer, eating one meal away from home each week could translate into roughly two extra pounds (0.9 Kg) per year.40 Despite the fact that the most recent systematic review of the association between eating-out and obesity found the link to be unclear,41 this review was confounded by the fact that many studies were not conducted in real-life restaurant settings, hence, the author concluded that despite modest effects, sufficient evidence exists for public health recommendations to limit fast-food consumption and facilitate healthier menu selection.

1.3. Nutritional quality of restaurant foods

In the lay press, restaurant foods have consistently been criticized for being unhealthy. However, at the commencement of this thesis there were only a small number of published studies—conducted in the United States and Australia—that reported on the nutrient composition and nutritional quality of restaurant foods.

1.3.1. Fast-food in Australia

The most comprehensive study of the nutritional quality of fast-food was conducted in Australia, by Dunford et al in 2009. 42 This study analyzed nutrient data for 584 fast-food menu- items from nine chain restaurants and found that the majority of fast-food products had high levels of sugar, total fat, saturated fat, and sodium, and did not meet the nutrient criteria established by the UK Food Standards Agency, to define ―healthy foods‖. Furthermore, their findings illustrated how diverse serving/portion sizes can make comparisons amongst fast-food meal items misleading. For example, on a per 100g basis, pizza had substantially higher sodium compared to hamburgers; however, when comparing these two products based on their serving size, the situation was reversed.

1.3.2. Children’s restaurant meals in the United States

The studies conducted in the US at the start of this thesis focused exclusively on children’s restaurant foods. For instance, a study investigating the nutritional quality of fast-food children’s meals from ten chains in Texas (using publicly available, industry reported nutritional data) compared meals to the criteria for the US National School Lunch Program which required that <30% of total energy come from total fat, <10% of total energy come from saturated fat, along with minimum requirements for protein, calcium, iron, vitamin C, and vitamin A. They found that only three-percent of children’s meals met all criteria.43 Furthermore, another US study that investigated 33 restaurants in Virginia compared fast-food and non-fast-food (sit-down restaurant) children’s menu items and found that fast-food restaurants provided smaller serving sizes, fewer calories, less total fat, saturated fat, protein, carbohydrates and fiber compared to sit- down restaurants.44 However, when controlling for portion sizes these differences disappeared.

1.3.3. Restaurant foods in Canada

At the commencement of this thesis there was no published literature on the nutritional quality of restaurant foods in Canada. The only data on restaurant foods in Canada at the time of this thesis was found in the online ―Fatabase‖ created by the Vancouver Sun newspaper in 2009. The ―Fatabase‖ contained nutrition information for over 65 chain restaurants in British Columbia. This online database provided nutrition information to consumers, and allowed users to search for foods by restaurant, menu item or key words to find the calories, fat, trans fat, sodium, carbohydrates, and sugars in chain restaurant foods. In 2011 the ―Fatabase‖ was updated.45

During the course of this thesis, two studies comparing the nutritional quality of restaurant foods globally were published. The first, conducted by Dunford et al compared salt levels in six countries (Australia, Canada, France, New Zealand, the United Kingdom and the United States) in 2010.46 They found that salt levels varied by food category, food company and country. Salt levels in Canada were occasionally higher compared to other countries (such as the United Kingdom which currently has a government-industry agreement to lower sodium levels), however, all comparisons were based on salt density (salt per 100g) and did not account for potential portion size differences across countries. Furthermore, they only included foods from six fast-food chains. In addition, a study by Hobin et al compared energy, total fat, saturated fat and sodium levels in children’s fast-food from four fast-food chains in Australia, Canada, New Zealand, the UK, and the US.47 They also found variation across companies and countries and concluded that this variability illustrates potential for improvement.

Overall, at the start of this thesis, studies investigating the nutritional quality of restaurant foods only contained a small number of chains, focused almost exclusively on fast-foods, and did not include menu-items from sit-down restaurants (restaurants with table service). This is a major oversight considering that sit-down restaurants account for nearly half, 22 of the 49.2 billion dollar, restaurant industry in Canada.48

1.4. Menu-Labelling

Menu-labelling refers to a type of food labelling where information about the nutritional composition of a food/beverage is displayed at the point-of-purchase on a restaurant menu,

6 menu-board or drive thru display. Numerous government agencies (such as the Institute of Medicine, Food and Drug Administration etc.) as well as health professional organizations (including the Ontario Medical Association, American Medical Association etc.) in Canada and the United States have recommended menu-labelling as a public health policy to help change the food environment and therefore assist in combating the rising rates of obesity and diet-related disease.49-53 However, at the start of this thesis, there was limited adoption of voluntary menu labelling initiatives and no existing legislation requiring the disclosure of nutrition information in restaurants or other food venues in Canada.

1.4.1. The Rationale for Calorie-Labelling

Lack of awareness about calories is a large part of the rationale for calorie labelling, as several studies have shown that consumers cannot estimate calorie content.54-58 For instance, one study showed that 90% of consumers underestimated the calorie content of unhealthy menu items by approximately 642 calories, and for certain items they underestimated by over 2000 calories. Even trained nutritionists had difficultly estimating calories in restaurant foods.59 In addition, there has been strong consumer demand for calorie-labelling, as research has shown that consumers want to see nutrition information on restaurant menus, even if they don’t consistently use it.60-62 A public health surveillance telephone survey (n=1,699), conducted by Toronto Public Health in late 2011 and early 2012, showed that 78% of respondents said they would use menu-labelling information if it was present.63

1.4.2. Menu-Labelling in the Canadian Context

1.4.2.1. The history of menu-labelling in Canada and Ontario

In Canada there have been several unsuccessful menu-labelling bills at both the provincial and federal level. The earliest menu-labelling bill in Canada (Bill C-398, An Act to Amend the Food and Drugs Act) was introduced in 2003, following the introduction of mandatory nutrition labelling on packaged foods.64 This was a federal bill introduced by Tom Wappel, an NDP MP, that proposed the disclosure of calorie, sodium, saturated fat and trans fat information on chain restaurant menus. The bill was unsuccessful, and was reintroduced in 2006 (Bill C-283), to no avail.65

Interest in menu-labelling in Ontario began in 2009, when an NDP MPP, France Gelinas, introduced the first of a series of private member bills calling for menu-labelling. Her first two bills (Bill 156 and Bill 90, introduced in 2010) recommended calorie labelling on restaurant menus.66,67 However, in 2012, she reintroduced an expanded version of her bill (Bill 86), that recommended the disclosure of calorie information in addition to sodium warning labels, to highlight high sodium foods on restaurant menus.68 This bill was stalled in fall 2012 when the government was prorogued and was reintroduced in spring 2013 (Bill 59).69

Around the same time, on April 29th 2013, Toronto Public Health issued a report entitled ―What’s on the Menu?‖ and recommended that if the province did not move forward with menu- labelling, then the city of Toronto should.70 Toronto Public Health’s recommendation differed from the two most publicized menu-labelling laws enacted in New York City and proposed in the US Federal Patient Protection and Affordable Care Act (both of which were calorie labelling laws),71 because it called for the disclosure of sodium information, in addition to calorie information.

A few months later, in December 2013, France Gelinas re-introduced her bill (Bill 47) calling for calorie labeling alongside high sodium warning labels;72 and in February 2014, the Provincial Liberal government introduced Bill 162, the Making Healthier Choices Act, calling for calorie labelling.73 Both bills were halted as a result of the spring 2014 election. In November 2014, both bills were reintroduced and on May 29th 2015, the first menu-labelling bill in Canada, Bill 45 The Making Healthier Choices Act (which includes the Health Menu Choices Act), was passed.74 This bill requires the disclosure of calorie information in all restaurants across Ontario that have 20 or more locations in the province, but does not require the disclosure of sodium information.

At the outset of this thesis, there was no research investigating Canadian consumer’s attitudes or use of menu-labelling.

1.4.2.2. The CRFA voluntary nutrition disclosure program for restaurants

In 2005, the Canadian Restaurant and Foodservice Association (CRFA, presently known as Restaurants Canada) developed a voluntary program to make nutrition information for restaurant foods available online.75 The CRFA set standards for which nutrients were required to

8 be disclosed (the thirteen core nutrients on the Nutrition Facts table, plus calories and serving size) and established a consistent format for displaying the information in the same order and style. In total, 33 establishments, representing approximately 60% of food service sales revenue, participated. Nevertheless, a study in eight locations representing four major restaurant chains (McDonald's, Burger King, Starbucks, and Au Bon Pain) showed that on average, only 0.1% of consumers sought out nutrition information when it was on-premise (on a wall poster or in pamphlets) but not present at the point-of-purchase.76

1.4.2.3. The Informed Dining Program

In British Columbia, Informed Dining, a voluntary program for disclosing nutrition information in restaurants was established in 2012 by the provincial government in collaboration with the restaurant industry.77 Restaurants participating in the Informed Dining program are required to provide their guests with nutrition information (including the 13 core nutrients) for all standard menu items in brochures on their counters. The 10,000 participating outlets must display the informed dining logo on their menu or menu board alongside a directional statement telling customers that nutrition information is available on-site. Restaurants can choose how they provide this information (in the form of a menu insert, a supplemental nutrition menu, a nutrition brochure, or a poster). Some of Canada’s largest chains including McDonalds, Subway, Pizza Pizza, Tim Hortons and A&W, have implemented the informed dining program in their restaurants across the country.78

1.4.2.4. Menu-labelling research in Canada At the start of this thesis there was no menu-labelling legislation in Canada, therefore, there were no natural experiments evaluating this policy in a Canadian context. However, a cross-sectional study in a hospital cafeteria (n=1003), using exit-surveys showed that Canadian consumers who saw calorie, sodium, and fat content on digital menu boards (showcasing four possible menu choices) consumed 21% fewer calories, 23% less sodium, 33% less saturated fat and 37% less total fat, compared to consumers who did not see nutrition information in a control cafeteria site.79

Furthermore, a randomized trial (n=635) in southwestern Ontario investigated calorie labeling on menus and compared a control (no labelling) with numerical calorie labelling, traffic

9 light (red, yellow, green) calorie labelling, and traffic light labelling of calories, sodium, fat and sugar.80 In this study they showed participants experimental menus and then tested their recall of nutrition information, knowledge of calorie content, and assessed the calorie content of their order. This study found that the calorie content of the meals ordered across conditions did not differ, however, the actual amount of calories consumed was significantly lower among participants who saw calorie information. Furthermore, as expected, participants who saw calorie information were more likely to recall the calorie content of meals and to report using nutrition information when ordering.

1.4.3. Menu-labelling in the United States

1.4.3.1. The history of menu-labelling in the United States

In 2008, New York City was the first jurisdiction in the world to implement mandatory calorie labelling on restaurant menus, menu boards and drive-through displays.71, 81 According to the regulation, the font and format of the calorie number must be at least as large as the font size of the name and price of the menu-item. The regulation applies to all restaurants with 15 or more locations and thus covers about one-third of all restaurant traffic in New York City.82

Following New York City, several other jurisdictions in the US enacted menu-labelling laws. For instance, in 2009, King County, Washington, implemented a labelling law that required the disclosure of calorie, sodium, carbohydrate and saturated fat information on menus and menu boards.83, 84 And in January 2010, Philadelphia enacted a similar law that required calorie labelling on menu-boards and the disclosure of calories, sodium, saturated fat, trans fat and carbohydrates on printed menus.85

In addition, in March 2010, mandatory calorie labelling was included in the 2010 US Health Reform legislation (the Patient Protection and Affordable Care Act, also known as ―Obamacare‖). This law requires calorie labelling on all menus, menu-boards and drive-thru displays at chain restaurants with 20 or more outlets nationally, in addition to several other requirements.86 Restaurants must include a succinct statement with the suggested total daily calorie intake on menus and menu boards, to provide context about the amount of calories consumed in an average daily diet. In addition, restaurants must also disclose written information for the thirteen core nutrients found on a nutrition facts table and include a visible statement

10 indicating the availability of the written information on menus and menu-boards. Additionally, these calorie labelling regulations also apply to operators who own or operate twenty or more vending machines. At this point in time, the US federal menu-labeling law, has yet to be implemented.87

1.4.3.2. Natural experiments evaluating calorie labelling in New York City

Evaluations of New York City’s menu-labelling regulations have yielded mixed results. For example, Vadiveloo et al compared food and beverage purchases before and after menu- labelling was implemented in four popular chains in New York City, and used Newark, New Jersey as a control for difference-in-difference analysis.88 By surveying customers as they exited the restaurants they collected data on types of food purchases (ex. type of beverage purchased, type of salad dressing purchased, whether French fries or salad was purchased, whether or not cheese was added to a menu item, the number of desserts ordered and their frequency of fast food consumption each week). Overall, they found no significant favourable differences and detected some unfavourable changes. Moreover, they showed that 41% of patrons reported noticing the information, but only 14.5% reported actually using the information. However, this study only investigated purchases in fast-food restaurants, as opposed to sit-down restaurants. It was cross-sectional in design, short in duration (four weeks) and had low power. Not to mention, it did not investigate changes in calories, but rather, only looked at changes in the types of foods purchased. Lastly, as with all menu-labelling studies to date, it did not record foods (calories) consumed, but only measured foods (calories) ordered.

Bollinger et al analyzed purchases from Starbucks loyalty card members and found that mandatory calorie labelling resulted in a 6% decrease in calories per transaction (falling from 247 to 232 calories) that persisted over 10 months.89 This decrease was due entirely to decreases in calories from food (which fell by 14%), as they observed no changes in beverages. Overall, Bollinger found that three-quarters of the reduction in calories per transaction were due to consumers buying fewer items, as opposed to consumers substituting for lower calorie items. Despite this, there was no effect on Starbucks’ profits; and compared to their competitor, Dunkin Donuts, menu-labelling may have increased Starbucks’ revenue.

Dumanovsky et al investigated 7309 adult consumers before and 8489 consumers after menu-labelling was implemented and found that overall there was no decrease in calories ordered. However, when controlling for restaurant chain, poverty level, sex of customer, type of purchase and inflation there was a small decrease from 847 to 827 calories per purchase.29 Furthermore, this study illustrated how the efficacy of menu-labelling can vary depending on the chain. For example, several major chains (notably McDonalds, Au Bon Pain and KFC) all saw significant decreases of 44 to 80 calories per purchase. Meanwhile, the increase in calories observed at Subway was more likely due to the $5 footlong promotion, illustrating how commerce can trump health. Overall, this study found that the 15% of people who saw and used the information ordered 106 fewer calories than those who didn’t see/use the info.

In an investigation of the effect of restaurant calorie-labelling on consumer’s ability to estimate the recommended amount of calories and the actual amount of calories in their meal, Elbel et al found that menu-labelling had little impact on consumer’s knowledge of how many calories they should be consuming or their knowledge about the amount of calories in fast foods.90 However, they did find that the number of low-income consumers who could correctly estimate the number of calories in a fast-food meal increased from 15% to 24%.

1.4.3.3. Long-term evaluation of menu-labelling in New York City

The most recent follow-up evaluation of menu-labelling in New York City (five years after it was implemented) was conducted at Burger King, McDonalds, Wendy’s and KFC by surveying 7699 customers who were exiting these chains at lunchtime.91 The study saw no change in calories purchased and reported a decline in menu-labelling use. However, this does not necessarily mean that the policy failed, as this follow-up included only a limited number of restaurants. Furthermore, to date, no study has ever investigated the effect of menu-labelling in sit-down restaurants in New York, which is a major gap in the current literature.

Furthermore, New York’s menu-labelling policy has been criticized because it lacks a contextual statement informing customers of the daily calorie recommendations. Research has shown that a contextual statement that includes the daily recommended calorie intake level makes the information translatable and results in an additional savings of 250 calories over the course of the day.92, 93 Furthermore, many restaurants post ranges of calories (that can vary by up

12 to 500 kcal) for their menu combos that come with different side dishes.88 These factors, along with the previously mentioned methodology limitations of evaluations to date, may be contributing to the lack of observed effects.

1.4.3.4. Evaluation of menu-labelling in sit-down/full-service restaurants To date, only two studies have evaluated menu-labelling in sit-down or full-service restaurants, defined by the presence of a waiter/waitress. A study comparing a voluntary menu- labelling program in six full-service restaurants in Pierce County Washington before and after implementation, found that menu-labelling led to 15 fewer calories and 45 fewer milligrams of sodium being ordered per entrée.94 Meanwhile, the 20.4% of consumers who reported using the information ordered entrées containing 75 fewer calories.

A cross-sectional study evaluating Philadelphia’s menu-labelling legislation in two outlets (containing labelling) compared to five control (non-labelled) outlets of the same restaurant chain, found that after controlling for confounders (gender, age, race, income, education, body size, day of the week, and frequency of eating at full-service restaurant chains) customers at labeled restaurants purchased meals with 151 fewer calories and 224 fewer milligrams of sodium, while those who reported using the information purchased 400 fewer calories and 370 fewer milligrams of sodium.95

1.4.4. Experimental tests of calorie-labelling in non-restaurant settings

Many studies have tested calorie labelling in non-restaurant settings such as school, work, or hospital cafeterias,79, 96-98 in doctor’s offices,99 in experimental lab settings80, 92, 100, 101 and using internet surveys.85, 86 The most recent systematic review of 19 studies investigating calorie labelling in restaurant and non-restaurant settings found a marginal decrease of only 18.3 calories,41 however, data from non-restaurant settings (especially surveys and lab settings) should be interpreted with caution, as the results may not reflect real-life purchases.

1.4.5. Calorie-labelling in independent (non-chain) restaurants and cafeterias

Currently, the existing menu-labelling legislation in Ontario and the United States applies exclusively to chain restaurants with a minimum of 20 or 15 locations, respectively.71, 102 Menu-

13 labelling is a feasible policy option for chain restaurants because chains have consistent menus. Furthermore, this is a minimally invasive form of labelling that only requires a single numerical addition to menus. However, there are many settings where calorie labelling is not feasible, such as independent restaurants which comprise 37.8% of the restaurant market share.103 A survey of independent restaurant owner’s attitudes towards menu-labelling revealed that 72% were not interested in providing nutrition information to their customers.63 In addition, 76% expressed that adjusting their menus to provide nutrition information would be too expensive an undertaking and 64% said that they were too busy to ―figure out‖ nutrition information.

In addition to independent restaurants, cafeterias are a source of food procured outside- the-home16 where menu-labelling may not feasible due to an eternally rotating menu cycle, or due to the lack of resources to calculate the calorie content of the numerous foods offered.104 As a result, there are many dining locations outside-the-home where calorie-labelling may not be a feasible or efficient way to promote healthier choices and therefore, alternative forms of labelling/education may be more desirable.

1.5. Alternative forms of nutrition labelling

Many studies in non-restaurant settings (such as cafeterias, experimental lab settings and surveys) have provided opportunities to test a multitude of different education interventions, including alternative forms of labelling that include additional information beyond calories (such as saturated fat and sodium);57, 79 interpretive forms of labelling, such as traffic light labelling;105, 80 or even innovations in the manner in which the menu-items are presented, such as ranking the items according to calorie content.106 One example of a novel type of labelling that has been tested in non-restaurant settings is Physical Activity Calorie Equivalent labelling.107, 108

1.5.1. Physical Activity Calorie Equivalent (PACE) Labelling

Physical Activity Calorie Equivalent labelling (PACE labelling) is a form of labelling that illustrates the calorie content of a food according to the minutes of physical activity that are required to burn the calories in that food. This provides a simplified representation of calories in a contextual manner that may be more meaningful and relatable for consumers. To date, only four studies have tested the effect of PACE labelling on consumer food choices and three of

14 these studies have utilized web-based methodologies.107-110 An online survey testing the effect of PACE labelling showed that while calorie labelling led consumers to reduce their calorie intakes by 380 calories, including calorie information alongside the number of minutes of walking to burn those calories, led consumers to save 440 calories, an additional 60 calories compared to calorie-labelling alone.107 A survey of 1000 parents showed that PACE labelling may influence parent’s decisions about fast-food choices for their children.109 Furthermore, 40% of consumers said that PACE labels were very likely to influence their food choice, while another survey showed that 82% of participants preferred PACE labels over calorie-only labels.108 A randomized study in a metabolic kitchen and graduate student residence compared exercise labels to control and found that exercise labels led to a 139 calorie decrease in calories ordered.110 However, they found no significant difference between exercise labels and calorie based labels.

1.5.2. Non-Numerical forms of labelling/education

Beyond the numerical forms of labelling—such as calorie labelling and PACE labelling—countless other interventions have shown promise.111-115 Notably, alternative forms of labelling that direct consumers to a preferred choice without requiring any kind of comparison or interpretation may be beneficial as it has been shown that reducing the cognitive load and simplifying food choices with a clear nudge towards the preferred food choice can be a good strategy for promoting healthier choices.116,117

For example, one study showed that simply labelling healthier choices with a shelf-tag that says ―fuel your life – healthy campus‖ results in a higher percentage of total sales coming from the tagged items.118 Furthermore, another study showed that point-of-purchase messaging to encourage healthier choices (such as: revitalize yourself by snacking on a fresh basket of crisp red peppers, juicy tomatoes and crunch carrots‖ or ―need a sweet, fast energy boost? Get long- lasting energy and tons of flavour by grabbing a yogurt‖) led to increased purchases of the promoted items.119

Another example is an intervention that focused on reducing saturated fat. This intervention demonstrated the benefit of providing education in the form of one single message at a time, to avoid confusing and overwhelming customers. They found that by educating people

15 with a simple message to switch from high fat milk to low fat milk, they were able to increase sales by 18%, and thus decreased saturated fat intakes, without ever mentioning the word ―saturated fat‖.120

That being said, research has shown that simply giving people health information (ex. soy has important phytochemicals) is not the most efficient way to change behaviour.121 Instead, giving people self-relevant personal health consequences (ex. soy will help lower your risk of heart disease) is more likely to change their behaviour. In addition to strategically wording health messages, the context of where the health message is conveyed may be equally important, and it has been suggested that the best place to educate consumers about nutrition is in settings where food selection actually occurs.122

These data show that in addition to calorie labelling, there are countless other ways to promote healthy eating in other settings where information—beyond just calories—can be feasibly displayed. Sodium is a particular nutrient that has received attention as a nutrient that could be labelled on restaurant menus,

1.6. Sodium-related public health initiatives in restaurants

1.6.1. Introduction to sodium as a public health issue

At the start of this thesis, sodium was a particularly visible public health issue due to the recent release of the Canadian Sodium Working Group’s recommendations.51 A high dietary sodium intake is a causal risk factor for hypertension,123 which affects 26% of adults,124 and is the leading preventable risk factor for death worldwide.125 According to the Institute of Medicine, the daily Adequate intake levels for sodium is 1500 mg and the daily upper tolerable intake level is 2300 mg.126 However, Canadians on average consume 3400 mg per day,127 with 70% of their sodium intake coming from processed and restaurant foods.128

1.6.2. The Canadian Sodium Working Group

In 2007 the Canadian Sodium Working Group (SWG) was struck to develop and oversee the implementation of a population health strategy for reducing Canadians’ sodium intake.129 This was a multi-stakeholder working group comprised of academic, clinical, consumer/health

16 non-governmental organizations (NGOs) and industry leaders. In 2010, the SWG published its recommendations—for all levels of government, non-governmental organizations, consumers, industry and relevant stakeholders. The recommendations had three primary foci: 1) structured voluntary reduction of sodium levels in processed food products and foods sold in food service establishments; 2) education and awareness of consumers, industry, health professionals and other key stakeholders; and 3) research, in addition to a fourth component focusing on monitoring and evaluation, which encompassed all three areas.

The SWG’s interim goal was to lower the Canadian population’s average intake of sodium to 2300 mg per day by 2016, with an ultimate goal to have most individuals (more than 95% of the population) with intakes below the tolerable upper intake level (UL of 2300 mg per day). The structured voluntary approach to sodium reduction (for packaged foods and restaurant foods) was intended to include published sodium reduction targets, with defined timelines, a mechanism for public commitment to the targets, a plan for third-party monitoring of progress, and an independent evaluation of success. However, before any of this could be implemented, the SWG was disbanded in late 2010.

Subsequently, in June 2012 Health Canada published their ―Guidance for the Food Industry on Reducing Sodium in Processed Foods.‖129 This document included benchmark targets for sodium reduction in processed foods. Targets were set for 15 major food categories and sub-categories. Targets were based on sales weighted mean average of sodium per 100g and apply to the sales weighted average sodium level for a company’s products. The 2016 targets were based on 25-30% reductions from the 2009/2010 baseline sales-weighted averages. This produced a target that is below the 10th percentile of the 2009/2010 sodium range. There were two intermediary phases for sodium reduction established representing one-third and two-thirds of the reduction required to meet the 2016 benchmarks. In addition, a maximum recommended level for sodium was set for 2016. This represented the 75th percentile of sodium levels in each category. The goal is that by the end of 2016 for all companies to lower sodium levels so that 1) their product portfolio sales weighted average meets or is below the 2016 SWG recommendation, and 2) all individual products meet or are below the maximum benchmark for each category. At this point in time, sodium reduction targets for restaurant foods have never been established.

It is important to note that the SWG’s recommendation 1.8 called for the disclosure of sodium information in restaurants and stated: ―The Working Group recommends that the Food and Drug Regulations and applicable provincial regulations be amended to require the on-site disclosure of nutrition information in a consistent and readily accessible manner for standardized menu items prepared and assembled on-site at restaurants and food services establishments, where feasible (i.e., in establishments with a high degree of standardization).‖51 At the time of this recommendation there was little research on the benefit of including sodium information on restaurant menus. However, a mailed survey in the United Kingdom (n=786) showed that 49.9 to 60.8% of survey panelists said they would like to see salt information on restaurant menus.130 Furthermore, in this survey, a greater proportion of respondents reported that they would like to see information on salt, compared to fat or energy.

In 2013 MP Libby Davies introduced a private-member bill calling for the implementation of the SWG’s sodium reduction strategy for Canada, however, this bill was defeated.131

1.6.3. The United States’ National Salt Reduction Initiative

Similarly, in the US there has been interest in lowering population sodium intakes. As a result, in 2009, the New York Department of Health and Mental Hygiene started the National Salt Reduction Initiative, a public-private partnership with an aim to reduce rates of heart disease and stroke by lowering sodium intake from packaged and restaurant foods.132

The NSRI set voluntary targets for sodium levels in 62 categories of packaged foods and 25 categories of restaurant foods to be reached by 2012 (the interim deadline) and 2014.133 The targets are sales weighted averages for each category. Therefore, not all foods are required to meet the target; however, there is an upper limit for individual products (1500 mg by 2012 and 1200 mg per item by 2014).

Currently, 27 companies have committed to the National Salt Reduction Initiative (NSRI) including: Au Bon Pain, Bertucci's Italian Restaurant, Black Bear European Style Deli, Boar's Head Provisions Co., Butterball, Campbell Soup Company, Delhaize America, Dietz & Watson, FreshDirect, Furmano's, Goya Foods, Hain Celestial, Heinz, Ken's Foods, Kraft Foods, LiDestri Foods, Mars Food US, McCain Foods, Premio, Red Gold, Inc., Snyder's-Lance, Inc., Starbucks

Coffee Company, Subway, Target Corporation, Unilever, Uno Chicago Grill and White Rose.134 The NSRI is monitoring companies’ progress through their Heart Follow-Up study, which is assessing sodium intakes in NYC residents using 24-hour urine collection,135 and through their menu stat program, where they annually collect nutrition information for foods from the leading chains.136

In 2011, the IOM recommended that the FDA reconsider the generally recognized as safe (GRAS) status of salt (which was granted in 1958) and decrease the daily value for sodium for 1500 mg per day. In response to this, the FDA opened a docket to request comments, data and information on ways to reduce sodium intake in the population. The FDA says that it is developing voluntary targets for sodium reduction, however, to date, no targets, recommendations, or changes have been made.137

1.7. Summary Overall, at the start of this thesis there was an established association between eating-out and risk for disease. Meanwhile, there was limited research on the nutritional quality of restaurant foods in Canada. Furthermore, the Canadian Sodium Working Group’s plan to monitor sodium was abandoned and despite interest in menu-labelling policies, there was no legislation and no research on menu-labelling in Canada. It was these factors that led to the objectives of this thesis.130

Chapter 2 Objectives and Hypothesis of Thesis

2.1. Objectives

The overarching objective of this thesis was to generate food supply and consumer data to inform public health policy debates around issues concerning food consumed outside-the-home.

There were three primary objectives:

1) The first objective focuses on the restaurant food supply with an aim to investigate the nutritional quality of Canadian fast-food and sit-down chain restaurants with a focus on calorie and sodium levels. And to monitor longitudinal changes in sodium levels over time.

2) The second objective focuses on restaurant menu-labelling policies with an aim to understand Canadian consumers’ attitudes towards menu-labelling and to experimentally test consumer’s use of calorie, sodium and calorie density labelling on restaurant menus.

3) The third objective focuses on labelling/education in a cafeteria setting and aims to test an alternative form of nutrition labelling/education (e.g. PACE labelling) designed to encourage university students to consume more fruits and vegetables and fewer sugar- sweetened beverages.

2.2. Specific hypotheses

I hypothesized that:

1) Menu-items from Canadian chain restaurants, much like their US counterparts, would have a wide range of calories and sodium levels and many foods would exceed the daily recommended sodium intake levels. Over time, restaurants would reformulate their products to lower the sodium levels in their foods.

2) Calorie and sodium labelling on restaurant menus would encourage Canadian consumers to choose meals with fewer calories, less sodium and a lower calorie density.

3) A labelling/education intervention designed to increase the consumption of fruits and vegetables, and decrease the consumption of sugar-sweetened beverages would positively influence the food choices of students’ in a University dining-hall.

2.3. Preview of chapters

Objective 1 was examined in chapters 3 to 5 using a national database of over 9000 menu-items from Canadian fast-food and sit-down chain restaurants, which I developed and created in 2010, and subsequently updated in 2013. In chapter 3, the range of calorie levels in various food categories and restaurants were analyzed, and the relative role of serving size and calorie density as determinants of total calories in restaurant foods was investigated in order to better understand how factors that determine calorie content may potentially influence the effectiveness of calorie labelling. In chapter 4, the sodium levels in Canadian chain restaurant foods were compared to the recommended daily AI and UL as well as to the US National Sodium Reduction Initiative Targets. Results from chapter 4 were followed-up in chapter 5 where the data collected in 2010 was recollected in 2013. In this longitudinal study, sodium levels were compared over a three year period and the proportion of foods that increased, decreased or stayed the same was investigated. In addition, the prevalence and magnitude of change in each food category and restaurant establishment were investigated and the proportion of foods exceeding the AI and UL was determined.

Although the major focus of objective 1 was on calories and sodium due to the relevance of these nutrients to current policy debates, we also investigated additional nutrients such as added sugars in supplementary manuscripts found in Appendix A and Appendix B.

Objective 2 was investigated in chapter 6 using an online, national consumer survey that included a repeated measures experiment in which consumers were asked to select what they would typically order from four mock-restaurant menus. Subsequently, consumers were randomly allocated to see one of three menu-labelling treatments (calories; calories and sodium; or calories, sodium and serving size) and were given the chance to change their order. Changes

21 in calories and sodium levels at the population-level and among the sub-set of consumers who used the information were analyzed and consumer attitudes were explored using open-ended questions.

Objective 3 was addressed through a quasi-experimental study to measure the effect of a population-level nutrition labelling/education intervention study in a campus cafeteria, and the results are presented in chapter 7. Posters encouraging students to ―fill half your plate with fruits and vegetables‖ and ―drink water when you are thirsty‖ were installed alongside a comparison of the minutes of jogging required to burn each of the available beverages. Changes in students’ beverage as well as fruit and vegetable choices were measured using inventory data and direct- measured observational data.

Together, these studies provide data on the nutritional quality of restaurant foods in Canada and examined two different interventions to promote healthier choices when consumers are eating-out. Overall, this thesis provides evidence to inform policy decisions that address issues concerning foods consumed outside-the-home.

2.4. Student Contribution The original idea for the restaurant database belongs to Dr. Mary L’Abbé. We collaboratively developed and created it. I was responsible for collecting, entering, validating and categorizing the data in 2010. In 2013, I developed the protocol for the second collection and received assistance with data entry from two undergraduate volunteers, whose work I personally verified.

The study concepts and designs in this thesis were all co-conceived by myself and Dr. L’Abbé. Data analysis was primarily conducted by me with some guidance from Dr. Paul Corey and Dr. Wendy Lou. I was primarily responsible for data interpretation with some input from co- authors. I personally drafted all manuscripts. Further details about contributions to each manuscript can be found at the end of each manuscript.

Chapter 3 Restaurant menus – calories, caloric density, and serving size (AJPM, 2012)

This manuscript has been published: Scourboutakos MJ, L’Abbé M. 2012. Restaurant menus – calories, caloric density and serving size. American Journal of Preventive Medicine, 43(3):249- 255. [http://www.ajpmonline.org/article/S0749-3797(12)00382-0/abstract]

3.1. Abstract

Background: The increasing trend towards eating-out along with concerns about the adverse nutritional profile of restaurant foods has prompted the introduction of calorie labeling. However, the calories available in food from sit-down and fast-food restaurants have not been analyzed.

Purpose: The calorie content of restaurant foods was analyzed in order to better understand how factors that determine calorie content may potentially influence the effectiveness of calorie labeling.

Methods: Nutritional information was collected from the websites of major (n=85) sit-down and fast-food restaurants across Canada in 2010. A total of 4178 side dishes, entrées and individual items were analyzed in 2011.

Results: There was substantial variation in calories both within and across food categories. In all food categories, sit-down restaurants had significantly higher calories compared to fast-food restaurants (p<0.05). Both serving size and caloric density were positively correlated with calories, however serving size was more strongly correlated (r=0.62) compared to caloric density (r=0.29).On average, items that were higher in calories had a significantly larger serving size compared to items that were lower in calories (p<0.05); however, they were often not significantly different in terms of caloric density.

Conclusion: Variation in calories per serving was seen when comparing different types of food, different types of establishments and the specific establishments that the foods were from.

Compared to caloric density, serving size was shown to be a more important driver of calories per serving in restaurant foods.

3.2. Background

Over the past 30 years the prevalence of food consumed outside-the-home has increased, while obesity rates have simultaneously risen.31, 138 Presently in Canada, on any given day, 17.7 million people (approximately half of the population), and visit 80,800 food-service establishments.15-17 Eating outside-the-home has been associated with increased caloric intake,21, 139 as well as increased risk for insulin resistance and obesity.140-142 Currently, 24.1% of Canadians are obese, while 34.4% of Americans are obese.143

In light of these concerns, mandatory calorie labeling on menus, menu boards and drive- through displays has been introduced as a policy option to address this situation. Menu-labeling was included in the 2010 US Health Reform legislation (though it is yet to be implemented nationally),71a and was recently proposed in Ontario, Canada.68b Though one study showed that calorie labeling decreases calorie intakes at certain chains,11 and another suggested that calorie labeling could decrease calories ordered for children,12 others have shown no significant effects.14, 15, 16 While many studies have demonstrated the prevalence of large portion sizes and their role in increasing energy intakes in the restaurant sector;17, 18 to date, the relative role of serving size and caloric density as determinants of total calories in restaurant foods, as well as the impact of other factors that may influence calorie content, have not been investigated.

The objective of this study was to use the company-provided nutrition data on chain restaurant websites to analyze the calories in restaurant foods. Specifically the aim of this study was to: 1) compare how calories, caloric density and serving size vary according to the type of food and the type of establishment that a food is from; 2) determine if the average calories in sit- down restaurants (SDR) differ among different chains; 3) determine the relative role of serving

a On March 9th 2016, the US national menu-labelling legislation was delayed from its implementation date on December 1st 2016, to an undetermined date that will be one year after the FDA issues its final Level 1 guidance on menu-labelling. b After this manuscript was published, the Ontario provincial government passed menu-labelling legislation on May 29th 2015.

24 size and caloric density as determinants of calories, and 4) compare the serving size and caloric density of items that are high in calories versus items that are low in calories.

We hypothesized that calorie content is not only influenced by the type of food, serving size and caloric density, but also the type of establishment (whether it is a sit-down restaurant or fast-food restaurant), and the specific establishment that a food is from. Exploring these topics may provide insight into the potential usefulness of calorie labeling as a means to compare the calories in menu items and thus promote the consumption of lower calorie items.

3.3. Methods

3.3.1. Data Collection

The study consisted of a systematic survey of Canadian sit-down and fast-food restaurants. A database containing nutrition information (including serving size and the thirteen nutrients commonly found on food labels) for over 9000 items from 85 chains across Canada was constructed. Any establishment that provided publicly available Canadian nutrition information online or in-store and that had twenty or more outlets in Canada (according to the 2010 Directory of Restaurant and Fast Food Chains in Canada) was included.19 Data were collected from September to December 2010 (with the exception of four establishments whose data were retrieved in early 2011). Data were analyzed and verified in 2011.

3.3.2. Construction of the Database

Establishments were categorized according to whether they were sit-down or fast-food restaurants, with sit-down restaurants distinguished by the presence of table service. Foods were categorized by food type, and sub-categorized according to various characteristics (i.e. salads with dressing/without dressing). Foods were further sub-categorized according to whether they were considered side dishes, main entrées including side dishes, main entrées without sides or single-items that could be purchased individually. All categorizations were based on the data provided by establishments on their websites. When necessary, establishments were contacted via phone and/or e-mail to verify categorizations. To ensure that the data were entered accurately, all serving sizes, calories, sodium and trans fat information were confirmed using the original website sources. Sort and rank procedures and Atwater calculations were utilized to

25 check for outliers. When necessary, establishments were contacted to confirm suspicious data. A random sample of 5% of the data was checked against the original sources by a third party. When establishments could not be contacted regarding errors detected in the original website source, the data in question were not included in the analysis. When necessary, data were combined to avoid inconsistencies (ex. salad dressing was added to salads and sandwich components were compiled if establishments did not provide the data in this format).

3.3.3. Inclusion Criteria

Only categories containing more than 20 items and with representation from two or more establishments were analyzed. 28 food categories met the inclusion criteria. Eleven were represented in both sit-down and fast-food restaurants, and seven represented either sit-down or fast-food restaurants. Ten side dish categories that combined sit-down and fast-food restaurant data (due to the small sample size) were included. The number of items in each category and a description of each category can be found in Appendix C.

3.3.4. Exclusion Criteria

Beverages, combination meals (that combined entrées and side dishes), and condiments were not included in this analysis. When multiple serving sizes for the exact same item were present, only the medium/regular sized items were included in the analysis. If two sizes were provided, the larger option was analyzed.

3.3.5. Statistical Analysis

Mean serving size, calories and caloric density were calculated for each food category. Differences between similar meal items found in both sit-down and fast-food restaurants were tested using Wilcoxon rank sum tests. Differences in calories among different establishments and among different food categories and side dishes were compared using box-plots, relative standard deviation, Kruskal-Wallis ANOVA and post hoc multiple comparisons. Simple linear regression was performed using serving size and caloric density as predictors of calories. Spearman’s correlation was used to examine relationships between calories, caloric density and serving size. In order to further explore these relationships, all items were categorized into 100 calorie intervals. The mean caloric density and serving size were plotted and compared using

Kruskal-Wallis ANOVA and post-hoc multiple comparison analysis. All statistical analyses were performed using Statistica version 10 [StatSoft Inc, Tulsa, Oklahoma] software. A p value <0.05 was considered statistically significant.

3.4. Results

A total 4178 meal items, side dishes and single items that can be purchased individually, from 20 sit-down restaurants and 65 fast-food restaurants (Appendix C) met the inclusion criteria and were analyzed.

Within all sit-down restaurant establishments, calories per serving ranged from 61 kcal to 2486 kcal (Figure 3.1). The minimum calories in sit-down restaurants were fairly consistent across establishments. All establishments had items exceeding 1000 calories per serving; however the number of items exceeding this threshold, as well as the maximum calories differed depending on the restaurant.

Sit-down restaurants had significantly higher calories per serving in all food categories when compared to fast-food restaurants (p<0.05 to p<0.001 depending on the category) (Table 3.1). For hamburgers, pasta, and fries, both calories and serving size were significantly higher in sit-down compared to fast-food restaurants. In salads with meat, sandwiches/wraps and stir fry entrées; calories, serving size and caloric density were significantly higher in sit-down compared to fast-food restaurants. In salads, tacos/burritos and soup, differences in calories and caloric density between sit-down and fast-food restaurants were statistically significant, yet there was no significant difference in serving size. Breakfast and chicken items from sit-down restaurants had a significantly lower caloric density compared to fast-food restaurants; nevertheless they still had a significantly larger serving size and significantly more calories.

Overall, even though some categories contained significantly more calories per serving compared to others, there was vast variation in calories per serving within all food categories. Among sit-down restaurant items, relative standard deviations ranged from 30% in stir fry dishes to 58% in soups. In sit-down restaurants, sandwiches/wraps, breakfast items, pasta entrées, hamburgers, stir fry dishes, meat/seafood dishes and ribs were significantly higher in calories per serving compared to salad entrées, chicken and seafood (p<0.05) (Figure 3.2.a). However, due to

27 the variation in calories per serving within food categories, at least 50% of salads, chicken and seafood entrées contained more calories per serving compared to the lowest calorie sandwiches/wraps, breakfast, pasta, stir-fry, meat/seafood and rib dishes. Furthermore, immense variation was seen when comparing the difference between the minimum and maximum calories per serving within a single food category. For example, the range of calories per serving within sit-down restaurant categories varied from 595 for tacos/burritos (2.5 fold different in calories per serving) to 2156 for rib entrées (7.5 fold difference).

In fast-food restaurants, pasta, sandwiches/wraps, tacos/burritos, stir fry entrées and hamburgers were significantly higher in calories per serving compared to pizza, chicken, salads and breakfast items (p<0.05) (Figure 3.2.b). Amongst fast-food restaurant items, relative standard deviations ranged from 35% in sandwiches/wraps to 57% in soups. Nevertheless, due to the variation within categories, at least 50% of pizza, chicken, salads and breakfast items contained more calories per serving than the lowest ranking pasta, sandwiches/wraps, tacos/burritos, stir fry meals and hamburgers. The range of calories per serving in fast-food restaurants varied from 380 in sushi (2.6 fold difference) to 1145 in breakfast items (25 fold difference).

Within side dishes, onion rings, fries (from sit-down and fast-food restaurants), baked potatoes with toppings, and fries with toppings were significantly higher in calories per serving compared to vegetables, coleslaw and soup (p<0.05) (Figure 3.2.c). The smallest range of calories within side dish categories was amongst baked potatoes (250 calories per serving, a 2.2 fold difference) while the largest range was found in sit-down restaurant-fries (690 calories per serving, a 3.7 fold difference).

All food categories exhibited immense variation in serving size (Table 3.1), with relative standard deviations ranging from 21% in fast-food restaurant salads with meat to 59% in fast- food restaurant tacos/burritos. Among sit-down restaurant sandwiches/wraps there was a seven fold difference in serving size, while in hamburgers there was almost an eight fold difference. Pearson’s correlation showed that both serving size and caloric density were positively correlated with calories (p<0.05). However, serving size was more strongly correlated (r=0.62) compared to caloric density (r=0.29).

In order to further examine the relative effects of caloric density and serving size on calories per serving, foods were grouped according to intervals of 100 calories and the serving size and caloric density of each calorie grouping is shown in Figure 3.3. This figure illustrates that as calories increased, serving sizes markedly increased while caloric density did not (with the exception of the first and second intervals). When comparing calorie intervals containing 600-1000+ calories, there were statistically significant differences in serving size (with higher calorie intervals having a significantly larger serving sizes compared to lower calorie intervals), however there were no significant differences in caloric density. For example, items with 1000+ calories were 62% larger than items containing 600-700 calories (p<0.001), but were not significantly different with respect to caloric density. Similar results were found when comparing mean serving size and caloric density among items within the 200-700 calorie interval range, as those containing 600-700 calories were significantly larger than all intervals containing 200-500 calories (p<0.001), but were not significantly different in terms of caloric density.

Table 3.1. Comparison of serving size, calories per serving and caloric density in sit-down and fast-food restaurantsa

Caloric Density Calories per (Calories/ n Serving size (g) Serving (kcal) 100g) Meal Items Mean±SD Range Mean±SD Mean±SD Breakfast SDR 108 453±156 (85-865) 850±299 195±528 FFR 232(176) b 155±81*** (35-658) 357±163*** 247±74*** Chicken SDR 39(31) 269±129 (115-665) 509±287 195±83 FFR 55(51) 129±61*** (41-368) 279±143*** 223±68* Hamburgers SDR 65(59) 372±132 (99-770) 926±284 281±123 FFR 81(71) 235±83*** (80-481) 635±229*** 263±25 Pasta Entrées SDR 140(128) 548±155 (215-1096) 946±322 171±43 FFR 30(26) 297±127*** (100-420) 448±178*** 162±43 Salad Entrées SDR 44(36) 272±100 (71-612) 425±180 166±66 FFR 53(39) 281±93 (56-479) 297±141*** 116±62*** Salad Entrées with SDR 93(83) 423±126 (170-691) 584±230 138±60 Meat FFR 118(89) 378±80** (221-602) 453±188*** 114±46*** Sandwiches/Wraps SDR 159(122) 391±131 (114-810) 756±260 202±64 FFR 513(356) 254±87*** (85-555) 469±163*** 189±59* Stir Fry SDR 25(24) 678±109 (469-869) 1017±308 145±36 FFR 39 468±57*** (342-543) 500±109*** 107±23*** Tacos/Burritos SDR 10 311±79 (190-410) 687±227 220±36 FFR 116(115) 304±178 (78-662) 507±278* 172±57*** Sides Fries SDR 22(19) 232±57 (142-341) 558±180 240±66 FFR 17(16) 173±66** (113-312) 425±161* 255±59 Soup SDR 98(79) 280±50 (223-454) 197±115 67±34 FFR 204(175) 308±172 (100-964) 162±93** 52±20*** FFR, fast-food restaurants; SDR, sit-down restaurants a Only categories with ten or more items in both SDR and FFR were included in the table. b (n) represents the n for caloric density and serving size because some establishments did not provide serving size data *p<0.05, **p<0.01, ***p<0.001 between SDR and FFR

Figure 3.1. Median and range of calories per serving in menu items (including main items and side dishes, excluding desserts) from sit-down restaurants (n=1390)

Note: Values do not depict the caloric content of an entire meal. Bars represent the interquartile range and lines represent the minimum and maximum values. Kruskal- Wallis ANOVA was significant (P < 0.001), multiple comparisons confirmed significant differences between establishments.

Figure 3.2.a. Median and range of calories in different food categories from sit-down restaurants (n=1390) arranged in increasing order by median calorie intake

Note: Bars represent the interquartile range and the lines represent the minimum and maximum values. Kruskal-Wallis ANOVA was significant (P < 0.001), multiple comparisons confirmed significant differences between different food categories.

Figure 3.2.b. Median and range of calories in different food categories from fast-food restaurants (n=2800) arranged in increasing order by median calorie intake

Figure 3.2.c. Median and range of calories in different side-dish (n=840) food categories, arranged in increasing order by median calorie intake

Figure 3.3. Mean serving size and caloric density of meal items and side dishes grouped according to 100 calorie intervals

Note: Kruskal-Wallis ANOVA was significant (p<0.001), multiple comparisons confirmed some significant differences in serving size and caloric density amongst calorie intervals. 3.5. Discussion

The most important finding from this study is that there is tremendous variation in calories per serving between and within food categories, restaurant establishments and types of establishments. Furthermore, serving size, rather than caloric density, is the main factor influencing the calories per serving in restaurant foods.

This variation among food choices within these restaurants indicates that customers are faced with a wide range of calorie options when making food selections and that following general rules for healthy eating does not guarantee that a customer will be able to make the lowest calorie choice. For example, although salads are perceived as containing fewer calories, and on average would be expected to contain fewer calories compared to a stir fry or rib meal; 50% of salads contained more calories per serving compared to the lowest ranking stir fry or rib meals.

Furthermore, due to the strong influence of serving size on calorie content, the introduction of calorie labeling may prompt restaurants to reduce the size of their items as a

35 means to decrease calories and thus make their menu items appear healthier. This would be a positive change due to the aforementioned prevalence of large serving sizes,20 and the fact that large serving sizes in restaurants have been shown to contribute to excess energy intakes.21 The data from this study show that, many lower calorie foods in restaurants contain fewer calories compared to higher calorie foods by virtue of their size. Therefore, the results of this study demonstrate that the most feasible approach to decreasing calories in restaurant foods would be to decrease serving sizes; as serving size, not caloric density, is the largest determinant of differences in calories. In addition, these results suggest that including serving size information alongside calorie labeling may be useful, as it would indicate why one item may be higher in calories compared to another, or in some circumstances it would help customers decide between items that may be similar in calories, but different in terms of serving size.

Although these data suggest that reducing serving size is necessary to decrease calories in the restaurant sector, focusing only on reducing serving size as a means to decrease calories could potentially be problematic as it neglects the fact that calories are not the only determinant of the healthfulness of a food. Additionally, a potentially negative consequence of reductions in serving size is that it could fuel the prevalence of small, calorie dense meal items. This is worrisome because consuming a smaller portion of an energy dense food, so as to consume fewer calories, is not the most effective strategy for weight-loss. Studies have shown that a diet consisting of low-energy-dense foods is advantageous because it enables the consumption of larger satisfying portions.22, 23 Thus, these cautionary observations suggest that in some circumstances reducing serving size could undermine weight-loss/weight-maintenance efforts.

3.5.1. Study Limitations

There are a few limitations of the present study, such as the lack of data for independent/privately owned restaurants and incomplete information for some establishments. It should also be noted that the accuracy of the findings presented in this study are dependent upon the accuracy of the data provided by the establishments. However, a recent study by Urban et al22 provides some justification for the accuracy of the calorie data as they showed that despite the existence of substantial inaccuracies for some items, stated energy content of restaurant foods were accurate overall, when compared to analyzed data. It is also important to address the caloric differences that were detected between sit-down and fast-food restaurants. Although the data

36 presented in Table 3.1 suggests that meal items from fast-food restaurants are lower in calories, this may not necessarily be the case for all choices. Many sit-down restaurant categories, particularly those that were low in calories, such as seafood and beef entrées, were not included in the comparison between sit-down and fast-food restaurants, because comparable items were not sold in fast-food restaurants. Finally, this study only analyzed the calorie content of foods, irrespective of other healthy or unhealthy nutrients. More research is needed to assess the overall nutritional profile of restaurant meals and to test whether a lower calorie content has any predictive value in identifying healthier, more nutritious foods.

3.6. Conclusion

Overall, there was immense variation in the caloric content of menu items between and within food categories, restaurant establishments and types of establishments. In addition, serving size was found to be the most important contributor to variation in calories per serving in foods from sit-down and fast-food restaurants. Thus, if calorie labeling encourages restaurants to reduce the calorie content of their menu items, the results of this study demonstrate that decreasing serving size is an important strategy for decreasing calories in the restaurant sector. Nevertheless, because calories are not the only determinant of the healthfulness of a food item, more research is needed to investigate other methods of menu labeling that incorporate information beyond just calories. In conclusion, even though calories are not a fool-proof indicator of the healthiest items in sit-down and fast-food restaurants, the variation demonstrated in this study indicates that calorie labeling is a major first step in the right direction towards informing customers of the wide range of calories available when dining outside the home.

Contributions Mary Scourboutakos and Dr. Mary L’Abbé conceived the study, designed its protocol, and interpreted the data. Mary Scourboutakos acquired and analyzed the data, and was responsible for drafting the manuscript.

Chapter 4 Sodium levels in Canadian sit-down and fast-food restaurants (CJPH, 2013)

This manuscript has been published: Scourboutakos MJ, L’Abbé MR. 2013. Sodium levels in Canadian fast-food and sit-down restaurants. Canadian Journal of Public Health, 104(1)e2-e8. [http://journal.cpha.ca/index.php/cjph/article/view/3683]

4.1. Abstract

Objective: To evaluate the sodium levels in Canadian restaurant and fast-food chain menu items.

Methods: Nutrition information was collected from the websites of major sit-down (n=20) and fast-food (n=65) restaurants across Canada in 2010 and a database was constructed. Four thousand and forty-four meal items, baked goods, side dishes and children’s items were analyzed. Sodium levels were compared to the recommended adequate intake level (AI), tolerable upper intake level (UL) and the US National Sodium Reduction Initiative (NSRI) targets.

Results: On average, individual sit-down restaurant menu items contained 1455 mg sodium/serving (or 97% of the AI level of 1500 mg/day). Forty percent of all sit-down restaurant items exceeded the AI for sodium and more than 22% of sit-down restaurant stir fry entrées, sandwiches/wraps, ribs, and pasta entrées with meat/seafood exceeded the daily UL for sodium (2300 mg). Fast-food restaurant meal items contained on average, 1011 mg sodium (68% of the daily AI), while side dishes (from sit-down and fast-food restaurants) contained 736 mg (49%). Children’s meal items contained on average, 790 mg/serving (66% of the sodium AI for children of 1200 mg /day); a small number of children’s items exceeded the children’s daily UL. On average, 52% of establishments exceeded the 2012 NSRI density targets and 69% exceeded the 2014 targets.

Conclusion: The sodium content in Canadian restaurant foods is alarmingly high. A population- wide sodium reduction strategy needs to address the high levels of sodium in restaurant foods.

4.2. Introduction

High dietary sodium intake, a causal risk factor for hypertension,123 is the leading preventable risk factor for death worldwide.125 Sixty-two percent of strokes and 49% of coronary heart disease are attributed to hypertension.144 While it is recommended that individuals should aim to consume around 1500 mg of sodium and no more than 2300 mg each day,145 the average Canadian presently consumes 3400 mg per day.127 Reducing Canadians’ dietary sodium intake by 1800 mg/day has been estimated to result in annual health care savings of 2.33 billion dollars.146

Seventy-seven percent of dietary sodium is derived from processed and restaurant foods.128 Hence, eating-out has been shown to be associated with higher dietary sodium intakes.21 Presently, approximately 25% of Canadians eat something prepared in a fast- food outlet on any given day and an additional 21% eat something prepared in a sit-down restaurant, cafeteria or other food venue.15, 16 In response to this situation, Canada’s Sodium Working group created a plan to track sodium reductions;51 however, the Federal Health Minister prematurely disbanded the sodium working group before it could establish the sodium monitoring system. Furthermore, in 2011 the Federal Health Minister rejected the sodium reduction monitoring plan proposed by federal and provincial officials.147 With these set-backs, the current sodium levels in Canadian restaurant foods remain unknown.

In the US, the National Salt Reduction Initiative (NSRI) was the first to develop voluntary targets for sodium reduction in restaurant foods.148 However, there have been no published analyses examining the current sodium levels in relation to these targets. Therefore, considering the prevalence of eating outside the home,15 as well as the pervasiveness of hypertension and its associated health risks,149 characterizing the sodium content of food items from sit-down and fast-food restaurants is exceedingly important. Furthermore, because of the lack of progress towards reducing sodium levels in Canadian restaurant foods, a comprehensive baseline assessment of current sodium levels in restaurant foods is necessary in order to create reformulation strategies and to monitor progress.

The objective of this study was to systematically evaluate the sodium levels in a wide variety of meal items, baked goods, side dishes and children’s items from Canadian sit-down and

39 fast-food restaurants. Sodium levels were evaluated in relation to the AI (daily adequate intake level, 1500 mg per day) and UL (daily tolerable upper intake levels of 2300 mg per day);145 in addition, the number of restaurants that exceeded the NSRI targets was determined. We hypothesize that sodium levels in food from restaurant and fast food chains will be high and may exceed the daily recommended intake levels.

4.3. Methods

4.3.1. Database construction

This study was a systematic, cross-sectional survey of sodium levels in Canadian restaurant foods. Using the 2010 Directory of Restaurant and Fast-Food Chains in Canada,150 172 fast-food and sit-down restaurants were identified as having 20 or more locations nationally. The website of each of these restaurants was visited, in order to determine if nutrition information was available online, or if the restaurants indicated that information was available in-store. When an establishment had a .ca and .com web-address, data were derived from the .ca address. In total, 95 establishments provided nutrition information online. Of the top 50 restaurants (according to number of locations),150 42 (84%) provided nutrition information online and thus were included in the study. Restaurants that were not included due to lack of data, tended to be smaller chains. Of the 95 restaurants included, four were excluded because their data were specific to the US, three were excluded because they were coffee shops that only provided data for a limited number of generic beverages, and lastly, one cafeteria supplier was also excluded. Data were collected between September and December 2010 (with the exception of four establishments, whose data were retrieved in early 2011). In total, over 9000 food items from 65 fast-food restaurants (FFR) (including fast-casual and coffee shops) and 20 sit-down restaurants (SDR, defined by the presence of table service) were included in the database. Most establishments provided data for the thirteen nutrients commonly found on the Nutrition Facts Table as well as calories and serving size. Food items were categorized according to the establishment and sub-categorized according to the type of food, as well as whether the food was considered a side dish, main entrée including side dishes, main entrée without side dishes or single items that could be purchased individually. When necessary, establishments were contacted via phone and/or e-mail to verify categorizations. To ensure that the data were entered

40 accurately by the first author, sodium levels were compared to the original website sources. In addition, sort and rank procedures were used to check for outliers, 5% of the database was checked by a third party and calculations were done using Atwater Factors to check for potential errors. When necessary, establishments were contacted to confirm suspicious outliers. Further details concerning the food categories and a list of establishments included in this study have been described elsewhere.151

4.3.2. Exclusion and Inclusion Criteria

Food categories containing 10 or more items were included, with the exception of: children’s meal categories, in which all categories containing 3 or more items were included, and side-dishes, in which categories with 8 or more items were included. In total, the analysis included 20 SDR and FFR categories, 5 baked-good categories, 14 side-dish categories and 19 children’s meal categories. Side-dishes, main entrées without side dishes, and single-items that can be purchased individually were included in the analysis, while main entrées including side- dishes were excluded, as were beverages, appetizers and condiments. When items were available in multiple-sizes, all sizes were included. As a result, a total of 4044 food items were included in the study. Frozen dessert and beverage categories were not reported as these foods are not common sources of sodium.

4.3.3. NSRI targets

Foods were compared to the US NSRI targets because restaurant sodium reduction targets or guidance values have not yet been established in Canada. The NSRI targets were set in 2009 and provide recommendations for sodium reductions per serving and per 100g, referred to as ―serving targets‖ and ―density targets‖, respectively. Targets that were established for 2012 and 2014 were used to benchmark the current levels in Canadian restaurants. The density targets recommended a mean sodium per 100g for all items in each food category at each establishment. These were based on a percent reduction from the market share-weight mean sodium density (calculated using the restaurant’s total sales volume as of 2008) in 25 different food categories. Serving targets were also set for the maximum amount of sodium per serving for all items (1500 mg/serving for 2012, and 1200 mg/serving for 2014).148

Food items in the database were sub-categorized according to the 25 food categories for which NSRI targets have been established. When necessary, the establishment’s menus were consulted and in some instances, establishments were contacted (via telephone or e-mail) to ensure accurate categorization. Several items were not included in the NSRI target analysis, for example, those that were in categories for which targets had not been established (such as pasta and salads), children’s items, and foods whose densities could not be calculated because the serving size was unknown.

4.4.4. Data Analysis

Descriptive statistics, including mean sodium per serving, sodium per 100g and %AI (sodium per serving divided by 1500 mg) were calculated for all menu items (not including accompaniments), side dishes, children’s meal items and children’s side dishes. The proportion of menu items that exceeded the daily sodium AI and UL were tabulated. The mean sodium density in each sodium reduction target category was calculated for both SDR and FFR establishments and the number of establishments whose mean sodium density exceeded the US NSRI targets was tabulated. The number of establishments with items that exceeded the serving targets was also tabulated. All data were analyzed using Statistica 10 software (Tulsa, Oklahoma).

4.5. Results

4.5.1. Sodium in single menu items compared to recommended daily intake levels

The mean, percentiles and range of sodium levels for each category from SDR as well as the mean sodium as a percentage of the AI and percentage of items in the category that exceeded the daily sodium AI and UL were determined (Table 4.1). On average, 40% of SDR menu items exceeded the AI for sodium, while 13% exceeded the UL. More than 22% of SDR sandwiches/wraps, ribs, and pasta entrées with meat/seafood exceeded the UL for sodium. Categories where the mean sodium per serving exceeded the daily recommended AI of 1500 mg were stir fry entrées 2360 mg, (157% AI); sandwiches/wraps, 1826 mg (122%); ribs, 1775 mg (118%); pasta entrées with meat/seafood, 1760 mg (117%); multiple-meat and/or seafood entrées

1746 mg, (116%); tacos/burritos, 1530 mg (102%); and hamburgers, 1517 mg (101%). On average, SDR meal items (not including side dishes) contained 1455 mg sodium/serving, or 97% of the adult daily AI. Categories with the lowest sodium levels per serving were seafood, 939 mg (63%); beef, 889 mg (59%); and salad entrées, 856 mg (57%).

Table 4.2 shows that similar trends were seen in FFR, although sodium levels tended to be lower. On average, FFR meal items provided 68% (1011 mg per serving) of the AI (daily recommended amount of sodium). The highest categories were stir fry entrées, 1953 mg (130% AI); poutine/fries with toppings, 1547 mg (103% ); nachos, 1402 mg (93%); tacos/burritos, 1322 mg (88%); sandwiches/wraps, 1287 mg (86%) and salads with meat/seafood, 1282 mg (85%). The range of sodium per serving within menu item categories varied from a 2 fold difference among stir fry entrées to a 78 fold difference among sandwiches/wraps. Despite the fact that SDR had more sodium per serving compared to FFR, additional analysis (not shown) did not show a clear trend explaining this finding. In some instances it was due to a larger serving size, while in other instances it resulted from a higher sodium density, or a combination of both larger serving size and higher sodium density.

On average, side dishes provided 49% (736 mg) of the AI (daily recommended amount of sodium) (Table 4.3). A number of side dishes, including some fries, soups and salads also exceeded the daily sodium UL. Side dishes that had the lowest sodium per serving were coleslaw (382 mg), vegetables (262 mg) and baked potatoes (165 mg).

4.5.2. Sodium in children’s meal items compared to recommended daily intake levels

On average, children’s menu items provided 790 mg sodium per serving, or 65% of the AI (daily recommended amount of sodium, 1200 mg for children aged four to eight) while children’s side dishes contained 377 mg (31% of the AI). In SDR, 33% of chicken items, as well as 18% of pizza meals and 17% of hamburgers exceeded the children’s daily sodium AI (Table 4.4). Very few menu items exceeded the children’s UL (1900 mg per day), although sodium levels were exceedingly high (>2000 mg) in some SDR children’s pizza items.

4.5.3. Sodium levels in restaurant foods compared to NSRI targets

A total of 1759 items from the 25 NSRI target categories were available for analysis (Table 4.5). On average, 52% of establishments exceeded the 2012 density targets in any given category. Categories where the majority of establishments exceeded the 2012 density targets were bakery products such as sweet yeast breads and cookies (where 75% of items exceeded targets), fried potatoes (73%), other sandwiches (72%), french fries (71%), sandwiches with luncheon meat (65%) and pizza (62%). On average, 69% of establishments exceeded the 2014 density targets. Categories where the majority of items exceeded the 2014 density targets were: cookies (92%), soups (87%), tacos (84%), bone-in breaded chicken (83%) and fried potatoes and onion rings (82%). With respect to serving targets, categories where the majority of establishments contained items that exceeded the targets were cheeseburgers (80% exceed 2012 target and 88% exceed 2014 target), chicken and fish sandwiches (67%, 91%), sandwiches with ham and cheese (84%, 97%), and burritos (80%, 100%) respectively.

Table 4.1. Sodium levels in Canadian sit-down restaurant menu items compared to the daily dietary reference intake (DRI) recommendations

% of meal menu items exceeding Sodium levels per category per serving (mg/serving) the daily sodium DRI levels† 10th 25th 50th 75th Food Category n Mean±SD %AI* Min Percentile Percentile Percentile Percentile Max AI: 1500mg UL: 2300mg Stir fry entrées 22 2360±1102 157 687 840 1705 2210 2922 4380 77 45 Sandwiches/wraps 164 1826±928 122 170 850 1223 1612 2270 6523 57 23 Ribs 29 1775±1031 118 320 520 1040 1430 2090 4599 45 24 Pasta entrées (including 111 1760±799 117 380 840 1150 1570 2260 4940 56 23 meat/seafood) Multiple-meat and/or 27 1746±918 116 495 760 1030 1557 1997 4000 59 11 seafood entrées Tacos/burritos 12 1530±455 102 941 971 1244 1430 1780 2330 42 8 Hamburgers 65 1517±568 101 733 940 1140 1412 1810 3880 40 8 Breakfast 120 1473±670 98 210 625 1014 1355 1932 3180 43 13 Pasta (just containing 75 1411±644 94 439 750 880 1310 1700 3360 35 12 sauce/cheese) Salads with 96 1242±584 83 110 590 849 1191 1536 3243 28 3 meat/seafood Chicken entrées 57 1130±682 75 127 310 680 980 1540 3210 26 7 Seafood entrées 50 939±736 63 119 196 306 755 1260 2980 24 6 Beef entrées 79 889±751 59 55 160 311 700 1300 3780 18 5 Salad entrées 44 856±448 57 170 301 544 812 1130 2260 7 0 * Mean sodium level in category, expressed as a percentage of the daily adequate intake (AI) for adults (1500 mg per day) as defined by the IOM (4). † Percentage of products in the category that exceeded the daily adult AI (1500 mg/day) or UL (2300 mg/day) per serving.

Table 4.2. Sodium levels in Canadian fast-food restaurant menu items compared to the daily dietary reference intake (DRI) recommendations

% of menu items exceeding Sodium levels per category per serving (mg/serving) the daily sodium DRI levels† 10th 25th 50th 75th Food Category n Mean±SD %AI* Min Percentile Percentile Percentile Percentile Max AI:1500mg UL:2300mg Meal items Stir fry entrées 38 1953±312 130 1396 1673 1695 1953 2090 2740 95 18 Poutine/Fries with 51 1547±680 103 610 650 990 1380 2090 2760 41 18 Toppings Nachos 10 1402±878 93 435 493 600 1313 2108 3100 40 10 Tacos/burritos 144 1322±304 88 304 440 682 1207 1742 3996 33 10 Sandwiches/wraps 611 1287±595 86 46 625 862 1200 1565 3600 28 8 Salads with 116 1282±405 85 380 830 953 1208 1539 2313 29 1 meat/seafood Hot dogs 20 1158±268 77 810 835 880 1124 1390 1618 5 0 Hamburgers 81 1131±366 75 510 680 850 1150 1400 2300 12 1 Sushi 23 790±239 53 380 490 680 710 940 1355 0 0 Stir fry entrées (low 63 790±381 53 26 307 440 850 1110 1510 0 0 sodium/no sauce) Chicken 65 733±392 49 0 250 430 690 960 1850 4 0 Salad entrées 51 732±383 49 50 355 484 637 900 1945 8 0 Breakfast items 219 713±427 48 0 250 420 620 960 2240 6 0 Pasta entrées 32 675±317 45 200 320 432 635 820 1360 0 0 Pizza (one slice of a 369 475±213 32 180 270 340 430 560 1740 0 0 medium pizza) Baked Goods Muffins 117 423±171 28 80 240 330 387 507 1110 0 0 Other Baked Goods (tea biscuits, brownies, 63 383±270 26 45 110 180 310 530 1090 0 0 tarts, scones, loafs) Donuts 95 326±133 22 41 220 255 286 360 997 0 0 Pastries 31 307±93 21 95 170 240 320 360 510 0 0 Cookies 73 179±101 12 70 100 120 140 200 630 0 0 *Mean sodium level in category, expressed as a percentage of the daily adequate intake (AI) for adults (1500 mg per day) as defined by the IOM (4). †Percentage of products in the category that exceeded the daily adult AI (1500 mg/day) or UL (2300 mg/day) per serving.

Table 4.3. Sodium levels in Canadian fast-food and sit-down restaurant side dishes compared to the daily dietary reference intake (DRI) recommendations

% of menu items exceeding Sodium levels per category per serving (mg/serving) the daily sodium DRI levels† 10th 25th 50th 75th Food Category n Mean±SD %AI* Min Percentile Percentile Percentile Percentile Max AI:1500mg UL:2300mg SDR soup 98 1060±553 71 161 510 745 975 1240 3200 13 5 FFR soup 204 1029±438 69 455 750 810 920 1080 3630 8 2 SDR fries 28 884±457 59 45 85 659 900 1151 1758 7 0 Mashed potatoes 16 834±389 56 350 432 555 820 920 1660 13 0 Onion rings 15 749±426 50 226 330 437 626 830 1760 13 0 Roasted potatoes 8 720±439 48 110 110 414 680 1007 1450 0 0 FFR fries 37 719±489 48 108 165 400 610 980 2120 8 0 Rice 25 622±623 41 0 12 140 630 841 2535 12 4 Salad 93 487±374 32 25 125 280 424 574 2810 1 1 Baked potato with 14 464±370 31 40 80 170 449 660 1482 0 0 toppings Coleslaw 15 382±205 25 67 160 230 360 520 890 0 0 Vegetables 45 262±370 17 0 16 42 140 341 2150 2 0 Baked potato 9 165±236 11 1 1 28 40 220 666 0 0 *Mean sodium level in category, expressed as a percentage of the daily adequate intake (AI) for adults (1500 mg per day) as defined by the IOM (4). † Percentage of products in the category that exceeded the daily adult AI (1500 mg/day) or UL (2300 mg/day) per serving.

% of menu items exceeding the Sodium levels per category per serving (mg/serving) daily sodium DRI levels† 10th 25th 50th 75th Food Category n Mean±SD %AI* Min Percentile Percentile Percentile Percentile Max AI:1200mg UL:1900mg Sit-down restaurant children's meal items Tacos/Burritos 3 1231±592 103 670 670 670 1174 1850 1850 33 0 Pizza 11 1076±529 90 470 545 849 990 1199 2100 18 18 Chicken 6 1021±633 85 430 430 525 865 1420 2020 33 17 Sandwiches/Wraps 16 932±284 78 330 710 771 877 1123 1640 5 0 Chicken Nuggets/Strips 13 888±287 74 210 628 744 880 1040 1380 8 0 Hamburgers 12 815±376 68 237 465 595 699 1020 1580 17 0 Breakfast 17 798±288 67 170 424 587 790 1020 1250 6 0 Pasta 28 705±493 59 115 220 345 640 945 2423 11 4 Seafood 4 520±440 43 70 70 154 499 886 1012 0 0 Fast-food restaurant children's meal items Sandwiches/Wraps 30 768±219 64 276 390 552 710 894 1106 0 0 Hamburgers 13 623±143 52 400 470 550 630 650 940 0 0 Chicken Nuggets/Strips 7 567±237 47 210 210 408 520 800 900 0 0 Tacos/Burritos 6 465±199 39 300 300 320 390 592 800 0 0 Children's side dishes Fries (SDR) 7 536±378 45 67 67 250 469 890 1127 0 0 Soup (SDR) 9 507±159 42 120 120 520 540 580 680 0 0 Fries (FFR) 5 372±237 31 71 71 270 299 560 660 0 0 Potaotes (SDR) 6 388±198 32 129 129 206 405 531 650 0 0 Salads (SDR) 11 358±313 30 75 122 125 280 420 1190 0 0 Vegetables (SDR) 10 165±214 14 1 1 15 80 255 550 0 0 *Mean sodium level in category, expressed as a percentage of the daily adequate intake (AI) for children (1200 mg per day) as defined by the IOM (4). † Percentage of products in the category that exceeded the children’s daily adult AI (1200 mg/day) or UL (1900 mg/day) per serving.

Table 4.5. Canadian and US restaurant sodium levels in comparison to the US National Salt Reduction Initiative (NSRI) Targets

Meal/ 2012 2014 Restaurants menu Mean sodium Level 2014 Sodium Density Serving Serving in Category items (mg/100g) 2012 Sodium Density targets targets targets targets

% Canadian % Canadian restaurants restaurants % Canadian % Canadian with items with items Target restaurants > Target restaurants > >1500mg/ >1200mg n n Canada* US† (mg/100g)‡ target§ (mg/100g)‡ target§ serving|| /serving|| HAMBURGERS Hamburgers 19 41 413 403 380 42 330 68 20 40 Cheeseburgers 17 85 507 539 460 59 410 71 80 88

CHICKEN Boneless Breaded Chicken 18 32 674 738 670 50 590 56 32 47 Bone-in Breaded Chicken 6 23 626 688 620 50 550 83 17 33 Bone-in Chicken without 11 32 428 518 440 36 390 45 27 36 Breading

SEAFOOD Breaded Seafood 8 10 655 751 680 37 560 50 38 38

SANDWICHES Chicken and Fish Sandwiches 33 154 520 572 520 45 460 67 67 91 Sandwiches with Ham and 32 204 677 628 590 41 500 62 84 97 Cured Meat Sandwiches with Luncheon 14 61 549 503 480 64 430 64 57 79 Meat Other Sandwiches (Ex. Cheese steak, grilled cheese, 32 175 458 415 390 72 370 75 56 78 tuna)

BREAKFAST Breakfast Sandwiches on a 4 28 637 836 770 0 630 25 25 25 SANDWICHES Biscuit Breakfast Sandwiches not on 13 61 509 657 560 46 520 46 46 62 a Biscuit

Pizza (Cheese Pizza and PIZZA 8 16 470 530 460 62 390 75 13 0 cheese pizza base)

MEXICAN Burritos 5 66 393 601 510 20 450 40 80 100 Tacos 6 49 427 464 410 50 350 84 33 50

POTATOES French Fries 31 60 440 347 290 71 240 71 13 26 Fried Potatoes & Onion Rings 11 18 640 518 460 73 380 82 36 27

SOUP Soups 30 267 359 395 340 57 280 87 32 55

Savoury Yeast Breads BAKERY without Salty Additions 3 4 471 457 410 67 360 67 0 0 PRODUCTS (Bagels, croissants etc.) Savoury Yeast Breads with Salty Additions (Ex. cheese, 3 5 561 543 470 67 410 100 0 0 meat) Sweet Yeast Breads 12 49 358 290 280 75 250 75 0 0 (danishes, donuts etc. ) Sweet Quick Breads (muffins, 27 171 309 288 280 48 250 59 0 7 scones, cake etc.) Pies and Turnovers 18 34 196 231 220 50 200 50 0 0 Biscuits 3 7 733 932 800 33 700 67 0 0 Cookies 12 58 381 354 310 75 260 92 0 0 * In some categories the Canadian mean sodium density includes multiple sizes of the same items. † US data are from the New York City Department of Health and Mental Hygiene’s ―National salt reduction initiative restaurant food categories and targets‖ (12). The American mean sodium density is market share weighted, Canadian data are not. ‡ The targets are market share weighted and apply to the mean sodium in all of a restaurant's items in each food category. § Density targets apply to the mean sodium among all of the restaurant's items in each food category. This data represents the percentage of restaurants where the mean for the items in the category for a particular restaurant exceeded the target for that category. || Serving targets represent the maximum sodium that is allowed per serving for all items. This data represents the percentage of restaurants that had items that exceeded serving targets.

4.6. Discussion

This study provides the first systematic assessment of sodium levels in a wide variety of menu items from Canadian sit-down and fast-food restaurants, which can be used as a baseline to assess progress in this sector, as it appears that Health Canada has discontinued further work in this area. The results showed that the average sodium levels in Canadian restaurant foods were extremely high. A large number of menu items (that in most cases do not constitute an entire meal, such as a hamburger not including the fries that may be consumed alongside it), and even a number of side dishes and children’s items, exceeded the daily recommended AI and UL, and also exceeded the US NSRI sodium reduction targets that have been established for 2012 and 2014. Considering the prevalence of food consumed outside the home,15 along with the high rates of hypertension and cardiovascular disease,149 and the associated economic consequences,146 the results of this study demonstrate the need for increased efforts focusing on restaurants as a key area where sodium reduction is necessary and has to date been overlooked.

It has been recommended that many groups including: individuals that are 51+ years old, African Americans, as well as individuals with chronic kidney disease, high blood pressure (approximately 19% of the Canadian population)152 or diabetes (26% of Canadians have diabetes or prediabetes)153 should consume no more than 1500 mg daily.126 Thus, considering that 40% of SDR menu items and 18% of FFR menu items (that in many cases do not constitute an entire meal) exceeded this daily cut-off (1500 mg), our data suggests that eating-out on a regular basis can be harmful, particularly among the large proportion of at-risk adults.

The large range of sodium within food categories supports Dunford et al’s conclusion that this variation demonstrates both the technical feasibility and taste acceptability of lower sodium products.154 Large ranges in sodium content also suggest that general advice about selecting low sodium options is not sufficient without on-site menu-labeling. For example, SDR salad entrées had the lowest mean sodium (856 mg per serving), yet some items still contained up to 2260 mg per serving. Therefore, because of the wide range of sodium per serving, on-site sodium labeling in restaurants may be necessary in order to clearly inform customers of the sodium content of menu options and enable them to make healthy choices.

In their study of fast food purchases, Johnson et al, found that excess sodium was not only the result of large portion sizes but was also due to a high sodium density.30 The data in table 4 showed that the average sodium density in many restaurants exceeded the recommended targets for reduction. Therefore, because of the high sodium density, reduction strategies cannot exclusively rely on decreasing portion sizes as a means to decrease sodium and therefore must emphasize decreases in sodium density as a means to decrease sodium levels per serving.

The high sodium levels within children’s items (meal items on average contained 32% of the daily recommended amount of sodium while children’s side dishes contained 16%) suggests the need to establish reduction targets specifically for children’s items, as it has been shown that frequent consumption of fast-food among adolescents, may alter taste perception and promote an increased preference for salt.155 Nevertheless, it has been demonstrated that small to moderate sodium reductions that are introduced slowly are not easily detected,156 and can lead to a preference for a lower sodium diet.157

The Canadian and American mean sodium densities were often similar even though the Canadian means were not market-share weighted (Table 5). Whether this produced inflated or conservative results is uncertain. A study of the sodium content in processed foods in the UK showed that purchase-weighted mean sodium was 18-35% higher than un-weighted mean sodium levels.158 This indicates that future research is needed to determine the degree to which market share influences these results. Our study examined sodium levels in chain restaurants and did not include independent establishments. Furthermore, the findings presented in this study are dependent upon the accuracy of the data provided by the establishments. In some instances sodium levels could vary and the validity of the industry reported sodium data has not been verified. The use of sodium’s AI and UL as benchmarks for sodium content was a very conservative approach, as these represent daily sodium intake levels and are not intended to be applied to single meal items. Furthermore, our study did not combine main entrées with the side dishes that would often accompany them, as they would typically be consumed, which would further increase the amount of sodium consumed when eating-out. More research is needed to demonstrate the sodium levels in whole meals and combos served at SDR and FFR as the data suggest that the sodium levels in complete meals would be dangerously high. Finally, even though our data were collected in 2010 and early 2011, and were compared to 2012 and 2014

52 targets, it is unlikely that there have been major decreases in sodium levels over the past one and a half years, as Canada has not yet established targets or implemented a reduction strategy for the restaurant sector.

A strength of this study is that percentiles were reported so that the data can be used by Health Canada to set sodium reduction targets for restaurant foods, a large gap in the current Canadian sodium reduction guidance to industry that was published earlier this year. In this study, data have been reported in a similar format as that used by Health Canada for setting reduction targets for packaged foods;159 therefore, the results presented here can be used to establish baseline sodium levels in Canadian restaurants and can guide target setting for this sector, reformulation strategies, and future longitudinal studies to assess sodium reduction progress in this sector. Given the large variability in sodium levels in restaurant foods, these results also demonstrate the value of menu-labelling in aiding consumers to select lower sodium menu options when eating out.

4.7. Conclusion

In conclusion, the large number of individual restaurant menu items that exceeded the daily AI and UL along with the small number of establishments that meet the US NSRI targets demonstrates the need for a Canadian sodium reduction strategy that also emphasizes reductions in restaurant foods along with packaged foods. Because of the prevalence of eating-out, as well as the high rates of hypertension and cardiovascular disease, addressing the exceedingly high sodium levels in restaurant foods is essential in order to decrease the burden of chronic disease.

Contributions Mary Scourboutakos and Dr. Mary L’Abbé conceived the study, designed its protocol and interpreted the data. Mary Scourboutakos acquired and analyzed the data, and was responsible for drafting the manuscript.

Chapter 5 A longitudinal study of changes in sodium levels in a sample of Canadian chain restaurant foods from 2010 to 2013 (CMAJ Open, 2014)

This article has been published: Scourboutakos MJ, L’Abbé MR. 2014. Changes in sodium levels in chain restaurant foods in Canada (2010-2013): a longitudinal study. CMAJ Open, 2(4):E343-E351.

5.1. Abstract

Background: Several restaurant chains have committed to reducing sodium in their foods; however the extent to which sodium levels have changed over the past few years is unknown. The objective was to measure changes in sodium in restaurant foods from 2010 to 2013.

Methods: Sodium, serving size and calorie data for 3878 foods were collected from the websites of 61 Canadian restaurant chains in 2010 and 2013. A longitudinal study of changes in sodium among foods present in 2010 and 2013 (n=2198) was conducted. Levels in newly reported and discontinued foods were also investigated.

Results: Sodium levels (mg/serving) decreased in 30% of foods, increased in 16% of foods, and were unchanged in 54% of foods. The average change among foods whose sodium level decreased was 220±303 mg/serving (a 19±17% decline), while the average change among foods whose sodium level increased was 251±349 mg/serving (a 44±104% increase). The prevalence and magnitude of change varied depending on the restaurant and food category. Overall there was a small, yet significant decrease in sodium per serving (-25±268mg, p<0.001), however, the percentage of foods exceeding the daily sodium adequate intake (1500 mg) and tolerable upper intake level (2300 mg) remained unchanged.

Interpretation: The observed increases and decreases in sodium illustrate that industry efforts to voluntarily decrease sodium levels in Canadian restaurant foods have produced inconsistent results. While the lower levels seen in some foods show that sodium reduction is possible, the

54 simultaneous increase in other foods demonstrates the need for a national sodium reduction strategy.

5.2. Background

Excessive sodium consumption is a causal risk factor for hypertension,123 which is the leading preventable risk factor for death worldwide.125 Currently, 85% of Canadian men and 63- 83% of women have sodium intakes that exceed the daily recommended tolerable upper intake level (UL, 2300 mg).127 Reducing Canadians’ sodium intakes to the recommended level could produce an annual health care savings of $2.33 billion by decreasing the rate of hypertension and cardiovascular disease events by 30 and 13%, respectively.146 Research has shown that 77% of dietary sodium is derived from processed and restaurant foods. This may be partly due to the prevalence of eating outside-the-home,15 as well as the high sodium levels found in restaurant foods.160, 161

Efforts such as the United States’ National Sodium Reduction Initiative have been established to promote sodium reduction.133 In addition, several restaurants and food companies have made voluntary commitments to lower sodium levels.162, 163 Although Canada’s Sodium Working Group created a plan to track sodium reductions,51 they were disbanded before a monitoring system could be implemented.147 Research conducted in 2010 showed alarmingly high sodium levels in Canadian restaurant foods,160, 161 however, according to the Canadian Restaurant and Food Service Association, many restaurants have decreased their sodium levels since this time.164 Nevertheless, no systematic studies have been conducted to investigate the accuracy of such assertions by the restaurant industry. Therefore, the objective of this study was to measure changes in sodium levels in Canadian restaurant foods from 2010 to 2013.

5.3. Methodology

We conducted a longitudinal follow-up to a previous report investigating the sodium levels in Canadian fast-food and sit-down restaurants.160

5.3.1. Restaurants Included in the Study

The restaurants included in the study were identified using the Canadian Directory of Restaurant and Fast-Food Chains.150 Of the 172 restaurants that had >20 locations across Canada (20 was selected because this is the cut-off for mandatory menu-labelling according to the Affordable Care Act in the United States165), 95 provided nutrition information online in 2010 (data for three restaurants were collected in early 2011). Data was downloaded and compiled into the University of Toronto Restaurant Database, which contains nutrition information for over 9000 a la carte entrées, side dishes, beverages, desserts and condiments. Details concerning the database have been described elsewhere.151

In May 2013 all 95 restaurants’ websites were revisited, and when available, their nutrition data were downloaded. 34 restaurants were excluded for the following reasons: they did not provide serving size data (n=8), they only provided data for beverages or ice cream (n=9), they only provided American data (n=7), they no longer provided publicly available nutrition information (n=4), they did not provide sodium data (n=3), they substantially changed the format of the data they provided (n=2), or they were a cafeteria supplier (n=1).

Table 5.1. Characteristics of the 61 restaurants included in the study

Total number of foods that satisfied the study inclusion criteria and for which Total data was Number of number of provided in Number of Newly Total number of outlets in 2010 and Discontinued Reported Restaurant outlets in 2010 2013 2013 Foods Foods Type 241 Pizza 96 95 10 0 9 FFR* A&W 701 772 34 4 6 FFR Arby’s 112 53 28 10 6 FFR Baton Rouge 28 30 16 0 0 SDR Bento Nouveau 31 31 5 11 19 FFR Boston Pizza 333 345 106 30 50 SDR† Burger King 305 317 51 10 3 FFR Casey's Bar and Grill 37 27 46 28 17 SDR Coffee Time 237 175 30 0 0 FFR Country Style 533 485 38 93 0 FFR Dagwoods Sandwiches and 27 25 38 8 9 FFR Salads Dairy Queen 507 501 18 19 3 FFR Denny's 50 53 124 0 0 SDR Druxy's Deli 48 45 6 0 4 FFR Earl's Restaurant 61 60 28 34 9 SDR East Side Marios 114 89 45 15 12 SDR Edo Japan 86 103 18 2 19 FFR The Extreme Pita 104 207 43 0 0 FFR Flying Wedge Pizza 20 18 14 5 14 FFR Harvey’s 269 250 25 0 0 FFR Jack Astor's 29 34 24 65 19 SDR Joey's Restaurants 78 74 20 19 37 SDR Jugo Juice 105 130 5 3 0 FFR Kelsey’s 103 97 44 28 27 SDR KFC 731 668 24 31 3 FFR Little Caesars 131 179 14 0 0 FFR Manchu Wok 77 77 14 0 0 FFR McDonald’s 1419 1417 46 32 33 FFR Mikes Restaurants 90 85 80 59 55 SDR Mmmuffins 33 23 46 0 75 FFR Montana’s 87 91 75 17 24 SDR Mr. Greek 26 21 35 4 3 SDR/FFR‡

Mr. Sub 390 339 49 13 5 FFR Mrs. Vanelli's Fresh Italian Foods 65 47 28 0 0 FFR New Orleans Pizza 73 68 20 6 35 FFR New York Fries 173 130 5 0 3 FFR Opa! Souvlaki of Greece 63 87 6 3 9 FFR Orange Julius 67 81 7 1 0 FFR Panago 174 182 171 0 64 FFR Pita Pit 114 149 9 21 0 FFR Pizza 73 79 89 22 2 9 FFR Pizza Delight 98 91 34 20 21 SDR Pizza Hut 339 316 73 65 49 SDR Pizza Nova 120 130 11 0 0 FFR Pizza Pizza 550 604 3 0 0 FFR Pizzaville 63 73 25 1 0 FFR Robin's Donuts 130 140 23 0 0 FFR Scores Rotisserie 34 42 30 11 7 SDR Shoeless Joe’s 34 32 18 39 12 SDR Subway 2467 2896 51 38 60 FFR Swiss Chalet 198 208 55 8 0 SDR Taco Bell 191 196 32 9 7 FFR Taco Del Mar 61 33 27 27 23 FFR Taco Time 118 117 18 12 0 FFR Teriyaki Experience 104 109 74 0 0 FFR The Great Canadian Bagel 37 30 85 0 0 FFR Tim Hortons 2995 3437 70 15 24 FFR Treats 86 75 6 12 18 FFR Van Houtte's Bistro 54 64 24 0 0 FFR White Spot Legendary 64 63 52 30 16 SDR Restaurant White Spot Triple O's 45 36 20 0 2 FFR *FFR, fast-food restaurants, encompassing fast-casual restaurants, quick-service restaurants and coffe shops. †SDR, sit-down restaurants, defined as restaurants with table service. Even though many sit-down restaurants also have a take-away option, they were classified as sit-down restaurants nevertheless. ‡In one instance, the restaurant provided separate data for their sit-down and quick-serve menus, because both sets of data were included in the analysis, they were classified as both.

Despite providing nutrition information online, the following restaurants were excluded from the study because they did not meet the study's inclusion criteria for the following reasons: Did not provide sodium data: Freshslice, Starbucks, Topper's Pizza; Did not provide serving size data: Booster Juice, Captain Sub, Good Earth Coffeehouse, The Keg, Milestones, Pizza Salvatore, Wendy's, Quiznos, Only provided data for beverages and/or ice cream: Baskin Robbins, Blenz Coffeehouse, Esquires Coffeehouse, Freshly Squeezed, La Cremiere, Marble Slab Creamery, Second Cup, TCBY, Yogen Fruz; Substantially changed the presentation of the data: Smitty's, Mucho Burritto; Only provided American-specific data: Applebees, Cinnabon, Domino's Pizza, Godfather's Pizza, Nandos Flame Grilled Chicken, Popeyes Chicken and Biscuits, Papa John's Pizza; They reported nutrition information in 2010 but not in 2013: Jimmy the Greek, Lick's, Mega Wraps, Mary Brown's Chicken;They were a cafeteria supplier: Aramark.

The following restaurants contained >20 outlets in 2010 and did not report nutrition information online in 2010: abc Country Restaurant, Au Vieux Duluth Express, Bean Around the World Coffees, Beaver Tails, Bourbon Street Grill, Cactus Club Café, Café on the Go, Café Presse, Café Supreme, Café Vienne, Canadian 2 for 1 Pizza, Casa Grecque, Chez Ashton, Chez Coras, Chicken Chef, Chicken Delight, Coffee Culture Café, Crabby' Joe's Tap and Grill, Cultures, Delta Hotels, Dixie Lee Chicken and Seafood, Double Double Pizza & Chicken, Double Pizza, Eggsquis, Family Pizza, Fit for Life, Flaming Wok, Free Topping Pizza, Gabriel Pizza, Gino's Pizza, Golden Griddle, Grandma Lee's Bakery and Café, Greco Pizza-Donair, Humpty's Family Restaurant, Husky House Restaurant, Jungle Jim's, Koryo Korean BBQ, Koya Japan, La Piazzetta, Le Muffin Plus, Les Rotisseries Fusee, Mamma's Pizza, Mandarin, Moxie's Bar and Grill/Classic Grill, Original Joe's, Pacini, Panzerotto Pizza, Pizza Depot, Pizza Shack, Pretzelmaker, Red Lobster, Restaurant Amir, Restaurant Normandin, Ricky's All Day Grill, Saint Cinnamon, Select Sandwich Co, Serious Coffee, Square Boys Pizza and Subs, St. Hubert , St. Louis Bar and Grill, Sukiyaki - A Japanaes Delight, Sunset Grill, Sushi Shop, Tea Shop 168, The Bagel Stop, The Firkin Group of Pubs, The Pantry Restaurant, Tiki-Ming, Tony Roma's A Place for Ribs, Tutti Fruitti Dejeuners, Valentine, Vern's Pizza, Villa Madina, Wild Wing, Williams Fresh Café, Wimpy’s Diner.

5.3.2. Database construction

Data collected in 2013 was entered into the existing database where all foods were categorized according to the restaurant, type of restaurant (fast-food vs. sit-down restaurant), food category (e.g. hamburgers, sandwiches, etc.) and type of food (e.g. entrée, side dish, kids’ item, etc.). All foods were categorized according to their ―food status‖ in 2013 (newly reported, discontinued or present in both 2010 and 2013). All data entry and categorization was double- checked against the original source and subjected to range and logic checks. Matched-pairs in 2010 and 2013 were examined to ensure plausibility. We excluded: all beverages/ice cream (because these foods don’t typically contain a large amount of sodium), appetizers (because this category is too heterogeneous to make reliable comparisons), sauces/condiments (because they were not reported by the majority of restaurants), meals combining entrées and side dishes (because we would be unable to determine what component of the meal decreased in sodium), size duplications (because per 100g comparisons would be redundant) and categories with less than ten items.

5.3.3. Statistical Analysis

Primary a priori analyses were the longitudinal changes in sodium levels (per serving and per 100g) overall, in each restaurant and in each food category. Secondary analyses included the percentage of foods meeting/exceeding sodium recommendations, simultaneous changes in calories and serving size, and the difference among foods that were newly reported, discontinued or present in 2010 and 2013. Among foods present in 2010 and 2013 (n=2198), descriptive statistics and pairwise t-tests were used to compare sodium levels per serving and per 100g. The data was right-skewed in both 2010 and 2013, therefore, monte-carlo simulations of the exact p- values were used to confirm the parametric findings. Additionally, general linear models and monte-carlo exact simulations that included ―restaurant‖ as a covariate were created to control for this potential confounder. The proportion of foods whose sodium level increased or decreased was calculated for each restaurant and food category. Chi-square tests compared the percentage of entrees whose sodium level (per serving) was less than 600mg (the healthy amount for restaurant meals and main dishes according to the FDA),166 or greater than the daily recommended adequate intake level (1500 mg) and tolerable upper intake Level (2300 mg)126 in 2010 versus 2013. A general linear model was constructed to investigate the effect of food status

(newly reported, discontinued or present in both 2010 and 2013) as a predictor of sodium levels while controlling for restaurant and food category. Statistical analyses were conducted using SAS version 9.3 software (2010; SAS Institute Inc.).

5.4. Results

Sixty-one restaurants, representing >15,500 food outlets 167 were included in this study (Table 5.1). A total of 3,878 a la carte entrée items, side dishes and baked goods from both regular and children’s menus were analyzed. There were 2,198 foods for which data was reported in 2010 and 2013, 860 discontinued foods, and 820 newly reported foods.

5.4.1. Overall change in sodium levels

Sodium levels (per serving) decreased in 30% of foods, increased in 16% of foods and remained unchanged in 54% of foods (Figure 5.1). Though the magnitude of decrease varied, the average change among foods whose sodium level decreased was 220±303 mg (per serving), which was on average a 19±17% decline (Table 5.2). Meanwhile, among foods whose sodium level increased, the average change was 251±349 mg per serving, which was a 44±104% increase. Overall, there was a small (-25±268mg), yet significant (p<0.001) decrease in sodium levels both per serving and per 100g.

5.4.2. Changes in sodium by food category

Table 5.3 illustrates that the change in sodium levels varied depending on the food category. Sodium levels (per serving) decreased significantly in some categories (sit-down pizza, sit-down pasta, sit-down kids’ entrées, fast-food chicken, and fast-food tacos/burritos, p<0.05), with the average percent decrease ranging from 7 to 26%.

5.4.3. Changes in sodium by restaurant

The degree to which sodium levels changed varied depending on the restaurant (Table 5.4). Though sodium levels were completely unchanged in 17 restaurants (28% of the sample), most restaurants had both increases and decreases within their menu. In certain restaurants (such as Subway, Pizza Hut, Taco Bell and Taco Time) sodium levels decreased by at least 20% in

61 more than 70% of foods surveyed. However in some restaurants, despite large decreases in certain foods, there were equally large increases in others.

Similar results were seen when sodium levels were standardized (mg per 100g) (Appendix D).

5.4.4. Proportion of foods exceeding sodium’s recommended daily intake levels

There was no significant change in the percentage of entrées exceeding the daily sodium adequate intake level (p=0.7) or tolerable upper intake level (p=0.4) (Figure 5.2). Restaurants decreased the sodium level in 39% of foods that exceeded the adequate intake level in 2010, however, they simultaneously increased the sodium level in 18% of foods that already exceeded the adequate intake level. Meanwhile, sodium levels were lowered in 51% of foods where sodium levels exceeded the tolerable upper intake level in 2010, however, sodium levels were increased in 12% of foods where sodium exceeded the tolerable upper intake level in 2010. Additionally, there was no significant change (p=0.5) in the percentage of entrées that contained a ―healthy‖ amount of sodium (<600mg per meal, according to the Food and Drug Administration166) in 2010 versus 2013.

Data for changes in sodium density (sodium per 100g) in each restaurant can be found in supplementary table 3.

5.4.5. Sodium levels in newly reported, discontinued and persisting foods

When controlling for restaurant and food category, there was no significant difference (p=0.3) in the sodium level of foods that were newly reported (983±730 mg/serving) in comparison to foods that were discontinued (993±706 mg/serving) or foods that remained on the menu in 2010 and 2013 (892±679 mg/serving) (Appendix D).

5.4.6. Changes in serving size and calories

Foods whose sodium level decreased from 2010 to 2013 also had a significant decrease in serving size, sodium per 100g and calories (p<0.0001) (Table 5.1). Meanwhile, foods whose sodium level increased, had a significant increase in serving size, sodium per 100g and calories

(p<0.05). Calories per 100g did not change among foods whose sodium level increased or decreased.

Figure 5.1. Percent change in sodium levels (per serving) in Canadian chain restaurant foods (n=2198) from 2010 to 2013

Percentage of restaurant fods (%) fods restaurant of Percentage

30 90 80 70 60 50 40 20 10

20 30 40 50 60 70 80 90

------

–

– – – – – – – –

100

>-90

–

>100

– – – – – – – –

–

11 21 31 41 51 61 71 81

1 1

21 21 81 81 71 61 51 41 31 11

------Percent (%) change in sodium level (per serving)

*This bin contains foods with no change in sodium from 2010 to 2013 †This bin contains foods with >0% to 10% change in sodium, and each subsequent bin has the same width (ex. >10 to 20%, >20 to 30% etc.)

Figure 5.2. Percentage of entrees (n=1004) exceeding the daily recommended Adequate Intake level (1500 mg) and Tolerable Upper Intake level (2300 mg) in 2010 and 2013

NS 35 p=0.7

20 NS p=0.5 15 NS p=0.4

10 Percentage of foods (%) foods of Percentage 5

0 <600 mg (FDA "healthy" >AI (1500 mg) >UL (2300 mg) amount) 2010 2013

NS=not significant according to chi-square tests; AI=daily adequate intake level (1500mg); UL=daily upper tolerable intake level (2300mg) as defined by the Institute of Medicine;[18] 600mg has been defined by the Food and Drug Administration as a "healthy level" of sodium for a restaurant meal.[17] It should be noted that this graph does not represent the sodium levels in restaurant meals, but rather is comparing the sodium level in a single entree (without the accompanying side dishes) to the healthy level for meals and main dishes. The types of entrees represented in this graph include chicken entrees, hamburgers, pasta entrees, ribs, salad entrees, stir fry entrees, sandwiches/wraps etc. Therefore, these proportions do not represent the amount of sodium in meals that would be typically consumed at chain restaurants. Note: This graph only included entrees for which data was available in 2010 and 2013 (n=1004), newly reported and discontinued entrees were not included in this graph because the different types of entrees were not equally represented in these two groups and therefore could bias the outcome. See Appendix D for data concerning discontinued and newly added foods.

Table 5.2. Descriptive statistics for sodium levels in 2010 and 2013 and the average change over time in 2198 Canadian chain restaurant foods

Sodium level (mg) Sodium (mg/100g) Serving Size (g) Calories (kcal/serving) Calories (kcal/100g) Mean±SD||

Average

Average percent change change n (%) 2010 2013 (mg) (%) 2010 2013 2010 2013 2010 2013 2010 2013

Overall 2198 917±694 892±679§ -25±268 1±48 401±194 390±185† 241±161 239±159* 438±287 433±288* 208±97 208±98

Foods whose sodium level (mg/serving)…

Decreased 662 (30) 1130±778 910±642 -220±303 -19±17 481±212 407±190§ 252±171 238±161§ 474±293 442±275§ 209±83 209±88

Increased 358 (16) 842±658 1093±858 251±349 44±104 353±175 431±197§ 246±163 254±167* 468±307 497±343‡ 205±71 206±71

Remained 1178 (54) 821±624 0±0 0±0 371±175 368±175§ 234±155 235±154‡ 410±284 409±274 209±110 207±110 unchanged *Paried t-tests, p<0.05, †p<0.01, ‡p<0.001, §p<0.0001, ||standard deviation Additional analysis using monte-carlo simulations of the exact p-value as well as general linear models to compare sodium (mg per serving) in 2010 versus 2013, resulted in the same p-value, even when controlling for the restaurant that a food is from. Note: The very large standard deviations seen in this table indicate the presence of extreme values within the sample. The "overall average percent change" is positive despite the fact that the mean sodium level in 2013 is lower than the mean in 2010 due to the presence of some foods with very large percent increases that skewed the data in the positive direction.

Table 5.3. Changes in sodium levels from 2010 to 2013 within each food category (n=1580)

Sodium level (mg/serving), Percentage Average percent Percentag Average percent of foods change in sodium e of foods change in sodium mean±SD§ that among foods that that among foods that decreased decreased, median increased increased, median (%) (25th, 75th (%) (25th, 75th n 2010 2013 percentile) percentile) Sit-Down Restaurant Food Categories Stir Fry Entrées 10 2198±986 2209±802 10 -50 40 25 (8, 67) Sandwiches/Wraps 73 1906±1083 1933±933 29 -8 (-23, -4) 30 21 (9, 49) Pasta Entrées 76 1822±738 1619±723‡ 64 -11 (-22, -7) 13 12 (3, 17) Entrées w/ multiple meats 12 1821±798 1862±802 8 0 25 6 (6, 14) and seafood (ex. surf n turf) Hamburgers 39 1653±657 1753±642* 33 -7 (-8, -6) 28 35 (21, 57) Breakfast Entrées (ex. pancakes, waffles, 74 1595±678 1573±681 26 -24 (-35, -8) 24 13 (2, 47) eggs&bacon) Rib Entrées 19 1594±712 1717±956 16 -4 (-4, -2) 5 137 Salad Entrée w/Meat 41 1218±447 1259±487 19 -11 (-16, -9) 32 20 (13, 29) Chicken Entrées 42 1145±669 1142±731 24 -28 (-44, -11) 21 25 (18, 41) Salad Entrées 19 863±434 908±394 37 -20 (-25, -9) 32 54 (18, 182) Steak and Beef Entrées 36 832±816 836±777 25 -15 (-39, -8) 17 135 (39, 217) Seafood Entrées 24 804±588 811±578 8 -38 (-40, -36) 12 48 (34, 64) Pizza (one medium slice) 93 575±204 446±163‡ 91 -17 (-12, 32) 7 17 (12, 32) Fast-Food Restaurant Food Categories Stir Fry 24 1766±181 1766±181 0 0 0 0 Poutine 15 1617±722 1505±476 40 -17 (-23, -6) 20 8 (1, 129) Tacos/Burritos 61 1389±852 1123±618‡ 79 -21 (-34, -11) 20 16 (4, 32) Sandwiches/Wraps 203 1354±554 1327±659 40 -18 (-25, -12) 16 23 (10, 56) Hot Dogs 13 1203±236 1163±264 61 -10 (-13, -6) 31 4 (1, 31)

Salads w/Meat 22 1094±282 1047±342 54 -11 (-28, -8) 14 60 (21, 67) Hamburgers 60 1061±307 1039±319 55 -6 (-11, -4) 8 20 (18, 73) Stir Fry (no sodium/low 37 1041±246 1041±246 0 0 0 0 sodium) Sushi 11 897±261 961±236 0 0 36 35 (14, 36) Chicken 47 777±338 689±301† 42 -20 (-25, -13) 6 7 (2, 97) Breakfast (ex. bagels, 113 721±427 710±397 28 -5 (-11, -1) 14 6 (2, 13) breakfast sandwiches) Pasta 25 718±335 718±335 0 0 0 0 Salads 20 588±222 646±311 10 -32 (-44, -20) 20 71 (37, 107) Pizza (one medium slice) 263 432±211 436±211 21 -9 (-17, -5) 29 14 (6, 22) Kid's Menu Items Sit-down restaurant kid's 68 760 (567, 1008) 730 (535, 972)* 33 -26 (-39, -13) 11 10 (6, 36) meal entrées Fast-food kid's meal entrées 13 752 (592, 900) 717 (686, 770) 23 -24 (-34, -22) 23 16 (11, 17) (ex. hamburger, nuggets) Sit-down restaurant kid's side 27 420 (75, 570) 380 (85, 520) 31 -30 (-32, -23) 15 140 (54, 258) dishes *Paried t-tests, p<0.05, †p<0.01, ‡p<0.001, §standard deviation Additional analysis using general linear models to predict the change in sodium while including establishment as a covariate, produced the same results and confirmed that establishment is not a covariate. Data for baked goods, desserts and side dishes has been provided in supplementary table 1. Data for changes in sodium density (sodium per 100g) in each category are provided in supplementary table 2.

Table 5.4. Percentage of foods (n=1603) at each restaurant whose sodium level increased or decreased, and the average percent change

Sodium (mg/serving) Average percent change in Average percent sodium among change in sodium foods that among foods Percentage of decreased, Percentage of that increased, foods that median‡ (25th, foods that median (25th, Restaurant n decreased (%) 75th percentile) increased (%) 75th percentile) Subway* 51 86 -20 (-27, -15) 8 2 (1, 2) Pizza Hut* 73 78 -37 (-44, -17) 4 4 (1, 41) Taco Bell* 32 78 -20 (-31, -13) 9 8 (2, 10) Taco Time 18 78 -35 (-50, -13) 22 8 (1, 73) Shoeless Joe's 18 72 -22 (-27, -12) 28 25 (25 , 39) Burger King 51 71 -3 (-7, -1) 2 33† Boston Pizza 106 68 -8 (-18. -5) 21 13 (10, 44)

Dagwood Sandwiches and Subs 38 66 -16 (-23, -13) 32 21 (8, 56)

White Spot Triple O's 20 65 -16 (-24, -15) 0 na§ Dairy Queen 18 61 -7 (-30, -6) 33 18 (16, 20) A&W 34 59 -7 (-12, -4) 29 11 (3, 15) East Side Mario's 45 59 -11 (-24, -6) 30 17 (8, 32) Taco Del Mar 27 56 -25 (-33, -11) 41 23 (9, 36) KFC* 24 54 -23 (-37, -19) 12 23 (7, 97) Joey's Restaurant 20 50 -13 (-54, -8) 50 66 (14, 96) Mike's Restaurant 80 47 -25 (-46, -12) 44 31 (10, 106) White Spot Legendary 52 44 -10 (-29, -5) 25 33 (4, 67) Restaurant

Arby's 28 43 -9 (-18, -5) 54 25 (16, 118) Kelsey's 44 39 -10 (-27, -10) 39 20 (9. 33) Panago 171 36 -9 (-17, -5) 43 16 (7, 25) Tim Hortons 70 34 -9 (-20, -4) 11 9 (6, 21) Jack Astors 24 33 -30 (-52, -17) 67 65 (20, 173) Montanas 75 29 -30 (-36, -25) 20 21 (13, 54) Mr. Greek 35 26 -1 (-43, -1) 6 21 (14, 54) Edo Japan 18 22 -9 (-10, -10) 0 na McDonald's* 46 26 -8 (-18, -7) 22 11 (6, 19) Harvey's 25 20 -14 (-19, 14) 0 na Casey's 46 17 -14 (-30, -4) 13 29 (12, 43) Scores Rotisserie 30 17 -12 (-24, -11) 20 5 (2, 18) Pizza Delight 34 12 -12 (-14, -9) 3 14 New Orleans Pizza 20 10 -40 (-69, -11) 85 52 (13, 51)

Earl's Restaurant 28 4 0 (0, 0) 7 23 (5, 42) Extreme Pita 43 0 na 0 na Mmmuffins 46 0 na 0 na Mrs. Vanelli's 28 0 na 4 13 The Great Canadian Bagel 85 0 na 0 na *Indicates restaurants that have made a voluntary commitment to reducing the sodium level in their products †When there was only one food that increased or decreased, the percent change in that food was presented without an indicator of variance ‡Medians were reported to prevent extreme values from skewing the average §not applicable The following restaurants reported no changes in sodium levels between 2010 and 2013: 241 Pizza, Baton Rouge, Coffee Time, Country Style, Denny's, Flying Wedge Pizza, Little Caesars, Manchu Wok, Mr. Sub, Pita Pit, Pizza 73, Pizza Nova, Pizzaville, Robin's Donuts, Swiss Chalet, Teriyaki Experience, Van Houtte's Bistro The following restaurants were excluded from this table because they had <10 menu items included in this study: Bento Nouveau, Druxy's Deli, Jugo Juice, New York Fries, Opa, Orange Julius, Pizza Pizza, Treats.

5.5 Interpretation

5.5.1. Main Findings

From 2010 to 2013, sodium levels decreased in 30% of foods, increased in 16% and remained unchanged in 54%. The percentage of foods that increased or decreased, and the magnitude of change varied depending on the restaurant and food category. Changes in sodium arose due to both altered serving size and sodium density. The number of menu items with unacceptably high amounts of sodium (exceeding the Adequate Intake level and Tolerable Upper Intake level) did not change. This study illustrates that industry efforts to voluntarily decrease sodium levels in Canadian restaurant foods have produced inconsistent results.

5.5.2. Comparison with other studies

Previous studies in the United States have similarly shown that both healthy and unhealthy changes in sodium are occurring simultaneously. When comparing identical restaurant foods in 2005, 2008 and 2011, Jacobson et al found a 2.6% increase in sodium levels.162 Meanwhile, Wu and Strum found no change in sodium levels between 2010 and 2011.168

Our data showed that many of the leaders in sodium reduction are restaurants that have made voluntary commitments to reduce the sodium level in their foods (Subway, Pizza Hut, Taco Bell, and KFC).162, 163 However, all of the restaurants that have made commitments still offer foods that exceed the Adequate Intake level for sodium, and some even increased sodium levels in certain products. This illustrates how voluntary, industry-designed commitments—such as those that aim to produce an average percent reduction in sodium across a restaurant’s entire menu—are not ideal because they allow large reductions in certain foods to mask increases in others.

Consuming excessive sodium has been shown to increase one’s taste preference for high sodium foods.155, 169, 170 However, it has been well established that gradually reducing sodium levels is an effective way to retrain consumers’ taste buds to prefer lower sodium foods.156, 157, 171, 172 Thus, this study shows that the first gradual step towards sodium reduction has been taken by some restaurants; nevertheless, further decreases are still needed to reduce sodium to an acceptable level.

Despite the creation of voluntary sodium reduction targets for Canadian grocery foods,159 and restaurant targets in the US;173 targets for Canadian restaurant foods are yet to be established. Canadian restaurant foods have been shown to have higher sodium levels when compared to countries such as the United Kingdom, which have a government-industry agreement to lower sodium levels. 46 Additionally, targets may not be the only way to motivate sodium reduction; research has shown that menu-labelling laws that require the disclosure of sodium information— such as those implemented in King County, Washington and proposed by Toronto Public Health70, 83, 85—may also promote sodium reduction.84

In this study, decreased sodium levels resulted from a combination of both reductions in serving size and sodium density. It is important to note that reduction targets and commitments must aim to lower sodium levels both per serving and per 100g to ensure that sodium reduction is not achieved solely via decreases in serving size. Additionally, given the large number of menu items with sodium levels exceeding the daily Adequate Intake level (22%) and Tolerable Upper Intake level (10%), maxima are needed to reduce sodium levels in products that are exceptionally high in sodium. Even though our study showed no change in calorie density, sodium reduction programs should also ensure that potentially adverse nutrients are not increased to compensate for decreases in sodium.

5.5.3. Limitations

This study included 64% of the top 50 restaurants in Canada (according to number of outlets),167 therefore, our sample may not be representative of the entire restaurant food supply due to the number of restaurants that did not disclose nutrition data (Table 1). Additionally, this study may not include all menu items from the restaurants represented in the sample. The extent to which the exclusion of restaurants that did not meet study criteria may have biased the results, in either direction, is unknown. Foods that were ―newly reported‖ in 2013 may not exclusively represent ―new menu items‖ because some restaurants may have reported nutrition information for more menu items in 2013. Thus, some newly introduced menu items may in fact be lower in sodium, despite the fact that the average did not show this. The accuracy of the findings presented in this study is dependent on the accuracy of the self-reported data provided by the establishments. Furthermore, this study represents the sodium level in foods available in restaurants and does not necessarily reflect consumption. A study in the UK investigating

71 purchase-weighted mean sodium in processed foods showed that mean sodium was 18-35% higher than unweighted sodium levels 158. More research is needed to understand the effect of market share or purchase weighting on this data. Finally, this study does not shed light on the means by which sodium levels have been decreased (ex. use of mineral salts (potassium chloride, magnesium sulphate), yeast extracts (hydrolyzed vegetable protein), amino acids, dendritic salt, or salt enhancers, etc.).172, 174

5.6. Conclusion

From 2010 to 2013 sodium levels in the majority of Canadian restaurant foods were unchanged. The large decreases seen in certain restaurants illustrates that sodium reduction is possible. Meanwhile the observed increase in some foods demonstrates that industry-wide commitments and a systematic monitoring program are needed. Therefore, this study demonstrates the importance of establishing sodium reduction targets for Canadian restaurant foods, and the need for a government enforced sodium reduction strategy and regular monitoring. In conclusion, due to the slow rate of progress over the past three years, alongside high rates of hypertension and cardiovascular disease, addressing the high sodium level in restaurant foods continues to be a public health priority, and an essential step towards decreasing the burden of diet-related chronic disease.

5.7. Acknowledgements Special thanks to Sahar Qassem and Zhila Semnani-Azad for assistance with data entry, as well as Dr. Paul Corey for his statistical guidance.

Contributions Mary Scourboutakos and Mary L’Abbé designed the study and interpreted the data. Mary Scourboutakos acquired the data with entry assistance from volunteers. Mary Scourboutakos analyzed the data and drafted the manuscript. Both authors critically reviewed the manuscript for important intellectual content.

Chapter 6 Restaurant menu-labelling – Is it worth adding sodium to the label? (CJPH, 2014)

This article has been published: Scourboutakos MJ, Corey PN, Mendoza J, Henson SJ, L’Abbé MR. 2014. Restaurant menu-labelling: Is it worth adding sodium to the label? Canadian Journal of Public Health, 105(5)e354-e361.

6.1. Abstract

Background: Many provincial and federal bills have recommended various forms of menu- labelling that would require information beyond just calories, however, the additional benefit of including sodium information is unknown.

Objective: To determine if sodium information on menus helps consumers make lower sodium choices, and to understand what other factors influence the effect of menu-labelling on consumer’s meal choices.

Methods: 3080 Canadian consumers completed an online survey that included a repeated measures experiment where consumers were asked to select what they would typically order from four mock-restaurant menus. Subsequently, consumers were randomized to see one of three menu-labelling treatments (calories; calories and sodium; or calories, sodium and serving size) and were given the option to change their order.

Results: There was a significant difference in the proportion of consumers that changed their order varying from 17-30% depending on the restaurant type. After seeing menu-labelling, sodium levels decreased in all treatments (p<0.0001). However, in three of the four restaurant types, consumers who saw calorie and sodium information ordered meals with significantly less sodium compared to consumers who only saw calorie information (p<0.01). Consumers who saw sodium labelling decreased the sodium level of their meal by an average of 171-384 mg, depending on the restaurant. Among the subset of consumers who saw sodium information and chose to change their order, sodium levels decreased by an average of 681-1360 mg, depending on the restaurant. Gender, intent to lose weight and the amount of calories ordered at baseline

73 were the most important predictors of who used menu-labelling. 80% of survey panelists wanted to see nutrition information when dining-out.

Conclusion: Including sodium information, alongside calorie information, may result in a larger decrease in the amount of sodium ordered by restaurant goers.

6.2. Introduction

In response to the growing obesity epidemic,138 and the prevalence of eating outside-the- home, 31 restaurant menu-labelling is a policy being explored as a means to address this situation. In the United States, several jurisdictions have enacted menu-labelling laws.71 Meanwhile, in Canada there have been several unsuccessful bills at both the provincial and federal level.64, 68 Recently, Toronto Public Health recommended legislation requiring the mandatory disclosure of calorie and sodium information in large restaurant chains.70

Toronto Public Health’s recommendation differs from the menu-labelling laws enacted in New York City and proposed in the Patient Protection and Affordable Care Act (which includes the US’s federal menu-labelling legislation),71 because it calls for the disclosure of sodium, in addition to calorie information. To date, only one county in Washington has legislated a menu- labelling law that includes information beyond calories, by also requiring the disclosure of saturated fat, carbohydrate and sodium content,83 while another county has a similar voluntary program.94

The inclusion of sodium information is important because research on the nutritional quality of restaurant foods has demonstrated that sodium levels are alarmingly high and that there is a wide range of sodium even among similar foods.160, 161 Because of this variation, there is no way for the consumer to determine which foods are higher or lower in sodium. This is concerning as dietary sodium is the leading preventable risk factor for hypertension,123 the leading risk factor for death worldwide.125

Studies show that people prefer forms of menu-labelling that include information beyond just calories. For example, Mackison et al found that 61% of consumers wanted to see sodium information on menus.130

To date, there is no published research investigating the effect of including sodium or serving size information on menus. Using a randomized controlled experiment embedded within an online survey, the objective of this study was to answer the following research questions:

1) Does the inclusion of sodium information on restaurant menus result in lower sodium choices? 2) Does the inclusion of serving size information result in the choice of meals with a lower calorie density (calories per 100g) and/or a lower sodium density (sodium per 100g)? 3) What factors (demographic factors, as well as the calorie and sodium content of the consumer’s meal choice) influence consumer’s use of menu-labelling?

6.3. Subjects and Methods

6.3.1. Participants

The Canadian Consumer Monitor (CCM) panel was used for this study. The CCM is a nationally representative consumer survey panel. It was recruited by a professional recruiting company to reflect the Canadian population (according to 2006 census data) for age, sex, education and region. Initially 31,223 Canadian adults were contacted by e-mail. Survey panelists were required to be the primary household grocery shopper and an initial invitation to participate was sent to all panelists to collect data on their demographic characteristics. 6,665 provided informed consent and completed the baseline questionnaire. Beginning in 2010, 15 minute surveys were administered to the CCM panel every 8-10 weeks 175, 176. Typically 2500- 7000 consumers participate in each survey. Due to attrition, 3080 consumers participated in this survey, which was administered in April 2012. Ethics approval was received from both the University of Toronto and the University of Guelph’s research ethics boards. Before being administered to the CCM, the survey was pilot tested on a small panel of 255 consumers from Guelph Ontario. The survey was administered using Snap 10 Professional Survey Software and Webhost (Snap Surveys, Portsmouth, NH).

6.3.2. Experimental Design and Survey Structure

A repeated measures randomized controlled experiment was embedded within the survey. The experiment used a parallel within-subject design where each consumer served as their own

75 control. Each consumer was asked to make a selection from each of four different restaurant menus. After making a selection, consumers were randomized to see one of three different types of menu-labelling (referred to as treatments): 1) calorie labelling, 2) calorie and sodium labelling, 3) calorie, sodium and serving size labelling. Randomization was based on the timing within the minute when the panelist started the survey. After randomization, consumers were shown the same series of menus, however, this time the menus were labeled with nutrition information (according to the treatment the consumer was assigned to). Upon seeing the labelled menu, consumers were given the option to change their order. This enabled pre-post comparisons where panelists who were influenced by the information and chose to change their order could be analyzed separately from consumers who did not change their order. Similar methods have been used in previous studies that analyzed consumers who report using the information separately from consumers who don’t use the information.29, 88

In addition, at the beginning of the survey, consumers were asked about their frequency of eating-out and whether they’re trying to lose weight. At the end of the survey, consumers were asked if the nutrition information they saw influenced what they ordered (with the option of answering yes, somewhat, or no), along with an open-ended response field to explain why.

6.3.3. Restaurant Menus

Four restaurant scenarios were tested in the survey: a fast-food hamburger restaurant, a sit-down breakfast restaurant, a sub shop, and a sit-down dinner restaurant (Appendix E). The restaurant menus were adapted from actual Canadian chain restaurant menus and were selected because they had a large range of menu offerings including both high and low calorie and sodium options. Multiple versions of each menu were created to reflect each of the treatments: no information, calorie labelling, calorie and sodium labelling, calorie, sodium and serving size labelling (Appendix E). The calorie, sodium and serving size information on the menu was based on the restaurant’s nutrition information disclosed online in 2010 and was retrieved from the University of Toronto’s restaurant database, however the restaurant’s identity was not presented to consumers.151 The labelled menus also provided consumers with information about the daily recommended amount of calories (2000 kcal) and the upper tolerable intake for sodium (2300

76 mg), as previous research has demonstrated the added benefit of including contextual statements with daily reference amounts. 88

6.3.4. Treatments/different types of menu-labelling that were tested

Three menu-labelling treatments were tested in this survey. Calorie labelling was tested because it is the most common form of menu-labelling.50, 165 Calorie and sodium labelling was tested because Toronto Public Health has recently recommended the disclosure of calorie and sodium information on restaurant menus.70 The third treatment, which includes calorie, sodium and serving size labelling, was used to determine if the addition of serving size information helps consumers choose meals with a lower calorie and/or sodium density.

6.3.5. Data Analysis

Primary outcomes included the difference in nutrient levels among treatments before versus after seeing labelling. Secondary outcomes include differences among the sub-set of panelists who opted to change their order, as well as the effect of serving size labelling on the nutrient density of meals ordered and demographic influences on menu-labelling. The analysis included both complete and incomplete surveys. Mean±SD for calories, sodium and serving size of meals ordered by consumers before and after seeing menu-labelling were calculated. Because data were not normally distributed, Monte Carlo simulations of the exact p-values were used to compare nutrient levels before and after seeing labelling. For the subset of panelists who changed their order, paired t-tests were used to compare the levels before and after seeing menu- labelling and to compare treatment one and treatment two. Panelists did not necessarily change their order in all four restaurants. Therefore data for the sub-set who changed their order is reported separately for each restaurant. Odds ratios were used to compare the proportion of consumers who changed their order in each restaurant and each treatment. ANOVA was used to compare the calorie density, sodium density and serving size of meals ordered by consumers who saw different treatments. To explore the role of socio-demographic factors on use of menu- labelling, the following predictors (age, gender, education, frequency of eating-out, intend to lose weight, treatment, and restaurant) were tested in a repeated measures logistic regression (using proc genmod). Separate models were constructed for each interaction term, and because there was interaction between the restaurant scenario and some demographic predictors, a separate

77 model was constructed for each restaurant. An additional model was constructed that incorporated income and Body Mass Index (BMI) data that was collected in the baseline questionnaire. Key themes in the open-ended questions were identified and responses were coded and quantified. Some responses were classified as having more than one theme. Only themes mentioned by a minimum of 5% of consumers were reported. Statistical analyses were conducted using SAS version 9.3 software (SAS Institute Inc., 2010).

6.4. Results

6.4.1. Participants

A total of 3080 panellists participated in the survey. Baseline characteristics of participants are reported in Table 6.1. More than 85% of respondents reported eating-out at least once a week.

6.4.2. Proportion of consumers who changed their order after seeing nutrition information Figure 6.1 shows the proportion of consumers, within each restaurant type, that changed their order after seeing labelled menus. There was a significant difference among restaurants, ranging from 17% in the sub shop to 30% in the breakfast restaurant. There was also a significant difference in the proportion of consumers in each treatment who changed their order. In the hamburger restaurant, sub shop and dinner restaurant scenarios, consumers who saw more information (serving size and/or sodium) were significantly (p<0.001) more likely to change their order, when compared to consumers who only saw calorie information.

6.4.3. Sodium level of meals ordered before, versus after seeing labelled menus (all consumers)

Table 6.2 shows the average calorie and sodium level of meals ordered before and after seeing labelled menus. After seeing menu-labelling, sodium levels decreased in all treatments (p<0.0001). However, in three of the four restaurant scenarios, consumers who saw calorie and sodium information ordered meals with significantly less sodium compared to consumers who only saw calorie information (p<0.01). The average decrease in sodium ranged from 56 to 290

78 mg, among panelist who saw calorie labelling, while the decrease ranged from 134 to 384 mg among panelists who saw calorie and sodium labelling.

6.4.4. Sodium level of meals ordered before versus after seeing labelled menus (sub-set of consumers who changed their order)

Table 6.2 shows the average calorie and sodium levels of meals ordered before and after seeing menu-labelling, among the sub-set of consumers who chose to change their order after seeing nutrition information. In this sub-set, the average difference in sodium ordered before and after seeing nutrition information was 471, 694, 396 and 1069 mg among consumers who only saw calorie information, while the difference was 681, 959, 626 and 1360 mg among consumers who saw calorie and sodium information, in the hamburger, breakfast, sub and dinner restaurants, respectively.

6.4.5. Effect of serving size information on the calorie and sodium density of consumer’s choices

Consumers who saw serving size information did not order meals with a lower calorie or sodium density compared to consumers who did not see serving size information (Table 6.3).

6.4.6. Consumer’s rationale for why the information influenced or did not influence their order

When asked, ―did the information influence what you ordered‖, 32% of consumers answered ―yes‖, 33% said it ―somewhat influenced their order‖ and 35% said it did not influence their order. There was no significant difference in the proportion of consumers from each treatment, who said that the information influenced their order. Table 6.4 shows that 56% of consumers who saw sodium information commented on the sodium level of the meals, and 15% specifically said that they were shocked by the sodium level. The most popular rationale for why consumers didn’t use the nutrition information was that they rarely eat out or they consider meals at restaurants to be a treat (35%). In addition, many consumers noted that they were already health conscious (17%) or that the information verified that they had already made a healthy choice (10%). 21% of consumers said that they didn’t care about the information and 6% said that other dietary restrictions govern their food choices.

6.4.7. Effect of demographics on whether or not the nutrition information influenced consumer’s decisions

Gender, intent to lose weight, and the amount of calories ordered at baseline were statistically significant predictors of who uses menu-labelling (Table 6.5). Because of the significant interaction between restaurant and various demographic predictors, at some restaurants, education, treatment, and frequency of eating-out were also statistically significant predictors (data not shown). In the secondary model that included income and BMI (data not shown), we found that neither of these additional predictors were significant.

6.4.8. Proportion of consumers that want to see nutrition information

80% of the surveyed consumers said that they would like to see nutrition information when dining-out. Specifically, 75% of wanted to see calories, 71% wanted to see sodium, 49% wanted to see total fat, 47% wanted to see sugar, 46% wanted to see trans fat, and 43% wanted to see saturated fat information.

Figure 6.1. Proportion of panelists in each restaurant who changed their order after seeing menu-labelling, and proportion of panelists from each treatment who changed their order 35

30 Panelists who saw 1.276* calorie, 25 (1.051, 1.551) sodium and 1.399* serving size (1.089, 1.798) information 20 Panelists 0.864 who saw 1.121 (0.715, 1.403) calorie and 15 1.342* 1.667* (0.918, 1.369) sodium (1.083, 1.664) (1.301, 2.137) labelling

changed their changed order 10 0.995 1.403* (0.824, 1.202) Panelists (1.130, 1.743) who saw Ref 5 Ref calorie labelling Ref Ref

Percentage of panelists in each restaurant who restaurant each in panelists of Percentage 0 Sub shop (n=2753) Hamburger restaurant (n=2972) Dinner restaurant (n=2949) Breakfast restaurant (n=2835) (cal=435±213, sod=1231±678)‡ (cal=740±337, sod=1556±550) (cal=930±306, sod=1793±981) (cal=977±322, sod=1537±1241) 2.040† Ref 1.332† 1.884† (1.168, 1.520) (1.658, 2.141) (1.798, 2.314)

Each bar represents the percentage of consumers in each restaurant who changed their order. Within each bar, the percentage of consumers from each treatment is shown. The black portion of the bar represents the proportion of consumers who saw calorie labelling, dark grey represents the proportion of consumers who saw calorie and sodium labelling, and light grey represents the proportion of consumers who saw calorie, sodium and serving size labelling. Data are presented as odds ratios (95% confidence intervals). There are two sets of odds ratios. The odds ratios within the bars are comparing the proportion of consumers in each treatment who changed their order, the reference group is the calorie treatment, therefore the odds ratios show the relative benefit of additionally including sodium information, or including sodium, and serving size information. * indicates when a group is significantly different from the reference group p<0.05. The odds ratios along the x axis are comparing the proportion of consumers in each restaurant who changed their order, † denotes when a restaurant is significantly different from the reference restaurant. ‡Baseline calories (kcal) and sodium (mg) ordered by panellists before seeing menu-labelling. Note: Note: Less than 1% of panelists opted to change their order to a meal that was higher in calories and/ or sodium.

Table 6.1. Demographic characteristics of the panelits in the study

Sample Characteristics Values Sample size (n) 3080 Age range [n (%)] 20-29 199 (7) 30-39 472 (15) 40-49 761 (25) 50-59 891 (29) 60-69 737 (24) Sex [n (%)] M 1012 (33) F 2058 (67) Education [n (%)] High school or less 622 (20) Trades 306 (10) College 1033 (34) University 1099 (36) Frequency of eating at fast-food restaurants [n (%)] Never 161 (5) Infrequently (once per month or less) 246 (8) Semi-Frequently (once per week) 778 (25) Frequently (more than once per week) 1892 (61) Frequency of eating at sit-down restaurants [n (%)] Never 33 (1) Infrequently (once per month or less) 321 (10) Semi-Frequently (once per week) 932 (30) Frequently (more than once per week) 1787 (58) Reported trying to lose weight [n (%)] 1525 (50) *Some demograhpic data was missing for certain variables (sex, 10 missing; education, 20; age, 20; frequency of eating fast-food, 3; frequency of eating at sit-down restaurants, 7; trying to lose weight 25) *Only 9 people said that they never eat out at sit-down or fast-food restaurants

Table 6.2. Comparing the calorie and sodium content of meals ordered by panelists who saw calorie or calorie and sodium labelling

Calories (kcal) Sodium (mg)

Before Change: Change: seeing After before vs Before seeing before vs menu- seeing after menu- After seeing after labelling menu- seeing labelling menu- seeing (baseline) labelling labelling (baseline) labelling labelling Restaurant Treatment n Mean±SD Mean±SD Mean±SD Mean±SD

Hamburger Calorie Labelling 1041 731±333 657±322 .-74* 1545±538 1452±490 .-93*† restaurant Calorie + Sodium Labelling 939 737±331 653±327 .-84* 1531±545 1360±571 .-171*

Calorie Labelling 998 968±323 793±324 .-175*† 1504±1198 1249±1048 .-255* Breakfast restaurant Calorie + Sodium Labelling 899 980±325 828±323 .-152* 1566±1289 1226±1052 .-340* All panelists Calorie Labelling 970 434±205 402±185 .-32* 1247±686 1191±640 .-56*† Sub shop Calorie + Sodium Labelling 868 432±216 390±189 .-42* 1203±670 1069±607 .-134*

Calorie Labelling 1034 937±301 816±301 .-121* 1811±979 1521±972 .-290*† Dinner restaurant Calorie + Sodium Labelling 933 917±314 798±310 .-119* 1773±1005 1389±982 .-384* Hamburger Calorie Labelling 204 585±328 485±233 .-100* 1737±631 1266±383 .-471*† restaurant Calorie + Sodium Labelling 238 783±333 447±216 .-336* 1637±523 956±431 .-681*

Sub-set of panelists Calorie Labelling 369 1133±228 659±263 .-474* 1820±1326a 1126±989 .-694* Breakfast restaurant who Calorie + Sodium Labelling 322 1115±273 685±262 .-430* 1955±1498a 996±920 .-959* changed their order after seeing Calorie Labelling 136 591±228 361±147 .-230* 1518±782 1122±510 .-396*† Sub shop labelled Calorie + Sodium Labelling 188 527±239 331±120 .-196* 1479±718 853±416 .-626* menus

Calorie Labelling 282 1079±238 636±196 .-443* 2181±796 1112±733 .-1069*† Dinner restaurant Calorie + Sodium Labelling 263 1034±267 611±186 .-423* 2206±824 846±600 .-1360*

Table 6.3. Comparing the calorie density ordered after seeing menu-labelling among those who opted to change their order

Serving size Calorie density Sodium density

Menu-labelling condition Menu-labelling condition Menu-labelling condition n Cal* Cal+Sod† Cal+Sod+SS‡ p Cal Cal+Sod Cal+Sod+SS p Cal Cal+Sod Cal+Sod+SS p Breakfast restaurant 988 455a 468a 472a 0.138 143a 144a 146a 0.2477 230a 202b 205b 0.0053 Sub shop 478 273a 246b 257ab 0.04 132a 134a 133a 0.2596 403a 367b 341b 0.0001 Dinner restaurant 853 408a 404a 387a 0.0902 164a 160ab 179c 0.0001 272a 207b 224b 0.0001 Note: Restaurant A was not included because it included beverages, which prevented the proper calculation of calorie and sodium density. Cal=calorie labelling treatment, Cal+Sod=calorie and sodium labelling treatment, Cal+Sod+SS=calorie, sodium and serving size labelling treatment. The n represents only those who changed their order and it includes panelists from all three treatments.

Panelist's responses n % Yes, the nutrition information influenced my decision... (n=762)

Sodium was too high, I tried to pick lower sodium meals, or any comment related to 331 67* the sodium level of the meal

Calories were too high, I tried to pick lower calories meals, or any comment related 463 61 to the calorie content of the meal

I was shocked or surprised by sodium level, I wasn't aware of how high the sodium 91 18* content was I was shocked or surprised by calorie level, I wasn't aware of how high the calorie 129 17 level was

The information helped me make a healthier choice, I changed something after 58 8 seeing the information, I am trying to be healthier

The information increased my awareness, I didn't realize how unhealthy my choices 36 5 were, it made me think more about what I was ordering

The information somewhat influenced my decision... (n=660) Sodium was too high, I tried to pick lower sodium meals, or any comment related to 160 41* the sodium level of the meal Calories were too high, I tried to pick lower calories meals, or any comment related 267 40 to the calorie content of the meal

The information helped me make a healthier choice, I changed something after 87 13 seeing the information, I am trying to be healthier

Shocked or surprised by sodium level, I wasn't aware of how high sodium was 41 11*

I rarely eat out, eating-out is a treat 55 8 Shocked or surprised by sodium level, I wasn't aware of how high the sodium content 54 8* was Other 45 7

The information helped me make a healthier choice, I changed something after 45 7 seeing the information, I am trying to be healthier

The information verified that I had already made a healthy choice 36 5

No, the information did not influence my decision... (n=528) I rarely eat out, eating-out is a treat 187 35

I don't care about the information, I eat what I want to eat, I am not concerned 113 21 about my weight I already eat in a healthy manner, I already know which choices are healthy or 90 17 unhealthy

The information verified that I had already made a healthy choice 52 10

Other 51 10 I have other dietary restrictions that govern my food choices (vegetarianism, 32 6 veganism, gluten intolerance etc.) *This percentage was calculated with 496 as the denominator because only two-thirds of panelists saw sodium information, therefore only those who were in treatment two or three were included in this percentage

Table 6.5. Odds ratios classified by restaurant and demographic predictor of who uses menu-labelling

Restaurant Odds Ratios 95% Confidence Interval p value Hamburger Gender female vs male 1.41 1.12-1.78 0.0037 restaurant* Intent to lose weight intent vs no intent 2.18 1.77-2.68 0.0001

Calories at baseline* (before 1.11 1.06-1.16 0.0001 seeing menu-labelling) Sodium at baseline* (before 1.03 1.01-1.06 0.0165 seeing menu-labelling) Breakfast Gender female vs male 1.74 1.39-2.18 0.0001 restaurant* Intent to lose weight yes vs no 1.91 1.57-2.32 0.0001

Calories at baseline (before seeing 1.38 1.32-1.44 0.0001 menu-labelling) Sodium at baseline* (before 0.99 0.98-1.02 0.4493 seeing menu-labelling) Sub shop* Gender female vs male 1.80. 1.36-2.38 0.0001 Intent to lose weight yes vs no 1.75 1.38-2.21 0.0001

Calories at baseline 1.38 1.25-1.53 0.0001

Sodium at baseline* 1.01 0.97-1.04 0.653

Dinner restaurant* Gender female vs male 1.45 1.15-1.84 0.0018 Intent to lose weight yes vs no 1.72 1.41-2.10 0.0001

Calories at baseline 1.23 1.18-1.29 0.0001

Sodium at baseline* 1.03 1.02-1.05 0.0001

*Calories at baseline refers to the amount of calories (kcal) in the meal ordered by the panelists before seeing menu-labelling. Sodium at baseline refers to the amount of sodium in the meal ordered by the panelist before seeing menu-labelling. Due to the observed interaction between restaurant and various demographic predictors, education and treatment were also significant predictors in the hamburger restaurant. In the breakfast restauarant frequency of eating-out and education were significant predictors. In the sub shop treatment, education and age were significant predictors. And in the dinner restaurant, treatment and frequency of eating-out were significant predictors.

6.5. Discussion

These results showed that when sodium information was provided on restaurant menus, consumers ordered meals with significantly less sodium, compared to consumers who only saw calorie information. However, the magnitude of the decrease varied depending on the restaurant.

Even when consumers only saw calorie information, the sodium content of their revised meal choices significantly declined. This is consistent with New York City’s rational for only labelling calories and not including sodium, as they showed that calories and sodium are positively correlated.81 However, our results confirmed that despite the inadvertent decrease in sodium that automatically results from decreasing calories, the inclusion of sodium information led consumers to choose meals containing significantly less sodium in comparison to meals selected by consumers who only saw calorie information.

In our study, 24% of consumers changed their order after seeing labelled menus. This finding is slightly higher than the findings in New York City which showed that approximately 15% of customers in New York city use calorie information, and slightly lower than other results which showed that up to 34% of consumers used the information provided on the menu.29, 88, 94

One of the most important findings was the heterogenous effect of menu-labelling according to the type of restaurant and the sodium/calorie level of the meal. This is consistent with the findings of Burton et al who showed that menu-labelling is more likely to influence consumers’ choices when the calorie content is less favourable than expected 58. This finding has important methodological and policy implications as it suggests that studies conducted in single settings, particularly if they are not high calorie settings, may not be a reliable indicator of the potential benefit of menu-labelling.

The results provided insight into the rationale for why some consumers choose not to use menu-labelling. Often consumer’s reasoning did not undermine the importance or relevancy of this potential policy, and only a small percentage of consumers didn’t care or did not want to see the information. This was consistent with previous research which showed that the public wants to see nutrition information on menus, even if they don’t use it every time.60-62, 177

In our study 42% of consumers who saw sodium information (and answered the open- ended question) said that the sodium level of the meal influenced their decision. This was much higher than Pulos et al’s findings which showed that when consumers saw calories, fat, sodium and carbohydrate information, only 7.8% of patrons said that they chose their entrée because it was lower in sodium.94 Previous research has shown that women, older and wealthier customers are more likely to use menu-labelling.29 Our results showed that women were more likely to use menu-labelling, however age and income were not significant predictors. Contrary to what might be expected, BMI was not a statistically significant predictor of use of menu-labelling, however this was due to the collinearity between intent to lose weight and BMI, as panelists that were trying to lose weight were more likely to be obese.

6.5.1. Strengths

Strengths included the large sample size, and the repeated measures design, which enabled us to detect within-subject effects. Furthermore the survey methodology enabled us to quantify the decrease in calories and sodium among consumers who actually used the information. This allowed us to measure the magnitude of the decrease at the level of the individual which to date has not been considered in most of the natural experiments and interventions conducted in real-life settings.

6.5.2. Weaknesses

The applicability of these results in a real-life setting are unclear, as this study only evaluated purchase intentions as opposed to purchasing behaviours, which can be affected by many other factors. Additionally, our results may be subject to social desirability bias,178 however the use of online surveys, has been shown to promote less bias than traditional interview methodologies.179 Furthermore, our sample was slightly older, more female and more educated than the 2006 Canadian census data, 180, 181 and therefore may be biased toward individuals who are more likely to use labelling, and therefore may not be representative of the Canadian population. The results should be confirmed with a real-life intervention that takes into consideration factors such as cost.

In addition, our study only investigated one of the many potential mechanisms through which menu-labelling can affect the nutritional content of consumer’s purchases. Therefore, it is

89 important to remember that in order to draw conclusions about the benefit of a policy such as menu-labelling, we must consider all of its potential benefits including its effect on promoting product reformulation and the introduction of new healthier menu offerings, as a recent study showed that 18 months after the implementation of menu-labelling the calorie, saturated fat and sodium level of restaurant meals were lowered.84

6.6. Conclusion

These results suggest that menu-labelling could have an impact on the nutrient content of meals ordered by some consumers when they are dining-out. Additionally, it showed that including sodium information may lead to lower sodium choices, compared to providing calorie information alone. Finally, this study shed light on the effect of context, and how the restaurant setting, and the nutritional quality of the foods being offered by the restaurant, has a large impact on the effect of menu-labelling on consumer choices. Thus, considering the prevalence of eating- outside the home, alongside the rising rates of diet related disease, and the alarmingly high calorie and sodium content of restaurant meals, it is important that menu-labelling interventions be considered by policymakers.

Contributions Mary Scourboutakos and Mary L’Abbé conceived the study and designed the survey. Spencer Henson and Julio Mendoza assisted with the administration of the online survey and provided input regarding the manuscript. Mary Scourboutakos and Paul Corey analyzed and interpreted the data. Mary Scourboutakos drafted the manuscript and all authors reviewed the manuscript for intellectual content.

Chapter 7 Does a healthy eating intervention in a buffet-style university dining hall change students’ food and beverage choices?

This article has been submitted for publication: Mary J. Scourboutakos, Catherine L. Mah, Sarah Murphy, Frank Mazza, Mary R. L’Abbé. 2016. Does a healthy eating intervention in a buffet- style university dining hall change students’ food and beverage choices. Public Health Nutrition.

7.1. Abstract Objective: To test the effect of a population-level nutrition intervention that included education and two labelling components—encouraging fruit and vegetable consumption while discouraging sugar-sweetened beverage consumption—on students’ food choices.

Design: This was a pre-post, quasi-experimental study to measure the effect of a population nutrition intervention in a university dining hall, between September 2014 and April 2015. Beverage choices and vegetable/fruit bar visits, were measured through direct-measured observational data collection, on six occasions before and six occasions after the intervention. Cafeteria inventory data was also collected. Pre-post data were compared .

Setting: This study was conducted in a buffet-style dining-hall in a University campus residence, where students pay a set price and can consume ―all they care to eat‖.

Results: 368 to 510 students visited the cafeteria on each of the data collection dates. There was a significant difference in the proportion of students selecting a sugar-sweetened beverage before (49%) versus after (41%) the intervention (p=0.004) and the proportion drinking water before (43%) versus after (54%) (p<0.001). There was a significant increase in the proportion of students taking fruit after (36%) the intervention (p<0.001) versus before (30% of the population). The number of students visiting the vegetable bar significantly increased from 60% to 72% (p<0.001). There was no correlation between the data collected via direct-measured observation and the inventory data.

Conclusions: This population level intervention may be one way to encourage healthy dietary choices in campus buffet-style dining halls.

7.2. Introduction The diets of university student populations have been shown to be low in fruits and vegetables,182-184 and high in calories, sugar, fat and sodium.184, 185 Young adults/adolescents on average consume 230 calories from sugar-sweetened beverages each day.186 Studies have demonstrated that sugar-sweetened beverages are a major determinant of weight-gain and obesity.187-189 Furthermore, the transition from high school to university has been identified as a critical period for weight-gain.190-196 Thus, understanding food choices at this life-stage is important for assessing the potential consequences for health later in life.197-200

In recent years, there has been a shift towards ―buffet style‖ dining halls in Universities to give students greater flexibility and more food choices.201 In these unstructured eating environments, the traditional cafeteria line-up, which often dictates a choice from each food group, has been replaced with a ―food court‖ style set-up that gives students freedom and flexibility in their food choices. Such settings override many of the barriers (such as cost, lack- of-access, and convenience) that prevent individuals from consuming fruits and vegetables.202 However, research has shown that students cite the readily available abundance of food in their campus dining halls as a major cause of weight gain.201

University students often have little knowledge about nutrition.203 For instance, one study showed that university students are completely unaware of the calorie content of their beverages.204 As a result, efforts to educate University students about nutrition have been studied in classroom205-209 and online settings210-212 and have produced significant changes in dietary habits. However, it is unrealistic to expect all students to participate in a nutrition program/course. Thus it has been suggested that educational programs that aim to encourage healthy food choices should take place in settings where food selection actually occurs, such as cafeterias or dining-halls.122

Numerous educational interventions have taken place in University cafeterias and it has been shown that labelling/educational intervention which direct consumers to a preferred choice without requiring any kind of comparison or interpretation may be beneficial because they reduce the cognitive load and simplify food choices with a clear nudge towards a preferred

92 choice.116,117,118,119 That being said, other research has suggested that giving people self-relevant personal health consequences (ex. soy will help lower your risk of heart disease) is more likely to change their behaviour than simply telling them ―soy has important phytochemicals‖.121

Physical Activity Calorie Equivalent labelling (PACE labelling) is a form of labelling that illustrates the calorie content of a food according to the minutes of physical activity that are required to burn the calories in that food. PACE labelling puts energy content into context and makes it more relatable, which is important because research shows that students are completely unaware of the calorie content of their beverages.204 Additionally, young people have said that shocking educational messages are needed to reduce their consumption of sugar-sweetened beverages,204 and preliminary research to date has demonstrated that PACE labelling can deter energy consumption.107, 213

The objective of this study was to test the effect of a previously unstudied population nutrition intervention—that included education to encourage fruit and vegetable consumption via direct messaging (fill half your plate with fruits and vegetables) and discouraged sugar sweetened beverage consumption via PACE labelling—on students’ food and beverage choices in a buffet-style University dining-hall.

7.3. Methodology

7.3.1. Study Setting This was a pre-post quasi-experimental design to measure the effects of an intervention implemented in Burwash Dining Hall, a buffet-style cafeteria in a campus residence, supplied by Gordon Food Services, at the University of Toronto. The dining-hall is independent from the University of Toronto and is run by staff led by the head chef at the residence. At Burwash dining-hall, the majority of students are on a meal plan (administered by the university) and have pre-paid for their meals. Students simply swipe their card and can consume however much they choose, with no external factors like price or convenience, to influence their decisions. Each dinner costs twelve dollars.

Students eating in the dining-hall live in residence, are generally between 18 and 23 years of age and come from a wide range of academic programs. The population is 69% female and 31% male.

At each dinner the dining-hall offers a number of entrées (such as chicken, lasagna, fajitas), including at least one vegetarian option or Halal option (such as Szechwan tofu, spinach sauce with rotini), a daily soup (such as mulligatawny, minestrone, black bean or cream of leek), salad bar (including various kinds of lettuce, fresh vegetables, dressings), and a selection of fruits (apples, oranges, bananas) and breads, along with multiple side dishes, beverages, and desserts. Foods served during the study collection dates are described in Appendix F.

The beverage cups in the dining hall hold 8 fluid ounces. Students were free to take as many cups, of as many beverages as they wish. The 19 different beverage options available at the cafeteria are listed in Table 7.1. As for fruit, fresh apples, bananas and oranges were available daily. Pears were available occasionally. A fresh vegetable bar offering various types of lettuce and salad toppings (cucumbers, peppers, carrots etc.), as well as cooked vegetable options at an adjacent hot food bar were available each day.

7.3.2. The Intervention This intervention was designed and implemented by the study authors, and is described in detail below. The study took place between September 2014 and April 2015. The pre phase lasted from October 1st to January 31st and the post phase lasted from February 1st (when the intervention was implemented) to April 30th. Inventory data was collected each month from October to April and direct-measured observational data was collected on six menu occasions at dinner during the pre-phase (October 28th, October 29th, November 11th, November 12th, January 6th and January 7th) and on six matched-menu occasions during the post-intervention phase (February 10th ,February 11th, February 24th, February 25th, March 10th, and March 11th). The study protocol was approved by the research ethics board at the University of Toronto.

7.3.3. Beverage Education

The first part of the intervention was a beverage education campaign that encouraged students to ―drink water when you are thirsty‖ (via a 3x2.5 foot poster, Figure 1E) and used labelling that illustrated

94 the number of minutes of jogging it takes to burn the calories in the different beverage options offered in the dining-hall (via a 1x6 foot poster, Figure 7.1.d).

This idea was adapted from an info graphic produced by ―LiveWell Omaha‖, a public health NGO led campaign that has not been previously evaluated.214 The infographic illustrated a comparison of the number of packets of sugar in different beverages with the statement ―drink water when you are thirsty‖. However, in this study, instead of packets of sugar, minutes of jogging, a form of physical activity calorie equivalent labelling,107, 213 was used in combination with a positive ―drink water when you are thirsty‖ message.

The minutes of jogging on the poster was based on the total calories in each beverage according to the manufacturers’ data provided by Gordon Food Services. On the poster, and in the study analysis, the 19 beverages were categorized into the following groups: juice, soft drinks, chocolate milk, flavoured coffee/hot chocolate, diet soft drinks, coffee/tea and plain water (Table 1). Regular (plain) milk was not labelled because we did not want to discourage or encourage milk consumption. The minutes of jogging for each beverage category was averaged and estimated using the Metabolic Equivalent of Task equation [1 MET = kcal/(Kilograms/hour)].215 The calculation was based on a ―general jogging‖ MET value of 7 and the mean weight of a Canadian adult, 75.6 Kg, according to the Canadian Health Measures Survey (CHMS) (2009-2011).216

Signs that said ―did you know this water is filtered‖ (10x30 inches in size) were installed in September 2014, before the study began. They were present throughout the pre and post phase to ensure that ―fear of tap water‖ was not deterring water consumption during the study.

7.3.4. Fruit and Vegetable Education This part of the intervention was a nutrition promotion campaign promoting fruit and vegetable consumption. It was comprised of educational posters hung at the entrance of the dining-hall (not shown), at the ―fruit station‖ (4x5 feet in size, Figure 1C), and posted on top of the vegetable bar (8.5x11 inches, Figure 1A). In addition, a banner giving students ―100 reasons to eat fruits and vegetables‖ was hung in the entranceway to the servery (2x20 feet, Figure 1B). The positioning of the messaging within the dining-hall was strategically selected to make the information attention-grabbing and to maximize exposure to the intervention, by being immediately visible to everyone entering the dining-hall.119

To encourage students to eat more fruits and vegetables, text stating ―fill half your plate with fruits and vegetables‖ and ―this is what a balanced meal looks like‖ was posted alongside Harvard’s

Healthy Eating Plate info graphic.217 The Harvard Plate was used because it is an ―infographic‖ food guide that can quickly convey information to cafeteria diners. Furthermore, the message to ―fill half your plate…‖ is currently being used by numerous governments and organizations.218-220 This messaging was combined with a rationale for fruit and vegetable consumption (via the ―100 reasons to eat fruits and vegetables‖ banner). The content of the posters was based on principles from the literature which showed that positive health messages (―eat this‖) framed in terms of the benefits, are more effective than negative messages (―don’t eat that‖) when the target population is University students who eat on campus. 221 222

7.3.5. Intervention Outcomes We measured two primary outcomes: beverage consumption and fruit and vegetable consumption. The intervention aimed to decrease consumption of sugar-sweetened beverages as well as increase fruit and vegetable consumption. These outcomes were chosen for two reasons: 1) data on students’ selection of sugar-sweetened beverages, fruits, and vegetables could be feasibly collected via direct-measured observation and reliably monitored via the dining-hall’s inventory [with the exception of vegetables which could not be monitored via inventory data]; and 2) these are commonly over-consumed and under-consumed components of University students’ diets,182-184, 187-189 respectively.

7.3.6. Data Collection

7.3.6.1. Direct-measured Observational Data Collection In-person, direct observational data collection took place during dinner from 4:30 to 7:30 pm, on six specific menu occasions before the intervention and on the same six menu occasions after the intervention. The dining-hall has a five-week rotating menu cycle, therefore, data collection dates (listed in Appendix F) were selected to ensure that data were collected on days when the same foods were being served before and after the intervention (―matched menus‖). We collected data on Tuesday and Wednesday evenings (because these were the most highly attended weekdays) during three of the five menu cycles.

A two-person data collection team, including the lead author (MS) and a research assistant trained by MS, was stationed in the dining-hall for the duration of dinner on the twelve data collection dates. One surveyor was stationed in the beverage area. Using a clipboard and a pencil, this surveyor recorded the gender and beverage choice of every student who selected a beverage, in addition to the measured cup volume (measured visually in quarter cup intervals, e.g. full cup, ¾ cup, half-cup, ¼ cup) of beverage taken. This observer was conspicuously located beside the beverage station, thus, students were aware

96 that a study was taking place. When they inquired about the study (<1% of students inquired throughout the entire study duration), students were told that the study was investigating food waste.

The second data collector was stationed near the fruit and vegetable bar to measure fruit and vegetable choices. Using a clipboard and pencil, this surveyor recorded the frequency of students taking each whole fruit (ex. applies, bananas, oranges and pears) as well as the frequency of students taking vegetables from the fresh or cooked vegetable station. They also recorded the gender of each student taking fruits or vegetables. The denominator for fruit and vegetable analyses was based on the total number of students visiting the cafeteria at each dinner during which data were collected. This information was provided by cafeteria staff who routinely record the number of attendees.

A description of direct observational data collection dates included in the final analysis can be found in Appendix F. Direct observational data collection for fruits on February 11th was excluded from the final analysis because a Valentine ’s Day dessert buffet confounded students’ fruit selections on that date. In addition, the first two data collection dates were collected using a different methodology and were excluded for consistency. Rate of fruit disappearance was the original data collection method, however it yielded inaccurate results and was thus changed to direct-measured observation.

7.3.6.2. Cafeteria Inventory Data Collection Inventory data, representing long-term secular trends in the cafeteria’s supply purchasing were collected as a secondary data source. The inventory data analysis was originally intended to verify that the data collected via direct observation was not the result of social desirability bias. Inventory data were provided by dining-hall staff via their suppliers on a monthly basis. Inventory data represented consumption during breakfast, lunch and dinner for the entire study duration. The number of cases of beverages ordered each month from September 2014 to April 2015 was provided. The number of cases of fruit ordered each month during the study period was also collected. Inventory data for vegetables were not available. Data for September, December and April were excluded from inventory data analyses because these months were not full academic months and were consequently under attended, hence ordering was atypical.

7.3.7. Analysis The total number of students choosing each beverage (e.g. soft drinks, juice etc.) was tabulated and calculated as a proportion of total visits to the beverage station. Beverage volume was converted from the cup measure (full cup, ¾ cup, half cup etc.) to an exact volume (mL) and further converted to calories for all caloric beverages. Logistic regression was used for primary analyses comparing the proportion of

97 students selecting each beverage before and after the intervention, while controlling for covariates including gender and the menu served on each data collection date. In secondary analyses, data for all sugar-sweetened beverages (juice, soft drinks, chocolate milk, flavoured coffee, hot chocolate) taken before and after the intervention were aggregated and compared. The average calories or milliliters (for non-caloric beverages) of each beverage taken by students before and after the intervention was calculated and compared using ANCOVA (proc glm) and monte-carlo simulation of the exact p-value (because data were not normally distributed) while controlling for gender and menu. The proportion of students visiting the salad bar and the proportion taking fruit before versus after the intervention was compared using logistic regression while controlling for menu and gender as a covariate.

Inventory data for fruits were not transformed. Inventory data for beverages were converted into liters of beverage. The number of litres of each individual beverage ordered each month from the suppliers was adjusted to a weekly average to account for missed weeks such as reading week, and final exams week when cafeteria attendance was lower than usual. Proc correlation was used to test the agreement between the inventory and direct-measured observational data.

All analyses were conducted using SAS version 9.3 (SAS Institute Inc.).

7.4. Results A range of 368 to 510 students visited the dining-hall for each dinner meal on the nights when data were collected [Appendix F]. Overall, a total of 6412 visits to the beverage station were recorded (3232 pre and 3180 post), totaling 8570 cups filled with beverages (table 2). 56% of beverages were taken by females and 44% were taken by males. A total of 3668 visits to the vegetable bar were recorded with 36% of visits made by males and 64% made by females. A total of 954 visits to the fruit station were recorded with 44% of visits made by males and 56% made by females.

Inventory data for the average liters of each beverage and the number of cases of fruit ordered each month were included for the month of October, November and January in the pre-phase as well as February and March in the post-phase of the intervention.

7.4.1. Beverages (direct observational collection) At baseline, the most popular beverage choices among students were water (40% of cups were filled with water), juice (18%), soft drinks (14%), and chocolate milk (8%) (table 2). After the intervention was implemented, there was a small decrease in the proportion of students taking soft drinks, juice, chocolate milk, flavoured coffee and hot chocolate, however, this trend was not significant. When

98 all sugar-sweetened beverages were combined, there was a significant decrease (p=0.004) in the proportion of students selecting a sugar-sweetened beverage before (49% of the population) versus after (41%) the intervention (Figure 7.2). The average amount of calories or milliliters (for non-caloric drinks) of each beverage being taken before versus after the intervention was not significantly different (Appendix F).

There was a significant increase (p<0.001) in the proportion of students drinking water before (43%) versus after (54%) the intervention (Figure 7.2).

7.4.2. Fruits and Vegetables (direct observational collection) There was a significant difference (p<0.001) in the proportion of students taking fruit before (30% of the population) versus after (36%) the intervention (figure 3). The proportion of students visiting the vegetable bar significantly increased from 60% to 72% (p<0.001).

7.4.3. Inventory Data Results Inventory data for beverages showed some trends towards decreases (among chocolate milk, apple juice, orange juice and flavoured coffees) however, there were equally large increases observed among other beverages, such as cranberry cocktail and lemonade (table 3). Inventory data for the number of fruit cases ordered each month showed no discernable trend (figure 4). With the exception of milk, inventory data for juices, chocolate milk, flavoured coffee/hot chocolate, tea/coffee, apples, oranges, bananas and pears were not correlated with direct observational data collection findings.

Table 7.1. Calorie content of beverages at the cafeteria and estimated minutes of jogging using in the Physical Activity Calorie Equivalent (PACE)* labelling

Estimated Minutes of Beverage Calories per Beverage Jogging* category 8oz cup

Water Water 0 0 Coke Zero 0 0 Diet Soft Drinks Diet Coke 0

Coffee & Tea Coffee & Tea 0 0 Ginger Ale 87 Sprite 95 Soft Drinks 16 Orange Pop 100 Coke 104 Orange Juice 95 Lemonade Juice 95 Juice Cranberry Juice 104 16 Peach Juice 109 Apple Juice 114 Flavoured Swiss Mocha Capuccino 149 Coffee/Hot French Vanilla Coffee 162 25 Chocolate Hot Chocolate 162 Chocolate Milk Chocolate Milk 170 27 Plain Milk** 2% Milk 123 na *Physical Activity Calorie Equivalent labelling (PACE labelling) is a form of labelling that illustrates the calorie content of a food according to the minutes of physical activity that are required to burn the calories in that food. Preliminary research has demonstrated that PACE labelling can deter energy consumption.(38,39) **Regular (plain) milk was not labelled in the intervention because we did not want to discourage or encourage milk consumption.

100

Figure 7.1. Images of the beverage, fruit and vegetable labelling/education intervention

A C D

B E

A-Harvard's Healthy Eating Plate infographic displayed on top of the vegetable bar (8.5x11"); B-Large banner giving students "100 reasons to eat fruits and vegetables" at the entraceway of the cafeteria (2x20 feet); C-Harvard's Healthy Eating Plate signage at the fruit station (4x5 feet); D-Signage illustrating PACE (physical activity calorie equivalent) labelling illustrating the number of minutes of jogging it takes to burn the calories in each of the beverage options offered (1x6 feet); E-A view of the beverage station including signage encouraging students to "drink water when you are thirsty" (3x2.5feet).

101

Table 7.2. Sample characteristics at baseline including visits to the beverage and fruit station and frequency of beverage and fruit consumption

n Total recorded visits to the beverage station* 6412 Visits pre 3232 Visits post 3180 Total cups filled with each beverage 8570 Proportion of cups filled with each beverage at baseline

Water 1640 (40%) Juice 739 (18%) Soft Drinks 591 (14%) Chocolate Milk 321 (8%) Coffee/Tea 305 (7%) Milk 267 (3%) Diet Soft Drinks 138 (7%) Flavoured Coffee/Hot Chocolate 105 (3%) Total recorded visits to the vegetable bar* 3668 Total recorded visits to the fruit station** 954 Frequency of each fruit taken

Apples 393 (41%) Bananas 292 (31%) Oranges 194 (20%) Pears 75 (8%) *Based on 12 data collection dates (6 pre and 6 post intervention) **Based on 9 data collection dates (4 pre and 5 post intervention)

102

Figure 7.2. Proportion of students who chose each beverage before and after the intervention

60 Pre * * Post

20 chose each beverage (%) beverage eachchose

Proportion of students students of Proportion who 10

0 Soft Drinks Juice Chocolate Milk Flavoured All SSBs Milk Diet Soft Coffee/Tea Water Coffee & Hot Combined Drinks Chocolate

Sugar-sweetened beverages (SSBs) Unsweetened beverages

*p<0.001 Beverage choices

103

Figure 7.3. Proportion of students who visited the fruit station and vegetable bar before and after the invention

100

80 *

(%) * 40

10 the vegetable bar and fruit station andfruit bar the vegetable Proportion of students who visited visited students of Proportion who 0 pre post pre pre n=1691 n=1247 n=2981 n=2613 Fruits Vegetable

s *p<0.001

104

Table 7.3. Inventory data illustrating the weekly average number of litres of each beverage taken before and after the intervention

Percent Pre* Post† change Mean Mean litres per litres per % Beverage week week 2% Milk 313 303 -10 Chocolate Milk 245 217 -9 Cranberry Cocktail 85 128 15 Lemonade 104 123 12 Apple Juice 184 110 -6 Orange Juice 224 208 -9 Peach Juice 120 151 13 Total Juice 717 720 10 French Vanilla Cappuccino 26 19 -8 Swiss Mocha Cappuccino 23 5 -2 Hot Chocolate 34 49 14 *Pre data was based the weekly average over 12 weeks †Post data was based on the weekly average over 7 weeks. Reading week and three weeks in April were excluded due to decreased attendance at the dining- hall. Note: Inventory data for soft drinks was not provided by the suppliers and therefore was not included in this analysis

105

Figure 7.4. Inventory data trends in the cases of fruit ordered each month

70 PRE POST

50 Bananas

40 Apples

30 Oranges 20

10 Pears

Number of Cases Ordered Cases of Number 0 Sept October Nov Dec Jan Feb March April Month Apples Bananas Pears Oranges

106

7.5. Discussion

This study showed that a population nutrition intervention that included education and two labelling components—including ―100 reasons to eat fruits and vegetables‖, as well as messaging to ―fill half your plate with fruits and vegetables‖, ―drink water when you are thirsty‖ and a comparison of the minutes of jogging it takes to burn each of the beverages available—could result in a modest decrease in the number of students choosing sugar-sweetened beverages, and a modest increase in the number of students choosing water, and visiting the fruit and vegetable bar.

These findings were similar but not always consistent with previous studies discouraging sugar- sweetened beverage consumption in various related contexts using various comparable interventions. For instance, the results of the Dutch Obesity Intervention in teenagers—which also utilized a multi- component intervention to discourage sugar-sweetened beverages and promote water—similarly found a decrease in sugar-sweetened beverages, but unlike our results, they did not see a simultaneous increase in water.223

A previous study employing physical activity based labelling resulted in consumers ordering meals with 93 fewer calories,107 while another study in females showed decreases of 162 to 206 calories.213 In our study, we found that students who continued to choose sugar-sweetened beverages after the intervention were on average consuming the same number of calories as before the intervention. Therefore, the intervention decreased the number of students choosing sugar-sweetened beverages, but did not change the amount of calories the habitual drinkers were consuming.

Another increasingly popular disincentive for sugar-sweetened beverage consumption is taxation. For example, a meta-analyses of taxation found that it led consumers to replace sugar-sweetened beverages with lower-calorie alternatives such as fruit juice, whole milk, and diet soft drinks, i.e., switching behaviour.224 In our study, where price was not an influence on choice, students did not appear to compensate for a reduction in sugar-sweetened beverage consumption with a switch to lower-calorie alternatives, as we did not observe any increases in fruit juice, diet soft drinks or milk during the post phase. We postulated whether the supplementary statement that we included in the intervention, ―drink water when you are thirsty‖, may have encouraged students to specifically replace sugar-sweetened beverages with water. However, this is not certain. Further research is needed to disentangle the relative effects of labelling, nutrition promotion messages such as the ―drink water when you are thirsty‖ statement, and other factors such as price to discern whether one component of the intervention is mainly responsible for switching effects, or whether there was a synergy by having multiple components present.

107

Previous research in a university convenience store showed that simply tagging healthy options with a label that says ―fuel your life healthy campus‖ can increase purchases of tagged items by 3.6%,118 which was a smaller increase than was observed in our study where the number of students choosing fruit increased by 6%, while water increased by 11%, and vegetables increased by 12%. The higher increase observed in our study could be explained by research which has shown that educational strategies that link knowledge about the attributes of a food with self-relevant health consequences are more likely to produce behaviour change.121, 221 Our study did this by combining the directional statement ―fill half your plate with fruits and vegetables‖ with a banner highlighting ―100 reasons to eat fruits and vegetables‖ aligning with previous research which has shown that providing students with educational information about the benefits of a food may be superior to simply highlighting healthy choices.121

The lack of correlation between the inventory data and the data collected via direct observation suggests that results should be interpreted with caution. However, there are a number of potential explanations for this lack of correlation. First, the data collected via direct observation only represented choices made at dinner, while inventory data represented all choices made during breakfast, lunch and dinner. Second, there were a limited number of data points in the inventory data, as September, December and April had to be excluded because they were incomplete months, leaving only five months (three pre and two post). Third, inventory decisions at the cafeteria were retrospective rather than prospective, i.e., using past consumption as a prediction for future consumption. For example, if the dining-hall ordered an excess of a certain beverage in September, they would order less in October. These limitations illustrate some of the problems of using inventory data in a program intervention setting, even though estimates of food availability in terms of commodity supply and disappearance are a commonly used economic metric for consumption, and in evaluation of policy interventions.225 Essentially inventory data are an availability measure and are perhaps more appropriate as a measure of environmental change rather than a proxy for consumption. For instance, in our study, the sharp drop in oranges observed during the month of October represented a supply shortage, not a decrease in consumption. Inventory data would also be more useful when used for longitudinal analysis in time series.

Another reason we used inventory data was as a second data source to assess potential social desirability bias in data collected through direct observation. It remains unclear as to whether we were able to meet this objective. In theory, the observed decrease in sugar-sweetened beverages was more likely to be influenced by the effect of social desirability bias because the surveyor was very conspicuous. Meanwhile, the observed increase in fruits and vegetables was less likely to be influenced, because the recorder was less conspicuously located. Despite this, the magnitude of change was similar for both

108 beverages and fruits/vegetables. Therefore, the impact of social desirability bias may not have been a major factor influencing the observed results in this study.

7.5.1. Limitations It has been shown that 60% of university students are interested in receiving nutrition information.226 Therefore, these results are not generalizable to other populations as this study was conducted on a population more likely to be responsive to the intervention. Furthermore, these findings may not be generalizable to other settings because research has shown that food selection behavior is more likely to be influenced in situations with a limited range of choices—similar to the beverage station at the dining-hall—where labelling can contrast the options available and provide a clear cue to the preferred choice. 97 Lastly, generalizability was further compromised by the fact that price was not a factor in decision making in this dining-hall.

Other limitations include our choice of outcome measure: we used food and beverages taken to measure choice, and did not correlate this with actual food eaten or food waste. Additionally, it should be noted that although we measured whole fruits and vegetable items taken, this underestimated produce consumption, particularly total vegetables taken by students because often vegetables were mixed in with the hot entrées and thus could not be accounted for via the data collection methodology employed.

7.5.2. Strengths An advantage of this study is that it tested the intervention over a long period of time; however, as a result, temporal changes may have biased food choices. For example, previous research has shown that university student’s cafeteria snack purchases become less healthy with each passing week of the semester;227 this was not observed in our study. Another strength was that efforts were made to minimize the effect of seasonal-based food choices. For instance, the intervention was started in February to prevent January ―weight-loss‖ resolutions from impacting results. Interestingly, in our study, food/beverage choices in January were no different from any other month. While a randomized controlled trial would be a future direction to overcome the limitations of the pre-post design, many of the challenges faced in this study—including the orange shortages and special food days such as chicken wing night, and the Valentine’s dessert buffet—would still be problematic, even in a more rigorous study design. Furthermore, the long duration of the study illustrated that a wash-out effect was not observed two months after the interventions were implemented.

109

7.6. Conclusion This study tested an intervention comprised of nutrition education and two labelling components (focused on beverages and fruits/vegetables) that we have shown may be a useful strategy to decrease students’ consumption of sugar-sweetened beverages, and increase consumption of water, fruits, and vegetables. Young adulthood is a transitional time where the prevalence of obesity has been shown to double.228 While obesity prevention requires multiple interventions—that not only address student food choices but also the food environment, physical activity, stress management and other factors that contribute to unhealthy lifestyle choices and weight-gain—addressing the poor dietary choices of young adults in University settings is one means by which obesity and other diet-related diseases can be prevented later in life. This particular intervention, that capitalized on University student’s intellect and provided educational information alongside labelling in a setting that minimized other barriers to healthier food choices such as price, may be one way to influence dietary choices effectively in campus buffet-style dining halls.

Contributions

Formulation of the research question (MS, ML, CM); study design (MS, ML, CM, NB, BM); data collection (MS, SM, FM, NB, BM); data analysis (MS); data interpretation (MS, ML, CM); drafting of the manuscript (MS); critical review of the manuscript (MS, CM, SM, FM, NB, BM, ML).

110

Chapter 8 Overall Discussion

At the outset of this thesis there were no data on the nutritional quality of Canadian chain restaurant foods and there was no menu-labelling legislation in Canada. Therefore, the objectives of this thesis were to 1) investigate calorie and sodium levels in the restaurant food supply; 2) explore consumers’ use of menu-labelling; and 3) test the potential of alternative forms of labelling in non-chain restaurant settings, with an ultimate aim to inform policy decisions around issues concerning foods consumed outside-the-home.

The results of this thesis produced four key findings. First, it was shown that there was wide variation in calorie levels in restaurant foods, and that portion size, as opposed to calorie density, was the most important driver of calories. Second, we found that the sodium levels in menu items from chain restaurants often exceeded daily recommendations and despite stated efforts by the Canadian restaurant industry to improve, as of 2013, reductions were minimal. Third, including sodium information on restaurant menus encouraged consumers to choose lower sodium foods; however the benefit of menu-labelling varied greatly depending on the restaurant setting, with the greatest effects usually seen in restaurants with the highest calorie items. And finally, the fourth key finding is that in non-restaurant settings, such as a University dining-hall, there may be potential to increase fruit and vegetable consumption, and decrease sugar- sweetened beverage consumption among university students using a labelling/education intervention. The following sections will discuss these findings in detail and when applicable, examine their impact on policy decisions.

8.1. The restaurant food supply

8.1.2. Calories

In Chapter 3 we investigated the calorie levels in chain restaurant foods and found a wide range within each restaurant and food category. This illustrates that in chain restaurants, there is no rule of thumb for selecting low-calorie foods. Hence, these data provide a strong rationale for calorie labelling because without labels, there is no way for consumers to predict which items are high in calories and which items are low in calories.

111

The results in Chapter 3 also demonstrated that serving size, as opposed to calorie density, was the most important driver of total calories in restaurant foods. This novel finding may have implications for the usefulness of calorie labelling as it showed that when consumers select a low-calorie option, they are likely selecting a menu-item with a smaller portion size, and they are not necessarily selecting a menu-item with a lower calorie-density. The potential consequences of this were illustrated by Roberto et al, who showed that after consumers used calorie labelling to select lower calorie meals, they later compensated for the calories that were ―saved‖ by eating more later in the day.92 These findings provided motivation to test the benefit of labelling portion size (in grams) in the survey described in Chapter 6 and discussed further below.

Lastly, the studies in Chapter 3 and Chapter 4 illustrated that foods from sit-down restaurants were on average higher in calories and sodium when compared to similar foods from fast-food restaurants. This had been previously shown in children’s meals from the United States.44 This is an important finding because it contradicts popular perceptions of fast-food and furthermore, it has implications for the interpretation of the menu-labelling evaluations currently being conducted exclusively in fast-food restaurants in New York City. This will be discussed further in the ―menu-labelling‖ discussion below.

8.1.3. Sodium

At the start of this thesis, Canada’s Sodium Working Group had just been disbanded and their plan to set targets and monitor sodium reductions from baseline (2010) through to 2016 was abandoned. This provided the impetus for the study in Chapter 4, which comprehensively documented the sodium levels in Canadian fast-food and sit-down chain restaurants in 2010. The results of this study illustrated that sodium levels were alarmingly high and often exceeded the daily AI and UL. Hence, these results emphasized the importance of implementing a sodium reduction strategy in Canada and the need to include targets and timelines for the restaurant sector.

The original plan for this thesis did not include a longitudinal follow-up study of sodium levels. However, when the data from Chapter 4 was published, the President and CEO of Restaurants Canada (formerly known as the Canadian Restaurant and Foodservice

112

Association), Gareth Whyte, commented in the Globe and Mail ―my response would be, a lot has been done since 2010.‖164 This assertion provided motivation for the longitudinal study of changes in sodium levels from 2010 to 2013, presented in Chapter 5. The results showed that over a three year period, there was on average, only a 25 milligram decrease (a <3% reduction) in sodium per serving. Furthermore, sodium levels increased in 16% of foods and decreased in only 30%, while the remainder (54%) stayed the same. These findings are significant because they illustrate that in the absence of sodium reduction targets and active monitoring of this sector, restaurants were not motivated to make substantial reductions. The SWG’s targets recommended 25-30% reductions from baseline in 2010 to 2013 and according to this study; mid-way progress was not on track to meet these reductions.

That being said, research from the United States suggests that simply establishing targets may not be sufficient to produce meaningful decreases in sodium. For example, a similar study of longitudinal changes in sodium levels in 78 American fast-food restaurants from 2005 to 2011, showed that despite the existence of a National Sodium Reduction Initiative (established in 2009, with specific targets for restaurant foods), sodium levels increased by 2.6% over this time period.162 This has implications for Canada, as it suggests that voluntary targets may be insufficient. Thus, while our results demonstrate that the creation of Canadian sodium reduction targets specifically for restaurant foods is needed, these targets would likely be futile without a transparent, public monitoring system, with strong government enforcement, as the SWG recommended.51

8.1.3.1. The Challenge of Reducing Sodium

The potential success of a publicly monitored reduction strategy was previously demonstrated with trans fats, as our research showed that nearly all restaurant foods currently meet the trans fat limits.229 However, reducing sodium is not as straightforward as the removal of trans fats, where vegetable oils could simply be substituted for trans fat.

Ideally, a gradual reduction in sodium across the food supply is the most efficient way to remove sodium as it has been shown that repeated exposure to lower sodium foods makes humans more sensitive to the taste of salt.157, 171, 230 However, a gradual reduction in sodium requires a food supply-wide approach, which is currently lacking in Canada. Therefore, in an

113 ideal scenario, salt substitutes would not be necessary. However, under a voluntary reduction strategy it is possible that restaurants may be more likely to resort to salt substitutes, because in the absence of a food supply-wide approach, the population may not be able to recalibrate their taste buds.

There are many technical challenges associated with salt substitutes such as potassium chloride—the most popular salt substitute—where levels are limited by the fact that it can have a bitter and metallic taste, and therefore it can only replace a certain percentage of the salt in a food because salt is still needed to be present to mask its taste.172 Furthermore, the expense of sodium replacements is another challenge, as salt is the cheapest way to add flavor.172,155

Our research did not investigate the prevalence of salt substitutes in the restaurant food supply. At this point in time, restaurants are not required to disclose ingredients lists; therefore the prevalence of salt replacements in the restaurant food supply is currently unknown. However, the market share for salt substitutes is increasing every year, and has been projected to grow to $1.24 billion by 2020.174 Therefore, if a sodium reduction strategy was implemented, it would be important to monitor the prevalence of these substitutes in order to understand the mechanism of how restaurants and the food supply in general are achieving reductions. Additionally, the long term consequences of higher intakes of salt substitutes (including magnesium sulfate, choline chloride amino acids (arginine, lysine, glutamate, lactate), alapyridain, umamix, and salt enhancers,172, 155, 231 ) are unknown, and future research will be needed to determine whether these replacements truly improve the healthfulness of these foods.

8.1.4. The big picture of restaurant nutrition…beyond calories and sodium

While sodium reduction is a priority, the nutritional issues associated with restaurant foods extend beyond sodium. For example, our research in Appendix 1 (―Restaurant meals: Almost a full day’s worth of calories, fats and sodium‖) and Appendix 2 (―Sugar levels in kids’ meals from Canadian chain restaurants‖) demonstrated that these foods are also high in other nutrients of public health concern such as saturated fat, and added sugars.161 Meanwhile other studies have similarly shown that restaurant foods lack fiber, contain few fruits and vegetables, and are highly palatable.232 Considering this, one might conclude that the hallmark of restaurant foods is the presence of potentially unhealthy nutrients in unhealthy amounts that can elicit

114 hedonic responses.140, 233 Therefore, it is uncertain whether any degree of reformulation will ever make restaurant food ―healthy‖ as there are many issues, beyond merely sodium, that would need to be addressed.

Furthermore, some restaurants have responded to the call for improvement by introducing seemingly healthier options like salads and yogurt to their menu. However, our research showed that a ―healthy‖ salad could still contain more calories than a hamburger,151 and furthermore, we found that the yogurt that has been added to ―healthy‖ kids’ meals contained high amounts of added sugars.234 Therefore, public health efforts to improve the population’s diet should focus on promoting home cooking, as the restaurant industry’s efforts to introduce healthier options and reformulate their products, may only offer marginal improvements.

8.2. Menu-Labelling

8.2.1. Serving size labelling

Based on the findings in Chapter 3—which showed that portion size, as opposed to calorie density, is the most important predictor of total calories in restaurant foods—the study in Chapter 6 sought to test whether including portion size information (labelled as the total grams per serving) on restaurant menus would help consumers identify situations where foods are lower in calories simply because they are small. This was the first study ever to investigate the benefit of including serving size information on menus, and the hypothesis was that including this information would lead consumers to select meals with a lower calorie density. However, the results showed that this information did not lead consumers to select meals with a lower calorie density.

A possible explanation for this finding may be the fact that this was a survey and we did not observe the actual amount of food being consumed. A previous study found that while calorie labelling did not influence consumer’s choices, it did lead consumers to eat less of their meal with more food being leftover.80 This illustrates that the findings in Chapter 6 should be interpreted with caution because real-life purchases and the actual amount of food consumed were not measured. Furthermore, this also highlights a limitation of the menu-labelling

115 evaluations conducted to date in New York City, as these studies only consider foods purchased, as opposed to foods actually consumed.88, 235, 236, 91

8.2.2. Sodium labelling

The menu-labelling study in Chapter 6 was novel because it tested the benefit of including sodium information, alongside calorie information, on restaurant menus. This specific combination of labelling sodium in addition to calories had not been previously investigated, despite that fact that it was recommended by the SWG. Furthermore, this was a relevant research question in Canada because the original federal menu-labelling bills first introduced in 2003 and 2006 recommended the inclusion of sodium information on menus. Our results showed that including sodium information on restaurant menus could lead to an average decrease of 171 to 384 mg of sodium per person; furthermore, those who actually use the information could save between 681 to 1360 mg.

When New York City implemented their calorie labelling bill in 2008 they chose not to label sodium because they found that calorie and sodium levels in restaurant foods were correlated.81 Therefore, when consumers choose a lower calorie meal, they are automatically also choosing a lower sodium meal. However, results in Chapter 6 showed that while calorie labelling produces an inadvertent decrease in sodium, including sodium information alongside calorie information produces an additional reduction of approximately 200 mg. This suggests that there may be a benefit to also including sodium information on restaurant menus.

These results informed Toronto Public Health’s technical report entitled ―What’s on the Menu – Making Key Nutrition Information Readily Available in Restaurants‖ and their recommendation to label both calorie and sodium on restaurant menus, which was presented to the Toronto Board of Health on April 29th 2013. Despite the fact that the food supply and menu- labelling data outlined in this thesis was presented to the Standing Committee on Government at Queen’s Park when the Making Healthier Choices Act was being debated, the Ministry of Health chose to only require the disclosure of calories in the Making Healthier Choices Act, which was passed on May 29th 2015. As a result, when the bill is implemented in January 2017, only calorie information will appear on chain restaurant menus in Ontario. The Ministry has said that it may consider adding additional nutrients in the future,237 however, it should be noted that this bill

116 revoked the jurisdictional authority of municipalities to label additional nutrients beyond calories. Therefore, despite their recommendation to label calories and sodium, Toronto no longer has the ability to craft their own municipal menu-labelling legislation that includes sodium labelling.

8.2.2.1. Sodium warning labels

Since 2012, Ontario MPP France Gelinas has sponsored numerous unsuccessful bills that called for sodium warning labels on restaurant menus. On June 10th 2015, a similar proposal was brought forth by the New York Department of Health and Mental Hygiene, which would require mandatory high sodium warning labels on chain restaurant menu for items containing more than 2300 mg of sodium. On September 9th 2015 this proposal was passed and on December 1st 2015 the warning labels (which depicts a salt shaker inside of a triangular yield sign) were implemented in New York City.238

Figure 8.1. New York City’s sodium warning label for restaurant menus

Our findings in Chapter 5 suggest that high sodium warning labels may not be the most efficacious way to inform customers of the sodium content of their menu items because according to our data, only 9% of individual menu items exceeded 2300 mg and would thus be required to display the label. However, additional research (Appendix 1), which evaluated restaurant entrées in combination with their accompanying side dishes, showed that 56% of sit-

117 down restaurant meals would qualify for the label.161 This demonstrates how a menu’s layout could influence the success of high sodium warning labels. If entrées and side dishes are listed separately, it’s likely that few items will be required to display the label, however, if whole meals are listed, a larger number of warning labels will be displayed. Therefore, our results suggest that warning labels may not be efficacious in fast-food restaurants, because few individual menu items exceed 2300 mg, however in sit-down restaurants, warning labels would likely appear beside many meals.

The New York Department of Health and Mental Hygiene chose to make 2300 mg the cut-off for the label so that consumers would clearly know when they’re consuming too much sodium, without having to make any assumptions.239 However, if the cut-off was at sodium’s daily AI (1500 mg), data in Chapter 5 showed that 32% of individual menu items would carry the label, while 84% of meals from sit-down restaurant meals would carry the label (Appendix 1).161 Thus, our results show that if the warning label cut-off was set at the AI, a substantial number of menu-items and meals would be required to display the label. This could be advantageous, because it would clearly warn customers of the high sodium levels in restaurant foods. However, our results in Chapter 6 illustrated that this could be disadvantageous because in many restaurant settings, consumers on average decreased the sodium level of their meals from 1819 mg to 913 mg, therefore numerical labelling helps consumers lower sodium levels even if they’re choosing meals that are lower than the UL. Meanwhile, warning labels will only deter consumption of meals exceeding the UL, and fail to give consumers a way to distinguish between menu-items that are 2300 mg versus menu-items that are 3300 mg or between menu- items containing 800 versus 2200 mg of sodium.

8.2.3. The gap in current evaluations of calorie labelling in New York City

The evaluations of menu-labelling in New York City to date, have demonstrated that this policy did not markedly lower the calorie level of consumers’ purchases.41, 88, 90, 91 However, the results of this thesis shed light on one of the limitations of the research conducted thus far in New York City.

Data in Chapter 6, showed that one of the predictors of menu-labelling was the calorie and sodium content of meals ordered at baseline. Hence, in restaurants where consumers are

118 more likely to get higher calorie and higher sodium meals, consumers are more likely to be surprised by the levels, and are thus more likely the use the labels to make a healthier choice. This agrees with previous research,56 such as Bollinger et al’s finding that individuals who ordered more than 250 calories per transaction at Starbucks decreased the calorie level of their order by 26%, which was far greater than 6%, the average reduction overall.89 Furthermore, data from King County Washington (described in Toronto Public Health’s technical report on menu- labelling) also found that despite very little change in total calories per transaction, a year and a half after the implementation of their legislation, consumers were less likely to make high calorie purchases, defined as meals exceeding 667 calories.70

Furthermore, data in Chapters 3 and 4 showed that sit-down restaurants on average have higher calorie and sodium levels compared to fast-food restaurants. Hence, it is understandable that the results in Chapter 6 showed that consumers were more likely to use menu-labelling in sit-down restaurants, where they saved on average 87 more calories compared to fast-food restaurants by having calorie labels present.

These findings are the first to suggest that menu-labelling is likely to be most efficacious in sit-down restaurants. This is important because to date, there has been only one evaluation of menu-labelling legislation in full-service restaurants in Philadelphia,95 and there have been no evaluations of calorie labelling in in sit-down restaurants in New York City, where evaluations of this policy have focused solely on fast-food chains.88 This is an issue, because the results of this thesis illustrate that the benefit of menu-labelling is context dependent and therefore, current evaluations of menu-labelling are not being conducted in the settings where labelling is likely to be most useful. Thus, present conclusions regarding the efficacy of calorie labelling are premature as they only focus on fast-food settings. Therefore, further evaluations are needed in sit-down restaurants.

8.3. Labelling/education in a cafeteria setting

Our investigation in Chapter 7 contributes to an existing body of literature which demonstrates that various nutrition labelling/education interventions have potential to influence food choices in cafeteria settings.119,118, 79, 111, 114, 115 The results showed that a population intervention that includes messaging to ―fill half your plate with fruits and vegetables‖, ―drink

119 water when you are thirsty‖ and a comparison of the minutes of jogging required to burn each of the available beverages (e.g. PACE labelling) could have a modest effect on discouraging sugar- sweetened beverage consumption, while increasing consumption of water, fruits, and vegetables. However, these findings should be interpreted with caution because the effect size was small and the results were found in a University student population that is likely to be more responsive to the intervention.

The results which demonstrate the effect of PACE labelling on decreasing sugar- sweetened beverage consumption provide evidence to contribute to the emerging debate around the implementation of ―activity equivalent‖ calorie labelling. At the start of this thesis, there were only four investigations of PACE labelling and three of the four were conducted using online surveys.107-110 This was the first study to investigate PACE labelling in a real-life setting. Furthermore, in April 2016, Shirley Cramer, the Chief Executive of the Royal Society for Public Health in London, England, published a commentary in the British Medical Journal which called for the introduction of ―activity equivalent‖ calorie labelling, with symbols showing how many minutes of several different physical activities are equivalent to the calories provided in the food product.240 The rationale for this is the belief that food packaging should not only provide nutritional information, but should also help motivate people to change their behaviour. The findings of this thesis can help inform future debates about this novel form of labelling.

Finally, one of the most important lessons learned from this study did not come from the results themselves, but rather from the experience executing the study. The lesson was that interventions in real-life are limited by the constraints of the setting and the collaborative relationship between the researcher and the cafeteria operator. Therefore, the design and execution of real-life interventions are constrained by the fact that they must honour the limitations of the real-life context over and above the scientific literature. For example, research has shown that putting healthier foods like fruits and vegetables at the start of the cafeteria line results in most consumers choosing these foods,113 however, cafeteria operators make decisions based on practicality and often choose to put these foods at the end of the line to avoid ―bottle necks‖ because these foods take longer to serve.241 Hence, real-life interventions must combine evidence with practicality to suite the challenges of each unique setting. Therefore, this study sheds light on the fact that in non-chain restaurant settings where policies like menu-labelling are not applicable, research may not be able to ascertain a single ―ideal intervention‖ to promote

120 healthy choices, because the ideal intervention for each setting is context dependent and requires a collaborative approach that balances evidence and practicality.

8.4. Future Directions

With the recent passing of the first calorie-labelling legislation in Ontario, future research in this area should focus on evaluating this policy in the Canadian context. The results of this thesis showcase many of the important research questions that should be addressed in studies evaluating this policy. Overall, the evaluation should focus on two primary outcomes, first, the effect of calorie labelling on consumer’s choices and consumption, and second, the effect of calorie labelling on the food supply.

8.4.1. Evaluate the effect of calorie-labelling on Ontario consumers’ food choices First and foremost, natural experiments investigating the types of foods Ontario consumers are ordering and the calorie content of their orders, before and after the policy is implemented, are essential. Lessons learned from New York City, and the research in this thesis, demonstrate that these studies should be conducted in fast-food and sit-down restaurants with a range of calorie offerings. Furthermore, my research also suggests that evaluations of menu- labelling should not only consider how many calories are being ordered, but also how many calories are being consumed, as it has been shown that even when calorie labelling does not influence the calories a consumer orders, it can still influence the total amount of calories that the consumer chooses to eat.80

Data in Chapter 3 illustrates that evaluations of Ontario’s menu-labelling policy should consider the calorie density and portion sizes of the meals being ordered before and after the policy, as this could provide insight into how consumers are using the labels and whether the policy is helping consumers choose meals with a lower calorie density, or whether it is resulting in consumers ordering meals that are lower in calories merely because they are smaller.

Data in Chapter 6 illustrated how the benefit of calorie labelling for the individuals who actually use the information is often masked when evaluations only consider the population effect. Thus, while ideally, the policy will have a measurable effect on population-wide calorie

121 levels, it will be important to separately quantify the magnitude of decrease in calories among the sub-set of the population who use the information, as we did in our survey, and as Dumanovsky and Vadiveloo did in their evaluations of calorie labelling in New York.29, 88 Furthermore, evaluations should also investigate the characteristics of the consumers who use the information to understand whether the policy is equitable and whether it affects consumers of different ages, ethnic backgrounds, socio-economic classes, weights, genders etc. Evaluations via natural experiments in real-life setting will improve upon the data reported in Chapter 6 because they will include factors such as cost and actual purchases, as opposed to purchase intentions.

In addition, research on the frequency of eating-out and the venues where people are eating-out—such as sit-down restaurants versus fast-food restaurants, chains versus independent restaurants—will also be important to understand if the policy impacts where people choose to eat. Lastly, because Ontario is a large jurisdiction it will offer opportunities to test whether there are differences in the use of menu-labeling in large cities versus small towns, or urban versus rural settings, as geographic differences affecting the efficacy of such policies have not previously been investigated.

8.4.2. Evaluate the effect of calorie-labelling on the nutritional quality of the Canadian restaurant food supply It is possible that the greatest benefit of menu-labelling may not come from its effect on consumers, but rather, may be its effect on promoting product reformulation. This is because reformulation could improve the nutritional intakes of everyone who eats-out at chain restaurants, irrespective of whether they use the calorie labels. In King County Washington, where calories, carbohydrates, saturated fat, and sodium are disclosed on menus, research showed that 18 months after the implementation of the policy, there was a 41±156 calorie decrease and a 108±541 mg decrease in sodium in the entrées being served at sit-down restaurants as a result of this policy.84

Monitoring the effect of Ontario’s menu-labelling policy on the calorie levels in Canadian chain restaurant foods is essential. If changes are detected, it will be important to ascertain whether changes are due to calorie density or portion size, as it is possible that restaurants may simply decrease portion sizes as a means to create the illusion of healthier, lower calorie offerings.

122

With the lack of sodium labelling in Ontario, it will be important to investigate if sodium levels in chain restaurant foods are increased to compensate for any decreases in calories that may result from the policy. Finally, it will be important to compare consumer choices and the nutritional quality of restaurant foods in Ontario with other jurisdictions, like New York, that have sodium warning labels to see if New York experiences larger decreases in the sodium level of their restaurant food supply as a result of the warning labels.

Additionally, beyond sodium and calories, it will be important to investigate changes in the overall nutritional quality of restaurant menu offerings from restaurants subject to the new labelling laws. Furthermore, the nutritional quality of new products, in comparison to discontinued products may also provide insight into the industry’s response to this policy. Lastly, market share weighted analysis of the nutritional quality of the restaurant food supply would provide further insight into the popularity of a menu-item in relation to its nutritional quality.

A systematic review of the impact of menu-labeling across socioeconomic groups suggested that menu-labelling may increase disparities in health among people of a higher and lower socio-economic status.242 Therefore, it is essential to monitor the effect of labelling on the nutritional quality of menu-offerings, as such changes may decrease health disparities, because if menu-labelling has effects on the food supply, everyone who eats-out will benefit, irrespective of their socioeconomic class.

8.4.3. Monitor Data Availability Furthermore, it will be important to monitor data availability in Canada. During the course of this thesis, the number of restaurants disclosing nutrition information decreased from 2010 to 2013. It is important to note that part of the US legislation requires the full disclosure of information for the thirteen core nutrients, plus calories and serving size;71 however, the legislation passed in Ontario, does not require the disclosure of any nutrients besides calories.102 Therefore, the future availability of the publicly available data that was utilized in this thesis is uncertain and thus should be monitored.

123

8.3.4. Investigate the nutritional quality of ready-made meals from grocery stores In addition to fast-food and sit-down restaurants, the prevalence of ―ready-made meals‖ at grocery stores indicates that this additional source of food prepared ―outside-the-home‖ may also be contributing to Canadians’ diets. A study in Luxemboug showed that ready-made meals from grocery stores account for 7% of total energy intake in their population and demonstrated an association between ready meals and abdominal obesity.243 Furthermore, at study of ready- made meals in England showed that they were high in saturated fat and salt, while being low in sugar.244 To date there is no research on the prevalence or consumption of ready-made meals, nor is there any research on the nutritional quality of these meals. Additionally, these meals are exempt from containing a Nutrition Facts table; therefore, future research should explore the potential benefit of labelling this other source of food from outside-the-home.

8.3.5. Evaluate changes in the prevalence of eating-out in the 2015 Canadian Community Health Survey In fall 2016, data from the 2015 collection of the Canadian Community Health survey will be released.245 This will provide opportunities to better understand the prevalence of food consumed outside-the-home in Canada, as well as the venues (fast-food versus sit-down) where people are eating-out. Longitudinally, this data will illustrate whether people are eating-out more or less and contribute to our understanding of the implications of eating-out on Canadians’ diets.

124

Chapter 9 Summary of Key Findings

The following section recapitulates the key findings from this thesis and the potential implications of these findings:

Key Finding #1: There is wide variation in calorie levels within each restaurant and food category, furthermore, portion size, as opposed to calorie density, is the most important driver of this variation. Implications: This finding provides a strong rationale for calorie labelling because without labels, there is no way for consumers to predict which items are high in calories and which items are low in calories. Additionally, this finding may have implications for the usefulness of calorie labelling as it shows that when consumers select a low-calorie option, they are likely selecting a food with smaller portion size, and they are not necessarily selecting a food with a lower calorie- density.

Key Finding #2: The sodium levels in menu items from chain restaurants often exceed daily recommendations and despite efforts to improve, as of 2013 reductions were minimal. Implications: This illustrates that the current voluntary approach to reducing sodium in the food supply is not working. Therefore, these findings demonstrate the importance of implementing a national sodium reduction strategy in Canada with reduction targets for restaurant foods and public monitoring of adherence. The high sodium levels that were observed illustrate that sodium warning labels, if implemented, would be present on a large number of foods, particularly if meal combinations are identified. Furthermore, considering the absence of significant reductions, public health efforts to improve the population’s diet, particularly their sodium intakes, should focus on promoting home cooking and less frequent eating-out.

Key Finding #3: There is a benefit to labelling sodium information, in addition to calorie information on restaurant menus, as this information may encourage consumers to choose lower sodium foods. However, the benefit of labelling varies depending on the restaurant setting. Implications: This finding demonstrates the missed opportunity in Ontario with the implementation of a calorie-only menu-labelling bill. Additionally, the variable effect of

125 labelling, particularly the fact that its rate of use and magnitude of impact are context dependent, highlight the major limitation of the current evaluations of menu-labelling in New York City which have all been conducted in fast-food restaurants where consumers are less likely to benefit from the information.

Key Finding #4: In University cafeteria settings, there may be potential to increase fruit and vegetable consumption, and decrease sugar-sweetened beverage consumption among students using a labelling/education intervention that includes physical activity calorie equivalent (PACE) labelling. Implications: The plethora of studies demonstrating similar findings in combination with the experience of conducting this study illustrate that there is no single ―ideal intervention‖ to promote healthy choices, because the ideal intervention for each setting is context dependent. Finally, the observed benefit of PACE labelling provides evidence to support the recent call for energy labelling in London, England.

126

Chapter 10 Conclusions

This thesis demonstrated that there is wide variation in calorie levels within each restaurant and food category. In addition, it was found that portion size, as opposed to calorie density, is the most important driver of variation in calories. With respect to sodium, it was found that levels in restaurant foods are alarmingly high and often exceed daily recommendations. Furthermore, efforts to decrease sodium levels, as of 2013, have been minimal.

This thesis also showed that menu-labelling policies that require the disclosure of calorie and sodium information may encourage consumers to choose meals with fewer calories and less sodium, however, the effect of labelling varied greatly depending on the restaurant setting. Lastly, the addition of calorie density information on restaurants did not assist consumers in selecting meals with a lower calorie density.

Finally, this thesis illustrated that an alternative education/labelling intervention in cafeterias could have potential to encourage students to consume more fruits, vegetables and water and to consume fewer sugar sweetened beverages.

Despite these findings, the political advancements during the course of this thesis demonstrated that policy decisions are not always based on evidence, but are rather the product of multiple political, economic, and social influences. Thus, future research will be needed to evaluate the efficacy of current policies, as a means to inform future public health and research priorities concerning foods consumed outside-the-home.

127

References

1. World Health Organization. Diet, nutrition and the prevention of chronic diseases: Report of a Join WHO/FAO Expert Consultation., in WHO Technical Report Series, No. 916. 2003: Geneva, Switzerland. 2. World Health Organization. Global status report on noncommunicable disease 2014. 2014; Available at: http://www.who.int/nmh/publications/ncd-status-report-2014/en/ (Accessed: March 13 2016). 3. Health Canada. Eating Well With Canada's Food Guide. 2007; Available at: http://www.hc- sc.gc.ca/fn-an/food-guide-aliment/index-eng.php (Accessed: January 2 2016). 4. Twells LK, Gregory DM, Reddigan J, and Midodzi WK. Current and predicted prevalence of obesity in Canada: a trend analysis. CMAJ Open, 2014. 2(1):E18-26. 5. Tjepkema M. Adult obesity, S. Canada, Editor. 2006. 6. Navaneelan TJ, T. Adjusting the scales: Obesity in the Canadian population after correcting for respondent bias, S. Canada, Editor. 2014. 7. Bleich S, Cutler D, Murray C, and Adams A. Why is the developed world obese? Annu Rev Public Health, 2008. 29:273-95. 8. Statistics Canada. Leading causes of death, by sex. 2014; Available at: http://www.statcan.gc.ca/tables-tableaux/sum-som/l01/cst01/hlth36a-eng.htm (Accessed: December 1 2015). 9. Public Health Agency of Canada. Centra for Chronic Disease Prevention: Strategic Plan 2016- 2019 Improving Health Outcomes - A Paradigm Shift. 2015; Available at: http://www.phac- aspc.gc.ca/cd-mc/ccdp-strategic-plan-2016-2019-plan-strategique-cpmc-eng.php (Accessed: March 14 2016). 10. Public Health Agency of Canada. Against the Growing Burden of Disease. 2015; Available at: http://www.ccgh-csih.ca/assets/Elmslie.pdf (Accessed: March 14 2016). 11. Guthrie JF, Lin BH, and Frazao E. Role of food prepared away from home in the American diet, 1977-78 versus 1994-96: changes and consequences. J Nutr Educ Behav, 2002. 34(3):140-50. 12. United States Department of Agriculture-Economic Research Service. Food Expenditures. 2015; Available at: http://www.ers.usda.gov/data-products/food-expenditures.aspx (Accessed: January 2 2016). 13. United States Department of Agriculture. Food Away-From-Home. 2014; Available at: http://www.ers.usda.gov/topics/food-choices-health/food-consumption-demand/food-away-from- home.aspx#nutrition (Accessed: January 4 2015). 14. Health Canada. Canadian Community Health Survey Cycle 2.2 Nutrition - A Guide to Accessing and Interpreting the Data 2006. 15. Garriguet D. Canadians' eating habits. 2007; Available at: http://www.statcan.gc.ca/pub/82-003- x/2006004/article/9609-eng.htm (Accessed: Dec 4 2015). 16. Garriguet D. Nutrition: Findings from the Canadian Community Health Survey Overview of Canadians' Eating Habits, H.S. Division, Editor. 2004: Ottawa. 17. Canadian Restaurants and Foodservice Association. Consumer Spending Research. 2011; Available at: www.crfa.ca/research/ (Accessed: September 21 2011). 18. Statistics Canada. Survey of household spending, 2013. 2013; Available at: http://www.statcan.gc.ca/daily-quotidien/150122/dq150122b-eng.htm (Accessed: 5 January 2016). 19. Slater J, Green CG, Sevenhuysen G, Edginton B, O'Neil J, and Heasman M. The growing Canadian energy gap: more the can than the couch? Public Health Nutr, 2009. 12(11):2216-24. 20. Canwest News Service. Pizza snares big slice of fast-food pie, in canada.com. 2007.

128

21. Paeratakul S, Ferdinand DP, Champagne CM, Ryan DH, and Bray GA. Fast-food consumption among US adults and children: Dietary and nutrient intake profile. Journal of the American Dietetic Association, 2003. 103(10):1332-1338. 22. Bowman SA, Gortmaker SL, Ebbeling CB, Pereira MA, and Ludwig DS. Effects of fast-food consumption on energy intake and diet quality among children in a national household survey. Pediatrics, 2004. 113(1 Pt 1):112-8. 23. Diliberti N, Bordi PL, Conklin MT, Roe LS, and Rolls BJ. Increased portion size leads to increased energy intake in a restaurant meal. Obes Res, 2004. 12(3):562-8. 24. Ello-Martin JA, Ledikwe JH, and Rolls BJ. The influence of food portion size and energy density on energy intake: implications for weight management. Am J Clin Nutr, 2005. 82(1 Suppl):236S- 241S. 25. Nielsen SJ and Popkin BM. Patterns and trends in food portion sizes, 1977-1998. JAMA, 2003. 289(4):450-3. 26. Bauer KW, Hearst MO, Earnest AA, French SA, Oakes JM, and Harnack LJ. Energy content of U.S. fast-food restaurant offerings: 14-year trends. Am J Prev Med, 2012. 43(5):490-7. 27. Guenther PM, Kirkpatrick SI, Reedy J, Krebs-Smith SM, Buckman DW, Dodd KW, et al. The healthy eating index-2010 is a valid and reliable measure of diet quality according to the 2010 dietary guidelines for americans. J Nutr, 2014. 144(3):399-407. 28. Hearst MO, Harnack LJ, Bauer KW, Earnest AA, French SA, and Michael Oakes J. Nutritional quality at eight U.S. fast-food chains: 14-year trends. Am J Prev Med, 2013. 44(6):589-94. 29. Dumanovsky T, Huang CY, Nonas CA, Matte TD, Bassett MT, and Silver LD. Changes in energy content of lunchtime purchases from fast food restaurants after introduction of calorie labelling: cross sectional customer surveys. British Medical Journal, 2011. 343:d4464. 30. Johnson CM, Angell SY, Lederer A, Dumanovsky T, Huang C, Bassett MT, et al. Sodium content of lunchtime fast food purchases at major US chains. Archives of Internal Medicine, 2010. 170(8):732-734. 31. Lin B, Frazao E, and Guthrie J. Away-from-home food increasingly important to quality of american diet: economic research service, U.S. Department of Agriculture Economic Research Service, Editor. 1999: Washington, DC. 32. Brownell KD. Fast food and obesity in children. Pediatrics, 2004. 113(1 Pt 1):132. 33. Thompson OM, Ballew C, Resnicow K, Must A, Bandini LG, Cyr H, et al. Food purchased away from home as a predictor of change in BMI z-score among girls. Int J Obes Relat Metab Disord, 2004. 28(2):282-9. 34. Duffey KJ, Gordon-Larsen P, Jacobs DR, Jr., Williams OD, and Popkin BM. Differential associations of fast food and restaurant food consumption with 3-y change in body mass index: the Coronary Artery Risk Development in Young Adults Study. Am J Clin Nutr, 2007. 85(1):201- 8. 35. Niemeier HM, Raynor HA, Lloyd-Richardson EE, Rogers ML, and Wing RR. Fast food consumption and breakfast skipping: predictors of weight gain from adolescence to adulthood in a nationally representative sample. J Adolesc Health, 2006. 39(6):842-9. 36. Satia JA, Galanko JA, and Siega-Riz AM. Eating at fast-food restaurants is associated with dietary intake, demographic, psychosocial and behavioural factors among African Americans in North Carolina. Public Health Nutr, 2004. 7(8):1089-96. 37. Kant AK, Whitley MI, and Graubard BI. Away from home meals: associations with biomarkers of chronic disease and dietary intake in American adults, NHANES 2005-2010. Int J Obes (Lond), 2015. 39(5):820-7. 38. Fulkerson JA, Farbakhsh K, Lytle L, Hearst MO, Dengel DR, Pasch KE, et al. Away-from-home family dinner sources and associations with weight status, body composition, and related biomarkers of chronic disease among adolescents and their parents. J Am Diet Assoc, 2011. 111(12):1892-7.

129

39. Bowman SA and Vinyard BT. Fast food consumption of U.S. adults: impact on energy and nutrient intakes and overweight status. J Am Coll Nutr, 2004. 23(2):163-8. 40. Todd JM, L; Lin B-H; . The Impact of Food Away From Home on Adult Diet Quality, United States Department of Agriculture - Economic Research Service, Editor. 2010. 41. Long MW, Tobias DK, Cradock AL, Batchelder H, and Gortmaker SL. Systematic review and meta-analysis of the impact of restaurant menu calorie labeling. Am J Public Health, 2015. 105(5):e11-24. 42. Dunford E, Webster J, Barzi F, and Neal B. Nutrient content of products served by leading Australian fast food chains. Appetite, 2010. 55(3):484-9. 43. O'Donnell SI, Hoerr SL, Mendoza JA, and Tsuei Goh E. Nutrient quality of fast food kids meals. Am J Clin Nutr, 2008. 88(5):1388-95. 44. Serrano EL and Jedda VB. Comparison of fast-food and non-fast-food children's menu items. J Nutr Educ Behav, 2009. 41(2):132-7. 45. The Fatabase - Nutritional information for over 80 BC restaurant chains. 2011; Available at: http://www.vancouversun.com/life/food/fatabase/index.html (Accessed: January 1 2016). 46. Dunford E, Webster J, Woodward M, Czernichow S, Yuan WL, Jenner K, et al. The variability of reported salt levels in fast foods across six countries: opportunities for salt reduction. CMAJ, 2012. 184(9):1023-8. 47. Hobin E, White C, Li Y, Chiu M, O'Brien MF, and Hammond D. Nutritional quality of food items on fast-food 'kids' menus': comparisons across countries and companies. Public Health Nutr, 2014. 17(10):2263-9. 48. Shaw H. Fast-food restaurants eating into full-service sector market share, in The Globe and Mail. 2015. 49. Institute of Medicine. Local Government Actions of Prevent Childhood Obesity. 2009; Available at: https://iom.nationalacademies.org/Reports/2009/Local-Government-Actions-to-Prevent- Childhood-Obesity.aspx (Accessed: 5 January 2015). 50. The Healthy Kids Panel. No Time to Wait: The Healthy Kids Strategy. 2013: Queen's Printer for Ontario. 51. Health Canada. Sodium Reduction Strategy for Canada. Recommendations of the Canadian Sodium Working Group. 2010. 52. Ontario Medical Association. Ontario’s Doctors Commend Legislation Revealing Calories in Restaurant Foods. 2014; Available at: https://www.oma.org/Mediaroom/PressReleases/Pages/OntDocsLegCalsinRestFoods.aspx (Accessed: 5 January 2015). 53. Administration FaD. Calories Count - Report of the Working Group on Obesity. 2004; Available at: http://www.fda.gov/food/foodscienceresearch/consumerbehaviorresearch/ucm081696.htm (Accessed: 5 January 2015). 54. Wansink B and Chandon P. Meal size, not body size, explains errors in estimating the calorie content of meals. Ann Intern Med, 2006. 145(5):326-32. 55. Block JP, Condon SK, Kleinman K, Mullen J, Linakis S, Rifas-Shiman S, et al. Consumers' estimation of calorie content at fast food restaurants: cross sectional observational study. BMJ, 2013. 346:f2907. 56. Burton SC, EH. What consumers don't know can hurt them: consumer evaluations and disease risk perceptions of restaurant menu items. Journal of Consumer Affairs, 2004. 38:121-145. 57. Burton S, Creyer EH, Kees J, and Huggins K. Attacking the obesity epidemic: the potential health benefits of providing nutrition information in restaurants. Am J Public Health, 2006. 96(9):1669- 75. 58. Burton S, Howlett E, and Tangari AH. Food for Thought: How Will the Nutrition Labeling of Quick Service Restaurant Menu Items Influence Consumers’ Product Evaluations, Purchase Intentions, and Choices? Journal of Retailing, 2009. 85(3):258-273.

130

59. Backstrand JW, MG; Young, LR; Hurley, J. Fat Chance - A survey of dietitians' knowledge of the calories and fat in restaurant meals. 1997; Available at: http://www.portionteller.com/pdf/cspistudy97.pdf (Accessed: January 3 2015). 60. Lando AM and Labiner-Wolfe J. Helping consumers make more healthful food choices: consumer views on modifying food labels and providing point-of-purchase nutrition information at quick-service restaurants. J Nutr Educ Behav, 2007. 39(3):157-63. 61. O'Dougherty M, Harnack LJ, French SA, Story M, Oakes JM, and Jeffery RW. Nutrition labeling and value size pricing at fast-food restaurants: a consumer perspective. Am J Health Promot, 2006. 20(4):247-50. 62. Bleich SN and Pollack KM. The publics' understanding of daily caloric recommendations and their perceptions of calorie posting in chain restaurants. BMC Public Health, 2010. 10:121. 63. Mah CL, Vanderlinden L, Mamatis D, Ansara DL, Levy J, and Swimmer L. Ready for policy? Stakeholder attitudes toward menu labelling in Toronto, Canada. Can J Public Health, 2013. 104(3):e229-34. 64. Wappel T. Bill C-398 An Act to Amend the Food and Drugs Act (food labelling), Legislative Assembly of Ontario, Editor. 2003: Ottawa, Canada. 65. Wappel T. Bill C-283 - An Act to amend the Food and Drugs Act (food labelling). 2006; Available at: https://openparliament.ca/bills/39-1/C-283/?page=2 (Accessed: 8 January 2016). 66. Gelinas F. Bill 156, Healthy Decisions for Healthy Eating Act, 2009. 2009; Available at: http://www.ontla.on.ca/web/bills/bills_detail.do?locale=en&BillID=2160&isCurrent=&detailPag e=bills_detail_status&Intranet (Accessed: 8 January 2016). 67. Gelinas F. Bill 90, Healthy Decisiosn for Healthy Eating Act, 2010. 2010; Available at: http://www.ontla.on.ca/web/bills/bills_detail.do?locale=en&BillID=2385&detailPage=bills_detai l_the_bill (Accessed: 8 January 2016). 68. Gelinas F. Bill 86, Healthy Decisions for Healthy Eating Act, Legislative Assembly of Ontario, Editor. 2012: Toroto, Ontario. 69. Gelinas F. Bill 59, Healthy Decisions for Healthy Eating Act, 2013. 2013; Available at: http://www.ontla.on.ca/web/bills/bills_detail.do?locale=en&BillID=2780&detailPage=bills_detai l_status (Accessed: 8 January 2016). 70. Toronto Public Health. What's on the Menu? Making Key Nutrition Information Readily Available in Restaurants. 2013: Toronto, Ontario. 71. Food and Drug Administration. New menu and vending machine labeling requirements. 2010; Available at: http://www.fda.gov/food/ingredientspackaginglabeling/ucm217762.htm (Accessed: 5 January 2015). 72. Gelinas F. Bill 149, Health Statute Law Amendment Act (Healthy Decisions Made Easy), Legislative Assembly of Ontario, Editor. 2014: Toronto, Ontario. 73. Matthews D. Bill 162, Making Healthier Choices Act, 2014. 2014; Available at: http://www.ontla.on.ca/web/bills/bills_detail.do?locate=en&Intranet&BillID=2939 (Accessed: 8 January 2016). 74. Damerla D. Bill 45, Making Healthier Choices Act, 2015. 2015; Available at: http://www.ontla.on.ca/web/bills/bills_detail.do?BillID=3080 (Accessed: 8 January 2016). 75. Canadian Restaurant and Foodservice Association. Restaurant Chains Participating in the CRFA's Nutrition Information Program. 2005; Available at: http://www.crfa.ca/resources/nutritioninformationprogram/participatingchains.asp (Accessed: 3 October 2010). 76. Roberto CA, Agnew H, and Brownell KD. An observational study of consumers' accessing of nutrition information in chain restaurants. Am J Public Health, 2009. 99(5):820-1. 77. Healthy Families BC. Informed Dining for Restaurants. 2012; Available at: https://www.healthyfamiliesbc.ca/home/informed-dining-restaurants (Accessed: January 1 2015).

131

78. Healthy Families BC. Informed Dining National. 2012; Available at: https://www.healthyfamiliesbc.ca/about-us/informed-dining-national (Accessed: 11 January 2016). 79. Vanderlee L and Hammond D. Does nutrition information on menus impact food choice? Comparisons across two hospital cafeterias. Public Health Nutr, 2014. 17(6):1393-402. 80. Hammond D, Goodman S, Hanning R, and Daniel S. A randomized trial of calorie labeling on menus. Prev Med, 2013. 57(6):860-6. 81. Farley TA, Caffarelli, Anna, Bassett, Mary T., Silver, Lynn, Frieden, Thomas R.,. New York City's Fight Over Calorie Labeling. Health Affairs, 2009. 28(6):w1098-w1109. 82. Angell SY. Sodium Warning Label: Proposed Amendment of Article 81 - Notice of Intent. 2015; Available at: http://www.nyc.gov/html/doh/downloads/pdf/cardio/sodium-warning-label-81.pdf (Accessed: 5 January 2015). 83. Finkelstein EA, Strombotne KL, Chan NL, and Krieger J. Mandatory menu labeling in one fast- food chain in King County, Washington. Am J Prev Med, 2011. 40(2):122-7. 84. Bruemmer B, Krieger J, Saelens BE, and Chan N. Energy, Saturated Fat, and Sodium Were Lower in Entrées at Chain Restaurants at 18 Months Compared with 6 Months Following the Implementation of Mandatory Menu Labeling Regulation in King County, Washington. Journal of the Academy of Nutrition and Dietetics, 2012. 112(8):1169-1176. 85. Public Health City of Philadelphia. Menu Labeling Requirements, Philadelphia City Code. 2010; Available at: http://www.phila.gov/health/services/Serv_MenuLabeling.html (Accessed: January 13, 2013. 86. Federal Register - The Daily Journal of the United States Government. Food Labeling: Nutrition Labeling of Standard Menu Items in Restaurants and Similar Retail Food Establishments. 2014; Available at: https://www.federalregister.gov/articles/2014/12/01/2014-27833/food-labelingnutrition-labeling-of-standard-menu-items-in-restaurants-and-similar-retail-food#h-11 (Accessed: March 14 2016). 87. U.S. Food and Drug Administration. FDA Statement on Menu Labeling Enforcement - March 9, 2016. 2016; Available at: http://www.fda.gov/Food/IngredientsPackagingLabeling/LabelingNutrition/ucm217762.htm (Accessed: March 14 2016). 88. Vadiveloo MK, Dixon LB, and Elbel B. Consumer purchasing patterns in response to calorie labeling legislation in New York City. Int J Behav Nutr Phys Act, 2011. 8:51. 89. Bollinger B LP, Sorenson A. Calorie posting in chain restaurants. National Bureau of Economic Research, 2010. 90. Elbel B. Consumer estimation of recommended and actual calories at fast food restaurants. Obesity (Silver Spring), 2011. 19(10):1971-8. 91. Cantor J, Torres A, Abrams C, and Elbel B. Five Years Later: Awareness Of New York City's Calorie Labels Declined, With No Changes In Calories Purchased. Health Aff (Millwood), 2015. 34(11):1893-900. 92. Roberto CA, Larsen PD, Agnew H, Baik J, and Brownell KD. Evaluating the impact of menu labeling on food choices and intake. Am J Public Health, 2010. 100(2):312-8. 93. Fitch RC, Harnack LJ, Neumark-Sztainer DR, Story MT, French SA, Oakes JM, et al. Providing calorie information on fast-food restaurant menu boards: consumer views. Am J Health Promot, 2009. 24(2):129-32. 94. Pulos E and Leng K. Evaluation of a voluntary menu-labeling program in full-service restaurants. Am J Public Health, 2010. 100(6):1035-9. 95. Auchincloss AH, Mallya GG, Leonberg BL, Ricchezza A, Glanz K, and Schwarz DF. Customer responses to mandatory menu labeling at full-service restaurants. Am J Prev Med, 2013. 45(6):710-9. 96. Milich R, Anderson J, and Mills M. Effects of visual presentation of caloric values on food buying by normal and obese persons. Percept Mot Skills, 1976. 42(1):155-62.

132

97. Davis-Chervin D, Rogers T, and Clark M. Influencing food selection with point-of-choice nutrition information. Journal of Nutrition Education, 1985. 17(1):18-22. 98. Cinciripini P. Changing food selections in a public cafeteria: an applied behaviour analysis. Behaviour Modification, 1984. 8(4):520-539. 99. Tandon PS, Wright J, Zhou C, Rogers CB, and Christakis DA. Nutrition menu labeling may lead to lower-calorie restaurant meal choices for children. Pediatrics, 2010. 125(2):244-8. 100. Gerend MA. Does calorie information promote lower calorie fast food choices among college students? J Adolesc Health, 2009. 44(1):84-6. 101. Harnack LJ, French SA, Oakes JM, Story MT, Jeffery RW, and Rydell SA. Effects of calorie labeling and value size pricing on fast food meal choices: results from an experimental trial. Int J Behav Nutr Phys Act, 2008. 5:63. 102. Legislative Assembly of Ontario. Bill 45, Making Healthier Choices Act, 2015. 2015; Available at: http://www.ontla.on.ca/web/bills/bills_detail.do?locale=en&BillID=3080&detailPage=bills_detai l_status (Accessed: September 8 2015). 103. Finance; GCF. 2014 Canadian Chain Restaurant Industry Review. 2014; Available at: http://www.restaurantinvest.ca/site/restaurant_invest/assets/pdf/2014__canadian_chain_restaurant _industry_review.pdf (Accessed: 17 January 2016). 104. Burwash Dining Hall - Victoria University. 2016; Available at: http://www.vicu.utoronto.ca/hospitality/Food_Services/Burwash_Dining_Hall.htm (Accessed: 23 January 2016). 105. Morley B, Scully M, Martin J, Niven P, Dixon H, and Wakefield M. What types of nutrition menu labelling lead consumers to select less energy-dense fast food? An experimental study. Appetite, 2013. 67(0):8-15. 106. Liu PJ, Roberto CA, Liu LJ, and Brownell KD. A test of different menu labeling presentations. Appetite, 2012. 59(3):770-7. 107. Antonelli R and Viera AJ. Potential Effect of Physical Activity Calorie Equivalent (PACE) Labeling on Adult Fast Food Ordering and Exercise. PLoS One, 2015. 10(7):e0134289. 108. Dowray S, Swartz JJ, Braxton D, and Viera AJ. Potential effect of physical activity based menu labels on the calorie content of selected fast food meals. Appetite, 2013. 62:173-81. 109. Viera AJ and Antonelli R. Potential effect of physical activity calorie equivalent labeling on parent fast food decisions. Pediatrics, 2015. 135(2):e376-82. 110. James A, Adams-Huet B, and Shah M. Menu labels displaying the kilocalorie content or the exercise equivalent: effects on energy ordered and consumed in young adults. Am J Health Promot, 2015. 29(5):294-302. 111. Vyth EL, Steenhuis IH, Heymans MW, Roodenburg AJ, Brug J, and Seidell JC. Influence of placement of a nutrition logo on cafeteria menu items on lunchtime food Choices at Dutch work sites. J Am Diet Assoc, 2011. 111(1):131-6. 112. Mayer JA, Brown TP, Heins JM, and Bishop DB. A multi-component intervention for modifying food selections in a worksite cafeteria. Journal of Nutrition Education, 1987. 19(6):277-280. 113. Hanks AS, Just DR, Smith LE, and Wansink B. Healthy convenience: nudging students toward healthier choices in the lunchroom. J Public Health (Oxf), 2012. 34(3):370-6. 114. Turconi G, Bazzano R, Roggi C, and Cena H. Helping consumers make a more conscious nutritional choice: acceptability of nutrition information at a cafeteria. Public Health Nutr, 2012. 15(5):792-801. 115. Ries CP and Schoon DK. Evaluation of a cafeteria-based education program for college students. Journal of Nutrition Education, 1986. 18(3):107-110. 116. Auchincloss AH, Young C, Davis AL, Wasson S, Chilton M, and Karamanian V. Barriers and facilitators of consumer use of nutrition labels at sit-down restaurant chains. Public Health Nutr, 2013:1-8.

133

117. Cohen DA and Babey SH. Contextual influences on eating behaviours: heuristic processing and dietary choices. Obes Rev, 2012. 13(9):766-79. 118. Freedman MR and Connors R. Point-of-purchase nutrition information influences food- purchasing behaviors of college students: a pilot study. J Am Diet Assoc, 2010. 110(8):1222-6. 119. Buscher LA, Martin KA, and Crocker S. Point-of-purchase messages framed in terms of cost, convenience, taste, and energy improve healthful snack selection in a college foodservice setting. J Am Diet Assoc, 2001. 101(8):909-13. 120. Reger B, Wootan MG, Booth-Butterfield S, and Smith H. 1% or less: a community-based nutrition campaign. Public Health Rep, 1998. 113(5):410-9. 121. Wansink B, Westgren RE, and Cheney MM. Hierarchy of nutritional knowledge that relates to the consumption of a functional food. Nutrition, 2005. 21(2):264-8. 122. Zifferblatt SM, Wilbur CS, and Pinsky JL. A new direction for public health care: changing cafeteria eating habits. J Am Diet Assoc, 1980. 76(1):15-20. 123. Intersalt: an international study of electrolyte excretion and blood pressure. Results for 24 hour urinary sodium and potassium excretion. Intersalt Cooperative Research Group. BMJ (Clinical research ed.), 1988. 297(6644):319-328. 124. Kearney PM, Whelton M, Reynolds K, Muntner P, Whelton PK, and He J. Global burden of hypertension: analysis of worldwide data. Lancet, 2005. 365(9455):217-23. 125. Lopez AD, Mathers CD, Ezzati M, Jamison DT, and Murray CJ. Global and regional burden of disease and risk factors, 2001: systematic analysis of population health data. Lancet, 2006. 367(9524):1747-1757. 126. Institute of Medicine. Dietary Reference Intakes for Water, Potassium, Sodium, Chloride, and Sulfate. Panel on Dietary Reference Intakes for Electrolytes and Water, Standing Committee on the Scientific Evaluation of Dietary Reference Intakes. 2004. 127. Health Canada. Canadian Community Health Survey, Cycle 2.2, Nutrition Nutrient intakes from food: provincial, regional and national summary tables. 2007, Health Canada: Ottawa. 128. Mattes RD and Donnelly D. Relative contributions of dietary sodium sources. Journal of the American College of Nutrition, 1991. 10(4):383-393. 129. Health Canada. Guidance for the Food Industry on Reducing Sodium in Processed Foods. 2012; Available at: http://www.hc-sc.gc.ca/fn-an/legislation/guide-ld/2012-sodium-reduction-indust- eng.php (Accessed: December 30 2015). 130. Mackison D, Wrieden WL, and Anderson AS. Making an informed choice in the catering environment: what do consumers want to know? J Hum Nutr Diet, 2009. 22(6):567-73. 131. Davies L. An Act respecting the implementation of the Sodium Reduction Strategy for Canada. 2013; Available at: http://www.parl.gc.ca/HousePublications/Publication.aspx?DocId=5821105&File=14 (Accessed: May 14 2014). 132. Hygiene; NYCDoHaM. National Salt Reduction Initiative. 2010; Available at: http://www1.nyc.gov/site/doh/health/health-topics/national-salt-reduction-initiative.page (Accessed: 13 January 2016). 133. New York City Department of Health and Mental Hygiene. National salt reduction initiative restaurant food categories and targets. 2009; Available at: http://www.nyc.gov/html/doh/downloads/pdf/cardio/cardio-salt-nsri-restaurant.pdf (Accessed: February 23 2014). 134. New York City Department of Health and Mental Hygiene. National Salt Reduction Initiative - Restaurant Food. 2016; Available at: http://www1.nyc.gov/site/doh/health/health-topics/salt- initiative-restaurantfood.page (Accessed: 24 January 2016). 135. The New York City Department of Health and Mental Hygiene. Heart Follow-Up Study. 2015; Available at: http://www.nyc.gov/html/doh/html/diseases/hfus.shtml (Accessed: December 30 2015). 136. MenuStat. 2012; Available at: http://menustat.org/ (Accessed: 13 January 2016).

134

137. Watson E. CSPI sues FDA for dragging its feet over sodium reduction; FDA still working on voluntary targets. 2015; Available at: http://www.foodnavigator-usa.com/Regulation/CSPI-sues- FDA-for-dragging-its-feet-over-sodium-reduction (Accessed: 18 March 2016). 138. Public Health Agency of Canada-Canadian Institute for Health Information. Obesity in Canada - A joint report from the Public Health Agency of Canada and the Canadian Institute for Health Information. 2011. 139. Rosenheck R. Fast food consumption and increased caloric intake: a systematic review of a trajectory towards weight gain and obesity risk. Obes Rev, 2008. 9(6):535-47. 140. Isganaitis E and Lustig RH. Fast food, central nervous system insulin resistance, and obesity. Arterioscler Thromb Vasc Biol, 2005. 25(12):2451-62. 141. Pereira MA, Kartashov AI, Ebbeling CB, Van Horn L, Slattery ML, Jacobs DR, Jr., et al. Fast- food habits, weight gain, and insulin resistance (the CARDIA study): 15-year prospective analysis. Lancet, 2005. 365(9453):36-42. 142. French SA, Harnack L, and Jeffery RW. Fast food restaurant use among women in the Pound of Prevention study: dietary, behavioral and demographic correlates. Int J Obes Relat Metab Disord, 2000. 24(10):1353-9. 143. Margot Shields MDC, Cynthia L. Ogden. Adult Obesity Prevalence in Canada and the United States. 2011. 144. Organization WH. Reducing risks, promoting healthy life, in World Health Report. 2002. 145. Institue of Medicine. Dietary reference intake tables. 2011. 146. Joffres MR, Campbell NR, Manns B, and Tu K. Estimate of the benefits of a population-based reduction in dietary sodium additives on hypertension and its related health care costs in Canada. The Canadian journal of cardiology, 2007. 23(6):437-443. 147. Schmidt S. Federal inaction on limiting Canadians’ salt intake criticized. Postmedia News, 2012. 148. New York City Department of H and Mental H. National Salt Reduction Initiative Restaurant Food Categories and Targets. 149. Tu K CZLL. Prevalence and incidence of hypertension from 1995 to 2005: a population-based study. Canadian Medical Association Journal, 2008. 178(11):1429-1435. 150. Monday Report on Retailers. Directory of restaurant and fast food chains in Canada. 2010, Toronto: Rogers Media Inc. 151. Scourboutakos MJ and L'Abbe MR. Restaurant menus: calories, caloric density, and serving size. Am J Prev Med, 2012. 43(3):249-55. 152. Wilkins K, Campbell NR, Joffres MR, McAlister FA, Nichol M, Quach S, et al. Blood pressure in Canadian adults. Health reports / Statistics Canada, Canadian Centre for Health Information = Rapports sur la sante / Statistique Canada, Centre canadien d'information sur la sante, 2010. 21(1):37-46. 153. Public Health Agency of C. The Face of Diabetes in Canada. 2008. 154. Dunford E, Webster J, Woodward M, Czernichow S, Yuan WL, Jenner K, et al. The variability of reported salt levels in fast foods across six countries: opportunities for salt reduction. CMAJ : Canadian Medical Association journal = journal de l'Association medicale canadienne, 2012. 184(9):1023-1028. 155. Kim GH and Lee HM. Frequent consumption of certain fast foods may be associated with an enhanced preference for salt taste. Journal of human nutrition and dietetics : the official journal of the British Dietetic Association, 2009. 22(5):475-480. 156. Rodgers A and Neal B. Less salt does not necessarily mean less taste. Lancet, 1999. 353(9161):1332. 157. Bertino M, Beauchamp GK, and Engelman K. Long-term reduction in dietary sodium alters the taste of salt. Am J Clin Nutr, 1982. 36(6):1134-44. 158. Ni Mhurchu C, Capelin C, Dunford EK, Webster JL, Neal BC, and Jebb SA. Sodium content of processed foods in the United Kingdom: analysis of 44,000 foods purchased by 21,000 households. The American Journal of Clinical Nutrition, 2011. 93(3):594-600.

135

159. Provincial-Territorial Ministers of Health. Reducing the Sodium Intake of Canadians: A Provincial and Territorial Report on Progress and Recommendations for Future Action. 2012 21 August 2012; Available at: http://www.gnb.ca/0131/pdf/w/SodiumReport.pdf (Accessed: February 23 2011). 160. Scourboutakos M and L'Abbe M. Sodium levels in Canadian fast-food and sit-down restaurants. Can J Public Health, 2013. 104(1):e2-8. 161. Scourboutakos MJ, Semnani-Azad Z, and L'Abbe MR. Restaurant Meals: Almost a Full Day's Worth of Calories, Fats, and Sodium. JAMA Intern Med, 2013. 173(14):1373-4. 162. Jacobson MF, Havas S, and McCarter R. Changes in sodium levels in processed and restaurant foods, 2005 to 2011. JAMA Intern Med, 2013. 173(14):1285-91. 163. Yum! Brands. Corporate Responsibility Report - Nutritional Improvement - Sodium. 2013; Available at: http://www.yumcsr.com/food/nutritional-improvement.asp (Accessed: January 13 2014. 164. Weeks C. Researchers 'shocked' by sodium levels at Canadian restaurants, in The Globe and Mail. 2013. 165. Stein K. A National Approach to Restaurant Menu Labeling: The Patient Protection and Affordable Health Care Act, Section 4205. Journal of the American Dietetic Association, 2010. 110(9):1280-1289. 166. Food and Drug Administration. Food labeling; nutrient content claims, definition of sodium levels for the term "healthy." Final rule. Fed Regist, 2005. 70(188):56828-49. 167. Monday Report on Retailers. 2013 Directory of Restaurant and Fast-Food Chains in Canada. 2013, Toronto: Rogers Publishing Ltd. 168. Wu HW and Sturm R. Changes in the Energy and Sodium Content of Main Entrees in US Chain Restaurants from 2010 to 2011. J Acad Nutr Diet, 2013. 169. Bertino M, Beauchamp GK, and Engelman K. Increasing dietary salt alters salt taste preference. Physiol Behav, 1986. 38(2):203-13. 170. Hayes JE, Sullivan BS, and Duffy VB. Explaining variability in sodium intake through oral sensory phenotype, salt sensation and liking. Physiol Behav, 2010. 100(4):369-80. 171. Girgis S, Neal B, Prescott J, Prendergast J, Dumbrell S, Turner C, et al. A one-quarter reduction in the salt content of bread can be made without detection. Eur J Clin Nutr, 2003. 57(4):616-20. 172. Dotsch M, Busch J, Batenburg M, Liem G, Tareilus E, Mueller R, et al. Strategies to reduce sodium consumption: a food industry perspective. Crit Rev Food Sci Nutr, 2009. 49(10):841-51. 173. New York City Department of Health and Mental Hygiene. National Salt Reduction Initiative Restaurant Food Categories and Targets. 174. Sodium Reduction Market by Ingredients, Applications & Georgephy - Global Trends & Forecasts to 2018. 2013; Available at: http://www.researchandmarkets.com/research/nkh4xt/sodium_reduction (Accessed: March 31 2014 2014). 175. Arcand J, Mendoza J, Qi Y, Henson S, Lou W, and L'Abbe MR. Results of a national survey examining Canadians' concern, actions, barriers, and support for dietary sodium reduction interventions. Can J Cardiol, 2013. 29(5):628-31. 176. Wong CL, Arcand J, Mendoza J, Henson SJ, Qi Y, Lou W, et al. Consumer attitudes and understanding of low-sodium claims on food: an analysis of healthy and hypertensive individuals. Am J Clin Nutr, 2013. 97(6):1288-98. 177. Center for Science in the Public Interest. Anyone's Guess: The need for nutrition labelling at fast- food and other chain restaurants. 2003, Center for Science inthe Public Interest: Washington, DC. 178. Hebert JR, Clemow L, Pbert L, Ockene IS, and Ockene JK. Social desirability bias in dietary self-report may compromise the validity of dietary intake measures. Int J Epidemiol, 1995. 24(2):389-98.

136

179. Kreuter F PS, Tourangeu R,. Social Desirability Bias in CATI, IVR and Web Surveys. Public Opinion Quarterly,, 2008. 72(5):847-865. 180. L Martel EM. 2006 Census: Portrait of the Canadian Population in 2006, by Age and Sex: Findings. 2006; Available at: http://www12.statcan.gc.ca/census-recensement/2006/as-sa/97- 551/index-eng.cfm (Accessed: May 13 2013). 181. Education Highlight Tables, 2006 Census. 2006; Available at: http://www12.statcan.ca/census- recensement/2006/dp-pd/hlt/97- 560/pages/page.cfm?Lang=E&Geo=PR&Code=01&Table=1&Data=Dist&Sex=1&StartRec=1& Sort=2&Display=Page (Accessed: May 13 2014). 182. Richards A, Kattelmann KK, and Ren C. Motivating 18- to 24-year-olds to increase their fruit and vegetable consumption. J Am Diet Assoc, 2006. 106(9):1405-11. 183. Chung S-J and Hoerr SL. Predictors of fruit and vegetable intakes in young adults by gender. Nutrition Research, 2005. 25(5):453-463. 184. Horacek TM and Betts NM. College Students’ Dietary Intake and Quality According to Their Myers Briggs Type Indicator Personality Preferences. Journal of Nutrition Education, 1998. 30(6):387-395. 185. Hampl JS and Betts NM. Comparisons of dietary intake and sources of fat in low- and high-fat diets of 18- to 24-year-olds. J Am Diet Assoc, 1995. 95(8):893-7. 186. Nielsen SJ and Popkin BM. Changes in beverage intake between 1977 and 2001. Am J Prev Med, 2004. 27(3):205-10. 187. Schulze MB, Manson JE, Ludwig DS, Colditz GA, Stampfer MJ, Willett WC, et al. Sugar- sweetened beverages, weight gain, and incidence of type 2 diabetes in young and middle-aged women. JAMA, 2004. 292(8):927-34. 188. Dhingra R, Sullivan L, Jacques PF, Wang TJ, Fox CS, Meigs JB, et al. Soft drink consumption and risk of developing cardiometabolic risk factors and the metabolic syndrome in middle-aged adults in the community. Circulation, 2007. 116(5):480-8. 189. Ludwig DS, Peterson KE, and Gortmaker SL. Relation between consumption of sugar-sweetened drinks and childhood obesity: a prospective, observational analysis. Lancet, 2001. 357(9255):505-8. 190. Anderson DA, Shapiro JR, and Lundgren JD. The freshman year of college as a critical period for weight gain: An initial evaluation. Eating Behaviors, 2003. 4(4):363-367. 191. Wengreen HJ and Moncur C. Change in diet, physical activity, and body weight among young- adults during the transition from high school to college. Nutr J, 2009. 8:32. 192. Hoffman DJ, Policastro P, Quick V, and Lee SK. Changes in body weight and fat mass of men and women in the first year of college: A study of the "freshman 15". J Am Coll Health, 2006. 55(1):41-5. 193. Vella-Zarb RA and Elgar FJ. The 'freshman 5': a meta-analysis of weight gain in the freshman year of college. J Am Coll Health, 2009. 58(2):161-6. 194. Jung ME, Bray SR, and Martin Ginis KA. Behavior change and the freshman 15: tracking physical activity and dietary patterns in 1st-year university women. J Am Coll Health, 2008. 56(5):523-30. 195. Gropper SS, Simmons KP, Gaines A, Drawdy K, Saunders D, Ulrich P, et al. The freshman 15-a closer look. J Am Coll Health, 2009. 58(3):223-31. 196. Butler SM, Black DR, Blue CL, and Gretebeck RJ. Change in diet, physical activity, and body weight in female college freshman. Am J Health Behav, 2004. 28(1):24-32. 197. Cunnane SC. Childhood origins of lifestyle-related risk factors for coronary heart disease in adulthood. Nutr Health, 1993. 9(2):107-15. 198. van Mechelen W, Twisk JW, Kemper HC, Snel J, and Post GB. Longitudinal relationships between lifestyle and cardiovascular and bone health status indicators in males and females between 13 and 27 years of age; a review of findings from the Amsterdam Growth and Health Longitudinal Study. Public Health Nutr, 1999. 2(3A):419-27.

137

199. Freedman DS, Khan LK, Dietz WH, Srinivasan SR, and Berenson GS. Relationship of childhood obesity to coronary heart disease risk factors in adulthood: the Bogalusa Heart Study. Pediatrics, 2001. 108(3):712-8. 200. Li S, Chen W, Srinivasan SR, Bond MG, Tang R, Urbina EM, et al. Childhood cardiovascular risk factors and carotid vascular changes in adulthood: the Bogalusa Heart Study. JAMA, 2003. 290(17):2271-6. 201. Nelson MC, Kocos R, Lytle LA, and Perry CL. Understanding the Perceived Determinants of Weight-related Behaviors in Late Adolescence: A Qualitative Analysis among College Youth. Journal of Nutrition Education and Behavior, 2009. 41(4):287-292. 202. Yeh MC, Ickes SB, Lowenstein LM, Shuval K, Ammerman AS, Farris R, et al. Understanding barriers and facilitators of fruit and vegetable consumption among a diverse multi-ethnic population in the USA. Health Promot Int, 2008. 23(1):42-51. 203. Matthews JI, Doerr L, and Dworatzek PD. University Students Intend to Eat Better but Lack Coping Self-Efficacy and Knowledge of Dietary Recommendations. J Nutr Educ Behav, 2015. 204. Block JP, Gillman MW, Linakis SK, and Goldman RE. "If It Tastes Good, I'm Drinking It": Qualitative Study of Beverage Consumption Among College Students. J Adolesc Health, 2013. 205. Ha E-J and Caine-Bish N. Effect of Nutrition Intervention Using a General Nutrition Course for Promoting Fruit and Vegetable Consumption among College Students. Journal of Nutrition Education and Behavior, 2009. 41(2):103-109. 206. Ha EJ, Caine-Bish N, Holloman C, and Lowry-Gordon K. Evaluation of effectiveness of class- based nutrition intervention on changes in soft drink and milk consumption among young adults. Nutr J, 2009. 8:50. 207. Ha EJ and Caine-Bish N. Interactive introductory nutrition course focusing on disease prevention increased whole-grain consumption by college students. J Nutr Educ Behav, 2011. 43(4):263-7. 208. Matvienko O, Lewis DS, and Schafer E. A College Nutrition Science Course As An Intervention To Prevent Weight Gain In Female College Freshmen. Journal of Nutrition Education, 2001. 33(2):95-101. 209. Emrich TE and Mazier MJ. Impact of nutrition education on university students' fat consumption. Can J Diet Pract Res, 2009. 70(4):187-92. 210. Gow RW, Trace SE, and Mazzeo SE. Preventing weight gain in first year college students: an online intervention to prevent the "freshman fifteen". Eat Behav, 2010. 11(1):33-9. 211. Harvey-Berino J, Pope L, Gold BC, Leonard H, and Belliveau C. Undergrad and Overweight: An Online Behavioral Weight Management Program for College Students. Journal of Nutrition Education and Behavior, 2012. 44(6):604-608. 212. Franko DL, Cousineau TM, Trant M, Green TC, Rancourt D, Thompson D, et al. Motivation, self-efficacy, physical activity and nutrition in college students: randomized controlled trial of an internet-based education program. Prev Med, 2008. 47(4):369-77. 213. Platkin C, Yeh MC, Hirsch K, Wiewel EW, Lin CY, Tung HJ, et al. The effect of menu labeling with calories and exercise equivalents on food selection and consumption. BMC Obes, 2014. 1:21. 214. Live Well Omaha - Make the Healthy Choice the Easy Choice. 2015; Available at: http://livewellomaha.org/ (Accessed: November 17 2015).

215. Compendium of Physical Activities. 2011; Available at: https://sites.google.com/site/compendiumofphysicalactivities/ (Accessed. 216. Statistics Canada HSD. Canadian Health Measures Survey: Cycle 2 Data Tables 2009 to 2011. 2013. 217. Harvard School of Public Health. Healthy Eating Plate and Healthy Eating Pyramid. 2011; Available at: http://www.hsph.harvard.edu/nutritionsource/healthy-eating-plate/ (Accessed: November 18 2015).

138

218. United States Department of Agriculture. Choose MyPlate.gov. 2016; Available at: http://www.choosemyplate.gov/ (Accessed: 26 January 2016). 219. Canadian Produce Marketing Association. Half Your Plate. 2016; Available at: http://www.halfyourplate.ca/what-is-half-your-plate/ (Accessed: 26 January 2016). 220. National Health Service. The eatwell plate. 2015; Available at: http://www.nhs.uk/Livewell/Goodfood/Pages/eatwell-plate.aspx (Accessed: 26 January 2016). 221. Rothman AJ and Salovey P. Shaping perceptions to motivate healthy behavior: the role of message framing. Psychol Bull, 1997. 121(1):3-19. 222. Wansink B and Pope L. When do gain-framed health messages work better than fear appeals? Nutr Rev, 2015. 73(1):4-11. 223. Veitch J, Singh A, van Stralen MM, van Mechelen W, Brug J, and Chinapaw MJ. Reduction in sugar-sweetened beverages is not associated with more water or diet drinks. Public Health Nutr, 2011. 14(8):1388-93. 224. Cabrera Escobar MA, Veerman JL, Tollman SM, Bertram MY, and Hofman KJ. Evidence that a tax on sugar sweetened beverages reduces the obesity rate: a meta-analysis. BMC Public Health, 2013. 13:1072. 225. United States Department of Agriculture Economic Research Service. Food Availability (per capita) Data System. 2015; Available at: http://www.ers.usda.gov/data-products/food- availability-(per-capita)-data-system/food-availability-documentation.aspx (Accessed: December 28 2015). 226. Boehm S, Selves EJ, Raleigh E, Ronis D, Butler PM, and Jacobs M. College students' perception of vulnerability/susceptibility and desire for health information. Patient Education and Counseling, 1993. 21(1–2):77-87. 227. Wansink B, Cao Y, Saini P, Shimizu M, and Just DR. College cafeteria snack food purchases become less healthy with each passing week of the semester. Public Health Nutr, 2013. 16(7):1291-5. 228. Gordon-Larsen P, The NS, and Adair LS. Longitudinal trends in obesity in the United States from adolescence to the third decade of life. Obesity (Silver Spring), 2010. 18(9):1801-4. 229. Arcand J, Scourboutakos MJ, Au JT, and L'Abbe MR. trans Fatty acids in the Canadian food supply: an updated analysis. Am J Clin Nutr, 2014. 100(4):1116-23. 230. Shepherd R, Farleigh CA, and Land DG. Preference and sensitivity to salt taste as determinants of salt-intake. Appetite, 1984. 5(3):187-97. 231. Ryan C. Israeli firm aims to reduce sodium in meats. Global Meat News.com 2015; Available at: http://www.globalmeatnews.com/Industry-Markets/Israeli-firm-aims-to-reduce-sodium-in-meats (Accessed: November 17 2015). 232. Kirkpatrick SI, Reedy J, Kahle LL, Harris JL, Ohri-Vachaspati P, and Krebs-Smith SM. Fast- food menu offerings vary in dietary quality, but are consistently poor. Public Health Nutr, 2014. 17(4):924-31. 233. Garber AK and Lustig RH. Is fast food addictive? Curr Drug Abuse Rev, 2011. 4(3):146-62. 234. Scourboutakos MS-A, Z; L'Abbe, MR;. Added sugars in kids' meals from chain restaurants. Preventive Medicine Reports, 2014. In-Press. 235. Elbel B, Kersh R, Brescoll VL, and Dixon LB. Calorie labeling and food choices: a first look at the effects on low-income people in New York City. Health Aff (Millwood), 2009. 28(6):w1110- 21. 236. Elbel B, Gyamfi J, and Kersh R. Child and adolescent fast-food choice and the influence of calorie labeling: a natural experiment. Int J Obes (Lond), 2011. 35(4):493-500. 237. Ogilvie M. Sodium should be included on menus alongside basic calories, in The Toronto Star. 2014. 238. New York City Department of Health and Mental Hygiene. National Salt Reduction Initiative - Sodium Warning Labels. 2015; Available at: http://www1.nyc.gov/site/doh/health/health- topics/national-salt-reduction-initiative.page (Accessed: 23 January 2016).

139

239. Scourboutakos M. High Sodium Warning Labels…New York’s Latest Public Health Policy. 2015; Available at: https://www.nutrition.org/asn-blog/2015/11/high-sodium-warning-labelsnew-yorks- latest-public-health-policy/ (Accessed: 23 January 2016). 240. Cramer S. Food should be labelled with the exercise needed to expend its calories. BMJ, 2016. 353:i1856. 241. Chef Nathanel Barrett. Discussion Regarding Burwash Dining Hall Cafeteria, M. Scourboutakos, Editor. 2014. 242. Sarink D, Peeters A, Freak-Poli R, Beauchamp A, Woods J, Ball K, et al. The impact of menu energy labelling across socioeconomic groups: A systematic review. Appetite, 2016. 99:59-75. 243. Alkerwi A, Crichton GE, and Hebert JR. Consumption of ready-made meals and increased risk of obesity: findings from the Observation of Cardiovascular Risk Factors in Luxembourg (ORISCAV-LUX) study. Br J Nutr, 2014:1-8. 244. Howard S, Adams J, and White M. Nutritional content of supermarket ready meals and recipes by television chefs in the United Kingdom: cross sectional study. BMJ, 2012. 345:e7607. 245. Statistics Canada. Canadian Community Health Survey - Nutrition (CCHS). 2014; Available at: http://www23.statcan.gc.ca/imdb/p2SV.pl?Function=getSurvey&SDDS=5049 (Accessed: 18 March 2016). 246. Te Morenga L, Mallard S, and Mann J. Dietary sugars and body weight: systematic review and meta-analyses of randomised controlled trials and cohort studies. BMJ, 2013. 346:e7492. 247. Yang Q, Zhang Z, Gregg EW, Flanders W, Merritt R, and Hu FB. Added sugar intake and cardiovascular diseases mortality among us adults. JAMA Internal Medicine, 2014. 248. Kell KP, Cardel MI, Bohan Brown MM, and Fernandez JR. Added sugars in the diet are positively associated with diastolic blood pressure and triglycerides in children. Am J Clin Nutr, 2014. 100(1):46-52. 249. Te Morenga LA, Howatson AJ, Jones RM, and Mann J. Dietary sugars and cardiometabolic risk: systematic review and meta-analyses of randomized controlled trials of the effects on blood pressure and lipids. Am J Clin Nutr, 2014. 100(1):65-79. 250. World Health Organization. Media Center - WHO opens public consultation on draft sugars guideline. 2014; Available at: http://www.who.int/mediacentre/news/notes/2014/consultation- sugar-guideline/en/ (Accessed: March 9 2014). 251. Heart and Stroke Foundation of Canada. Heart and Stroke Foundation of Canada Position Statement - Sugar, Heart Disease and Stroke. 2014; Available at: http://www.heartandstroke.com/site/c.ikIQLcMWJtE/b.9201361/k.47CB/Sugar_heart_disease_an d_stroke.htm (Accessed: Nov 8 2014). 252. Johnson RK, Appel LJ, Brands M, Howard BV, Lefevre M, Lustig RH, et al. Dietary sugars intake and cardiovascular health: a scientific statement from the American Heart Association. Circulation, 2009. 120(11):1011-20. 253. Ervin R and Ogden C. Consumption of added sugars among US adults, 2005-2010, US Department of Health and Human Services, Editor. 2013, Centers for Disease Control and Prevention: United States. 254. Health Canada. Canadian Nutrient File (CNF). 2012; Available at: http://webprod3.hc- sc.gc.ca/cnf-fce/index-eng.jsp (Accessed: March 23 2014). 255. United States Department of Agriculture. USDA National Nutrient Database for Standard Reference. 2011; Available at: http://ndb.nal.usda.gov/ (Accessed: March 23 2014). 256. U.S. Department of Health and Human Servies and U.S. Department of Agriculture. Dietary Guidelines for Americans 2005. 2005: Washington, DC: U.S. Government Printing Office. 257. Ordinance amending Article 8 of the San Francisco Health Code by adding Sections 471.1 through 479.9 to set nutritional standards for restaurant food sold accompanied by toys of other youth focused incentive items. 2010; Available at: http://www.sfbos.org/ftp/uploadedfiles/bdsupvrs/committees/materials/lu092710_101096.pdf (Accessed: March 8 2014).

140

258. Otten JJ, Hekler EB, Krukowski RA, Buman MP, Saelens BE, Gardner CD, et al. Food marketing to children through toys: response of restaurants to the first U.S. toy ordinance. Am J Prev Med, 2012. 42(1):56-60. 259. Hobin EP, Hammond DG, Daniel S, Hanning RM, and Manske S. The Happy Meal(R) Effect: the impact of toy premiums on healthy eating among children in Ontario, Canada. Can J Public Health, 2012. 103(4):e244-8. 260. Health Canada. Improving Nutrition Information on Food Labels: Health Canada's Proposed Changes to the Look of the Nutrition Facts Table and the List of Ingredients. 2014; Available at: http://www.hc-sc.gc.ca/fn-an/label-etiquet/nutrition-facts-valeur-nutritive-fs-fr-eng.php (Accessed: July 28 2014). 261. U.S. Food and Drug Administration. Proposed Changes to the Nutrition Facts Label. 2014; Available at: http://www.fda.gov/food/guidanceregulation/guidancedocumentsregulatoryinformation/labelingn utrition/ucm385663.htm (Accessed: July 28 2014 2014).

141

Appendix A

Restaurant meals - Almost a full day’s worth of calories, fats and sodium (JAMA Intern Med, 2013)

This research letter has been published: Scourboutakos MJ, Semnani-Azad Z, L’Abbé MR. 2013. Restaurant meals – almost an entire day’s worth of calories, fats and sodium. JAMA Internal Medicine, 173(14):1373-1374.

Background:

Due to the prevalence of eating-out, the connection between fast-food consumption and disease- risk has garnered widespread attention.1 However, less attention has been given to the disease promoting potential of meals from sit-down restaurants (referred to as SDR, and defined by the presence of table service) which account for a larger share of total away-from-home food spending, and whose share is expected to rise over the next decade.2

To date, no study has systematically documented the nutrient levels in meals from SDR. In particular, nutrients of concern include calories, fat, saturated fat and sodium, whose excess consumption is associated with: obesity, hypertension, heart disease, diabetes and cancer.3,4 The objective of this study was to analyze the nutritional profile of breakfast, lunch and dinner meals from SDR.

Methods:

23 chain restaurants that provided nutrition information online and had 10 or more locations were identified using the 2010 Directory of Restaurant and Fast-food Chains in Canada.5 Menus were retrieved from the restaurant websites in 2012. All breakfast, lunch and dinner meals whose constituents were present in the U of T restaurant nutrition database (constructed in 2010/2011) were included. Four restaurants were excluded because less than 60% of their meals could be calculated using the database. The nutritional profile of every potential meal combination (entrée, plus side dish(es) and sauces that are customarily served with the meal) was calculated. With a few minor exceptions, the majority of ―upgrades‖ that cost extra (such as upgrading from regular fries to sweet potato fries) were excluded.

142

In total, 3507 different variations of 685 meals, as well as 156 desserts from 19 SDR were included. Data was weighted so that meals with many different variations were not overrepresented. Nutrient values were calculated as a percentage of the daily recommended intake level (%DV). Descriptive statistics were calculated for: calories, fat, saturated fat, trans fat, cholesterol, sodium, percentage of total calories derived from fat and the percentage of total fat derived from saturated fat, using Statistica version 10 software (Tulsa, Oklahoma) and Microsoft Excel 2002.

Results:

On average, breakfast, lunch and dinner meals from 19 chain SDR contained: 1129 calories (56% of the average daily 2000 calorie recommendation), 151% of the amount of sodium an adult should consume in a single day (2263 mg), 89% of the recommended daily fat intake level (58 g), 83% of the recommended daily saturated and trans fat intake level (16 g saturated fat and 0.6 g trans fat) and 60% of the daily value for cholesterol (179 mg) (Table).

With respect to sodium, more than 80% of meals exceeded the daily recommended intake level (1500mg), with more than 50% exceeding the daily tolerable upper intake level (2300 mg). Only 1% of meals had less than 600 mg of sodium, the ―healthy level‖ for meals, according to the FDA.6 Almost 50% of meals exceeded the daily recommendations for fat (65g) and 25% exceeded daily recommendations for saturated/trans fat and cholesterol.

Furthermore, a dessert, if ordered, would add an additional 549 calories, 27 g of fat (43% DV), 13 g of saturated fat, 0.6 g of trans fat (68% DV), and 46 g of sugar.

Meals identified by the restaurants as being ―healthy‖ contained on average 474 calories, 13 g of fat (20% DV), 3 g of saturated fat (17% DV) and 752 mg of sodium (50% AI).

143

Table. Nutrient levels in sit-down restaurant meals All meals (n=3507) a Breakfast (n=150) Lunch (n=533) Dinner (n=2824) Mean±SE b 95% CL for mean %DVc Mean±SE 95% CL %DV Mean±SE 95% CL %DV Mean±SE 95% CL %DV Calories (kcal) 1129±12 (1105, 1152) 56 1126±24 (1092, 1186) 56 1025±26 (973, 1077) 51 1153±14 (1126, 1180) 58 Fat (g) 58±1 (56, 60) 89 52±3 (46, 58) 80 53±2 (49, 57) 81 60±1 (58, 62) 92 Saturated Fat (g) 16±0.3d (16, 17) 17±1 (15, 19) 14±1 (13, 16) 17±0.4 (16, 17) 83 88 73 87 Trans Fat (g) 0.6±0.02d (0.5, 0.6) 0.7±0.1 (0.5, 0.9) 0.6±0.04d (0.5, 0.7) 0.5±0.02d (0.5, 0.6) Cholesterol (mg) 179±4 (171, 188) 60 421±29 (364, 479) 140 130±6 (118, 143) 43 168±4 (160, 176) 56 Sodium (mg) 2263±30 (2204, 2322) 151e 2027±141 (1748, 2305) 135 2206±71 (2067, 2345) 147 2297±34 (2231, 2364) 152 Fat % total 45±0.4d (44, 46) NA 39±2 (36, 42) NA 46±1 (44, 47) NA 45±1 (44, 46) NA calories Sat fat % total fat 24±0.4d (23, 25) NA 32±1 (30, 35) NA 27±1 (25, 29) NA 28±0.4d (28, 29) NA Abbreviations: NA, not applicable. %DV, mean nutrient level expressed as a percentage of the daily value as defined by the FDA.6 a n represents the maximum n. Some nutrient’s n may be smaller due to missing data. b Means were weighted so that meals with many different variations were not overrepresented c Daily values are as follows; Calories: 2000 kcal, Fat: 65g, Saturated+Trans Fat: 20g, Cholesterol: 300mg d Note: an additional decimal place was included in the SE for clarity from non-zero values e Sodium was calculated as %AI, which is the mean sodium level in the category, expressed as a percentage of the daily adequate intake (AI) for adults (1500 mg per day) as defined by the IOM4

144

Comment:

This study presents the average nutrient levels in a variety of different breakfast, lunch and dinner meals from the major SDR chains. On average, meals contained more than a full day’s worth of sodium and nearly a day’s worth of fat and saturated fat.

The high level of saturated fat is worrisome because according to the IOM, intakes of saturated fat should be kept as low as possible.3 Furthermore, though recommendations suggest that approximately 20-35% of energy should come from fat; in this study, 45% was derived from fat.3

On a positive note, the results showed that trans fat levels were commendably low. Furthermore, meals advertised as being ―healthy‖ were substantially healthier compared to average meals.

Presently there is no data on the nutritional profile of meals from SDR. Previous research on meals purchased from fast food chains reported an average of 1751 mg of sodium and 881 calories,8 which is lower than the levels seen in SDR meals in this study.

Limitations include the fact that the data represented meals available in restaurants, and did not reflect actual meals consumed by restaurant patrons. Furthermore, beverages, appetizers and condiments that are often added by the consumer, and would further increase intake levels, were not accounted for.

Overall, the results of this study demonstrate that calorie, fat, saturated fat and sodium levels are alarmingly high in breakfast, lunch and dinner meals from multiple chain sit-down restaurants. Therefore, addressing the nutritional profile of restaurant meals should be a major public health priority.

Contributions:

Mary Scourboutakos and Mary L’Abbé had full access to all of the data in the study and take full responsibility for the integrity and accuracy of the data analysis. Mary Scourboutakos and Mary L’Abbé were responsible for the study concept and design, as well as interpretation of data. Mary Scourboutakos and Zhila Semnani-Azad were responsible for acquisition of the data. Mary Scourboutakos was responsible the analysis of data, and received assistance from Dr. Wendy

145

Lou. Mary Scourboutakos drafted the manuscript and all authors reviewed the manuscript for intellectual content.

146

Appendix B

Sugar levels in kids’ meals from Canadian chain restaurants (Prev Med Reports, 2015)

This article is currently in-press: Scourboutakos MJ, Semnani-Azad Z, L’Abbé MR. 2015. Sugar levels in kids’ meals from Canadian chain restaurants, (In-Press, Preventive Medicine Reports).

Abstract:

Objective: To analyze the added sugars in kids’ meals from Canadian chain restaurants in relation to the World Health Organization’s proposed sugar recommendation (less than 5% of total daily calories should come from added sugars) and current recommendation (less than 10% of total daily calories should come from added sugars).

Methods: Total sugar levels were retrieved from the websites of 10 fast-food and 7 sit-down restaurants in 2010. The added sugar levels in 3178 kids’ meals from Canadian chain restaurants were calculated in 2014 (in Toronto, Canada) by subtracting all naturally occurring sugars from the total sugar level.

Results: The average amount of added sugars in restaurant kids’ meals (25±0.36 g) was equivalent to the WHO’s proposed daily recommendation for sugar intake. There was a wide range of added sugar levels in kids’ meals ranging from 0g to 114g. 50% of meals exceeded the WHO’s proposed daily sugar recommendation, and 19% exceeded the WHO’s current daily sugar recommendation.

Conclusion: There is a wide range of sugar levels in kids’ meals from restaurants, and many contain more than a day’s worth of sugar.

Introduction

It is well recognized that excessive sugar consumption is associated with dental caries and obesity.246 However, additional adverse effects have recently emerged.247 While the Institute of Medicine recommends that up to 25% of total calories can safely come from added sugars

147

(sugars that are not naturally occurring in a food), Yang et al showed that cardiovascular disease risk increases exponentially as the percentage of total calories coming from sugar increases.247 Additionally, studies have shown that added sugars can increase risk for high blood pressure and high triglycerides, even in kids.248, 249

Recently, the World Health Organization (WHO) released draft sugar guidelines recommending that added sugars (including all monosaccharides [such as glucose, fructose] and disaccharides [such as sucrose or table sugar] that are added to food by the manufacturer, the cook or the consumer, as well as sugars that are naturally present in honey, syrups, fruit juices and fruit concentrates),250 should contribute less than 10% of total energy intake per day. Furthermore, they recommended that intakes totaling less than 5% of total energy may have additional benefits.250 The Heart and Stroke Foundation of Canada also recommends that an individual’s total intake of free sugar should not exceed 10% of total calories, and ideally should be less than 5%.251 Additionally, the American Heart Association recommends that women should consume no more than 100 calories from added sugars per day, while men should consume no more than 150 calories for added sugars, with an aim to keep levels at approximately 5% of total calories depending on energy intake.252

Thirty-three percent of added sugars are consumed outside-the-home.253 While restaurant foods have been shown to have a poor nutritional quality,151 there are few published studies examining sugar levels.43 Thus, the objective of this study was to analyze levels of added sugar in kids’ meals from chain restaurants.

Methods

Data was derived from the Food Label Information Program for Restaurants (FLIP-R) database which was created in 2010/2011.151 The database includes nutrition information for all restaurants that provided publicly available nutrition information online and had 20 or more locations in 2010. Of the 85 chain restaurants whose data was included in the database, 33 offered a kids’ menu and provided sufficient data for their kids’ meals to be calculated. Additional details concerning the construction of the database can be found elsewhere.151

148

Kids’ meals typically consisted of an entrée, side dish, beverage and dessert. All potential kids’ meal combinations were calculated. For example, each entrée was calculated with every potential side dish, beverage and dessert that can be ordered for the set price.

Restaurants reported ―total sugar‖ levels which include both naturally occurring and added sugars. Therefore, estimates of added sugars were calculated by subtracting all naturally occurring sugars (from fruits, vegetables, juices, and dairy foods) from the total sugar data. Estimates for the amount of naturally occurring sugar in a meal was determined using the Canadian Nutrient File and the USDA National Nutrient Database for Standard Reference.254, 255

Data was weighted so that each restaurant was equally represented. Descriptive statistics for added sugars were calculated for all meals and for each category of meal components (entrée, side dish, beverage and dessert). Because data was weighted, standard errors were the most appropriate measure of variance. Added sugar levels were evaluated as a percentage of total energy based on the daily estimated energy requirement recommendations for a sedentary 4-8 year old boy, which is equivalent to the recommendations for a moderately active 4-8 year old girl (1800 kcal/day). This estimate is from the "Dietary Guidelines for Americans" which derive these estimates using Estimated Energy Requirements from the Institute of Medicine Dietary Reference Intake macronutrients report.256 Based on the WHO’s recommendation this would be 45 g of added sugar per day according to the 10% recommendation and 22.5 g based on their proposed (5%) recommendation. Lastly, the proportion of meals that exceeded the WHO’s current recommendation (<10% of total energy should come from added sugar) and proposed recommendation (<5% of total energy should come from added sugar) was tabulated.250 All statistical analysis was conducted in 2014 using SAS version 9.3 software (2010; SAS Institute Inc.).

Results

3178 meals from 10 fast-food and 7 sit-down restaurants were analyzed. The amount of added sugars in an average kids’ meal (25±0.36 g) was equivalent to the WHO’s proposed daily recommendation for sugar intake (Table 1). There was a wide range of added sugar levels in kids’ meals ranging from 0g to 114g.

149

50% of meals exceeded the WHO’s proposed daily sugar recommendation (<5% of total energy should come from added sugar), and 19% exceeded the WHO’s current daily sugar recommendation (<10% of total energy should come from added sugar) (Figure 1). Beverages on average contained 16±20 g. Beverages with the highest sugar content were typically soft drinks, fruit juices with added sugars and chocolate milk. Desserts on average contained 12±7 g of added sugar. Entrées on average contained 3±6 g. The highest ranking entrées were chicken nugget meals containing honey mustard dipping sauces, as well as rib meals and sandwiches with sweet sauces. Side dishes on average contained 0.6±2 g of added sugars.

Contributions

Mary Scourboutakos and Mary L’Abbe were responsible for the study concept and design as well as the interpretation of data. Mary Scourboutakos and Zhila Semanani-Azad were responsible for data collection. Mary Scourboutakos conducted the statistical analysis and drafted the manuscript. All authors reviewed the manuscript for important intellectual content.

150

Table 1 Added sugars in kids' meals from chain restaurants Added sugars (g) % of total 25th 75th n Mean±SE/SD* energy† Minimum percentile Median percentile Maximum

Entire Meal (entrée + side dish + beverage + dessert) 3178 25±0.4 6% 0 9 22 37 114 Components of the Meal Entrées 85 3±6 0.5% 0 0 1 3 33 Side Dishes 56 0.6±2 0% 0 0 0 0.1 12 Beverages 33 16±20 4% 0 0 11 28 73 Desserts 35 12±7 3% 0 8 14 16 30 *SE/SD, For the "entire meal" standard error was reported because this is a weighted mean (data was weighted so that each restaurant is equally represented), however, standard deviation was reported for the "components of the meal" because this data was not weighted. †Daily estimated energy requirements were conservatively estimated based on the recommendations for a sedentary 4-8 year old boy, which is equivalent to the recommendations for a moderately active 4-8 year old girl (1800 kcal/day). This estimate is from the "Dietary Guidelines for Americans" which derive these estimates using Estimated Energy Requirements from the Institute of Medicine Dietary Reference Intake macronutrients report.

Entrées include: hamburgers, chicken nuggets, tacos, sandwiches, hot dogs, macaroni and cheese. Side dishes include French fries, onion rings, applesauce, vegetables, salads, and soup. Beverages include soft drinks, juice, and milk. Desserts include ice cream, cookies, jello, and baked goods

151

Figure 1 Proportion of kids' meals that exceeded the WHO's current and proposed added sugar recommendations

50% of kids' meals exceeded the WHO's of meals proposed daily 50% recommendation* contained less than the WHO's daily recommendations 19% also exceeded the WHO's current daily recommendation*

*The WHO's proposed recommendation (<5% of total calories should come from added sugars) and current recommendation (<10%) were calculated using the daily estimated caloric requirements for a sedentary 4-8 year old boy or a moderately active 4-8 year old girl (1800 kcal/day), according to the "Dietary Guidelines for Americans" which derive these estimates using Estimated Energy Requirements from the Institute of Medicine Dietary Reference Intake macronutrients report. Proportions are based on weighted frequencies

Discussion

Many kids’ meals contained excessive amounts of added sugar and often exceeded the WHO’s daily sugar recommendation. Even though the majority of added sugars were found in beverages and desserts, this study showed that some entrées and side dishes also contained excessive amounts of added sugars.

152

We found that the average amount of added sugars in a kids’ beverage equals approximately 4% of total calories (almost an entire days’ worth of sugar). Meanwhile, the amount of sugar in an average kids’ dessert adds an additional 3%. Thus, the amount of sugar in an average kids’ beverage and dessert surpasses the 5% recommendation. Therefore, when kids select entrées and side dishes that also contain added sugars (such as those with sweet dipping sauces), their meal will far surpass the daily recommendation for added sugars.

Only one other study has investigated added sugar levels in restaurants, and similarly found that kids’ meals contained 22g of added sugars.43

In light of the childhood obesity epidemic, improving the nutritional quality of kids’ meals is an important priority. While restaurants have made efforts to offer healthier choices—that include fruit and milk instead of dessert and soft drinks—these healthier options are still being offered alongside unhealthy options. Whether or not kids are actually choosing the healthier menu options is unknown. Thus, efforts to improve this situation should not only focus on making healthy options available, but should also consider ways to incentivize healthier choices. Examples of this are the San Francisco and Santa Clara County toy ordinances, which restrict the distribution of toys with meals that fail to meet nutritional criteria.257-259

Currently, no chemical analytic methods exist to distinguish between added and naturally occurring sugars. Thus, our study conservatively estimated added sugar levels. Therefore, due to the lack of available data, the proposed inclusion of added sugars in the Nutrition Facts Table in both Canada and the United States will greatly aid future research in this area.260, 261

Overall, this study sheds light on the emerging issue concerning sugar in the food supply and demonstrates the need for policy action to address this situation.

153

Appendix C Supplementary tables from “Restaurant Menus – Calories, Caloric Density and Serving Size”

Supplemental Table 1 Description of restaurant food categories

Meal Items Establishment N N Notes on Categories Type Est.a Breakfast SDR 108 4 Includes breakfast entrees that combine multiple different items. QSR 232 19 Items such as bagels are plain (ex. butter has not been added). Chicken SDR 39 12 Includes chicken nuggets/strips, does not include sauces, quantity varies from 2-9 QSR 55 11 and is based on the quantity provided by the establishment.

Hamburgers SDR 65 15 Includes all varieties of hamburgers: single, double, with and without cheese QSR 81 9 and bacon Pasta Entrées SDR 140 16 Includes pasta dishes that contain meat and/or seafood. When data for a "full order" and "half order" was provided, QSR 30 7 only the "full order" was analyzed. Typically represents the entire meal, hence entrée designation.

Salad Entrées SDR 44 17 All salads were analyzed including dressing. When salad data was provided without dressing, dressing was added. To account for variation, salads were analyzed with the dressing containing QSR 53 22 the most calories as well as the dressing containing with the least calories. Only salads designated as being entrée or full size salads were analyzed.

Salad Entrées SDR 93 18 Does not include salads containing w/Meat bacon bits. QSR 118 21 Sandwiches/Wraps SDR 159 19 When multiple data for the same sandwich/wrap was provided (such as with and without cheese and sauces) the value with the extra toppings was QSR 522 38 included and the plain value was not. When multiple sandwich/wrap sizes were provided, only the "medium/regular" was analyzed. Stir Fry SDR 25 8 Stir fry entrees only contain sauces that were used as part of the cooking process, QSR 39 2 sauces that would be added by customers were not included.

154

Tacos/Burritos SDR 10 6 Also includes quesadillas and enchiladas. QSR 116 6 Beef SDR 75 12 Contains steak, prime rib, meatloaf and tenderloin. When multiple sizes were included, all data was analyzed. Meat/Seafood SDR 25 9 Contains dishes that contain a combination of more than one type of meat or meat in combination with seafood. Ribs SDR 29 13 Includes multiple portion sizes. Seafood SDR 50 15 Pizza QSR 399 14 All data represents one "medium" or "generic sized" slice of pizza. All varieties of crusts were provided. Pizza from SDR was not analyzed. Sushi QSR 23 3 Hot Dogs QSR 20 6 Includes hot dog toppings. Other Items Cookies QSR 75 15 Desserts SDR 177 19 Includes any dessert served at a restaurant. Donuts QSR 99 5 This category does not include tiny donuts. Frozen Dessert QSR 264 14 Contains ice cream etc. Fruit Smoothies QSR 186 10 Muffins QSR 119 11 Includes low-fat varieties. Other Baked Goods QSR 74 21 Pie, cake, brownies, scones, etc. Pastries QSR 34 12 Cinnamon rolls, croissants, etc. Sides

Fries SDR 22 18 For FF outlets, only "medium/regular size" fries were analyzed. When only two sizes were given the larger size was QSR 17 17 analyzed. Sweet potato fries were included.

Soup SDR 98 18 Data represents "bowl" values. For consistency "cup" values were not QSR 204 17 analyzed.

Baked Potato SDR/ QSR 9 8 Plain with no toppings. Baked Potato SDR/ QSR 14 8 Includes sour cream, cheese, butter, etc. w/Toppings Coleslaw SDR/ QSR 16 15 Fries w/Toppings SDR/ QSR 19 14 Includes poutine, fries with gravy, cheese or any kind of topping. Mashed Potatoes SDR/ QSR 16 14 Onion Rings SDR/ QSR 11 11 When multiple sizes were provided, only the medium size was analyzed. Rice SDR/ QSR 27 21 Roasted Potatoes SDR/ QSR 8 7 Salad SDR/ QSR 91 30 Includes salad dressing.

155

Vegetables SDR/ QSR 48 20 aN Est.=the number of establishments that are represented in each category

Supplemental Table 2 Sit-down and quick-service restaurants that were included in the database

Sit-Down Restaurants Quick Service Restaurants Baton Rouge (71)a 241 Pizza (10) Jugo Juice (37) Pizza 73 (21) Boston Pizza (83) A&W (43) KFC (53) Pizza Delight (47) Casey’s (58) Arby's (43) La Cremiere (37) Pizza Nova (10) Denny's (120) Baskin Robbins (83) Lick’s (17) Pizza Salvatore (7) Earl's Restaurant (63) Bento Nouveau (16) Little Caesars (14) Pizzaville (30) East Side Marios (57) Booster Juice (71) Manchu Wok (13) Quiznos (56) Jack Astors (81) Burger King (67) Marble Slab Robin's Donuts (65) Joey’s Restaurant (36) Captain Sub (47) Creamery (4) Starbucks (44) Kelsey's (67) Coffee Time (117) Mary Browns (12) Subway (95) Mike’s Restaurant (105) Country Style (131) McDonald's (85) Taco Bell (40) Milestones (64) Dagwood Sandwiches Mega Wraps (82) Taco Del Mar (54) Montanas (71) and Subs (46) Mmmuffins (46) Taco Time (29) Mr. Greek (47) Dairy Queen (105) Mr. Greek Express (24) TCBY (29) Pizza Hut (134) Druxy’s Deli (46) Mr. Sub (62) Teriyaki Experience (50) Scores Rotisserie (29) Edo Japan (20) Mrs. Vanelli's (28) The Great Canadian Shoeless Joe's (99) Esquires Coffee House Mucho Burrito (43) Bagel (85) Swiss Chalet (59) (4) Nando’s Flame Grilled Tim Hortons (85) The Keg (77) Extreme Pita (37) Chicken (18) Topper’s Pizza (26) White Spot Legendary Flying Wedge Pizza New Orleans Pizza (25) Treats (18) Restaurant (69) (19) New York Fries (5) Van Houtte’s Bistro (27) Freshly Squeezed (35) Opa (14) Wendy's (45) Freshslice (4) Orange Julius (49) White Spot Triple O’s Good Earth Coffeehouse Panago (167) (20) (58) Pita Pit (37) Yogen Fruz (6) Harvey's (25) Jimmy the Greek (12) a This represents the number of items from each establishment that were analyzed. The n listed for each establishment may differ from the n that is available on their website due to the exclusion of items that did not met the study’s inclusion criteria or due to ―combined values‖ where items such as salads were combined with dressings.

156

Appendix D Supplementary tables from “Changes in sodium levels in chain restaurant foods”

Supplementary Figure 1 Distribution of the sodium level (mg/serving) in Canadian chain restaurant foods (n=2198) in 2010 and 2013

250 2010 200

150

100

50 Numberofrestaurant foods

500 100 300 700 900

1500 1100 1300 1700 1900 2100 2300 2500 2700 2900 3100 3300 3500

300 >3600

250

200 2013

150

100

50 Numberofrestaurant fods

100 300 500 700 900

1100 1300 1500 1700 1900 2100 2300 2500 2700 2900 3100 3300 3500 >3600

157

Supplementary Table 1 Changes in sodium levels from 2010 to 2013 among baked goods, desserts and side dishes (n=803)

Sodium level (mg/serving), mean±SD§ Average percent Average change in percent sodium change in among sodium Percenta foods that Percenta among foods ge of decreased ge of that foods , median foods increased, that (25th, 75th that median decrease percentile) increase (25th, 75th n 2010 2013 d (%) || d (%) percentile)

Baked Goods and Desserts Other Baked 14 439±298 461±283 14 -6±1 29 48 (26, 94) Goods Muffins 87 432±159 423±150 8 -18±15 1 5 SDR Desserts 92 332±238 342±285 11 -47±28 16 20 (6, 90) Donuts 49 283±60 283±60 0 0 0 0 Pastries 16 273±91 304±92 19 -11±8 25 97 (53, 135) Cookies 45 179±95 174±97 22 -19±9 7 29 (19, 37) Side Dishes Soup (fast-food) 120 1087±535 1061±497‡ 15 -11 (-15, -9) 0 0 Soup (sit-down 51 979±431 873±470* 39 -26 (-56, -12) 16 21 (14, 42) restaurants) Fries (sit-down 15 818±523 892±593 33 -23 (-27, -7) 27 33 (21, 60) restaurants)

Fries (fast-food) 15 776±445 781±487 36 -30 (-33, -8) 7 237

Rice 13 684±704 660±705 15 -43 (-58, -29) 15 100 Potatoes (baked, 20 540±356 546±362 10 -16 (-19, -12) 10 26 (20, 28) roasted or mashed) Salad 45 378±175 402±187 22 -27 (-31, -19) 31 49 (18 ,121) Coleslaw 12 354±135 324±144 25 -37 (-47, -14) 0 0 Vegetables 23 232±250 220±220 17 -36 (-66, -21) 13 245 (156, 416) *Paried t-tests, p<0.05, †p<0.01, ‡p<0.001, §standard deviation ||Medians were reported to prevent extreme values from skewing the average

158

Supplementary Table 2 Changes in sodium density (mg/100g) from 2010 to 2013 in various food categories from Canadian chain sit-down and fast-food restaurants (n=2198)

Average Average Perc percent percent enta change in change in ge of sodium sodium food among Percen among s foods that tage of foods that that decreased, foods increased, Sodium density decr median that median (mg/100g), mean±SD§ ease (25th, 75th increas (25th, 75th n 2010 2013 d (%) percentile)|| ed (%) percentile) Sit-Down Restaurants Pizza 93 524±154 433±149‡ 90 -17 (-32, -6) 10 10 (8, 12) Sandwiches/Wraps 73 507±194 534±218 22 -9 (-21, -8) 34 12( 6, 35) Chicken 42 456±206 432±197 26 -21 (-44, -5) 19 19 (0.5, 42) Entrées w/ multiple meats and seafood 12 454±178 455±176 25 -3 (-3, -1) 8 12 (ex. surf n turf) Hamburgers 39 415±135 443±128 33 -41 (-45, -28) 28 43 (19, 94) Ribs 19 407±114 416±129 5 -131 16 18 (18, 35) Breakfast 74 389±143 389±142 27 -18 (-28, -5) 19 24 (8, 68) Pasta Entrée 76 334±113 305±110‡ 62 -9 (-22, -4) 18 8 (1, 17) Salad Entrée 19 325±152 340±151 32 -19 (-23, -18) 37 32 (3, 179) Steak and other Beef 36 319±301 320±288 19 -37 (-54, -21) 22 55 (27, 148) Entrees Stir Fry 10 319±135 347±149 10 -41 40 46 (14, 93) Seafood 24 311±217 303±211 17 -20 (-32, -8) 4 69 Salad Entrée w/Meat 41 294±97 306±103 17 -13 (-17, -10) 36 19 (3, 34) Fast-Food Hot Dogs 13 818±60 773±82 77 -7 (-11, -4) 15 11 (3, 20) Chicken 47 619±174 559±163‡ 42 -19 (-35, -12) 6 10 (10, 14) Poutine 15 599±196 568±144 47 -14 (-23, -5) 13 69 (8, 129) Sandwiches/Wraps 203 554±208 548±210 38 -16 (-24, -8) 20 27 (9, 61) Tacos/Burritos 61 522±264 424±124‡ 72 -22 (-35, -13) 28 15 (2, 28) Hamburgers 60 490±131 488±135 37 -7 (-15, -5) 27 5 (2, 8) Breakfast 113 474±191 459±181* 53 -12 (-15, -5) 13 2 (2, 8) Pizza 263 397±119 402±123* 26 -5 (-10, -3) 30 12 (5, 16) Sushi 11 342±71 388±112 0 na 36 35 (14, 36) Stir Fry 24 326±33 326±33 0 na 0 0 Salads w/Meat 22 317±80 313±92 45 -8 (-24, -3) 23 17 (3, 55) Pasta 25 275±152 275±152 0 na 0 0 Salads 20 208±94 218±90 47 -31 (-45, -16) 13 69 (34, 106) Stir Fry (no sodium/low 37 208±56 208±56 0 na 0 0 sodium) Kids' Menu Items

159

Fast-food kid's Meals 13 532±197 503±188 38 -23 (-23 , -4) 8 16 Sit-Down Kid's Meals 68 438±196 419±193 31 -18 (-34, -7) 15 7 (5, 43) Kid's Side Dishes 27 253±178 241±158 33 -22 (-28, -18) 28 183 (12, 263) Baked Goods and Desserts Other Baked Goods 14 425±183 456±177 0 na 43 17 (3, 52) Donuts 49 394±65 392±66 2 -23 0 0 Pastries 16 368±199 381±127 19 -3 (-46, -1) 25 67 (33, 96) Cookies 45 339±115 335±106 18 -21 (-26, -13) 11 37 (35, 60) Muffins 87 335±117 333±112 5 -20 (-29, -9) 5 10 (9, 14) SDR Desserts 92 172±113 173±115 9 -40 (-69, -15) 18 29 (4, 48) Side Dishes Fries (fast-food) 15 479±268 451±219 21 -33 (-45, -19) 14 84 (1, 168) Rice 13 369±321 355±314 23 -29 (-50, -1) 15 122 (100, 145) Fries (sit-down 15 362±227 425±316 27 -15 (-25, -3) 20 31 (6, 240) restaurants) Soup (sit-down 51 350±139 320±125* 41 -21 (-27, -11) 16 17 (5, 52) restaurants) 12 Soup (fast-food) 343±65 328±64‡ 23 -17 (-20, -10) 1 3 0 Coleslaw 12 305±94 284±103 17 -43 (-49, -37) 8 15 Potatoes (baked, 20 243±159 238±160 20 -19 (-29, -12) 5 15 (1, 29) roasted or mashed) Salad 45 243±125 247±120 31 -23 (-35, -15) 29 38 (16, 122) Vegetables 23 171±168 166±151 17 -29 (-60, -17) 13 245 (163, 416) *Paried t-tests, p<0.05, †p<0.01, ‡p<0.001, §standard deviation ||Medians were reported to prevent extreme values from skewing the average

160

Supplementary Table 3 Percentage of foods (n=1603) at each restaurant whose sodium density (sodium per 100g) increased or decreased, and the average percent change, in order of percentage of foods that decreased

Average Average percent percent change in change in sodium sodium (mg/100g) (mg/100g) among foods among foods Percentage that Percentage that of foods decreased, of foods increased, that median (25th, that median (25th, decreased 75th increased 75th Restaurant n (%) percentile) ‡ (%) percentile) Subway* 51 82 -16 (-23 ,-12) 10 2 (2, 2) Pizza Hut* 73 79 -29 (-36, -7) 3 1 (0.15, 1) Taco Time 18 78 -35 (-50, -13) 22 8 (1, 73) Taco Bell* 32 78 -20 (-31, -13) 9 8 (2, 10) A&W 34 73 -6 (-12, -5) 15 2 (1, 2) Boston Pizza 106 72 -9 (-20, -5) 18 10 (6, 15) Dairy Queen 18 72 -13 (-21, -4) 28 15 (1, 27) Shoeless Joe's 18 67 -20 (-27, -11) 28 27 (25, 76) White Spot Triple O's 20 65 -16 (-23, -15) 0 na KFC* 24 62 -23 (-36, -13) 4 9† Dagwood Sandwiches and 38 53 -14 (-25, -7) 45 37 (12, 58) Subs Burger King 51 51 -3 (-7, -1) 22 5 (3, 8) East Side Mario's 45 50 -12 (-25, -5) 41 9 (3, 19) Mike's Restaurant 80 46 -20 (-29, -12) 44 23 (7, 82) Kelsey's 44 45 -16 (-33, -10) 32 18 (3, 32) Arby's 28 43 -5 (-19, -2) 57 18 (5, 72) Panago 171 43 -6 (-12, -3) 44 12 (5, 16) White Spot Legendary 52 42 -10 (-28, -5) 23 23 (8, 47) Restaurant Mmmuffins 46 37 -19 (-20, -17) 0 na Montanas 75 37 -19 (-27, -4) 15 17 (8, 54) Taco Del Mar 27 30 -22 (-36, -13) 70 26 (9, 55) Tim Hortons 70 28 -12 (-21, -3) 19 9 (8, 14) The Great Canadian Bagel 85 27 -15 (-15, -15) 0 na Joey's Restaurant 20 25 -26 (-40, -9) 75 35 (15, 94) McDonald's* 46 30 -7 (-15, -6) 22 10 (1, 17) Edo Japan 18 22 -9 (-10, -9) 0 na Jack Astors 24 21 -36 (-50, -27) 79 78 (30, 144) Harvey's 25 20 -14 (-19, -14) 0 na Mr. Greek 35 17 -31 (-44, -1) 6 132 (20, 244) Scores Rotisserie 30 17 -12 (-22, -11) 20 8 (1, 19)

161

Casey's 46 11 -34 (-49, -18) 22 23 (18, 43) Pizza Delight 34 9 -4 (-7, -3) 6 13 (1, 25) New Orleans Pizza 20 5 -38 90 53 (14, 74) Earl's Restaurant 28 0 na 7 22 (3, 42) Extreme Pita 43 0 na 2 73 Mrs. Vanelli's 28 0 na 4 13 The following restaurants reported no changes in sodium level between 2010 and 2013: 241 Pizza, Baton Rouge, Coffee Time, Country Style, Denny's, Flying Wedge Pizza, Little Caesars, Manchu Wok, Mr. Sub, Pita Pit, Pizza 73, Pizza Nova, Pizzaville, Robin's Donuts, Swiss Chalet, Teriyaki Experience, Van Houtte's Bistro The following restaurants were exlcuded from this table because they had <10 menu items included in this study: Bento Nouveau, Druxy's Deli, Jugo Juice, New York Fries, Opa, Orange Julius, Pizza Pizza, Treats *Indicates restaurants that have made a voluntary commitment to reducing the sodium level in their products †When there was only one food that increased or decreased, the percent change in that food was presented without an indicator of variance ‡Medians were reported to prevent extreme values from skewing the average.

Supplementary Table 4 Sodium levels in foods that were discontinued, newly reported or present in 2010 and 2013 (n=3878)

sodium n (mg/serving) Discontinued Foods 860 993±706* Newly Reported Foods 820 982±730 Foods Present in 2010 and 2013 2198 892±679 *F-value=1.25, p-value=0.3, when controlling for restaurant and food category, there was no significant difference among foods that were present in 2010 and 2013 in comparison to foods that were discontinued or newly reported.

162

Appendix E Supplementary Files from “Restaurant menu-labelling: Is it worth adding sodium to the label”

Supplementary Figure 1A: Hamburger Restaurant Menu Mains Sides Drinks

HAMBURGERS FRENCH FRIES SOFT DRINKS (any kind) Hamburger Small Small Cheeseburger Medium Medium Deluxe Burger Large Large Deluxe Cheeseburger DIET SOFT DRINKS (any kind) Double Burger ONION RINGS Small Double Cheeseburger Small Medium Veggie Burger Medium Large Large JUICE (any kind) SANDWICHES Small SIDE GARDEN SALAD Fried Chicken Sandwich Medium Regular Dressing Grilled Chicken Sandwich Large Fish Sandwich Light Dressing MILKSHAKE Small CHICKEN POUTINE Medium 4 Chicken Nuggets/Strips MOZZARELLA STICKS Large 6 Chicken Nuggets/Strips COFFEE/TEA APPLE SAUCE SALADS Cream Chicken BLT Salad Milk Chicken Caesar Salad Sugar WATER

163

Supplementary Figure 1B: Breakfast Restaurant Menu Breakfast Menu Eggs Benedict French Toast Omelettes (2

eggs) Classic Eggs Multigrain French Deli Omelette

Benedict Toast Ham and Cheese Two poached eggs and bacon on an English Two slices of multigrain bread, with cinnamon muffin, topped with hollandaise sauce. fresh mixed berries and maple syrup. Spinach and Feta

Wild Mushroom Smoked Salmon Benedict Loaded French Toast Salmon on an English muffin topped with French toast topped with fruit, caramel, Veggie & Cheese two poached eggs and hollandaise sauce. candied pecans, vanilla frozen yogurt Avocado Omelette and pure maple syrup.

Florentine Benedict Cheddar Cheese Waffles Two poached eggs on pumpernickel Plain Waffle Classic Items Bread with cream cheese, spinach, smoked salmon and hollandaise sauce. Strawberry Waffle Bacon, 2 Eggs, Toast & Potatoes Asparagus and Brie Banana Waffle Steak, 2 Eggs, Toast & Benedict Mixed Berry Waffle Potatoes English muffin with two poached eggs, cheese, asparagus and hollandaise sauce. Warm Apple Waffle Granola Yogurt Parfait

164

Supplementary Figure 1C: Sub Shop Menu The Sub Shop

Black Forest Ham Sub Available in Turkey Sub Two Sizes: Veggie Sub Roast Beef Sub Tuna Sub 6 Inch Roasted Chicken Sub Sweet Onion Chicken Teriyaki Sub 12 Inch Turkey Breast & Black Forest Ham Sub Meatball and Tomato Sauce Sub All sandwiches are Italian Salami, Pepperoni and Ham Sub served with your Pepperoni Pizza Sub choice of lettuce, Italian Salami, Pepperoni and Cheese Sub tomato, cucumber or Chicken Pizza Sub onions

Steak & Cheese Sub Turkey, Roast Beef and Ham Sub Chicken, Cheese and Bacon Sub Turkey, Ham, Salami and Bologna Sub Turkey, Ham, Bacon and Cheese Sub

165

Supplementary Figure 1D: Dinner Restaurant Menu

***The Following Items are Served with Your Choice Entrées of One Side Dish***

Salad Entrées Sandwiches*** Caesar Salad Grilled Chicken on a Ciabatta Chicken Caesar Salad Roasted Chicken Quesadilla Warm Beet and Spinach Salad Chicken Tacos Santa Fe Chicken Salad Steak Sandwich Pasta Entrées Short Rib Beef Dip Penne Alfredo Chicken*** Mediterranean Linguini w/ Chicken Cajun Blackened Chicken Prawn and Scallop Linguini Roast Chicken with Dijon International Entrées Parmesan Pine Nut Chicken Chicken Curry Rice w/ Naan Bread Steak*** Pad Thai 9oz Top Sirloin Spicy Thai Curry with Shrimp 9oz Top Sirloin w/Peppercorn Sauce Kung Pao Stir Fry 12 oz New York Striploin Seafood Entrées 12 oz Blackened New York Striploin Fish and Chips Seafood*** Cedar Planked Salmon Side Dish Choices (pick one) Fingerling Potatoes w/Garlic Butter Garlic Mashed Potatoes Potato Salad Roast Potatoes Penne Alfredo Coleslaw Mediterranean Vegetables Mixed Green Salad with Vinaigrette French Fries

166

Supplementary Figure 2A: Menu-labelling Treatment 1: Calorie Labelling Mains Sides Drinks

HAMBURGERS Calories FRENCH FRIES Calories SOFT DRINKS Calories Hamburger 300 Small 270 (any kind) Cheeseburger 340 Medium 350 Small 160 Deluxe Burger 670 Large 440 Medium 230 Deluxe Cheeseburger 760 Large 320 Double Burger 910 ONION RINGS DIET SOFT DRINKS Double Cheeseburger 1000 Small 200 (any kind) Veggie Burger 310 Medium 320 Small 0 Large 380 Medium 0

SANDWICHES Large 0 SIDE GARDEN Fried Chicken Sandwich 640 JUICE (any kind) SALAD Grilled Chicken Sandwich 370 Small 180 Regular Dressing 200 Fish Sandwich 500 Medium 260 Light Dressing 40 Large 360 CHICKEN MILKSHAKE 4 Chicken Nuggets/Strips 160 POUTINE 740 Small 440 6 Chicken Nuggets/Strips 250 MOZZARELLA Medium 640 SALADS STICKS 350 Large 950 Chicken BLT Salad 540 COFFEE/TEA APPLE SAUCE 50 Chicken Caesar Salad 450 Cream 22 Milk 5 A 2000 calorie diet is used as the basis for general nutrition advice; Sugar 16 however, individual calorie needs may vary. WATER 0

167

Supplementary Figure 2B: Menu-labelling Treatment 2: Calorie and Sodium Labelling Mains Sides Drinks

HAMBURGERS Calories Sodium FRENCH FRIES Calories Sodium SOFT DRINKS Calories Sodium Hamburger 300 530 Small 270 600 (any kind) Cheeseburger 340 730 Medium 350 790 Small 160 35 Deluxe Burger 670 910 Large 440 1000 Medium 230 50 Deluxe Cheeseburger 760 1320 Large 320 70 Double Burger 910 980 ONION RINGS DIET SOFT DRINKS Double Cheeseburger 1000 1390 Small 200 390 (any kind) Veggie Burger 310 770 Medium 320 920 Small 0 0 Large 380 740 Medium 0 0

SANDWICHES Large 0 0 SIDE GARDEN Fried Chicken Sandwich 640 1420 JUICE (any kind) SALAD Grilled Chicken Sandwich 370 910 Small 180 70 Regular Dressing 200 425 Fish Sandwich 500 860 Medium 260 100 Light Dressing 40 665 Large 360 140 CHICKEN MILKSHAKE 4 Chicken Nuggets/Strips 160 310 POUTINE 740 2500 Small 440 300 6 Chicken Nuggets/Strips 250 470 MOZZARELLA Medium 640 400 SALADS STICKS 350 930 Large 950 590 COFFEE/TEA Chicken BLT Salad 540 1490 0 APPLE SAUCE 50 Cream 22 4 Chicken Caesar Salad 450 1420 Milk 5 5 A 2000 calorie diet, with no more than 2300 mg of sodium per day is used as Sugar 16 0 the basis for general nutrition advice; however, individual needs may vary. WATER 0 0

168

Supplementary Figure 2C:Menu-Labelling Treatment 3: Calorie, Sodium and Serving Size Labelling Mains Sides Drinks

Serving Serving Serving HAMBURGERS Size (g) Calories Sodium FRENCH FRIES Size (g) Calories Sodium SOFT DRINKS Size (oz) Calories Sodium Hamburger 126 300 530 Small 88 270 600 (any kind) Cheeseburger 138 340 730 Medium 116 350 790 Small 16 160 35 Deluxe Burger 290 670 910 Large 147 440 1000 Medium 22 230 50 Deluxe Cheeseburger 315 760 1320 Large 32 320 70 Double Burger 373 910 980 ONION RINGS DIET SOFT DRINKS Double Cheeseburger 398 1000 1390 Small 57 200 390 (any kind) Veggie Burger 175 310 770 Medium 91 320 920 Small 16 0 0 Large 108 380 740 Medium 22 0 0

SANDWICHES Large 32 0 0 SIDE GARDEN Fried Chicken Sandwich 235 640 1420 JUICE (any kind) SALAD Grilled Chicken Sandwich 197 370 910 Small 16 180 70 Regular Dressing 152 200 425 Fish Sandwich 180 500 860 Medium 22 260 100 Light Dressing 152 40 665 Large 32 360 140 CHICKEN MILKSHAKE 4 Chicken Nuggets/ Strips 62 160 310 POUTINE 330 740 2500 Small 16 440 300 6 Chicken Nuggets/Strips 92 250 470 MOZZARELLA Medium 22 640 400 SALADS STICKS 98 350 930 Large 32 950 590

Chicken BLT Salad 416 540 1490 COFFEE/TEA APPLE SAUCE 111 50 0 Chicken Caesar Salad 332 450 1420 Cream 22 4 Milk 5 5 Sugar 16 0 WATER 0 0

169

Appendix F

Supplementary tables/figures from “Does a Healthy Eating Intervention Change Students’ Food and Beverage Choices”

Supplementary Figure 1 Total calories/volume of each beverage taken before and after the intervention

350

Pre 300 Post 250

200

150

100

50 Calories (kcal) / Millilitres Calories Millilitres (kcal) / (ml)* 0 Soft Drinks Juice Chocolate Milk Flavoured All SSBs Milk Diet Soft Drinks* Coffee/Tea* Water* Coffee and Hot Chocolate

Sugar-sweetened beverages (SSBs) Unsweetened beverages

Beverage choices

170

Supplementary Table 1 Food served at the cafeteria on each date of data collection

Collection PRE:Tuesday October PRE: Wednesday PRE: Tuesday PRE: Wednesdsay PRE: Tuesday PRE: Wednesday Dates 28th, POST:Tuesday October 29th, POST: November 11th, November 12th, January 6th, POST: January 7th, POST: February 24th Wednesday February POST: Tuesday POST: Wednesday Tuesday March Wednesday March 25th February 10th February 11th 10th 11th Soup Butternut Squash Mulligatawny Indian Red Lentil Black Bean Indian Rasam Cream of Leek and Potato Broth Turkey and Orzo Lebanese Vegetable Minestrone Chicken Mulligatawny Whole Grain Rice Japanese Udon Soup with Vegetables Hot Table 5 Spice Chicken Wings Chunky Meat Sauce Beef Lasagna Roasted Jerk Pork Steak Fajitas Chicken Korma 1 with Spaghetti Hot Table Szechwan Tofu Spinach Sauce with Veggie Lasagna Chicken Roti or Veggie Fajitas Vegetable Korma 2 Rotini Vegetable Roti Rationale Egg Rolls with Plum Sauce Garlic Cheese Bread Garlic Bread Grilled Pineapple with Bruschetta Hot Naan Bread Centre Coconut Topping; Grilled Plantain Specialty Vietnamese Noodle Vietnamese Noodle Chinese Rice Bowl Japanese Rice Bowl Entrée Salad: Thai Entrée Salad: Bowls: Shanghai and Rice Bowls: Cantonese and with Tofu or with Tofu or Beef: Chicken Noodle Turkey Cobb Vermicelli noodles; pork Udon noodles; beef Chicken: jasmine whole grain and and tofu; sauces and and tofu; sauces and and brown; sauces white; sauces and toppings toppings and toppings toppings Self Serve Jasmine Rice, Rice Vegetable Pilaf, Boiled Roast Potatoes, Rice and Peas, Spiced Vegetable Risotto, Pulau Rice, Herbed Vegetable Noodles, Steamed Asian Red Skin Potatoes, White and Wild Rice Roasted Potatoes, Fluffy White Rice, Potatoes, Mixed s Vegetables, Spinach Collard Blend, Whole Green Collard Green Beans with Vegetables, Corn Greens/Spinach, Beans, Medley of Greens/Spinach Mushrooms Niblets, Parslied Carrot Coins Vegetables Poppadoms and Peas Dessert Carrot Cake Assorted Squares Double Chocolate Oreo Cheesecake Strawberry Carrot Cake Feature Bread Rhubarb Pie Available Fresh Fruit: apples, bananas, oranges and pears Daily Fresh vegetable bar: including various types of lettuce, salad toppers (cucumbers, chickepas, grated carrot, cauliflower), as well as seeds and an assortment of dressings Beverages: Water, coke zero, diet coke, coffee, tea, ginger ale, sprite, orange pop, coke, orange juice, lemonade, cranberry juice, peach juice, apple juice, swiss mocha capuccino, french vanilla coffee, hot chocolate, chocolate milk and 2% milk

171

Supplementary Table 2 Data collection dates and details

Number of Visits to students Beverage the Fruits Date Collection Dates dining in the choices vegetable taken cafeteria bar

October 28th 2014 499 * * Excluded October 29th 2014 425 * * Excluded November 11th 2014 410 * * * November 12th 2014 437 * * * January 6th 2015 414 * * * January 7th 2015 430 * * * February 10th 2015 414 * * * February 11th 2015 486 * * Excluded February 24th 2015 510 * * * February 25th 2015 368 * * * March 10th 2015 430 * * * March 11th 2015 405 * * *

172