Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The role of specific dieting practices in the development of eating disorders Mackay, Ellen Maria 1996

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1996-0404.pdf [ 8.49MB ]
Metadata
JSON: 831-1.0087211.json
JSON-LD: 831-1.0087211-ld.json
RDF/XML (Pretty): 831-1.0087211-rdf.xml
RDF/JSON: 831-1.0087211-rdf.json
Turtle: 831-1.0087211-turtle.txt
N-Triples: 831-1.0087211-rdf-ntriples.txt
Original Record: 831-1.0087211-source.json
Full Text
831-1.0087211-fulltext.txt
Citation
831-1.0087211.ris

Full Text

THE ROLE OF SPECIFIC DIETING PRACTICES IN THE DEVELOPMENT OF EATING DISORDERS by ELLEN MARIA MACKAY B.Sc, The University of British Columbia, 1988 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (School of Family and Nutritional Sciences) We accept this thesis as conforming to the required standards THE UNIVERSITY OF BRITISH COLUMBIA May 1996 © Ellen Maria Mackay, 1996 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada Date DE-6 (2/88) Abstract The purpose of this study was to investigate the relationship between dieting practices and the development of eating disorder symptomology in dieting female college students. A self-report questionnaire on dieting practices was administered at the beginning and end of the academic year to 179 dieting women (Age: 21 ± 3 years (mean ± SD); Wt: 62.5 ±11.7 kg; Ht: 165.3 ± 7.1 cm; BMI: 22.8 ± 3.5 kg/m2). A Food Frequency Questionnaire was administered at baseline. Data were obtained on the duration of dieting, methods used to lose weight, the macronutrient intake, restrictiveness of the diet, and the rate of weight lost/week dieting. Eating disorder symptomology was measured by the Health Information Questionnaire (HIQ) and was assessed based on DSM-IV criteria for eating disorders. Responses to the HIQ yielded an overall Symptom Intensity Score (SIS). Subjects with probable eating disorders at baseline were excluded from the analyses. Correlations were observed between the SIS and age at first diet (baseline: r = -.31, p<.01; 6 months: r = -.20, p<.01) and baseline long-term dieting (r = .16, p<.05). Of the dieting methods, correlations between the SIS and Food Avoidance (baseline: r= .51, p<.001; 6 months: r=.49, p<.001), Fat & Calorie Counting (baseline: r= .31, p<.001; 6 months: r=.30, p<.001), Use of Diet Aids (baseline: r= .27, p<.001; 6 months: r=.22, p<.01), and Number of Dieting Methods (baseline: r= .60, p<.001; 6 months: r=.49, p<.001) were also observed. A correlation between carbohydrate intake and Change in SIS was observed (r=-.16, p<.05). Restrictiveness of the diet did not correlate with SIS. However, rate of weight loss correlated with the SIS (baseline: r= .314, p<.01; 6 months: r=.332, p<.01). None of the dieting variables had a significant influence on the prediction of SIS; only 1-5% of the variance in the SIS was explained by the variables examined in this study. These results suggest that specific dieting practices may not predict disordered eating but may be associated with some disturbed eating practices. Further consideration of the frequency of dieting, rate of weight loss and the methods used to lose weight may be warranted as these dieting practices may perpetuate dieting and disturbed eating behaviours. i i Table of Contents ABSTRACT ii TABLE OF CONTENTS iii LIST OF TABLES vi LIST OF FIGURES ix LIST OF APPENDICES x ACKNOWLEDGMENTS xii I. INTRODUCTION 1 1. Rationale 1 2. Purpose of the Study 2 3. Objectives/Specific Aims 2 4. Hypotheses 3 II. LITERATURE REVIEW 4 1. Eating Disorders 4 A . Introduction 4 B. Prevalence of Eating Disorders 9 C . Etiology 10 2. Dieting 10 A . Definition of Dieting 10 C . Prevalence of Dieting in Women 12 D. Causes of Dieting 13 3. The Role of Dieting in the Development of Eating Disorders 15 A . Cl inical Studies 16 B. Dieting: The Risk vs. The Benefits 20 4. Dieting Practices as Risk Factors for Eating Disorders 22 A . Duration of Dieting 23 B. Dieting Methods 27 C . Macronutrient Intake 31 D. Restrictiveness of the Diet 37 5. Summary 43 III. EXPERIMENTAL DESIGN AND METHODS 45 i i i 1. Study Design 45 A . Overview 45 B . Subject Criteria 49 C . Sample Size 50 2. Methods 50 A . Anthropometric Measurements 50 B . Questionnaires 51 C . Pretesting of the Dieting Practices Survey and F F Q 61 3. Statistical Analysis 64 IV. RESULTS 67 1. Distribution of the Data 67 2. Description of the Sample 67 A . Sample Size 67 B . Sample Characteristics 68 C . Eating Disorder Symptomology 71 3. Relationship Between Subject Characteristics and Eating Disorder Symptomology 80 A . Correlation Between Subject Characteristics and Symptom Intensity Score 80 B. Correlation Between Subject Characteristics and Binge Eating and Compensatory Behaviours 80 4. Relationship Between Duration of Dieting and Eating Disorder Symptomology 84 A . Duration of Dieting 84 B . Correlation Between Duration of Dieting and Symptom Intensity Score 84 C . Correlation Between Duration of Dieting and Binge-eating and Compensatory Behaviours 84 D. Regression of Baseline Duration of Dieting on Symptom Intensity Score 87 5. Relationship Between Dieting Methods and Eating Disorder Symptomology 87 A . Dieting Methods 87 B . Factor Analysis of Dieting Methods 89 C . Correlation Between Dieting Method Factors and Symptom Intensity Score 92 D. Correlation Between Dieting Method Factors and Binge-Eating and Compensatory Behaviours 92 E . Regression of Dieting Method Factors on Symptom Intensity Score 95 6. Relationship Between Macronutrient Intake and Eating Disorder Symptomology 95 A . Macronutrient Intake 95 B. Correlation Between Macronutrient Intake and Symptom Intensity Score 97 C . Correlation Between Macronutrient Intake and Binge-Eating and Compensatory Behaviours 97 D. Regression of Macronutrient Intake on Symptom Intensity Score 101 7. Relationship Between Restrictiveness of the Diet and Eating Disorder Symptomology 101 A . Restrictiveness Variables 101 B . Correlation Between Restrictiveness of the Diet and Symptom Intensity Score 101 C . Correlation Between Restrictiveness of the Diet and Binge-Eating and Compensatory Behaviours 105 D. Regression of Restrictiveness of the Diet on Symptom Intensity Score 105 8. Vitamin/Mineral Supplementation 105 9. Vegetarianism 108 10. Comparison of Dieting Practices in Subjects According to Dieting Status at 6 Months 108 A . A g e , Weight and B M I 108 iv B. Eating Disorder Symptomology 108 C . Duration of Dieting 111 D. Dieting Method Factors 111 E . Macronutrient Intake 111 F. Restrictiveness of the Diet 111 V DISCUSSION AND CONCLUSIONS 117 1. Major Findings 117 2. Sample Characteristics 120 A . Sample Characteristics 120 B . A g e 120 C . Weight and B M I 120 D. Vegetarianism 123 3. Eating Disorder Symptomology 123 A . Eating Disorder Diagnoses 123 B . Previous Treatment 124 C . Symptom Intensity Score 124 D. Binge Eating and Compensatory Weight Loss Behaviours 125 4. Duration of Dieting 126 A . A g e at Onset of Dieting 126 B . L o n g - and Short-term Duration of Dieting 127 5. Dieting Methods 127 A . Sources of Information 127 B . Dieting Methods 128 6. Macronutrient Intake 129 7. Restrictiveness of the Diet 130 8. Dieting Status 134 9. Limitations 134 A . Questionnaires 135 B . Experimental Design 138 10. Conclusion 139 11. Clinical Application 140 12. Future Research 141 REFERENCES 143 APPENDICES 157 v List of Tables Table 1: • DSM-IV Diagnostic Criteria for Eating Disorders 5 Table 2: Correlations Between the Mean and Relative Macronutrient Intakes Derived from Food Frequency Questionnaires (FFQ) and Three-Day Food Records in a Sample of Female College Students 63 Table 3: Number of Dieting Female College Students Recruited From Participating Campuses 69 Table 4: Characteristics of Study Participants 70 Table 5: Physical Characteristics of Study Participants 70 Table 6: Probable Eating Disorder Diagnosis of Study Participants Based on Response to the Health Information Questionnaire 72 Table 7: Changes in Probable Eating Disorder Diagnosis from Baseline to 6 Months in Study Participants 73 Table 8: Symptom Intensity Scores and Compulsive Eating Score of Study Participants Based on Response to Health Information Questionnaire 74 Table 9: Mean Frequency of Binge Eating and Compensatory Behaviours in Study Participants 81 Table 10: Correlations Between Baseline Characteristics of Study Participants and Symptom Intensity Score 82 Table 11: Correlation between Baseline Characteristics of Study Participants and Use of Binge Eating and Compensatory Behaviours 83 Table 12: Self-Reported Duration of Dieting in Study Participants 85 Table 13: Correlation between Duration of Dieting Variables and Symptom Intensity Score in Study Participants 85 Table 14: Correlation between Baseline Duration of Dieting Variables and Use of Binge Eating and Compensatory Behaviours in Study Participants 86 Table 15: Multiple Regression Analysis of Baseline Duration of Dieting Variables on Change in Symptom Intensity Score in Study Participants 88 Table 16: Self-Reported Use of Dieting Methods by Study Participants 90 Table 17: 'Other' Dieting Methods Reported by Study Participants 91 vi Table 18: Correlation between Baseline Dieting Method Factors and Symptom Intensity Score and Change in Symptom Intensity Score in Study Participants 93 Table 19: Correlation between Use of Binge Eating and Compensatory Behaviours and Baseline Dieting Method Factors in Study Participants 94 Table 20: Multiple Regression Analysis of Dieting Method Factors on Change in Symptom Intensity Score in Study Participants 96 Table 21: Baseline Macronutrient Intake of Study Participants Determined by Food Frequency Questionnaire 98 Table 22: Correlation between Macronutrient Intake and Symptom Intensity Score of Study Participants 99 Table 23: Correlation between Use of Binge Eating and Compensatory Behaviours and Macronutrient Intake in Study Participants 100 Table 24: Multiple Regression Analysis of Macronutrient Intake on Change in Symptom Intensity Score in Study Participants 102 Table 25: Estimated Energy Requirement and Level of Energy Restriction in Study Participants 103 Table 26: Correlation between Restrictiveness Variables and Symptom Intensity Score in Study Participants 104 Table 27: Correlation between Use of Binge Eating and Compensatory Behaviours and Baseline Restrictiveness of the Diet and Rate of Weight Loss in Study Participants 106 Table 28: Multiple Regression Analysis of Restrictiveness of the Diet on Change in Symptom Intensity Score in Study Participants 107 Table 29: Comparison of Age and Baseline Weight, Height and B M I of Study Participants According to Dieting Status at 6 Months 109 Table 30: Comparison of Symptom Intensity Scores and Compulsive Eating Scores of Study Participants According to Dieting Status at 6 Months 110 Table 31: Comparison of Baseline Binge Eating and Compensatory Behaviours of Study Participants According to Dieting Status at 6 Months 112 Table 32: Comparison of Baseline Duration of Dieting of Study Participants According to Dieting Status at 6 Months 113 vii Table 33: Comparison of Baseline Dieting Methods of Study Participants According to Dieting Status at 6 Months 114 Table 34: Comparison of Baseline Macronutrient Intake of Study Participants According to Dieting Status at 6 Months 115 Table 35: Comparison of Restrictiveness Variables of Study Participants According to Dieting Status at 6 Months 116 viii List of Figures Figure 1: Self-Reported Use of Binge Eating by Study Participants 76 Figure 2: Self-Reported Use of Fasting to Lose Weight by Study Participants 76 Figure 3: Self-Reported Use of Diet Pills to Lose Weight by Study Participants 77 Figure 4: Self-Reported Use of Diuretics by Study Participants 77 Figure 5: Self-Reported 'Loss of Control Over Eating' by Study Participants 78 Figure 6: Self-Reported Use of Crash Dieting to Lose Weight by Study Participants 78 Figure 7: Self-Reported Use of Vomiting to Lose Weight by Study Participants 79 Figure 8: Self-Reported Use of Laxatives to Lose Weight by Study Participants 79 ix List of Appendices Appendix A: Certificate of Ethical Approval 158 Appendix B : Recruitment Poster 159 Appendix C: Consent Forms 160 Appendix D: Questionnaires Administered to Subjects at Baseline 164 Appendix D l : The General Information Sheet ..164 Appendix D2: The Dieting Practices Survey 165 Appendix D3: Food Frequency Questionnaire 169 Appendix D4: The Health Information Questionnaire 181 Appendix D5: The Health Information Questionnaire (Coding Key) 184 Appendix E: Questionnaires Administered to Subjects at 6 Months 185 Appendix E l : The Dieting Practices Survey (6 Month Version) 186 Appendix F: Letter to Subjects with 'Not In Service' Telephone Numbers at 6 Months 190 Appendix G: Follow Up Letter to Subjects Receiving a Probable Eating Disorder Diagnosis 191 Appendix H: Summary Letter to Subjects 192 Appendix I: Letter from Dr. Elizabeth Bright-See Granting Permission to Use the Food Frequency Questionnaire 194 Appendix J: Serving Size Key used in Coding the Food Frequency Questionnaire 195 Appendix K: Pretest Consent Form 199 Appendix L: Pretest Questionnaires 200 Appendix M : Self-Reported Use of Dieting Methods by Study Participants at Baseline 208 Appendix N : Self-Reported Use of Dieting Methods by Study Participants at 6 Months 209 x Appendix O: Self-reported Rate of Weight Loss (kg lost/week dieting) According to Dieting Status at 6 Months in Study Participants 210 xi Acknowledgments I consider myself fortunate to have worked with many special individuals during the course of this project. I would like to extent my appreciation to my supervisor, Linda McCargar, for sharing her research experience, not to mention her patience and optimism throughout every stage of this project. A special thank you to Gwen Chapman and Elliot Goldner for their invaluable input and guidance. I am also grateful to Susan Barr and Janet Le Patourel for reviewing my thesis. Sincere thanks to all the women who volunteered their time to participate in this study. Sharon Fagerlund from Simon Fraser University, Sara Dixon from Capilano College and many individuals at the participating colleges and universities were instrumental in assisting with the recruitment process. I would also like to extent my thanks to Sarah Cockell and Josie Geller for sharing their insight and statistical wonders. Thanks also goes to fellow graduate students for their enthusiasm and much needed survival skills. I gratefully acknowledge financial support from the British Columbia Medical Services Foundation (Vancouver Foundation) for funding this project and a summer studentship. Finally, a special thank you to my families for their never-ending support and to my best friend and husband, Stuart Cole, for his amazing rejuvenating effect, sense of humor, and constant supply of encouragement. xii Introduction I. Introduction 1. Rationale Dieting has long been considered a putative risk factor in the development of eating disorders (Kendler et al., 1991; Krenz et al., 1993; Patton et al., 1990; Story et al., 1991; Striegel-Moore et al., 1986). However, not everyone who diets develops an eating disorder. Specific characteristics of the diet may contribute to the development of disordered eating in women who are already dieting. Previous research has suggested that key dieting variables such as duration of dieting, dieting methods, macronutrient intake and restrictiveness of the diet may increase the risk of developing abnormal eating behaviours (Drewnowski et al., 1994; Striegel-Moore et al., 1986; Johnson et al., 1984; French et al., 1995; Polivy & Herman, 1985; Tuschl et al., 1990b; Rock & Yager, 1987). If these factors are responsible for promoting the development of eating disorders, individuals reporting a longer dieting duration, use of unhealthy dieting methods, an unbalanced macronutrient intake, and a highly restrictive diet, will experience an increase in disordered eating symptoms over time. The extent to which these dieting characteristics predict the development of eating disorders has not been examined prospectively. To investigate this question, the present study examined specific dieting characteristics in a sample of dieting women. A non-clinical sample of university and college women was chosen for several reasons. By examining a non-clinical population, it may be possible to isolate suspected risk factors before they become obscured by the physical and psychological changes associated with eating disorders (Patton, 1988). In addition, post-secondary female students have been considered to be at greater risk for eating disorders due to the stress commonly associated with academia and because they are at an age when eating disorders are frequently seen (Dickstein, 1989; Shisslack et al., 1987). This study, therefore, has the capacity to identify diet-related factors that may directly affect the development of disordered eating. I Introduction Defining the risk factors for eating disorders has obvious benefits. Early recognition of eating disorders has been associated with a more favorable outcome and fewer medical consequences and thus would be cost-effective in both human and economic terms (Hsu, 1990; Tonkin, 1994). Significant findings from this study may be useful in identifying individuals at risk for developing more serious eating behaviours. Sensitivity to the dieting behaviours, duration, and types of diet used may prevent the entrenchment of more severe eating and weight reducing behaviours. In addition, knowledge of the risk factors may assist in the development of preventative educational strategies and increase the potential to reduce the prevalence of eating disorders. 2. Purpose of the Study The purpose of this study was to describe the dieting practices of a sample of female college and university students who were dieting to lose or maintain body weight. A second purpose was to investigate the relative ability of specific dieting practices (the duration of dieting, dieting method, macronutrient intake, and restrictiveness of the diet) measured at the beginning of the academic year, to predict eating disorder symptomology, measured by the Health Information Questionnaire, at the end of the school year, six months later. 3. Objectives/Specific Aims 1) To measure, by questionnaire, the long- and short-term duration of dieting of subjects to determine whether the duration of dieting predicts the change in eating disorder symptomology six months later. 2 Introduction 2) To determine, by questionnaire, the dieting methods used to lose or maintain weight and to determine whether specific dieting methods predict the change in eating disorder symptomology six months later. 3) To determine, by food frequency questionnaire, the absolute and relative macronutrient intake of the diet and to determine whether absolute and relative macronutrient intake predict the change in eating disorder symptomology six months later. 4) To determine, by questionnaire, the restrictiveness of the diet (the energy intake, the extent to which the energy intake deviates from the energy requirement, and the rate of weight loss) and to determine whether the restrictiveness of the diet predicts the change in eating disorder symptomology six months later. 4. Hypotheses The null hypotheses tested in this sample were as follows: 1) The long- and short-term duration of dieting does not predict the change in eating disorder symptomology. 2) The dieting method used for the purpose of weight loss or maintenance does not predict the change in eating disorder symptomology. 3) The macronutrient intake, with respect to absolute and relative protein, carbohydrate and fat content, does not predict the change in eating disorder symptomology. 4) The level of restrictiveness of the diet does not predict the change in eating disorder symptomology. 3 Literature Review II. Literature Review The purpose of this chapter is to review the existing literature on the relationship between specific dieting practices and the development of eating disorders. First, eating disorders, including the prevalence and etiology of the disorders, will be reviewed. Second, the 'issue' of dieting will be considered, which includes the definition, prevalence, and causes of dieting. Third, the relationship between dieting and the development of eating disorders is discussed. Finally, specific dieting practices (duration of dieting, dieting methods, macronutrient intake and restrictiveness of the diet) are discussed with regard to the role that these dieting factors may play in the development of eating disorder symptoms. 1. Eating Disorders A. Introduction The term eating disorder encompasses a variety of clinical forms. The fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) recognizes three separate categories of eating disorders: anorexia nervosa (restricting type, binge eating/purging type); bulimia nervosa (purging type, non-purging type); and Eating Disorders Not Otherwise Specified, which includes binge eating disorder (American Psychiatric Association, 1994). Refer to Table 1 for the DSM-IV Diagnostic Criteria for Eating Disorders. Anorexia nervosa is characterized by a refusal to maintain a minimal healthy body weight, amenorrhea, and a morbid fear of gaining weight or becoming fat (even though underweight). In addition, one or more of the following criteria is also included: disturbance in the way in which one's body weight or shape is experienced, undue influence of body shape and weight on self-evaluation, and/or denial of the seriousness of the current low body weight. Characteristic features of bulimia nervosa include recurrent episodes of binge eating, a sense of lack of control over eating during the binge episode, recurrent compensatory behaviour after binge eating to prevent weight gain (vomiting, 4 Literature Review Table 1: DSM-IV Diagnostic Criteria for Eating Disorders* Anorexia Nervosa A. Refusal to maintain body weight at or above a minimally normal weight for age and height (e.g. weight loss leading to maintenance of body weight < 85% of that expected; or failure to make expected weight gain during period of growth, leading to body weight < 85% of that expected). B. Intense fear of gaining weight or becoming fat, even though underweight. C. Disturbance in the way in which one's body weight or shape is experienced, undue influence of body shape and weight on self-evaluation, or denial of the seriousness of the current low body weight. D. In post-menarchal females, amenorrhea, that is, the absence of at least three consecutive menstrual cycles. (A women is considered to have amenorrhea if her periods occur only following hormone (estrogen) administration.) Restricting Type: During the episode of anorexia nervosa, the person does not engage in binge eating or purging behaviour. Binge eating / purging Type: During the episode of anorexia nervosa, the person regularly engages in binge eating or purging behaviour. •Diagnostic Criteria For Eating Disorders. American Psychiatric Association: Diagnostic and Statistical Manual, Fourth Edit ion. Washington, D .C . , American Psychiatric Association, 1994 5 Literature Review Table 1 (continued) Bulimia Nervosa A . Recurrent episodes of binge eating. An episode of binge eating is characterized by both the following: 1. eating, in a discrete period of time (within any 2 hour period), an amount of food that is definitely larger than most people would eat during a similar period of time and under similar circumstances; and, 2. a sense of lack of control over eating during the episode (a feeling that one cannot stop eating or control what or how much one is eating. B. Recurrent inappropriate compensatory behaviour in order to prevent weight gain, such as: self-induced vomiting; misuse of laxatives, diuretics, enemas, or other medications; fasting; or excessive exercise. C. The binge eating and inappropriate compensatory behaviours both occur, on average, at least twice a week for three months. D. Self-evaluation is unduly influenced of body shape and weight. E . The disturbance does not occur exclusively during episodes of anorexia nervosa. Purging Type: Regularly engages in self-induced vomiting or the misuse of laxatives, diuretics or enemas. Non-purging Type: Uses other inappropriate compensatory behaviours, such as fasting or excessive exercise, but does not regularly engage in self-induced vomiting or the misuse of laxatives, diuretics or enemas. *Diagnostic Criteria For Eating Disorders. American Psychiatric Association: Diagnostic and Statistical Manual , Fourth Edit ion. Washington, D C , American Psychiatric Association, 1994 6 Literature Review Table 1 (continued) Eating Disorders, Not Otherwise Specified This category is for disorders that do not meet the criteria for any specific eating disorder. Examples include: 1. A l l of the criteria for anorexia nervosa are met except the individual has regular menses. 2. A l l of the criteria for anorexia nervosa are met except, despite significant weight loss, the individual's current weight is in the normal range. 3. A l l of the criteria for bulimia nervosa are met except binges occur at a frequency of less than twice a week or a duration of less than three months. 4. An individual of normal body weight regularly engages in inappropriate compensatory behaviour after eating small amounts of food. 5. An individual who repeatedly chews and spits out, but does not swallow, large amounts of food. 6. Binge Eating Disorder: Recurrent episodes of binge eating in the absence of the regular use of inappropriate compensatory behaviours characteristic of bulimia nervosa. •Diagnostic Criteria For Eating Disorders. American Psychiatric Association: Diagnostic and Statistical Manual, Fourth Edit ion. Washington, D . C . , American Psychiatric Association, 1994 7 Literature Review fasting, enemas, misuse of laxatives or diuretics, or excessive exercise), and over concern about weight and shape. The term Eating Disorders Not Otherwise Specified applies to those individuals who do not meet the strict criteria for anorexia nervosa or bulimia nervosa but have a clinically significant preoccupation with weight control and may engage in compensatory eating behaviours. Clinical cases of eating disorders represent only a small proportion of individuals who demonstrate a major preoccupation with weight and food intake and engage in unhealthy weight control practices. Partial syndrome eating disorders are more prevalent than full syndrome eating disorders, occurring at a ratio of 51 (Dancyger & Garfinkel, 1995). These eating disorders are considered to be not only a suspected risk factor for the development of a full clinical eating disorder, but an independent health concern (Dancyger & Garfinkel, 1995; Wilson & Eldredge, 1992). Individuals suffering from subclinical eating disorders may experience serious eating problems, psychological distress, and body image concerns (Shisslak et al., 1995). Ninety to 95% of eating disorder cases occur in women (Adams & Shafer, 1988; American Psychiatric Association, 1993; Whitaker, 1989). The mean age of onset for anorexia nervosa and bulimia nervosa is 17 years and 18 years, respectively (Garfinkel et al., 1995; Hsu, 1990), although the disorder may begin before age 10 or after age 40 (Hsu, 1990). Halmi et al. (1979) has suggested that these disorders may follow a bimodal pattern of onset with peak incidences at ages 14 and 18. It was once thought that eating disorders affected only Caucasian females from upper to middle social classes in developed countries. Eating disorders, however, now appear to be permeating different cultures and all socio-economic classes. Eating disorders result in a high incidence of physical complications. Dehydration and electrolyte imbalances may result from frequent purging, overuse of laxatives, enemas, diuretics, or emetics. In individuals with insulin-dependent diabetes, reduction or elimination of insulin 8 Literature Review may have the same result. Prolonged semi-starvation and weight loss can result in vitamin andmineral deficiencies, amenorrhea, organ dysfunction and osteopenia (American Psychiatric Association, 1993; Birmingham, 1989; Goldner & Birmingham, 1994; Tonkin, 1994). Eating disorders during adolescence can impede physical development. The mortality rate due to eating disorders has been estimated at approximately 3.3% and 3.1% for anorexia nervosa and bulimia nervosa, respectively, but reports vary between 0-22%, depending on the length of follow-up (Patton, 1988). The majority of deaths have been attributed to suicide or to cardiovascular complications, likely secondary to starvation (Krahn, 1991; Patton, 1988). A variety of emotional and psychological sequelae also result from eating disorders. These disorders may lead to the erosion of self-esteem, irritability, poor concentration, anxiety, apathy, emotional lability, fatigue, and social isolation (Garner & Garfinkel, 1980b). A high prevalence of anxiety and affective disorders have been reported in individuals with eating disorders (Goldbloom & Garfinkel, 1993; Garfinkel et al., 1995). Comorbid conditions may be present including depression and obsessive-compulsive behaviours. In addition, there is a strong relationship between food deprivation, eating disorders and alcohol and chemical dependence (Krahn, 1991). B. Prevalence of Eating Disorders The prevalence of eating disorders varies depending on the criteria used and the population studied. Recent studies have reported a decline in prevalence rates of anorexia nervosa and bulimia nervosa, possibly due to more rigorous classification criteria and improved methods of case detection. Lifetime prevalence of bulimia nervosa (proportion of individuals who receive the diagnosis at any time in their life) in Canadian women has been estimated at 1.1% (Garfinkel et al., 1995). In a review of interview-based epidemiological studies, Fairburn and Beglin (1990) estimated the point prevalence rate of bulimia nervosa among adolescent and young adult women at 0.9%. Point prevalence rates of anorexia nervosa has been estimated between 0.2% to 0.8% in 9 Literature Review adolescent and college women (Eisler & Szmulker, 1985; Patton, 1988b; Whitehouse & Button, 1988). Rates of eating disorders in certain subgroups, such as athletes, models and dancers, have been reported to be much higher than in the general population (Striegel-Moore et al., 1986). Subclinical forms of both eating disorders have been reported to occur in approximately 5% of all women (Button & Whitehouse, 1981; Goldbloom & Garfinkel, 1993). C. Etiology Anorexia nervosa and bulimia nervosa have been described as multidetermined disorders of eating (Garfinkel & Garner, 1992). However, there appears to be a consensus on some of the principal predisposing factors responsible for the development and/or maintenance of these disorders. In general, these factors can be categorized as individual (encompassing biological, cognitive, emotional, and personality variables), sociocultural, and familial factors (Garner & Garfinkel, 1980a; Johnson-Sabine et al., 1988). In a review of etiological factors, French et al. (1995) proposed that, in addition to female gender, specific risk factors for eating disorders included body dissatisfaction, low self-esteem, a high need for social approval, depression and a history of substance, physical or sexual abuse. Dieting is frequently described as a risk factor in the development of eating disorders (Polivy & Herman, 1985; Hsu, 1990; Patton et al, 1990). Some authors have identified dieting as one of the intermediates between the psychopathology of the disease and the manifestation of the eating disordered behaviours (Garfinkel et al., 1992; Garner & Garfinkel, 1980a; Cooper, 1995). Little is known, however, about the factors that combine with dieting to produce an eating disorder (Cooper, 1995). Thus, the role of dieting in the development of eating disorders in women will be the focus of the remainder of this literature review. 2. Dieting A. Definition of Dieting 10 Literature Review Dieting is the most widely used method to lose weight (Bennett, 1991; National Institutes of Health, 1992). There is, however, no consistent definition of dieting in the literature. The concept of dieting, or being on a diet, is affected by subjectivity and by cultural norms. Often the term dieting subsumes any behaviour that results in weight loss. As such, dieting may be characterized by a continuum of eating behaviours that range from making healthy changes in lifestyle, physical activity and food selection at one end of the continuum, to severe caloric restriction and fasting, use of purging techniques or diet pills at the other end (Brownell & Rodin, 1994a; French et al., 1995). Generally, dieting means a reduction in caloric intake for the purpose of weight loss. Even within this context, approaches differ. Dieters vary in the pattern of their daily eating behaviour and in the content of their intake. Some dieters eat the same foods they normally do but consume less quantity overall. Other people diet by prohibiting or reducing their intake of certain foods, food groups, or 'classes' of food (carbohydrate, fat or alcohol), or by changing the proportions of fat, protein and carbohydrate consumed. Still others completely alter their eating patterns and eat only restricted types of foods, in specific combinations, order, or time periods. Any of these dieting 'techniques' may be done casually or very stringently (Polivy & Herman, 1983). As diets vary widely in food content, in the length of time they are maintained and in their restrictiveness, it is important to determine the dieting practices that may be harmful or benign (Brownell & Rodin, 1994a; French & Jeffrey, 1994; Polivy & Herman, 1983). However, due to the variation in the definition of dieting, comparison of results across studies is limited (Moore, 1993). In addition, methodological issues must be considered. In studies based on self-reported data, it may not be clear whether the respondents understood the term dieting in the same way. Memory errors will also affect the validity of self-report depending on an individual's dieting history. That is, frequent dieters may respond more accurately than inexperienced dieters (French & Jeffrey, 1994). Unusual or abnormal dieting methods may be under-reported as such behaviours may be considered socially undesirable. 11 Literature Review There appears to be a consensus on some of the critical elements in the definition of dieting. The purpose of dieting is to lose weight or maintain current weight (Brownell & Rodin, 1994a; Levy & Heaton, 1993; Serdula et al., 1993; Williamson et al., 1992). The restriction of food intake or caloric intake is generally deliberate or voluntary. Therefore, it appears that dieting could logically be defined as: a conscious restriction of food intake for the purpose of losing weight or maintaining current body weight (Brownell & Rodin, 1994a; Johnson et al., 1983; Johnson et al., 1984; Levy & Heaton, 1993; Moore, 1993). Such a definition encompasses many dieting strategies and categories of dieters (for example, weight cyclers, chronic dieters, restrained eaters). C . Prevalence of Dieting in Women Thirty-nine percent of the women surveyed in Canada's 1990 Health Promotion Survey reported that they were currently trying to lose weight (Health and Welfare Canada, 1993). In the United States, 52% of surveyed women felt they were overweight, 40% were trying to lose weight, and 23% were trying to maintain their weight, primarily by dieting. (Horm & Anderson, 1993; National Institutes of Health, 1993; Serdula et al., 1993; St Jeor, 1993). Prevalence rates of dieting range from 31% to 66% in the adolescent female population and from 50% to 60% in college age women (Frank et al., 1991; Nevo, 1985; Patton et al., 1990; Rosen et al., 1987). In 1993, the British Columbia Adolescent Health Survey reported that 68% of the females in grade seven to twelve indicated they were actively trying to lose weight or maintain their current weight (Tonkin et al., 1993). Prevalence rates of dieting vary due to differences in the specificity of the survey question; broad questions regarding dieting status (for example, "Are you trying to lose weight?") will generate higher prevalence rates than more specific questions (such as, "Are you currently dieting to lose weight?") (French & Jeffrey, 1994). It is important to distinguish between actual attempts to lose weight and the desire to lose weight. 12 Literature Review Many women trying to lose weight are normal weight or are underweight (National Institutes of Health, 1993). The 1990 Canada's Health Promotion Survey found that 37% of surveyed women with an acceptable weight (BMI between 20 and 25) and 8% of women who were underweight (BMI < 20) were trying to lose weight (Health and Welfare Canada, 1993). In another survey, half of the women trying to lose weight had a B M I below 26 (Levy & Heaton, 1993). Stephenson et al. (1987) found that 19% of 'lean' (BMI < 20.5) and 54% of normal weight women (BMI between 20.5 and 27.2) between the ages of 18-29 years were trying to lose weight. Serdula et al. (1993) reported that 20% of adult women and 25% of female high school students who considered themselves to be the 'right weight' were trying to lose weight. Larger estimates were reported by Frank et al. (1991). Of those college age women who were underweight (as defined by the 1979 Metropolitan Life tables), 25% were on a diet to lose weight and of those whose weight was normal, 49% were on a diet to lose weight. Horm and Anderson (1993) found that between 1985 and 1990, the percentage of those normal weight men and women who considered themselves overweight increased from 0.7% to 4%. These studies suggest that a weight considered acceptable or 'normal' by health standards, may not necessarily be considered acceptable by the current sociocultural standard of beauty (Health and Welfare Canada, 1993). D. Causes of Dieting There appear to be many different reasons for dieting. Women often try to lose additional weight gained during pregnancy or that gained after cessation of smoking. Others may seek to improve their self-image through weight loss or they are motivated to lose weight because of the health concerns associated with obesity. One of the most widely held beliefs, however, is that sociocultural influences promote dieting (Davis et al., 1993; Polivy & Herman, 1983; Rodin et al., 1984). Since the 1960's, the feminine ideal of beauty appears to have shifted to a thinner, more tubular shape, and a body that is increasingly physically fit (Brownell et al., 1992; Garner et al., 1980). 13 Literature Review Thinness in women is associated with greater femininity, virtue, and success (Shisslak et al., 1987). This 'ideal' is promoted through the media and by the diet, cosmetic, fashion and fitness industries (Gilbert, 1986). Consequently, a women's perception of her attractiveness is strongly determined by her weight and body shape (Rodin et al., 1984). The increased availability of diet foods, and a wide range of no-fat, low-fat, and reduced-fat foods, conveys an acceptance of dieting as part of a normal lifestyle (Adams & Shafer, 1988). Indeed, this widespread preoccupation with weight and dieting appears to be encouraged and rewarded in Western society (Omichinski, 1995). This may account for why a substantial proportion of women who are dieting are of normal weight or are even underweight (Davis et al., 1993). In addition, society stigmatizes overweight individuals as 'bad', 'lazy', and lacking 'willpower' (Health and Welfare Canada, 1988a). Obesity is perceived to be associated with failure and a collapse of self-discipline, particularly in women (Rodin et al., 1984). The discrimination against obese individuals may affect not only their education and vocational success but their interpersonal relationships. Thus, the current cultural preoccupation with thinness in women and the strong social sanctions against obesity may promote the weight and body image dissatisfaction that lead to dieting (Polivy & Thomsen, 1988). Body dissatisfaction in women appears to take two forms: dissatisfaction with weight and with body shape. Weight indices (such as weight and BMI) have emerged from cross-sectional and prospective epidemiological studies as the variable most consistently associated with dieting (Emmons, 1994; Hsu, 1990; Killen et al., 1994; Patton et al., 1990; Story et al., 1991). However, weight dissatisfaction and dieting usually accompany subjective impressions of feeling fat even in the absence of realistic evidence for such feelings (Grodner, 1992; Striegel-Moore et al 1986). Dieting is also a frequent behaviour among women who report a high degree of body shape dissatisfaction (Grodner, 1992; Striegel-Moore et al 1986). Body dissatisfaction appears to be considerably site-specific. Young girls may consider the increased fat deposition that comes with 14 Literature Review normal pubertal development as unwanted weight gain and they may begin dieting (Adams & Shafer, 1988). As adolescent girls become proportionately heavier for their height, they become less satisfied with their waist and hip measurements. In fact, a negative correlation between the desire to lose weight and satisfaction with the hip measurement has been reported (Davis & Furnham, 1986). Similarly, body dissatisfaction in college age women is widespread and specific to the hips and buttocks (Bailey et al., 1990). Dieting behaviour in college age women has been shown to be influenced more by the degree to which their skeletal structure deviates from the 'thin-ankled, narrow-hipped standard of female beauty' than by their degree of adiposity (Davis et al., 1993). This lower hip girth dissatisfaction appears independent of total body fatness (Bailey et al., 1990). Site-specific body dissatisfaction suggests that women who lose weight may still be unhappy with their body as this dissatisfaction is focused, in part, on shape rather than size or weight. 3. The Role of Dieting in the Development of Eating Disorders Dieting has long been considered a precipitating factor in the development of eating disorders (Kendler et al., 1991; Krenz et al., 1993; Patton et al., 1990; Story et al., 1991; Striegel-Moore et al., 1986). Dieting is associated with and may predispose some individuals to binge eating and other symptoms of eating disorders either through cognitive or physiological mechanisms (Brownell & Rodin, 1994b; Patton, 1992; Wilson & Eldridge 1992). The onset of abnormal eating and eating disorder symptoms has frequently been reported to follow a period of restrictive dieting (American Dietetic Association Reports, 1994; Huon et al., 1987; Polivy & Herman, 1985). It has been argued that dieting and eating disorders lie on a behaviourial continuum (Nylander, 1971; Rodin et al., 1984). Due to the co-prevalence of both anorexic and bulimic behaviours in many patients with clinical eating disorders, the disorders themselves may be considered to exist on a continuum (American Psychiatric Association, 1993). Szmulker et al. (1985) supported the theory that dieting and eating disorders lie on a continuum by suggesting that dieting behaviour, fear of fatness, weight loss, subclinical and clinical cases of eating 15 Literature Review disorders occur on a severity spectrum. Both Szmulker et al. (1985) and Hsu (1990) contend that the prevalence of eating disorders is higher in populations where dieting is common. It is evident that not all dieters develop eating disorders. Some authors have suggested that dieting behaviour, in the presence of other psychiatric and personality factors, may lead directly to the onset of an eating disorder (Johnson-Sabine et al., 1988; Patton et al., 1990). Hsu (1990) argues that dieting is the catalyst for an eating disorder if 'intensified' by certain adolescent 'issues' and accompanied by certain risk factors. It is speculated that the pathogenesis of eating disorders is mediated by dieting to control or lose weight and that the factors that intensify or perpetuate this dieting indirectly increase the risk of developing an eating disorder (Garfinkel et al., 1992; Hsu, 1990). Specific precipitating factors, however, have yet to be established. A. Clinical Studies Support for the hypothesis that dieting is causally related to the development of eating disorders comes from three lines of evidence: retrospective studies of clinical populations; cross-sectional studies; and prospective studies. i . Clinical Populations Studies of predictors of eating disorders typically focus on retrospective study of those individuals with eating disorders. It has been reported that most eating disordered patients diet (Herman & Polivy, 1988; Polivy & Herman, 1985). In addition, dieting appears to precede disordered eating rather than following it (Herman & Polivy, 1988; Polivy & Herman, 1985; Huon et al., 1987). In cases where it was possible to reconstruct the temporal sequences of symptom development, dieting preceded binge eating by many months (Wilson & Eldridge, 1992). Retrospective study of clinical populations, however, may be biased by the 'psychopathology' of the illness; "recall of possible etiological factors may be altered both by time and the secondary 16 Literature Review psychological and physiological changes that accompany the eating disorder" (Patton et al., 1990). To eliminate this effect, non-clinical populations must also be examined prospectively to characterize the risk factors that predict the development of eating disorders. i i . Cross Sectional Studies Epidemiological evidence suggests that the population groups most concerned with dieting and weight are also the ones with the highest incidence of eating disorders. In professions or careers that emphasize slimness or external appearance, individuals may be more vulnerable to eating disorder onset. For example, the prevalence of dieting and eating disorders is higher among dancers, models, gymnasts, actresses and athletes, than in the general population (Beals & Manore, 1994; Gordon, 1989; Rodin et al., 1984; Shisslak et al., 1987). According to Garner and Garfinkel (1980), most of these individuals developed eating disorders after joining the professions and experiencing the pressure to be thin. Stress and a high level of pressure to achieve may also predispose some individuals to the development of eating disorders (Rosen et al., 1993; Shisslak et al., 1987). This occurs not only in competitive sports, but in the competitive environment of college or university. Students in stressful academic careers, such as medical school, may be particularly vulnerable (Shisslak et al., 1987). The current university environment may, through direct or indirect means, facilitate the development of eating disorders in susceptible female students (Dickstein, 1989). The pressures commonly associated with commencement of studies include financial concerns, academic demands, new relationships or the stress of transition and adjustment to college life (Megel et al., 1994; Shisslak et al., 1987). The university or college environment, therefore, may impose unique academic pressures and psychosocial stressors that may not have been previously encountered. i i i . Prospective Studies 17 Literature Review Perhaps the strongest evidence for the role of dieting in the development of eating disorders are the prospective studies of certain population sub-groups. Studies of subjects that are forced to diet are rare. However, one of the earliest prospective studies examining the physiological and psychological changes induced during semi-starvation and subsequent refeeding was performed by Keys et al. (1950). In this study, 36 male volunteers were subjected to a 24 week 'semi-starvation' period (1600 kcal/d) in which their body weight was reduced to approximately 74% of their normal weight. Subjects were later refed and they returned to their prestudy weight. During the 'semi-starvation' period, intense preoccupation with food and eating emerged. Some subjects developed bizarre eating habits, such as food hoarding, abnormal taste preferences, binge eating and other appetite disturbances. Subjects also experienced symptoms of depression, obsessionality, apathy, irritability and other personality changes. Even months after rehabilitation, 10 out of the 15 men that were seen for follow-up had put themselves on a weight reducing diet. This experiment suggested that prolonged semi-starvation resulted in symptoms similar to those seen in individuals with eating disorders. In a more recent study, Patton et al. (1990) examined the course of eating disorders and dieting in school girls in order to determine the relative importance of suspected risk factors in the development of eating disorders. A two-stage survey was used. Female high school students (n=1030), aged 15 years, at eight London schools were administered self-report questionnaires. Three groups of students (n=230) were later selected for a blinded interview: 1) an 'at risk' group which included subjects with abnormal eating attitudes; 2) a second 'at risk' group including those students with high "psychiatric morbidity" scores; and 3) a 'low risk' or control group. A l l three groups were interviewed to assess clinical status and putative risk factors for eating disorders. The subjects were diagnosed either as cases (subjects with anorexia nervosa, bulimia nervosa or the partial syndrome of anorexia nervosa), dieters or non-dieters. After one year, questionnaires were readministered to the high school students (n=735). One hundred and 18 Literature Review seventy six students in the interview group were reinterviewed to assess the development of eating disorders and the relative contribution of the risk factors to their development. The prevalence of dieting was lower at follow-up (initial: 34%; follow up: 29%) with considerable change in the composition of the dieting group. The majority of dieters (n=48, 79%) did not develop an eating disorder at follow up. However, 13 (21%) of the dieters at follow-up had developed diagnosable eating disorders compared to 3% of the non-dieters. This suggests that dieters had an eight-fold greater risk of developing an eating disorder than non-dieters. Most of the suspected putative risk factors for eating disorders were not found to be associated with the development of eating disorders but were associated with attempts to control weight alone. The authors suggested that while the great majority of dieters remained well or stopped dieting, dieting does predispose to later development of eating disorders and may, therefore, be an intermediate stage between suspected risk factors and the development of eating disorders. In another prospective study, Drewnowski et al. (1994) conducted a longitudinal survey to determine the distribution of dieting behaviours among female freshmen undergraduates. It was argued that since the severity and frequency of bulimic behaviours are distributed along a continuum, the DSM-UI-R criteria (an earlier version of the diagnostic criteria for eating disorders) for bulimia nervosa provided little information on the etiology and time course of eating disorders. The goal of the study was to develop a continuous scale of 'eating pathology' in order to assess the severity of weight control efforts and the persistence of symptoms of eating disorders. In the fall term, a questionnaire that included items that approximated the DSM-UI-R criteria for bulimia nervosa, was distributed to over 2000 female students. The data collected from the 902 respondents were used to construct a five-category Eating Pathology Scale. The sample was divided into five mutually exclusive categories: probable bulimic subjects; dieters at 19 Literature Review risk; intensive dieters; casual dieters; and non-dieters. In the spring term, six months later, the questionnaire was readministered and the sample was similarly divided into the five categories. Results were based on the subjects for whom both fall and spring data were available (n=557). The distribution of dieting behaviours was comparable between the two time points, shown as fall and spring, respectively: probable bulimic subjects, 3%, 3%; dieters at risk, 10%, 10%; intensive dieters, 31%, 25%; casual dieters, 42%, 41%; and non-dieters, 14%, 18%. Assignment to the more severe categories of the Eating Pathology Scale was associated with higher frequencies of dieting and a greater number of dieting strategies used during the past month. Movement between the different Eating Pathology Scale categories occurred primarily between adjacent categories. Forty-two percent of the women diagnosed with probable bulimia in the fall were still classified as having bulimia in the spring, while 58% of the women diagnosed with probable bulimia moved to the 'at risk' and the intensive dieter categories. No subject classified as a casual dieter or non-dieter in the fall developed bulimia by spring, although some subjects in these categories did move to the intensive dieter or the 'at-risk' category. However, 4% of the intensive dieters and 15% of the dieters at risk were classified as probable bulimic by spring. The results suggest that the 'severity' of dieting is strongly associated with the development of abnormal eating behaviours. B. Dieting: The Risk vs. The Benefits The struggle to achieve a weight and shape consistent with the societal ideal of beauty may lead to a variety of disordered eating patterns such as restrictive eating, dieting, excessive exercising, and binge eating and purging. These behaviours may mark the early stages of serious health problems such as anorexia nervosa or bulimia nervosa (Health and Welfare Canada, 1988). While only a minority of chronic dieters will develop the clinical syndrome of anorexia nervosa or bulimia nervosa, a larger group may experience negative health effects from chronic dieting. 20 Literature Review Health hazards associated with weight-loss in women include menstrual irregularities, weakness, persistent irritability, constipation, poor concentration, reduced resting metabolic rate, sleep difficulties and, in adolescence, growth retardation (French et al., 1995; Story et al., 1991; Rodin et a l , 1984). Evaluation of the risk of dieting in relation to development of an eating disorder is challenging. Severe dieting is a central feature of eating disorders (Garner & Garfinkel, 1985; Hsu, 1990; Wilson, 1993). However, eating disorders vary in severity and symptom profile. Also, eating behaviours of many adolescents and young adult females are highly variable by nature and not all dieters who use compensatory behaviours show clinical signs of eating disorders. Despite the aforementioned studies, support for the hypothesis that dieting leads to eating disorders is not unanimous (French & Jeffrey, 1994). Whether dieting precedes or follows an eating disturbance is unclear. In addition, suggesting that dieting lies on the continuum of eating disorders does not satisfactorily explain why only a minority of all dieters develop an eating disorder or why some dieters develop anorexia nervosa and others develop bulimia nervosa (American Dietetic Association Reports, 1994; Wilson, 1993). Dieting in some populations may be considered beneficial. In overweight and obese adults, the practice of dieting has been defended, suggesting that the dangers of dieting are far less than those of remaining obese (Garrow, 1991; McDonald, 1995). Obesity has been associated with elevated plasma lipids and lipoproteins, cholelithiasis, hypertension, diabetes, gout and osteoarthritis (Canadian Dietetic Association, 1988; Health and Welfare Canada, 1988). In addition, there is an association between excess weight and endometrial and breast cancer in postmenopausal women (Health and Welfare Canada, 1988). Dieting in overweight and obese individuals may avert serious medical problems and result in significant health benefits. 21 Literature Review Brownell and Rodin (1994a) have suggested that dieting in individuals who are not obese may be considered 'preventative dieting' which may reduce the weight gain associated with increasing age or to offset other tendencies to gain weight. They have argued that the prevalence of obesity may be even greater without the current prevalence of dieting. Some have suggested that chronic calorie restriction in the nonobese, both animals and humans, may have many positive health effects (Levin et al., 1995). These can be as far reaching as slowing of the aging process and reduced neoplasm development, to improved lipid profiles and blood pressure with no adverse effects on mental performance, feelings of hunger, satiety or mood. Dieting, however, is a practice not exclusive to the overweight or obese individual. Ironically, the pressure to be lean and the prevalence of dieting is greatest in young women. This is not the group in which the greatest risk of obesity occurs and it is not the group in which the benefits of dieting outweigh the risks. Arguments that dieting can be pathological in normal weight individuals, may have been used to discourage treatment in the overweight or obese (McDonald, 1995). It is important, therefore, to distinguish dieting in individuals who are at a normal weight from those whose weight may be considered a medical and/or psychological risk. 4. Dieting Practices as Risk Factors for Eating Disorders Available evidence suggests that dieters have a greater risk of developing an eating disorder than their non-dieting counterparts (Garner & Garfinkel, 1985; Hsu, 1990; Wilson, 1993). The mechanisms that influence the development of eating disorders by dieting, however, are not known. It is clear that not everyone who diets develops an eating disorder. Other predisposing influences must occur to precipitate an eating disorder. Factors that intensify dieting may indirectly increase the risk of developing an eating disorder (Hsu, 1990). In addition, key influences may operate to initiate or perpetuate dieting rather than the eating disorder itself (Garfinkel et al., 1992; Hsu, 1990). Additional prospective research is required to identify the 22 Literature Review factors that intensify dieting and promote the development of eating disorders. (French & Jeffrey, 1994). Previous research has identified several characteristics of the diet that may predispose a dieter to the development of abnormal eating behaviours and eating disorders. These include the duration of dieting, the methods used to lose weight, the macronutrient intake while dieting, and level of restrictiveness of the diet. Each of these factors are discussed separately below. A. Duration of Dieting The definition of dieting encompasses many dieting strategies (Wilson & Eldredge, 1992). Of these, chronic dieting and weight cycling have most frequently been cited as risk factors in the development of eating disorders (Johnson, 1984; Story et al., 1991). The criteria for chronic dieting vary from requiring a specified number of years dieting, a specified number of dieting attempts made over a particular time period, to the generalized concept of frequent intermittent dieting behaviour or 'always' dieting (Grodner, 1992; Levy & Heaton, 1993). The definition of weight cycling, likewise, includes frequent, intermittent dieting. Often a minimum number of years is specified in the definition. Weight cycling may also include several other key aspects such as a number of weight loss attempts in the past year and a requirement of notable weight loss and regain at least once per year in the past 5 years (McCargar et al., 1993; National Task Force on the Prevention and Treatment of Obesity, 1994). Both chronic dieting and weight cycling share a common feature, such as a longer duration of dieting behaviours. The length of time an individual is dieting appears to be one of the key elements that distinguish chronic types of dieting from other, more benign, forms of dieting. Several studies have examined the association between repeated or frequent dieting and abnormal eating behaviours. In Drewnowski's study of college women, a more serious eating pathology was associated with a higher reported frequency of dieting (Drewnowski et al., 1994). 23 Literature Review Forman et al. (1986) suggested that, depending upon the type and duration of the diet, dieting in the normal weight individual could result in adverse health consequences, including eating disorders. Westinhoeffer and Pudel (1993) found that, in both men and women, the greater the number of weight loss diets followed in the past, the higher the incidence of problem eating behaviours. Striegel-Moore et al. (1986) proposed that a prolonged history of repeated dieting attempts was a risk factor for the development of bulimia nervosa. Repeated weight gain and loss through weight cycling, may also have adverse psychological and physical effects (National Institutes of Health, 1992). In runners, a history of weight cycling was associated with higher levels of disturbed eating practices (Kiernan et al., 1992). And in a large multisite trial, obese binge eaters were more likely than nonbinge eaters to have a longer history of dieting and weight cycling (Spitzer et al., 1992). In a study of high school students, Story et al., (1991) showed chronic dieters were more likely than non-dieters to experience abnormal eating patterns and to report unhealthy weight loss methods, such as self-induced vomiting (relative risk (RR), 9.92), laxative use (RR, 7.18), ipecac use (a medication that stimulates vomiting) (RR, 8.33), and diuretic use (RR, 7.3). In addition, female chronic dieters were 6 to 9 times more likely to report purging behaviours (vomiting, laxatives, ipecac use) than nondieters. In another study of high school students, there was a significant relationship between duration of dieting and the reporting of bulimic behaviours: 33% of the bulimic group reported to be always dieting, compared to only 6% of the nonbulimic group; and 14% of the bulimic group reported dieting more than 10 times in the last year, compared to 3% of the nonbulimic group (Johnson et al., 1984). This suggests that those students with bulimic symptoms dieted more frequently and over a longer time period than the non-bulimic student group. Duration of dieting was also associated with a poor body image, fear of being unable to control eating and a more prevalent history of binge eating. It appears, therefore, that there exists a relationship between the duration of dieting and the symptoms associated with eating disorders. 24 Literature Review In the adolescent population, frequent dieting may be harmful not only due to the inherent physical risks associated with dieting in this population, but due to the association of frequent dieting with negative health behaviours. French et al. (1995) reported that in high school students an increased frequency of dieting was associated with an increased prevalence of suicide risk, sexual promiscuity, and physical or sexual abuse. French et al. also suggested that frequent dieters may share similarities with individuals with eating disorders such as preoccupation with body weight and shape, body dissatisfaction and perfectionistic tendencies. As such, frequent dieting and eating disorders may have common risk factors. Measurement of the duration of dieting in the evaluation of risk is challenging. Duration of dieting is often ill-defined or is alluded to in terms of frequency of dieting or number of dieting attempts. Such concepts are difficult to assess as duration of dieting is highly variable and number of dieting attempts does not take into consideration the actual time spent dieting. Few studies have specifically measured the length of time an individual is dieting or have examined the role of duration of dieting in relation to the development of abnormal eating behaviours. As such, our understanding of duration of dieting in relation to risk has not been advanced. Assessment of the risk of dieting in the development of eating disorders requires that the specific time frame and duration of dieting be addressed (French et al., 1995). To more specifically examine the role of dieting duration, it must be quantified by the number of days dieting over a set time period and number of years from the time that the dieting was initiated. Surveys of dieting practices in women have shown that on average, respondents reported that they had spent an average of 6 to 6.4 months over the previous 12 months trying to lose weight, and they had made an average of 2 to 2.5 attempts at weight loss over the same period (Levy & Heaton, 1993; National Institutes of Health, 1992). Another US survey found that, for women under the age of 50, the median duration of their current weight loss attempt was 4 weeks (95% C L = ± 0.2 weeks) (Williamson et al., 1992). The body weight of the dieter may impact on the frequency 25 Literature Review and duration of dieting. Individuals with higher B M I levels have been shown to engage in more frequent dieting behaviour of shorter duration than did those with lower B M I levels who tended to remain on a diet for longer periods (National Institutes of Health, 1992). None of these studies, however, attempted to relate the measured duration of dieting to the development of eating disorders. Few studies have considered the age when the dieting behaviour began. Nylander (1971) found that the frequency of dieting in high school students reached a maximum at age 18 and the frequency of dieting was positively correlated to age and body weight. Johnson et al. (1984) identified a group of high school students suffering from bulimia and found that this group began dieting at an earlier age. A significantly greater percentage of students with bulimic symptoms had started dieting by age 14 (69%) compared to the nonbulimic group of students (52%). In addition, Johnson et al. (1984) found that the group of students suffering from bulimia had dieted more frequently and over a longer time period than the non-bulimic student group. However, when the number of years since dieting first began was compared, the difference between the bulimic and non-bulimic groups was only marginal (t=1.67, p< 0.09). Mussell et al. (1995) found that the age of onset of significant dieting was 22 ± 10 years. (Significant dieting was defined by the age at which the subject first lost 10 pounds by dieting.) Frank et al. (1991) reported that the median age for beginning a restrictive diet was 16 years. Kuehnel and Wadden (1994) investigated the relationship between binge eating and dieting history in 70 obese women. Subjects were divided into categories of binge eaters, non-binge eaters, and problem eaters. Analysis of dieting history revealed a significant difference among groups in the age at which the subjects first dieted: Binge Eating Disorder, 18.5 + 7.0 years old; problem eaters, 20.3 ± 6.3 years old; nonbingers, 25.1 ± 9.7 years old. It has been suggested that frequent episodes of dieting may serve as a screening marker for more severe eating and weight loss behaviours (Story et al., 1991). From the previous research, it 26 Literature Review appears that greater long-term (years) and short-term (months or days) duration of dieting, and an earlier age of onset of dieting may increase an individual's susceptibility to the development of eating disorders. The measurement of the duration of dieting, however, is not well established. Few studies have looked specifically at quantitative measures of dieting duration, in terms of long- and short-term dieting, and its relationship to disordered eating. Identification of these variables in a population, (particularly a high risk group such as women currently dieting), and assessing the relationship with disordered eating, is an important research question. B. Dieting Methods It has been argued that specific methods of weight control are associated with the development of atypical eating behaviours and subsequent eating disorders (Krey et al., 1989; Kurtzman et al., 1989; Polivy & Herman, 1985). Currently, dieting is the most widely used method to lose weight and improve body shape (Bennett, 1991; McCargar et al., 1993; National Institutes of Health, 1992). However, not everyone who diets develops abnormal eating behaviours and not all approaches to dietary change are pathological. Therefore, grouping all approaches to weight loss and dietary change under one label may be misleading (Brownell and Rodin, 1994a). Weight loss methods such as lowering total fat intake, increasing intake of fruits and vegetables and increasing physical activity may have different outcomes than the more controversial approaches to weight control such as low-calorie (800 to 1500 kcal/d) and very-low-calorie (less than 800 kcal/d) diets or some advertised commercial programs. The challenge, therefore, is to be more specific in discussing methods of dietary change and to identify particular dieting methods that may be harmful. Despite the high prevalence of dieting in the Western culture, relatively few studies have examined the weight loss practices of adults and adolescents. The National Institutes of Health Technology Assessment Conference Panel (1992) reviewed four US national surveys. The most common dieting and weight loss methods include calorie restriction, exercise, behaviour 27 Literature Review modification, weight loss drugs or various combinations of these methods. To achieve weight loss, 80 to 84% of women were eating fewer calories and 60 to 80% were increasing physical activity. Other weight loss methods, cited in decreasing order from 28% to 3%, included use of vitamin supplements, meal replacements or over-the-counter [pharmaceutical] products, participation in a commercial weight loss program and diet supplements. Results from the 1992 Weight Loss Practices Survey were published by Levy and Heaton (1993). Of the 1030 adult women surveyed, dieting and exercise were the most prevalent weight loss practices used; 87% of the women dieted and 83% exercised. The use of vitamin or mineral supplements was the next most common weight-loss practice (33%), followed by the use of meal replacements (15%), weight loss pills (appetite suppressants, diuretics or thyroid pills) (14%), formal weight loss programs (including both commercially available programs and those sponsored by hospitals or health professionals) (13%), fasting (6%), diet supplements (fibre or protein supplements) (4%), and laxatives (3%). Behaviours relating specifically to the alteration of the diet were also examined. The most common practices included counting calories (25%), followed by skipping meals (21%), keeping a record of food intake (15%), using set food menus and eating more frequent meals (13%). In a survey of female college freshmen, the prevalence of unhealthy weight loss behaviours, such as diet pills, laxative and diuretic use and fasting and purging behaviours, were much less common, ranging from 16% for fasting to 1.4% for use of laxatives (Frank et al., 1991). The source of dieting information used by dieters has also been examined to a small extent. In a survey of US adults, 29% of women reported that healthy eating was discussed 'sometimes' or 'often' when visiting a doctor or other health professional for routine care (Stephenson et al., 1987). According to the Weight Loss Practices Survey, 40% of all the men and women surveyed had checked with their doctor before starting an attempt at weight loss (Levy & Heaton, 1993). Under one third of those surveyed reported seeking information from 'expert' sources, 28 Literature Review such as trained health professionals. The likelihood of consulting with a doctor or seeking information from an 'expert' increased with increasing BMI. Most of those surveyed received dieting and/or nutrition information from 'informal' sources, including family, friends, television, newspaper and magazines. Other authors have found similar results, suggesting that the majority of those people attempting to reduce their weight do so independently, using books and magazines as resources or diets provided by others, by joining exercise programs, or by enrolling in commercial or self-help programs (Brownell & Rodin, 1994a). Large scale surveys describing weight loss practices in adolescent populations are less common than adult surveys. Results from the 1990 Youth Risk Behaviour Survey (n=11500 students), found that female high school students reported using the following weight loss methods in the week preceding the survey: exercise (51%), skipping meals (49%), diet pills (4%), and self induced vomiting (3%) (Serdula et al., 1993). The percentage of female students, however, who reported ever using these methods was generally much higher: exercise (80%), diet pills (21%), and vomiting (14%). In the British Columbia Adolescent Health Survey, female high school students trying to lose or maintain their weight reported using the following weight loss methods in the week preceding the survey: exercise (46%), dieting (13%), dieting and exercising (30%), purging (4%), diet pills (2%), and purging and using diet pills (1%) (The McCreary Centre Society, 1993). Most weight control practices can be divided into two broad categories: dieting methods and weight loss behaviours. As previously discussed, dieting is defined as a conscious restriction in food intake for the purpose of weight loss or maintenance. Therefore, for the purpose of this discussion, a dieting method will be defined as: a specific modification made to food intake for the purpose of weight loss or weight maintenance. Other non-dieting weight loss methods, henceforth termed weight loss behaviours, include the use of physical activity, pharmacological aids (diuretics, diet pills, and laxatives) and purging for the purpose of controlling body weight. 29 Literature Review Dieting methods are frequently grouped together with weight loss behaviours such that differentiation of the risk that each contributes to the development of abnormal eating behaviours is unknown. There is evidence that some weight loss behaviours, such as purging, excessive exercise, and the use of weight loss drugs, are associated with the development of eating disorders (Frank et al., 1991; Kurtzman et al., 1989; Polivy & Herman, 1985). Other 'pathogenic' or unhealthy weight control techniques include vomiting and the use of laxatives (Story et al., 1991; Taub & Blinde, 1992). Such behaviours are often considered precursors of severe eating disorders. It has been suggested that these weight loss behaviours constitute a risk by being self-perpetuating (Story et al., 1991). While weight loss behaviours have been associated with abnormal eating, there is little data on the extent to which the dieting method, alone or in association with other weight-loss behaviours, contributes to the risk of developing abnormal eating behaviours. Fasting, crash dieting (eating only small amounts of food for a few days), taking diet supplements (such as high-protein or fibre supplements) and diet aids have also been considered by some to be unhealthy or 'pathogenic' dieting methods (Levy & Heaton, 1993; Story et al., 1991; Taub & Blinde, 1992). Skipping meals and a chaotic pattern of eating has been suggested to "weaken learned satiating contingencies" which enhances the tendency to binge (Tuschl, 1990a). A greater number of dieting methods may be associated with the development of eating disorders. In Drewnowski's longitudinal survey of dieting behaviours among female freshmen students, assignment to higher categories of the Eating Pathology Scale was linked to using a greater number of dieting strategies during the past month (Drewnowski et al., 1994). Grodner (1992) suggested that use of three or more 'fad diets' within a two-year time period may lead to the eventual development of abnormal eating behaviours. In addition, Frank et al. (1991) reported that women who were currently using purging behaviours usually employed multiple methods of weight control. 30 Literature Review The desire to lose weight may lead to various unsound or questionable dieting methods such as fasting, crash dieting or very-low-calorie diets. Definition of questionable practices is controversial. Dieting methods that do not permit an adequate or balanced nutrient and/or energy intake, that involve supplementation, that are available from a questionable source of information (such as some television programs and some magazines) or are unsupervised, have been labeled as questionable dieting methods (Brownell & Rodin, 1994a; Dwyer, 1980; Levy & Heaton, 1993). Such practices are classified as questionable based on the possibility of abuse (Levy & Heaton, 1993; Story et al., 1991; Taub & Blinde, 1992). These dieting methods may be associated with health problems and the development of disordered eating. However, some of these methods could be considered a reasonable element of a weight-loss plan if used moderately, infrequently and while under professional supervision. For example, it may be considered appropriate to use diet supplements to supplement a healthy diet, but it may be inappropriate to use diet supplements as a substitute for food to lose weight (Levy & Heaton, 1993). In addition, under controlled settings, diets, behaviour modification, exercise and weight-reducing drugs produce short-term, healthy weight loss with reasonable safety. However, adverse health effects may be seen in persons who are not overweight and who are using healthy weight-loss methods (National Institutes of Health, 1992). The role of specific dieting methods in the development of eating disorder symptoms has not been explored. Further research, therefore, is required to describe the dieting methods, both the type of method and the number of methods used by dieters and to determine the association of specific dieting methods with the development of abnormal eating behaviours. Such research has the capacity to identify particular dieting methods that may be harmful. C. Macronutrient Intake Current national nutrition guidelines focus primarily on the reduction of fat intake by Canadians to no more than 30% of energy intake as a means of preventing cardiovascular disease, some 31 Literature Review cancers and obesity (Health and Welfare Canada, 1990; Reid & Hendricks, 1994). Consequently, Canadians have become increasingly aware of the role of dietary fat in some diseases and have been altering their dietary practices. In 1989, a survey of Canadian adults revealed that fat intake was a concern to 71% of respondents (Beggs et al., 1993). More recently, the Grocery Products Manufacturers' survey of Grocery Attitudes of Canadians (1994) found that 77% of respondents reported that the fat content of food was one of the 'leading nutritional issues'. This was followed closely by concern over the number of calories (67%) and amount of sugar (57%) in food. In addition, the most common dietary approach to weight control among North American adults includes reducing fat and calorie intake (Bennett, 1991). In response to health and weight concerns, the food industry has increased the availability of diet foods, and a wide range of no-fat, reduced-fat and low-energy foods (Drewnowski, 1995; Omichinski, 1995). By mimicking the sensory qualities of fat without providing energy or dietary fatty acids, these foods are meant to help people lose weight, to maintain a reduced weight, and to assist in altering blood lipid profiles (Drewnowski, 1995). Fat-reduced foods are particularly interesting because they might help decrease the proportion of energy consumed as fat, consistent with current national recommendations. However, even in clinical settings and among very motivated individuals, reducing the amount of fat consumed has proven to be difficult (Drewnowski, 1995). Furthermore, adherence to weight-loss diets is poor due to reported 'cravings' by dieters for fat-rich, sweet desserts. Results from the Women's Health Trial also indicate that some dietary modifications are difficult to maintain over the long-term (Kristal et al., 1992). Despite intensive nutritional counseling to reduce dietary fat, the participants in the Health Trial who were randomized to the low fat diet found that the avoidance of adding fat to foods and in cooking, and eliminating desserts (such as cakes, cookies, and ice cream) were the most difficult habits to sustain. This suggests that the palatability of fat is hard to resist. Several concerns have been raised over the widespread availability of nutrient-modified foods. Such foods may convey an acceptance of dieting as part of a normal lifestyle (Adams & Shafer, 32 Literature Review 1988). Focusing on an awareness of the quantity of fat in foods may replace the monitoring of calories in food. Consequently, a dieter may replace counting calories with counting grams of fat. Also, studies of ingestive behaviours have suggested that a restriction in energy intake will be met with an energy compensation at a later time point such that total energy intake remains consistent (Louis-Sylvestre et al., 1994). Furthermore, use of low-fat products may result in individuals eating twice as much or overeating other foods (Rolls et al., 1994). The implications of the current Canadian nutritional recommendations are three fold. First, due to the possible energy compensation, reduced energy and fat foods may be ineffective in reducing total energy intake. This may have direct implications for dieters attempting to restrict their intake and may result in increased frustration with dieting. Secondly, individuals that successfully restrict their intake by modifying their fat intake may experience increased hunger and food cravings that may ultimately impact on food obsession. These experiences may increase the risk of developing eating disorders (Hill et al., 1993; Tuschl et al., 1990b). In addition, in an attempt to restrict fat intake for health or for weight reasons, dieters may not be meeting their nutritional requirements. Currently, little is known about the macronutrient intake of dieters. While the macronutrient composition of a weight loss diet may not necessarily influence the amount, composition or rate of weight loss, the composition of the diet will affect the degree of negative energy balance (Hill et al., 1993). Evaluation of the health consequences of dieting must address whether the nutrient composition of a dieter's food intake is actually different from non-dieters and, if so, whether it is healthy or unhealthy. It is possible that dieters may consume a more nutritious diet than non-dieters. In an attempt to restrict total energy intake, dieters may eliminate high fat or high calorie foods and replace them with lower fat or lower calorie food or foods rich in complex carbohydrate and fibre. Conversely, dieters may restrict intake of all foods such that their nutritional requirements are not met (French & Jeffrey, 1994). 33 Literature Review Several studies have examined the macronutrient intake of clinical populations. Food preferences of individuals with eating disorders and control subjects have shown significant differences when compared by macronutrient content. Both the preference for, and desire to eat, low-fat foods are greater in individuals with eating disorders (Stoner et al., 1996). Similarly, individuals with eating disorders have reported a taste aversion to dietary fat (Kaplan, 1993; Drewnowski, 1995). Consequently, individuals with eating disorders appear to consume a diet low in fat. In addition, when compared to controls, individuals with anorexia nervosa consumed a diet with a higher proportion of energy from protein and a lower proportion of energy from total and 'concentrated' carbohydrates (Kaplan, 1993). Both Kaplan (1993) and Gwirtsman et al. (1989) found that individuals with anorexia nervosa had significantly lower overall energy intake than the control subjects. Gwirtsman et al. (1989), however, found no difference in the percentage of energy derived from protein, carbohydrate or fat between women with anorexia nervosa and control subjects. Individuals with bulimia nervosa have been reported to consume less energy, and more protein and fibre on 'binge-free' days than individuals without bulimia (O'Connor et al., 1987; Rossiter et al., 1992). Rock and Yager have suggested that a low fat intake contributes to reduced meal satiety and to the likelihood of precipitating the urge to binge (Rock & Yager, 1987). Also, the most obvious difference in macronutrients between the so-called safe (non-binge) and forbidden (binge) foods for an individual with bulimia is the fat content. It is unknown, however, if the macronutrient composition of the diet, particularly the intake of fat, has a role in the development of eating disorders. The aversion to dietary fat may lead to the exclusion of meat by many individuals with eating disorders (Baken et al., 1993). Several authors have shown that 50 to 54% of women with anorexia nervosa practice vegetarianism (Baken et al., 1993; O'Connor et al., 1987). In these studies, vegetarianism was defined as the exclusion of red meat from the diet. In most cases of eating disorders, adoption of a vegetarian diet occurred after the onset of the disorder and was associated with a longer duration of the disorder and a lower BMI. Whether a vegetarian-type 34 Literature Review diet impacts on the development of abnormal eating behaviours is unknown, however, concern has been raised about the increasing use of vegetarian diets, in the name of health, as a means of weight loss (Kalucy, 1987). While not a clinical eating disorder, restrained eating is considered to be both a symptom of eating disorders and a precursor to the development of eating disorders (Striegel-Moore et al., 1986). Qualitative differences in macronutrient intake between restrained and unrestrained eaters have been reported. Restrained eaters have been shown, by some, to consciously eat less food and restrict calories more than their unrestrained counterparts (Herman & Mack, 1975; Tuschl, 1990a; Wilson, 1993). Restrained eaters also have been shown to consume a higher percentage of energy from protein (Laessle et al., 1989) and, by avoiding higher fat foods, a lower percentage of energy from fat (Tuschl et al., 1990b). It has been suggested that the consequence of such a pattern of food intake is a decrease in 'hedonic' pleasures, an impaired satiation, and a frustration that may increase the propensity to overeat (Laessle et al., 1989; Tuschl et al., 1990b). It should be noted, however, that significant differences in the energy or macronutrient intake between restrained and unrestrained women have not been consistently demonstrated (Barr et al, 1994a, 1994b). Others have echoed Tuschl's work suggesting that the most significant effects of the macronutrient composition of the diet is on the degree of behaviourial compliance with the diet (Hill et al., 1993). That is, stronger feelings of hunger and deprivation may result in a greater urge to break the diet. It is possible that the macronutrient composition of the diet will influence feelings of hunger and, thus, affect the ability to adhere to a weight loss diet, particularly over the long-term. Available data, however, are unclear. These feelings of hunger and deprivation, and the response to these feelings, could potentially be a factor in the development of disordered eating. 35 Literature Review Two key issues related to dieting and the potential effect the composition of the diet may have on development of eating disorders are that of energy and nutrient-specific compensation. Energy compensation is the adjustment in intake to maintain usual energy intake after a high energy or low energy preload. The issue of energy compensation is complex and is related to the time-frame in which the compensation is measured. Both short-term (Rolls et al., 1994) and long-term (Louis-Sylvestre et al., 1994) energy compensation has been demonstrated. However, consistent short-term energy compensation has been difficult to demonstrate in females, including those defined as obese, restrained or unrestrained, and males who were obese or categorized as restrained (Rolls et al., 1994). The issue of nutrient compensation, (compensation for a reduction in intake of a particular nutrient) has also been investigated. Studies in rat models suggest that food restriction that results in loss of body weight produces an increase in voluntary intake (Del Prete et al., 1994). This 'transient hyperphagia' appeared to be specific for carbohydrate. Evidence from both human and animal studies has suggested that weight cycling may increase preference for dietary fat (Brownell & Rodin, 1994b). However, Krahn (1991) has suggested that self-imposed food deprivation increases preference for sweet, high fat foods. Interestingly, these are the foods most often included in binges. The phenomena of energy or nutrient compensation suggest that dieting stimulates eating behaviours aimed at restoring food intake and maintaining weight. The effect of the macronutrient content of foods on subsequent food intake and satiety is unclear. In a review comparing the effects of fat and carbohydrate intake on subsequent food intake, hunger and satiety, Rolls and Hammer (1995) concluded that there is little difference in the effect of fat or carbohydrate on food intake or subjective ratings. This suggests that the consumption of reduced fat foods will , indeed, result in a reduction of the percentage of energy from fat. In many of the studies, however, the information about the fat content of the food was withheld from the 36 . Literature Review subject; providing information about the fat content of the food may alter these results (Rolls et al., 1994). The composition of the diet, in terms of macronutrient intake, is one of the many variables that can influence the success of weight loss. The effect of the macronutrient content of the weight loss diet on the subsequent eating behaviours is unclear. The physiological and behaviourial response to the diet depends, at least in part, on the level of deprivation and hence on the actual nutrient intake. It would appear that highly restrictive diets, particular those that restrict calories from fat, may influence hunger and food cravings leading to food obsession and possibly binge eating. It is unknown, however, whether the macronutrient distribution of a weight loss diet is associated with the development of abnormal eating behaviours. To date, the role of macronutrient intake in disordered eating has not been systematically examined. D. Restrictiveness of the Diet Weight control is the result of a balance between metabolizable energy intake and total energy expenditure (Hill et al., 1993; Schutz, 1995). As described above, the restriction of energy intake, by dieting, is the primary means of producing a negative energy balance (Levy & Heaton, 1993; National Institutes of Health, 1992). The reduction of caloric intake can take many forms. Some diets specify the total caloric level that should be consumed while others appear to have no caloric limits but, in fact, restrict energy intake by one technique or another, such as restricting specific foods or macronutrients. The end result is that the dieter must eat fewer calories (Dwyer, 1980; Herman & Polivy, 1988). Studies examining the links between dieting and development of eating disorders have suggested that restrictive dieting is linked to the development of eating disorders (Herman & Polivy, 1988; Hsu, 1990; Striegel-Moore et al., 1986). Keys et al. (1950) demonstrated that prior restriction of calories produced subsequent overeating, even after the weight lost during the deprivation phase 37 Literature Review had been regained. Nylander (1971) suggested that the 'intensity' of the diet leads to symptoms of anorexia nervosa. Experimental evidence has demonstrated strong links between 'rigid' dieting and binge eating (Huon & Wootton, 1987) Also, Polivy and Herman (1985) suggested that chronic restriction of food intake, or 'stringent' dieting, is associated with and precedes binge eating. These studies suggest that the quantitative nature of the caloric restriction may be associated with the subsequent development of an eating disorder. While restrictiveness is an assumed dimension of weight loss diets, it is unknown what exactly constitutes restrictive dieting and what level of caloric restriction, if any, is pathological. The caloric restriction of the diet is often described in terms such as 'restrictive', 'stringent', or 'intense'. Such terms imply a degree or level of calorie restriction (French et al., 1995). However, even within the DSM-IV criterion for eating disorders there is no explicit definition of 'strict dieting'. The implication of such vague terminology is that any amount of dietary or caloric restriction is potentially dangerous (Rossiter et al., 1992). Several questions arise from this. Does the caloric level of the diet impact on the progression of the disease? Is there a threshold caloric intake below which the risk of developing disordered eating is higher? Or is the risk related to caloric need versus absolute caloric intake? Extreme caloric restriction and restrictive eating are two of the primary features of eating disorders (Polivy & Herman, 1985). Some authors claim that severe caloric and food restriction usually precede the development of anorexia nervosa and bulimia nervosa (Krey et al., 1989). In anorexia nervosa, weight loss is achieved mainly through a reduction in total food intake. These individuals may begin by excluding energy dense foods from their diet. This may result in a very restrictive diet, sometimes limited only to a few foods. Caloric restriction also plays a role in bulimia nervosa. When not binge eating and purging, individuals with bulimia nervosa usually exhibit restraint over their caloric intake and select low-calorie or diet food while avoiding foods they perceive to be fattening or likely to trigger a binge (Polivy & Herman, 1985). In addition, a 38 ' Literature Review resumption in dietary restriction after successful treatment of bulimia nervosa has been considered a cause of relapse (Wilson, 1993). It is unclear, however, whether the dramatic dietary restrictions seen in eating disorders are a result of or a cause of the disorder. According to Nutrition Recommendations for Canadians, females at age 16 to 24 and 25 to 49 should consume 2100 kcal/d and 1900 kcal/d, respectively (Health & Welfare Canada, 1990). A nutritionally balanced reducing diet of at least 1200-1500 kcal/d may be considered safe and effective for moderate weight control for some individuals. Reducing diets which have caloric levels below 1200 kcal are usually inadequate to meet suggested Recommended Nutrient Intakes for most age and sex categories without reliance on nutritional supplements (Dwyer, 1980; Grodner, 1992). Repeated dieting with low calorie intakes, further increases the risk of nutrient deficiencies (McCargar et al., 1993). Low-calorie-diets (1000-1200 kcal), especially very-low-calorie-diets (VLCD) (<800 kcal), are often associated with a variety of short-term adverse effects, such as fatigue, hair loss, dizziness and other symptoms (National Institutes of Health, 1992). However, the data on short-term adverse health effects of weight loss come from programs that only include overweight persons. Some of these effects may be greater in persons who are not overweight but are severely restricting calories. In addition to negative health effects, food and calorie restriction appears to stimulate both physiological and cognitive symptoms that may precipitate abnormal eating behaviours. Physiologically, a reduced energy expenditure and reduced metabolism may result from calorie restriction. In addition, animal models have shown that just a few hours of food deprivation is sufficient to cause changes in hypothalamic neurotransmitters and their receptors (Leibowitz & Shor-Posner, 1986). When the animal is hungry the brain initially attempts to activate food seeking behaviours. In humans, ingestive behaviours are controlled by many integrating factors. The level of caloric restriction could be one factor that impacts both on the satiation cues from the gastrointestinal tract and neurotransmitter synthesis. 39 Literature Review Cognitive changes also result from calorie restriction. Dieting demands that normal hunger and satiety cues that regulate weight be ignored and be replaced with externally imposed rules designed to achieve a lower weight (Polivy & Herman, 1985; Omichinski, L. , 1995). The hunger associated with dieting can lead to psychological distress such as depression, anxiety and increased preoccupation with food (Keys et al., 1950; Grodner, 1992). Chronic energy restriction can exacerbate these psychological symptoms (Beals & Manore, 1994). Other reported consequences of food and calorie restriction include weight obsession, poor self-image, inappropriate exercise patterns, and disordered eating (Polivy & Herman, 1983). In addition, the dieter may lose the ability to eat in response to normal internal hunger and satiety cues. Polivy and Herman (1983) have suggested that the more drastic the calorie and nutrient restriction, (ie the smaller the variety and quantity of food eaten), the more serious the potential side effects appear to be. As this 'cognitive control' is susceptible to disinhibition, certain triggers, such as preoccupation with food, frustration, stress, or disruption of the diet through actual or perceived transgressions, may affect subsequent intake and dieting behaviours (Beals & Manore, 1994; Grodner, 1992; Tuschl et al., 1990b; Wilson, 1993). In varying degrees of intensity, this may lead susceptible individuals to overeating, a sense of being out of control, and binge eating (Polivy & Herman, 1985). Critics of restrictive diets often suggest that these diets will lead to feelings of deprivation and subsequent preoccupation and cravings for foods that are considered forbidden while on the diet (Weingarten & Elston, 1990; Krahn, 1991). The level of caloric restrictiveness of a diet may be one of the factors influencing the degree of preoccupation with food (Westenhoeffer & Pudel, 1993). Thus, food preoccupation and cravings may be one of the links between restrictive dieting and abnormal eating behaviours, particularly when sustained over long time periods. Several studies have examined the relationship between dieting and food cravings and food preoccupation in obese individuals in weight loss programs. Harvy et al. (1993) assessed 40 Literature Review changes in food cravings at two levels of caloric restriction (1000-1200 kcal/d and 400-500 kcal/d) over a 6 month weight loss program in 80 obese, type II diabetic men and women. The authors concluded that there was no evidence that eliminating certain foods from the diet increased cravings for those foods. Instead, complete restriction of a food may lower cravings for that type of food. Lappalainen et al. (1990) found conflicting results when assessing food cravings in 18 obese patients who were either fasting or consuming a 1200-1600 kcal diet during a three-week weight loss program. Their results suggest that severe caloric restriction (i.e. fasting) abolishes food cravings while moderate caloric restrictions may indeed maintain food cravings and food preoccupation. Binge eating is common in obese individuals trying to lose weight. In a large multi-site study, Spitzer et al. (1992) found that approximately 30% of obese patients trying to lose weight met the criteria for binge eating disorder. It is unclear, however, if binge eating was associated with the level of calorie restriction. Wadden et al. (1994) compared the frequency of binge eating at two levels of caloric restriction. Forty-nine obese women were randomly assigned to a 52-week behaviourial program combined with either moderate restriction (1200 kcal/d) or severe caloric restriction (420 kcal/d for 16 weeks followed by 1200 kcal/d thereafter). They found that reports of binge eating decreased in both groups of subjects over the course of the treatment. Likewise, Wilson et al. (1993a) found that caloric restriction did not differ between obese binge eaters and obese non-binge eaters; only a minority of obese binge eaters reported being on a strict diet at the time that binge eating began. Furthermore, there was no difference between binge eaters and non-binge-eaters in several diet-related activities, such as avoidance of eating or avoidance of specific foods to lose weight. Telch and Agras (1993), however, suggested that there may be a temporal relationship between calorie restriction and binge eating. They assessed the frequency of binge eating during a combined V L C D and behaviourial weight-loss program. When patients were reintroduced to food, 30% of those who had been identified as non-binge eaters before treatment subsequently reported binge eating episodes. 41 Literature Review Whether these results are generalizable to non-obese or non-diabetic populations is unclear. With regard to food cravings, it is possible that improved metabolic control of blood glucose in diabetic obese individuals may have contributed to the reduced cravings. It is difficult to determine whether the reduction in food cravings is related to the decrease in hunger frequently seen with extended fasting. In addition, it is unknown whether the subjects could distinguish between these two concepts. Dieting in nonobese populations may result in markedly different behavioural reactions to dieting than in obese populations (Grodner, 1992; Wilson, 1993). However, studies examining the effect of caloric restriction on the development of eating disorder symptoms in nonobese individuals are limited. In a study seeking to determine the etiology of eating disorders in female athletes, Sundgot-Borgen et al. (1992) found that prolonged caloric restriction was the most important "trigger" for the development of clinical eating disorders. Szmulker (1983) found that binge eating and self-induced vomiting were related to the level of food restriction in British school girls. Food avoidance scores were related to the frequency of binge eating and vomiting except for those with very high scores. In subjects scoring extremely high on the food avoidance scale, rates of vomiting and binge eating decreased, but were still present. Findings from Szmulker's work suggests that subjects able to severely restrict their energy intake were less susceptible to binge eating than those subjects restricting at an intermediate level. Food avoidance and food preoccupation were given a combined score based on the subjects' perception of severity and frequency of their occurrence. As such, the measure of restrictiveness in this study was subjective. From the aforementioned literature, it appears that the level of restrictiveness of the diet may influence the development of abnormal eating behaviours. However, it is difficult to define the degree of restriction in terms of caloric level, because restrictiveness usually relies on subjective interpretation by the interviewer or the subject. No direct measurements of absolute caloric 42 Literature Review intake while dieting, the amount of weight lost, or the level of caloric deprivation from the calorie requirement have been done. As such, the relationship between restrictiveness of the diet and disordered eating has not been fully elucidated. It may be speculated that the more restrictive the diet, both in absolute caloric intake and the extent to which the caloric intake deviates from caloric need, the greater the likelihood of developing abnormal eating behaviours. However this theory has not been systematically examined. 5. Summary Dieting is a common practice among college age women. The health implications of dieting are numerous and include negative effects on both physical and psychological well-being, particularly in the non-obese. In addition, it has been speculated that dieting may predispose some individuals to unhealthy weight loss practices. Evidence from clinical populations, cross-sectional and prospective studies have identified dieting to be a risk factor in the development of eating disorders in susceptible individuals. Despite these concerns, dieting in response to the sociocultural value of thinness is widespread in Western society, especially in female adolescents and young women. Not all women who diet develop eating disorders. The pathogenesis of eating disorders may be mediated by dieting and the factors that intensify or perpetuate this dieting may indirectly increase the risk of developing an eating disorder (Garfinkel et al., 1992; Hsu, 1990). Little is known about the specific factors which influence whether or not a dieter progresses to develop disordered eating. Previous research has suggested that key dieting variables such as the duration of dieting, the specific dieting methods, the macronutrient intake, and the level of calorie restrictiveness of the diet may play a role in the development of abnormal eating behaviours. The practice of dieting cannot be totally dismissed. Dieting remains an integral part of the medical treatment of some conditions (for example, obesity, hypertension, diabetes, 43 Literature Review cardiovascular disease) where the health risk of excess weight may exceed those risks associated with dieting. Thus, the challenge is to specify the characteristics of the diet and the dieting behaviours that are harmful. The extent to which dieting and the characteristics of the diet predict development of eating disorders is, therefore, an important but relatively unexplored question. To date this has not been systematically investigated. Diets vary significantly in their duration, method, macronutrient content, and level of calorie restrictiveness. Thus, a prospective study of these characteristics may provide a better understanding of the factors that influence the development of eating disorder symptoms. 44 Design & Methods III. Experimental Design and Methods The work described in this project was conducted as part of a collaboration between the School of Family and Nutritional Sciences, Division of Human Nutrition, The University of British Columbia, Vancouver, B.C. and the Eating Disorders Clinic, St. Paul's Hospital, Vancouver, B.C. This study received ethical approval from the University of British Columbia Behavioural Screening Committee, Vancouver, British Columbia, Canada. See Appendix A for a copy of the certificate of approval. 1. Study Design A. Overview A two-stage, prospective survey design was chosen to fulfill the objectives of the study. Subjects were recruited at the beginning of the academic year (September/October, 1994) from the female students attending six post-secondary campuses in the Greater Vancouver area: The University of British Columbia (UBC), Simon Fraser University (SFU), Capilano College, Langara Community College, and Vancouver Community College - King Edward Campus (KEC) and City Centre Campus (CCC). Notices which briefly described the study, eligibility criteria and prizes offered were posted at all campuses, included in all campus newspapers and, where possible, on video monitors (SFU and K E C only). See Appendix B for copies of the recruitment poster. After receiving permission from the respective departments and the individual instructors, announcements were made in the following courses at UBC: first year English (56 sections); Nursing (2 sections); Family and Nutritional Sciences (4 sections); and Women's Studies courses (4 sections). Announcements were also made in five of the first and second year courses at SFU (psychology, physics, biology and chemistry). Recruitment 'booths' were set up for four days during the supper hour outside student residence cafeterias at U B C (Place Vanier and Totem Park). Booths were also set up, for one to two days, outside each of the student 45 Design & Methods lounges at Capilano College, Langara Community College, K E C , C C C and SFU. Interested students were directed to call the telephone number provided and leave their name and telephone number. Students were screened for eligibility either at the time of recruitment or over the telephone. Four research assistants were involved in the recruitment process. Arrangements were made at each of the participating campuses to reserve a classroom or conference room. Appointments were made with each subject to arrive at the respective classroom or conference room at a set time. Subjects provided their written informed consent prior to commencing the study. Parental consent was received from two subjects who were under 18 years of age at the time of recruitment. See Appendix C for copies of the consent forms. To ensure that subjects were naive to the true purpose of the study, the consent form stated that the purpose of the study was to examine the eating and dieting habits of university women and how dieting practices may change over time. Each respondent was assigned a unique identification number to ensure confidentiality and to match data from baseline and 6 months. Addresses were obtained from subjects who wished to receive a summary of the results at the end of the study. After measurement of height and weight, the following questionnaires were administered to the subjects: (1) The Dieting Practices Survey; (2) A Food Frequency Questionnaire; (3) The Health Information Questionnaire; and (4) The General Information Sheet. (As this survey was part of a collaborative study, five other questionnaires were also administered to the subjects.) See Appendix D for a copy of the baseline questionnaires. The General Information Sheet provided basic demographic information. The Dieting Practices Survey was designed to elicit the duration of dieting, dieting methods used, rate of weight loss and pattern of food intake. (Data on the pattern of food intake were collected but were not analyzed in this report.) The Food Frequency questionnaire (FFQ) was used to quantify the 46 Design & Methods energy and macronutrient content of the diet. The Health Information Questionnaire was a self-report instrument assessing the frequency of binge-eating and compensatory behaviours. This questionnaire was based on DSM-1Y criteria for eating disorders (American Psychiatric Association, 1994) and yielded a probable eating disorder diagnosis, a Compulsive Eating Score, and an overall Symptom Intensity Score. Subjects receiving a probable diagnosis of anorexia nervosa or bulimia nervosa at baseline were excluded from all data analyses. Each questionnaire and the respective rating scale were reviewed with the subjects. Plastic food models, measuring cups and teaspoons, glasses and bowls illustrating the portion sizes used on the FFQ were also reviewed with the subjects. A l l subjects were encouraged to notify the researcher if any uncertainty or difficulty arose while completing the questionnaires. Baseline evaluation sessions required an average of 60 minutes. A l l efforts were made to ensure the privacy of the subjects during weighing and measuring and while completing the questionnaires. Due to scheduling difficulties at baseline, 14 subjects were mailed the questionnaires. Subjects were directed to complete the questionnaire in private and at one sitting. A l l directions were fully explained over the telephone. Subjects were encouraged to call the researcher if they had any questions regarding the completion of the questionnaires. Postage-paid, self-addressed envelopes were provided for the completed questionnaires. If the completed questionnaires were not received within two weeks, follow-up calls were made to verify that the questionnaires were received and to determine if any problems had arisen. For the subjects completing the questionnaires by mail, weight and height were self-reported. Assuming little variation in height over the study time period, self-reported height was substituted with the measured height obtained directly at the 6 month evaluation session. The self-reported weight, however, was used in the calculation of B M I at baseline. 47 Design & Methods The Health Information Questionnaire and a modified version of the Dieting Practices Survey were readministered at the end of the school year (March/April, 1995), six months later. See Appendix D4 for a copy of The Health Information Questionnaire and Appendix E for a copy of the 6 month Dieting Practices Survey. Once again, arrangements were made at each of the participating campuses to reserve a classroom or conference room. Subjects were contacted by telephone and follow-up appointments scheduled. Appointments were scheduled so that each subject was seen six months + 2 weeks after the baseline appointment. Subjects whose telephone numbers were not in service were mailed letters indicating that the second part of the study was soon to commence and there was difficulty reaching them by telephone. These subjects were directed to telephone the study phone number and provide their current phone number. See Appendix F for a copy of the letter. The procedure for weighing and measuring the subjects and administration of the questionnaires was similar to that reported at baseline. Once the questionnaires were completed, subjects were invited to ask any questions relating to the study. The follow-up sessions required approximately 25 minutes. Due to scheduling difficulties or subject re-location (subjects that had moved out of town or subjects attending out-of-town academic placements), 30 subjects were unable to attend the 6 month session and were mailed the questionnaires. The directions for completion were similar to those reported at baseline. The height measured at baseline was substituted for the self-reported height at 6 months. The self-reported weight, however, was used in the calculation of BMI at 6 months. Subjects were not paid for their participation. The identification numbers of subjects completing both phases of the study were placed in a lottery draw for a paid ski holiday weekend (cash value $600.00) and two cash prizes ($300.00 and $100.00). Three subject numbers were randomly drawn by the administrative assistant of the School of Family and Nutritional Sciences at the end 48 Design & Methods of April. The subjects were notified by telephone and arrangements made to forward the prizes to the subjects by mail. Subjects identified by The Health Information Questionnaire (HIQ) at baseline and 6 months as having a probable eating disorder diagnosis were sent a follow-up letter in June. (As the data from the subjects receiving a probable eating disorder were used by the collaborators of this study, information regarding the results of the HIQ was not released until after the study was completed.) The letter indicated that their responses to the questions on the HIQ were suggestive of irregular eating patterns and behaviours. A list of some community resources was included in the letter. Subjects were also encouraged to contact the candidate should they require any further information. See Appendix G for a copy of the letter. A summary of the study results was sent to each subject in November, 1995. See Appendix H for a copy of the summary. B. Subject Criteria The following inclusion criteria were specified for the purpose of assembling a sample with an increased risk of developing an eating disorder. Subject criteria included those individuals who were: . female 17-30 years of age at baseline . currently dieting at baseline. Dieting was self-reported and defined as the conscious restriction of food intake for the purpose of losing or maintaining weight (Brownell & Rodin, 1994a; Johnson et al., 1983; Johnson et al., 1984; Levy & Heaton, 1993; Moore, 1993) full-time university or college students. 49 Design & Methods Subject exclusion criteria included those individuals who had a probable diagnosis of anorexia nervosa or bulimia nervosa as defined by the DSM-IV criteria for eating disorders (American Psychiatric Association, 1994). C. Sample Size The present study hypothesized that specific dieting practices would not predict a worsening of eating disorder symptomology, as measured by the Change in Symptom Intensity Score. Determination of the sample size necessary to detect a worsening of eating disorder symptoms requires the estimation of effect size. For this one-sample study, effect size (d) is defined as the degree to which the correlation between dieting practices and Change in Symptom Intensity Score deviates from zero, the null hypothesis. Lack of previous research in this specific area and limited knowledge of the effect size that could be detected, precluded the use of standard sample size equations (Howell, 1992). Therefore, special conventions and power tables, as proposed by Cohen (1988), were used to calculate sample size for this study. Cohen has suggested that in new areas of research, where the phenomena under study is not under good experimental control, the effect size is anticipated to be small (i.e. the size of the 'effect' of the dieting practices on the development of eating disorders is small). A small effect size is operationally defined as d = 0.2. Thus, assuming a small effect size, a type I error of a < 0.05 and desired power of 80%, a sample size of 300 subjects was required (power > 90%). Assuming an attrition rate of approximately 30%, a sample of 400 dieters was to be assembled. 2. Methods A. Anthropometric Measurements i . Weight and Height 50 Design & Methods Body weight was determined to the nearest 0.1 kg on a beam-balance scale. Standing height was determined to the nearest millimeter using a stadiometer. Both measurements were made with the subjects wearing light clothing without shoes. i i . Body Mass Index Body Mass Index (BMI, kg/m2) was calculated as body weight (kg) divided by height (m) squared. Body size categories were determined as follows (Health and Welfare Canada, 1988a; Levy & Heaton, 1993; Williamson et al., 1992): . B M I < 20: underweight . B M I 20-25: normal or acceptable weight . B M I 26 < 29: overweight . B M I > 30: obese. B. Questionnaires i . The General Information Sheet The General Information Sheet is a brief demographic questionnaire used to obtain the subject's date of birth, year of study at college or university, and ethnic background. See Appendix D l for a copy of the General Information Sheet. i i . Dieting Practices Survey A self-report questionnaire administered within a semi-supervised setting was deemed the most suitable means to meet the objectives of the research, given the constraints of available time and resources. This combination of self-report and semi-supervision was chosen for several reasons. Self-administration removes interviewer bias and reduces the labour cost usually associated with individual interviews. Self-administration also permits a number of subjects to participate at one time. The administration of questionnaires in a controlled environment may have enhanced completion rates as the researcher would always be available to answer subjects' questions and review questionnaires for completeness. 51 Design & Methods The Dieting Practices Survey was developed specifically for this study. See Appendix D2 for a copy of the Dieting Practices Survey. This 7-item questionnaire consisted primarily of closed-ended questions and a rating scale format. A closed-ended response format was chosen to permit an accurate and rapid summary of the results. The Dieting Practices Survey used at 6 months was similar but included a question to determine the dieting status of the respondent at 6 months. See Appendix E l for a copy of the Dieting Practices Survey used at 6 months. In developing the Dieting Practices Survey, steps were taken to increase the credibility of the instrument. To address content validity, items were developed following a literature review and after consultation with several professors of nutrition at the School of Family and Nutritional Sciences, UBC. To assess test-retest reliability, several questions were repeated at 6 months; responses to questions such as 'age at first diet' and 'number of years dieting' were expected to remain stable over the study period. Internal consistency was also evaluated by assessing correlations between items on the Dieting Practices Survey and the Health Information Questionnaire that measured similar behaviours. The Dieting Practices Survey was designed to assess the duration of dieting, the dieting method, pattern of food intake, and the activity level of the subject. The definition and measurement of these variables are discussed separately below. (Pattern of Food Intake was assessed but not analyzed in this report.) a. Duration of Dieting The duration of dieting was divided into long- and short-term dieting in order to determine the impact of the duration of dieting on the development of disordered eating. Long-term dieting was defined as the cumulative duration of dieting, in years. Subjects were asked at what age they first began dieting and then were asked to estimate the number of years they had dieted. The total reported number of years dieting does not imply consecutive years of dieting. Only those 52 Desien & Methods years in which the individual had been dieting for a minimum of 14 days were included. Short-term dieting was defined as the total number of days the subject was dieting over the previous six months. To determine short-term dieting subjects were asked to estimate, over the last 6 months, the total number of days they were actively dieting. Responses to the questions were coded as continuous variables. If a range of ages, years or days was provided by the subject, the midpoint was recorded. If a number of months was reported, this response was converted into the appropriate number of days or fraction of a year(s). b. Dieting Methods Dieting method was defined as the specific modification made to food intake for the purpose of weight loss or weight maintenance. Other non-dieting weight loss methods, termed weight loss behaviours, were not evaluated in the Dieting Practices Survey as these behaviours were examined in the Health Information Questionnaire. [Weight loss behaviours included the use of pharmacological aids (diuretics, diet pills, and laxatives) and purging for the purpose of controlling body weight.] Subjects were provided with a list of nine dieting methods: A diet plan that is part of a commercial weight-loss program (for example, Weight Watchers® or Jenny Craig®) . A diet the participant made up herself A liquid formula or drink that replaces a meal or several meals The respondent eats the same way she usually eats but eats smaller portions and avoids or cut down on certain foods . Skipping meals Counting calories Avoiding eating when hungry Fasting (defined as going without food for more than 24 hours) 53 Design & Methods Counting grams of fat. Respondents were asked to rank their use, over the last six months, of each method on a five-point scale (never to always). Responses were then transformed into continuous scores according to the following scale: never (0), rarely (1), sometimes (2), often (3), always (4). Subjects were also requested to rank their use of the following three sources of dieting information: • A particular diet that the participant read about (for example in a book, magazine or newspaper) Information or advice received from a health professional (dietitian, nurse, doctor) A diet heard about on television, video or from friends or family. Responses were coded as described above. Participants were given an opportunity to list any other dieting method(s) they may use to lose weight. Response frequencies to this open-ended question were then tabulated but not analyzed. c. Activity Factor Estimation of an activity factor was required in the calculation of estimated energy requirement. The activity factor was derived from the response to a question on activity level in the Dieting Practices Survey ("Over the last six months, how often do you exercise for a period of at least one hour?"). Activity factors were coded as follows: less than once a week (sedentary): activity factor = 1.3; 1 to 3 times each week (moderate): activity factor =1.5; more than three hours each week (active): activity factor =1.7. (ESHA Research, 1990; Zeman & Ney, 1996). i i i . Food Frequency Questionnaire A food frequency questionnaire (FFQ) was used to measure the usual macronutrient intake of the sample with respect to average intake of energy (kcal/d), and absolute (g) and relative intake (%) of fat, carbohydrate and protein. 54 Design & Methods A FFQ is a self-administered questionnaire that requires the respondent to estimate the frequency of consumption for specified food items (Perkins, 1992). This method is considered a practical method to accurately estimate food consumption of a large group of people (Bright-See et al., 1994; Mullen et al., 1984). In the past, this method has been used to determine associations between nutrient intake and disease. Widespread use of the technique can be partially attributed to the advantages it has over the other dietary intake assessment methods, such as food records and 24-hour recall. The FFQ can be self-administered which markedly lowers labour costs relative to other dietary assessment methods (Kushi, 1994). A FFQ is generally easy to complete and the respondent burden often associated with personal interview is minimized (Margetts et al., 1989; Willet, 1994). The FFQ more accurately measures individual nutritional intake than methods based on household inventory or menu assessment (Mullen et al., 1984). In addition, a FFQ more accurately reflects usual intake and may provide a closer estimation of the usual diet than would a 24-hour recall (Margetts et al., 1989). Usual intake is defined as an individual's mean intake over an extended time period (Tarasuk & Beaton, 1992). Several studies have shown that the FFQ is valid and reliable when used to measure relative nutrient intakes for large subject groups (Kushi, 1994; Mullen et al., 1984; Perkins, 1992). Bright-See et al (1994) assessed the relative validity of the FFQ used in the 1990 Ontario Health Survey. Four day food records and FFQ from 147 subjects were compared. It was found that absolute values for most nutrients were greater on the FFQ than the food record. However, when expressed in relation to energy, there were essentially no differences between the FFQ and food records in reported intakes of carbohydrate, fat and protein. The FFQ used in the Ontario Health Survey was adapted for use in the present study. This particular FFQ also fulfilled the criteria of being relatively short and easy to complete by large numbers of subjects. Permission to use this FFQ was received from Dr. Elizabeth Bright-See, Department of Home Economics, Brescia College, London, Ontario. See Appendix I for a copy of the letter. 55 Design & Methods Several foods/food categories were added to the FFQ used in the present study. This was done to include foods that may be commonly available from school cafeterias or dormitory dining halls (sweetened cold cereal, pancakes, soups), vending machines (diet soft drinks), or might be specific to individuals following vegetarian diets (tofu). Alcohol was also included as consumption of alcoholic beverages may make a notable contribution to the energy intake of some college students. Four low fat dairy products (1% milk and low-fat cheese, yogurt and cottage cheese) were added as these foods may be consumed by dieting women (Tuschl et al., 1990b). See Appendix D3 for a copy of the FFQ. The format of the FFQ consisted of preselected foods from which the subjects were asked to indicate whether each food was consumed at least once a month or not eaten at all. The frequency they consumed the food was then estimated (daily, weekly, monthly). Serving size options were provided based on typical serving sizes or natural food portions (cups, items, slices, cubes, or number). Serving sizes used to quantify food intake, other than those specified on the FFQ, are presented in Appendix J. Plastic food models, measuring cups and teaspoons, glasses and bowls illustrating the portion sizes used on the FFQ were reviewed with the subject to help facilitate size estimations. Reported frequencies were converted into daily intakes and multiplied by the nutrient content of the specific food or food category. For example, an item marked as two cups per week was assigned a value of 0.28 cups per day. Mean daily intakes were calculated using the Canadian Nutrient File data base (ESHA Research, 1990). The nutrient composition of specific food items on the questionnaire was determined in one of two ways. The composition of a single food item was based on the database values for that specific food (e.g. 1% milk). The nutritional composition of food groupings (e.g. Plain Cereal) was calculated as the average of nutrients provided by several related foods in the food groupings. For example, the nutritional composition of 3/4 cups Plain Cereal was the average nutrients provided by 3/4 cups Corn flakes, Rice Krispies, and Cheerios. 56 Design & Methods The FFQ was used to assess the potential relationship of the restrictiveness of the diet and the macronutrient intake to the development of eating disorder symptoms. The definition and measurement of these variables are discussed separately below. a. Macronutrient Intake Total grams of protein, carbohydrate (CHO) and fat consumed per day over the previous 6 months was estimated by the FFQ. The percentage of energy derived from protein, CHO, and fat was calculated as the energy (kcal) from protein (g/d x 4 kcal/g), CHO (g/d x 4 kcal/g), and fat (g/d x 9 kcal/g) divided by the total energy intake, multiplied by 100. b. Restrictiveness of the Diet The restrictiveness of the diet was defined as the difference between estimated energy requirement and energy intake (estimated energy requirement (kcal/d) - energy intake (kcal/day)). The greater the difference between energy requirement and intake, the greater the restrictiveness of the diet. Energy intake was based on usual intake over the last 6 months as measured by the FFQ. Estimated energy requirement was calculated as the estimated resting metabolic rate (determined by a prediction equation) multiplied by the activity factor. 1) Resting Metabolic Rate In the absence of indirect calorimetry, many equations are available for predicting resting metabolic rate (RMR). Energy requirements are usually determined by estimating the R M R from a prediction equation, and by estimating energy expended in physical activity. (Overall, the magnitude of the thermic effect of food is small (about 6% to 10% of total ingested energy) and usually considered insignificant (Hill et al., 1993)). Most available equations were developed from a regression formula for predicting R M R from sex, age, height, and weight. The Harris-Benedict Equation (Harris & Benedict, 1919) is the equation most commonly used. However, several studies have shown that the Harris-Benedict equation overestimates R M R in healthy, 57 normal men and women by approximately 5-al., 1986; Owen et al., 1987). Design & Methods 15% (Daly et al., 1985; Mifflin et al., 1990; Owen et The Mifflin equation for the prediction of R M R was selected for use in the present study for several reasons. The regression equation was developed in a large sample of healthy, lean and obese men and women (251 men and 247 women) between the ages of 19 and 78 years. Similar prediction equations have been developed, however, the sample sizes were notably smaller (sample size varies from 44 to 103 female subjects) (Thompson & Manore, 1996). The Mifflin Equation was derived in a more 'modern' population (1990) as compared to the population used in the Harris-Benedict (1919). In addition, a recent study found that the equation accurately predicted the R M R in healthy, moderately overweight women (McCargar et al., in press). The resting metabolic rate (RMR) was, therefore, calculated using the following prediction equation for females (Mifflin et al., 1990): R M R = 9.99 (Weight in kg) + 6.25 (Height in cm) - 5 (Age in years) - 161 Weight and height used in the prediction equation were measured directly. The age of the subject was obtained from the General Information Sheet. 21 Rate of Weight Loss A second estimate of the restrictiveness of the diet was determined by the rate of weight (kg) lost during a one week period. Subjects were asked how much weight they usually lose during each week of dieting. If a range of weight lost per week was provided, the midpoint of the range was recorded as the response. Those who responded that they were maintaining their weight were given a value of '0'. Amounts specified as less than 1 pound (0.5 kg) were coded as 0.5 pound (0.23 kg). 58 Design & Methods c. Vegetarianism The FFQ was also used to determine if the subject was following a vegetarian diet. Vegetarian status was defined as: (1) lacto-ovo-vegetarian: a diet that excludes meat, poultry, fish but includes milk products and eggs; (2) partial vegetarian: a diet that excludes meat and poultry but includes milk products, fish and eggs; and (3) vegan: a diet that excludes all foods of animal origin (Dwyer, 1991). Vegetarian classification was based on actual intake rather than by the subject's comments. d. Vitamin/Mineral Supplementation Nutritional supplementation practices of the subjects was determined by the FFQ. Subjects were asked to determine the frequency of use of vitamin/mineral supplement over the past 6 months. Micronutrients contributed by the supplement were not included in the nutritional analysis. A participant was considered a regular users of vitamin/mineral supplements if a weekly or more frequent use was indicated. iv. The Health Information Questionnaire Eating disorder symptomology was measured by the Health Information Questionnaire (HIQ). See Appendix D4 for a copy of the HIQ. This questionnaire was based on the DSM-IV criteria for eating disorders (American Psychiatric Association, 1994). The HIQ was also based on other self-report surveys developed for similar purposes (Greenfeld et al., 1987; Kagan & Squires, 1983; Whitaker et al., 1989). It assessed weight and menstrual history, fear of weight gain, worries about eating habits and presence and severity of binge eating, perceived lack of control over eating and six compensatory behaviours (excessive exercise, fasting, use of diet pills, diuretics, vomiting and laxatives). The response scales were designed to assess DSM-IV threshold criteria as well as sub-threshold levels of severity for binge eating and all relevant compensatory behaviours (Geller et al., in press). Three measures were obtained from the HIQ: a Compulsive Eating Score; a Symptom Intensity Score; and a probable eating disorder diagnosis 59 Design & Methods (bulimia nervosa, purging and non-purging subtype; anorexia nervosa, restricting and binge-purge subtype; and Eating Disorder Not Otherwise Specified (EDNOS)). Cronbach's alpha for the HIQ has been calculated at .85 (Geller et al., 1996). The HIQ correlates well with other measures of eating disorder symptomology, such as the Eating Disorders Inventory. (The Eating Disorders Inventory is a 64-item self-report scale designed to measure attitudes, personality features, and eating disorder symptoms thought to be relevant to anorexia nervosa and bulimia nervosa (Garner & Olmstead, 1984; Garner et al., 1983)). The HIQ has previously been used as a classification instrument to detect clinical symptomology in university undergraduates (Geller et al., in press). The HIQ has also been shown to differentiate between normal controls (82 undergraduate women between the ages of 18 and 45 years) and individuals with known clinical eating disorders (48 women with diagnosed eating disorders) (Geller, 1996). Responses to HIQ items were entered into a computer program written specifically for this study (J. Geller, E. Goldner, R. Mackenzie, St. Paul's Hospital, Vancouver, B.C.). The Symptom Intensity Score can range between 0 and 80 with higher scores indicating greater eating disturbances. The Symptom Intensity Score is an additive score based on responces outlined in Appendix D5. The Compulsive Eating Score is a component of the Symptom Intensity Score and sums responses indicative of compulsive eating. The HIQ probable diagnosis is derived from groupings of responses that are consistent with the DSM-IV criteria for the five eating disorder categories. Subjects endorsing all relevant statements that met the essential DSM-IV criteria for a particular eating disorder received a probable diagnosis for that particular category. A probable diagnosis of EDNOS was given to subjects endorsing some but not all criteria for anorexia nervosa or bulimia nervosa, as defined in the DSM-IV. Data from any subject receiving a probable diagnosis of bulimia nervosa (purging or non-purging subtype) or anorexia nervosa (restricting or binge-purge subtype) at baseline were not analyzed in the present study. (These subjects, however, were maintained in the sample and were evaluated 60 Design & Methods again at 6 months as these data were analyzed by the study collaborators.) Due to the sub-threshold nature of the diagnosis of EDNOS, subjects receiving a probable diagnosis of EDNOS at baseline remained in the sample. A l l subject receiving a probable diagnosis of anorexia nervosa, bulimia nervosa or EDNOS were sent a follow-up letter in June. As the purpose of the study was to investigate the relative ability of specific dieting variables to predict eating disorder symptomology, the Symptom Intensity Score (a continuous measure) and the Change in Symptom Intensity Score (baseline score - 6 month score) were used as the dependent measures. C. Pretesting of the Dieting Practices Survey and FFQ The Dieting Practices Survey and Food Frequency Questionnaire were pretested in July and August, 1994. Despite extensive pretesting and validation of the FFQ used in the Ontario Health Survey, the relative validation does not necessarily extend to a Vancouver-based sample. There may be differences in regional food habits across Canada. In addition, the sample surveyed in Ontario was not made up exclusively of dieters. To date, no reference or 'Gold Standard' dietary assessment method has been devised, so the approach of 'relative validation' is widely used to compare different nutrient intake methods (Borrelli et al., 1989). As the comparison of FFQ and food records were likely to have the fewest correlation errors, the relative validity of the FFQ was assessed by comparing the FFQ data to that obtained by a three-day food record (Bright-See et al., 1994). Students attending a second year summer nutrition course and graduate students attending summer courses were approached for pretesting of the questionnaire. These students were considered to resemble the intended sample in terms of demographic background. The students, however, were not necessarily dieting. Announcements were made in classrooms and volunteers were asked to participate in the pretest study. Written informed consent was received from the subjects prior to administering the questionnaires. See Appendix K for a copy of the pretest consent form. Subjects were then provided with verbal and written instructions on how to 61 Design & Methods complete the Dieting Practices Survey, the FFQ and a Three Day Food Record. See Appendix L for a copy of the pretest questionnaires and Food Record. Questionnaires were completed at the subject's home and returned to the researcher approximately one week later. Pretest subjects were also provided an evaluation form and were encouraged to comment on the clarity and wording of the questions and ease of completion. The ease of administration was also a focus of the pretest. a. Dieting Practices Survey Thirteen completed Dieting Practices Surveys were obtained. Based on the responses and comments received, some items were reworded and several questions deleted. Seven revised items remained. b. Food Frequency Questionnaire Nine three-day food records and FFQ were obtained from the pretest sample. Reported frequencies were converted into daily intakes and multiplied by the nutrient content of the specific food. Daily mean intakes were calculated in a manner similar to that reported in section 2.B.iii.a. Table 2 presents the mean energy and macronutrient intakes calculated from the FFQ and three-day food records. Correlations between mean values obtained from the FFQ and three-day food records are also presented. Energy, protein, and fat intakes were significantly correlated while agreement between the FFQ and three-day food records for percent of energy from CHO was poor. The low correlation between measures of CHO may be due to the small sample size or due to within-person variability (Margetts et al., 1989). Despite the low correlation for percent of energy from CHO between the FFQ and three-day food records, the significant correlations for the other nutrients are indicative of relative agreement between the two sets of nutrient data. In addition, this particular FFQ has been extensively used in the Ontario Health Survey. For these reasons, the food frequency questionnaire was considered appropriate for use to estimate food 62 Design & Methods Table 2: Correlations Between the Mean and Relative Macronutrient Intakes Derived from Food Frequency Questionnaires (FFQ) and Three-Day Food Records in a Sample of Female College Students. Macronutrient FFQ Food Record r§ (n=9) (n=9) Energy(kcal/day) 2410 ±498* 2546 ±493 0.84b Protein (g/day) 92 ± 2 1 77 ± 14 0.75a CHO (g/day) 334 ± 70 401 ± 100 0.56 Fat (g/day) 84 ± 2 4 70 ± 2 1 0.69a % Energy from Protein 15 ± 2 12 + 2 0.34 % Energy from CHO 55 ± 5 62 ± 9 0.03 % Energy from Fat 31 ± 5 25 ± 6 0.31 §Pearson Product Moment Correlation (r) * M e a n ± S D Signif icant correlation: a p< 0.05, p<0.01 63 Design & Methods consumption in this study. However, due to the low correlation between measures of CHO, caution is advised in interpreting the results especially with regard to the CHO intake of the subjects. 3. Statistical Analysis Statistical Packages of the Social Sciences (SPSS) Professional Statistics for Windows, Version 6.0 (1993) was used for all data analyses. Results were considered statistically significant at a probability level of p<0.05. However, to control the family wise error rate for multiple comparisons (comparison of dieting methods and binge-eating and compensatory behaviours between baseline and 6 months), the Bonferroni correction method was used and the probability level was reduced by the corresponding number of comparisons in the analysis (Howell, 1992). A statistician was consulted throughout the research. A l l questionnaires used in the study were checked for completeness and coded prior to analysis. A. Description of the Sample Descriptive statistics were used to summarize sample characteristics, eating disorder symptomology and the key dieting variables (duration of dieting, dieting methods, macronutrient intake and restrictiveness of the diet). Normal probability plots (normal plot and detrended normal plot) were examined for each variable to determine that the data came from a normal distribution. Variables that were not normally distributed were transformed (log and square root transformation) prior to inferential statistical testing. B. Student's Paired Sample t-Statistic Student's paired sample t-statistic was used to test the significant difference in means between baseline and 6 months for paired measures (age at first diet, long- and short-term dieting, weight, height, BMI, rate of weight loss, dieting methods, Symptom Intensity Scores, Compulsive Eating 64 Design & Methods Scores and use of binge eating and compensatory behaviours). As some responses should not change over time (age at first diet, long-term dieting), paired t-tests were used to assess the test-retest reliability of subject responses. C. Factor Analysis of Dieting Methods The correlation matrix determined from the dieting methods resulted in a set of interrelated variables. The observed correlations between variables was assumed to be due to their sharing underlying constructs or factors. To identify underlying constructs and group dieting method variables, data from all dieting methods were evaluated using principal components factor analysis (equimax rotation). Adequacy of variable sampling was examined by Bartlett's test and Kaiser-Meyer-Olkin measure for sampling adequacy (SPSS Inc., 1993). Only factors with eigenvalues greater than one were considered for subsequent analysis. D. Correlation Between Dieting Variables and Symptom Intensity Score Pearson Product Moment correlation coefficients were calculated to determine the relationship between the key dieting variables measured at baseline and the dependent variables, Symptom Intensity Score (baseline and 6 months) and Change in Symptom Intensity Score. Change in Symptom Intensity Score was calculated as baseline Symptom Intensity Score minus the 6 months Symptom Intensity Score. An increase in Symptom Intensity Score (as indicated by a negative Change in Symptom Intensity Score) represented a worsening of symptoms. As the Compulsive Eating Score was a component of the Symptom Intensity Score, it was not considered for separate analysis. E. Correlation Between Dieting Variables and Binge Eating and Compensatory Behaviours The Symptom Intensity Score was derived from the frequency of binge eating and compensatory behaviours. To determine which specific behaviour(s) correlated with the dieting variables, further correlations were determined between the key dieting variables measured at baseline and the use of binge eating and compensatory behaviours. 65 Design & Methods F. Multiple Regression Analysis of Dieting Variables on Change in Symptom Intensity Score Multiple regression analysis, using simultaneous entry, was used to investigate the relative predictive ability of the key dieting variables, measured at the beginning of the university year, to predict the Change in Symptom Intensity Score at the end of the school year, six months later. G. Comparison of Dieting Variable in Subjects According to Dieting Status at 6 Months Multivariate analysis of variance (MANOVA), with Hotelling's T Squared correction for multiple comparisons, was used to compare differences in baseline measures of the key dieting variables between the group of subject who, at 6 months, identified themselves as dieters compared to the group of subject who identified themselves as non-dieters. 66 Results IV. Results 1. Distribution of the Data Normal probability plots (normal plot and detrended normal plot) and histogram plots for each variable were examined to determine that the data came from a normal distribution. Variables that were not normally distributed (age at first diet, number of years dieting, rate of weight loss, fat intake, protein intake, carbohydrate intake) were transformed (log and squareroot transformations). Distributions for age, body weight and B M I were slightly positively skewed; this was expected in view of the sampling strategy. Boxplots revealed the presence of outliers, most notably on weight and BMI. Data from these subjects were checked to determine if responses to other items were unreasonable. Nothing was evident that would make the legitimacy of the data questionable. Identified outliers, therefore, were maintained in the data set. 2. Description of the Sample A. Sample Size At baseline, 206 questionnaires were completed by the subjects. Thirteen questionnaires were excluded due to responses inconsistent with the inclusion criteria (i.e. not dieting (n=7), altering intake to prevent weight loss (n=2), or over 30 years old (n=4)). A further 7 were excluded based on responses to the Health Information Questionnaire which indicated a probable eating disorder diagnosis (bulimia nervosa, purging sub-type (n=2), non-purging sub-type (n=5)). In total, 186 questionnaires remained. At 6 months, 7 subjects could not be contacted. Therefore, the total number of subjects at 6 months was 179, which represents a response rate of 96 % (179/186). The estimated sample size of 400 was not achieved. However, based on power tables proposed by Cohen (1988), a sample size of 179 subjects resulted in a power of 0.83. 67 Results Baseline characteristics of the 7 subjects that did not complete the questionnaire at 6 months were compared with baseline characteristics of the subjects that completed the study. There were no significant differences between the groups. A l l results are expressed excluding data from the 7 subjects with probable eating disorder diagnosis at baseline. The number of subjects per analysis may vary as cases with missing values were excluded from the respective analysis. Missing values were due to illegible or ambiguous responses. Results are presented as the mean ± standard deviation. B. Sample Characteristics i . Demographic Characteristics Tables 3 and 4 present the demographic characteristics of the sample. The majority of participants were students at the University of British Columbia, were attending first year college or university and identified themselves as being of European and Canadian ethnic origin. Two subjects from other campuses, British Columbia Institute of Technology and Douglas College, also participated in the study. These students learned of the study from friends at the participating campuses. i i . Physical Characteristics Physical characteristics and age of the subjects are presented in Table 5. The mean weight at baseline was 62.5 ±11 .7 kg. However, of the total sample, 17.9 % were underweight (BMI < 20), 7.3 % were overweight (BMI 26 < 29), and 3.4 % were obese (BMI > 30). At 6 months, the mean weight was 62.9 ±11.8 kg. The distribution within the B M I categories at 6 months was not significantly different from baseline categories; of the total sample, 15.1% were classified as underweight, 7.8% overweight and 3.9% classified as obese. 68 Results Table 3: Number of Dieting Female College Students Recruited From Participating Campuses Participating Campuses n (%) (n = 179) University of British Columbia 91 (50.8) Simon Fraser University 35 (19.6) Capilano College 23 (12.8) Langara Community College 11 (6.2) Vancouver Community College! 17 (9.5) Other* 2 (1.1) *British Columbia Institute of Technology, Douglas College f Includes King Edward Campus and City Centre Campus Results Table 4: Characteristics of Study Participants Subject Characteristics n(%) (n= =179) Academic Year First 106 (59) Second 27 (15) Third 17 (10) Fourth 19 (11) Graduate 3 (2) Ethnicity Canadian 44 (25) British (Irish Scottish, Welsh) 30 (17) Oriental 19 (11) South Asian 5 (3) Other European 54 (30) Other 19 (11) Undefined 8 (5) Table 5: Physical Characteristics of Study Participants Baseline 6 Months (n=179) (n=179) Age (y) 21.1 + 3.2* — Weight (kg) 62.5 ± 11.7 62.9 ± 11.8 Height (cm) 165.3 ±7 .1 165.3 ±7 .0 BMI (kg/m2) 22.8 ±3.5 22.9 + 3.5 * Mean ± S D N o significant differences found in paired sample r-Test Results i i i . Activity Factor Subjects were asked to estimate their frequency of exercise (exercise for a period of at least one hour or longer). Forty-eight (27%) subjects reported exercising less than once per week, 92 (51%) subjects reported exercising between one and three times per week, 39 (22%) subjects exercised more than three times per week. C. Eating Disorder Symptomology i . Probable Eating Disorder Diagnosis Table 6 presents the probable eating disorder diagnosis as determined by the Health Information Questionnaire. At baseline, 26 (14.5%) subjects received a probable diagnosis of Eating Disorder Not Otherwise Specified (EDNOS). At 6 months, 21 (11.7%) subjects received a probable diagnosis of EDNOS, while 7 (4%) subjects appeared to progress to more clinical forms of eating disorders, anorexia nervosa and bulimia nervosa. Table 7 presents the changes in probable eating disorder diagnosis over the study period. Of those subjects receiving a classification of 'No Diagnosis Found' at baseline, 137 (89.5%) remained in the 'No Diagnosis Found' category at 6 months while 16 (10.5%) subjects progressed to other probable diagnostic categories: EDNOS: 11 (7.2%); bulimia nervosa, non-Purging Subtype: 4 (2.6%); anorexia nervosa, restricting sub-type: 1 (0.7%). Of those subjects receiving a diagnosis of EDNOS at baseline, 10 (38.5%) remained in this category at 6 months. However, 14 (53.9%) reverted to the 'No Diagnosis Found' category while 2 (7.7%) progressed to the category of bulimia nervosa, one each for the purging subtype and non-purging subtype. i i . Symptom Intensity Score and Compulsive Eating Score As seen in Table 8, no significant differences were observed in Compulsive Eating Scores between baseline and 6 months. However, a significant decrease in Symptom Intensity Score was 71 Results Table 6: Probable Eating Disorder Diagnosis of Study Participants Based on Response to the Health Information Questionnaire* Probable Eating Disorder Diagnosis Baseline (n=179) 6 Months (n=179) n(%) No Diagnosis Found 153 (85.5) 151 (84.4) Bulimia Nervosa - Purging Subtype 0 1 (0.6) Bulimia Nervosa - Non-Purging Subtype 0 5 (2.8) Anorexia Nervosa - Restricting Subtype 0 1 (0.6) Anorexia Nervosa - Binge-Purge Subtype 0 0 Eating Disorder Not Otherwise Specified 26 (14.5) 21 (11.7) * Based on D S M - I V criteria for Eating Disorders (American Psychiatric Association, 1994) 72 Results Table 7: Changes in Probable Eating Disorder Diagnosis from Baseline to 6 Months in Study Participants* Baseline Diagnosis n (n=179) 6 Month Diagnosis No Bulimia Nervosa Anorexia Nervosa EDNOSf Diagnosis Purging Non- Restricting Binge/ Found Subtype Purging Subtype purge Subtype Subtype n (n=179) No Diagnosis 153 Found EDNOS 26 137 - 4 1 -- 11 14 1 1 -- - 10 * Based on D S M - I V criteria for Eating Disorders (American Psychiatric Association, 1994) f Eating Disorders Not Otherwise Specified 73 Results Table 8: Symptom Intensity Scores and Compulsive Eating Score of Study Participants Based on Response to Health Information Questionnaire* Baseline 6 Months (n=179) (n=179) Symptom Intensity Score 17.3 ± 9.3f 16.1 ± 9.1 a Compulsive Eating Score 3.2 ± 2.2 3.2 ± 2.2 Corrected Symptom Intensity Score§ --- 1.27 + 6.9 *Based on D S M - I V criteria for Eating Disorders (American Psychiatric Association, 1994) t M e a n ± S D a Signif icantly different from Baseline: p < .05 (Paired sample t-Test) §Corrected Symptom Intensity Score = Baseline Symptom Intensity Score - 6 Months Symptom Intensity Score 74 Results observed. Overall, the Symptom Intensity Score of the sample decreased 1.3 ± 6.9 points over the six months. Over the study period, the Symptom Intensity Score increased in 66 subjects, did not change in 13 and decreased in 125 subjects. i i i . Previous Treatment for Eating Disorders Overall, 21 (11.7%) subjects at baseline and 14 (7.8%) subjects at 6 months, reported receiving treatment in the past for eating disorders. Of the 158 (88.3%) subjects who, at baseline, stated that they had not received any prior treatment for eating disorders, 3 (1.9%) subjects reported that they had received treatment at 6 months. However, of the 21 (11.7%) subjects who reported, at baseline, that they had received treatment in the past for eating disorders, only 11 (6.2 %) subjects reported receiving previous treatment for eating disorders at 6 months. i i i . Binge Eating and Compensatory Weight Loss Behaviours Self-reported use of binge eating and compensatory weight loss behaviours at baseline and 6 months are presented in Figures 1 to 8. At baseline, 27.4% of the sample reported binge-eating (once or more per week). Extreme dieting and weight loss behaviours were reported with much less frequency. With a combined frequency of once a week or more, 0.6% of the sample reported fasting, 1.2% reported using diet pills, 0.6% reported using diuretics, 4.5% reported vomiting, and 0.6% reported using laxatives for the purpose of losing weight. Loss of control over eating was experienced by 17.3% of the sample (frequency of 'often' to ' all the time') and loss of weight by 'crash dieting' was reported by 8.1% of the sample ('more than once in the last month'). At 6 months, frequency of binge-eating and compensatory behaviours appeared to be similar to baseline. With a frequency of weekly or more, 26.2% of the sample reported binge-eating, 1.7% reported fasting, 1.7% reported using diet pills, 1.7% reported using diuretics, 2.8% reported 75 Results 1 0 0 % i 9 0 % 8 0 % 70%-60%-5 0 % 4 0 % 3 0 % 2 0 % 1 0 % i o%-Figure 1: Self Reported Use of Binge Eating by Study Partiripants| I Baseline 1 6 Months n = 179 111 l a 6 ^ l | ] T W % " « * 1-1% 0,6% 0.0% Never Less than About once a About once a Between 2&6 Every day More than once/month month week times a week once everyday Frequency Figure 2: Self-Reported Use of Fasting to Lose Weight by Study Participants I Baseline i 6 Months n = 179 °- 6 % IIL..I!M . 0-0%-£L6%_ 0.0% 0.0% 0-0% 0-0% Never Less than About once a About once a Between 2&6 Every day More than once/month month week times a week once everyday Frequency 76 Results Figure 3: Self-Reported Use of Diet Pills to Lose Weight by Study Participants| 1 0 0% I Baseline I 6 Months n = 179 Never " S'tt.m.ymm. 0 0 % 0-0* 0.6% 0-0% 0-6% 1 j r % 0.0% 0.0% Less than once/month About once a About once a Between 2&6 month week times a week Frequency Every day More than once everyday Figure 4: Self-Reported Use of Diuretics by Study Participants^ I Baseline 1 6 Months n = 179 1 J % 0.6%_LJ%. 0.0% 0.0% 0.0% 0.6% 0.0% 0.0% Never Less than About once a About once a Between 2&6 Every day More than once/month month week times a week once everyday Frequency 77 Results Figure 5: Self-Reported Loss of Control Over Eating by Study Participants| 100%-90%-80%-70%-60%-50%-4 0 % i 30%~ 20%-j 10%" 0%+' Never Hardly Ever S E E SEP? Sometimes I Baseline I 6 Months n = 179 Often Very Often All the Time Frequency i o o % -90%" 80%-70%" 60%" 50%" 40%-30%" 20%-10%" 0%-r- 1 Figure 6: Self-Reported Use of Crash Dieting To Lose Weight by Study Participants! Never Once in the Last Week I Baseline 1 6 Months n = 179 More than Once in the Last Year Frequency Once in the Last Month More than Once in the Last Month 78 Results Figure 7: Self-Reported Use of Vomiting to Lose Weight by Study Participants j I Baseline I 6 Months n = 179 Never 3.4%JL5% ^ 2 . 2 % [ n% n r « < 0 . 0 % 0 . 0 % ^ 0 . 6 % 0 . 0 % Less than About once a About once a Between 2& 6 Everyday More than once/month month week times a week once everyday Frequency 100%-| 90%-Figure 8: Self-Reported Use of Laxatives to Lose Weight by Study Participants| TBTOfflCTWTTTWmillllllll 1  II lim*—111111111 i r r — r T r r — — — — — — — — — — — — — — — — — • — — n i l I Baseline 1 6 Months n = 179 0.6%Jj£L 0.0% 0.6% 0.0% 0.0% 0.6% 0.0% 0.0% 0.0% Never Less than About once a About once a Between 2&6 Every day More than once/month month week times a week once everyday Frequency 79 Results vomiting, and 0.6% reported using laxatives for the purpose of losing weight. Loss of control over eating was experienced by 16.8% of the sample (frequency of 'often' to ' all the time') and loss of weight by 'crash dieting' was reported by 8.4% of the sample ('more than once in the last month'). Mean response rate to binge-eating and compensatory behaviours at baseline and six months are presented in Table 9. Only frequency of 'Crash Dieting' was significantly lower between baseline and 6 months (t=3.2, p=.002). 3. Relationship Between Subject Characteristics and Eating Disorder Symptomology A. Correlation Between Subject Characteristics and Symptom Intensity Score Table 10 reports the correlations between baseline age, weight and B M I and the Symptom Intensity Scores at baseline and 6 months and the change in Symptom Intensity Score. A significant inverse relationship was observed between age and Symptom Intensity Score at baseline and 6 months. Baseline measures of weight and B M I were also correlated with Symptom Intensity Score. Change in Symptom Intensity Score was negatively correlated with weight. B. Correlation Between Subject Characteristics and Binge Eating and Compensatory Behaviours Table 11 presents the correlations between baseline age, weight and B M I and binge eating and compensatory behaviours at baseline and six months. Inverse correlations between age and 'crash dieting', and at 6 months, vomiting were observed. At baseline, significant correlations 80 Results Table 9: Frequency of Binge Eating and Compensatory Behaviours in Study Participants Binge-Eating and Compensatory Behaviours Baseline (n=179) 6 Months (n=179) Binge-eating 1.8±1.3* 1.7 ± 1.3 Fasting 0.3 + 0.6 0.2 ± 0.6 Diet Pills 0.2 + 0.6 0.2 ± 0.7 Diuretics 0.1 ±0 .4 0.1 ±0.6 Vomiting 0.3 ± 0.9 0.3 ± 0.7 Laxatives 0.1 ±0.5 0.1 ±0 .4 Loss of Control over Eating 1.6 ± 1.3 1.5 ± 1.2 Crash Dieting 1.2 ± 1.3 0.8 ± 1.3a Mean ± S D ; means derived from: 'Never' - 0 to 'More than Once Everyday' = 5 (Binge-eating, Fasting, Diet Pills, Diuretics, Vomit ing, Laxatives); 'Never' = 0 to 'A l l the time' = 5 (Loss of control over eating); 'Never' - 0 to 'More than Once in the Last Month' = 4 (Crash Dieting). a Significantly different from Baseline: p = .002 (Paired sample t-Test with Bonferroni Correction for multiple comparisons). 81 Results Table 10: Correlations Between Baseline Characteristics of Study Participants and Symptom Intensity Score Symptom Intensity Change in Score Symptom Baseline Measures Baseline 6 Months Intensity Score* (n=179) (n=179) (n=179) Age (y) -.26tb -.25b .04 Weight (kg) .17^  .04 -.17a BMI (kg/m2) .17a .07 -.13 *Corrected Symptom Intensity Score = Baseline Symptom Intensity Score - 6 Month Symptom Intensity Score. (Symptom Intensity Score determined from responses to the Health Information Questionnaire) t Pearson Product Moment correlations (r) S ign i f icant correlation: ap<.05, r jp<.01 82 Results Table 11: Correlation between Baseline Characteristics of Study Participants and Use of Binge Eating and Compensatory Behaviours Baseline Measures Binge Eating and Age (y) Weight (kg) BMI (kg/m2) Compensatory Behaviours (n=179) BASELINE Binge-eating -.02* .23b .20b Loss of Control over Eating -.12 .11 .13 Crash dieting -.25b .05 .04 Diet Pills -.12 -.04 -.03 Diuretics -.07 .08 .06 Fasting -.01 -.04 -.03 Laxatives -.07 -.06 -.04 Vomiting -J3 £ 3 .04_ 6 MONTHS Binge-eating -.12 -.02 -.03 Loss of Control over Eating -.08 .12 17a Crash dieting -.20b -.04 -.08 Diet Pills -.05 -.02 -.02 Diuretics .002 -.07 -.07 Fasting .04 -.10 -.12 Laxatives -.12 -.06 -.08 Vomiting -J5^ -JTL -.04 * Pearson Product Moment correlations (r) S ign i f icant correlation: a p<.05, bp<.01 83 Results between binge-eating and weight and B M I were also observed. In addition, B M I correlated with 'loss of control over eating' at 6 months. 4. Relationship Between Duration of Dieting and Eating Disorder Symptomology A. Duration of Dieting Table 12 reports the duration of dieting variables at baseline and 6 months. Overall, the subjects in this sample began dieting at age 15, had been dieting for four years and had dieted for approximately half of the previous six months. There were no significant differences between baseline and 6 months measures of 'age at first diet', 'long-term dieting' duration, and 'short-term dieting' duration. B. Correlation Between Duration of Dieting and Symptom Intensity Score Table 13 reports the correlations between the duration of dieting variables measured at baseline and the Symptom Intensity Scores at baseline and 6 months and the change in Symptom Intensity Score. The age at first diet variable was negatively correlated with Symptom Intensity Score at baseline and 6 months. Long-term dieting duration correlated with Symptom Intensity Score at baseline, but not at 6 months or with Change in Symptom Intensity Score. No significant correlation between short-term dieting and Symptom Intensity Score or Change in Symptom Intensity Score was observed. C. Correlation Between Duration of Dieting and Binge-eating and Compensatory Behaviours Table 14 presents the correlations between duration of dieting measured at baseline and use of binge-eating and compensatory behaviours at baseline and 6 months. At baseline, negative correlations were found between age at first diet and 'crash dieting' and feelings of 'loss of control over eating'. Only 'loss of control over eating' remained negatively correlated with age at first diet at 6 months. Correlations were observed between feelings of 'loss of control over eating' and 84 Results Table 12: Self-Reported Duration of Dieting Variables in Study Participants Dieting Duration Variables Baseline 6 Months (n=179) (n=179) Age at First Diet (y) 15.4 ±3 .5* 15.2 ±3 .4 Long-term Dieting (y) 4.1 ±3 .2 4.3 ±3 .2 Short-term Dieting (676 mo.) 85 ± 6 1 77 ± 6 2 Mean ± S D N o Significant difference found in paired sample t-Test Table 13: Correlation between Duration of Dieting Variables and Symptom Intensity Score in Study Participants Symptom Intensity Change in Score Symptom Baseline Dieting Baseline 6 Months Intensity Score Duration Measures (n=179) (n=179) (n=179) Age at first diet (y) -.31t b -.20 b .15 Long-term Dieting (y) .16 a .05 -.13 Short-term Dieting (d/6 mo.) 10 .03 -.10 Corrected Symptom Intensity Score = Baseline Symptom Intensity Score - 6 Month Symptom Intensity Score (Score determined from the Health Information Questionnaire), t Pearson Product Moment correlations (r) S ign i f icant correlation: ap<.05, bp<.01 85 Results Table 14: Correlation between Duration of Dieting Variables and Use of Binge Eating and Compensatory Behaviours in Study Participants Baseline Measures Age at First Long-term Short-term Binge Eating and Compensatory Behaviours Diet(y) Duration of Dieting (y) (n=179) Duration of Dieting (d/6 mo.) BASELINE Binge-eating -.10* .04 -.10 Loss of Control over Eating -.27b .14 -.01 Crash dieting -.23c .13 .05 Diet Pills -.02 -.01 .02 Diuretics -.08 .07 -.01 Fasting -.05 .07 .02 Laxatives .05 -.05 .17 Vomiting -.11 .12 .2lb 6 MONTHS Binge-eating -.08 .06 .01 Loss of Control over Eating -.18a .15a -.05 Crash dieting -.11 -.11 -.05 Diet Pills .09 -.08 -.05 Diuretics -.03 .06 -.03 Fasting -.01 .09 -.14 Laxatives .03 -.07 .04 Vomiting -.13 .10 .17a Pearson Product Moment correlations (r) S ign i f icant correlation: ap<.05, b p<.01, cp<.001 86 Results long-term dieting frequency at 6 months. Use of vomiting to control weight correlated positively with short-term dieting frequency at baseline and 6 months. D. Regression of Baseline Duration of Dieting on Symptom Intensity Score Multiple regression, using baseline age at first diet and long- and short-term dieting duration, was used to predict Change in Symptom Intensity Score. As seen in Table 15, none of the duration of dieting variables predicted Change in Symptom Intensity Score and the regression equation did not attain significance, F (3, 159) = 2.14, NS. 5. Relationship Between Dieting Methods and Eating Disorder Symptomology A. Dieting Methods Appendices M and N present a summary of the dieting methods used by the subjects at baseline and 6 months, respectively. At baseline, 7.9% of the sample (combined frequency of 'Often' to 'Always') obtained dieting information from books, magazines or newspapers, 20.5% received dieting information from a health professional and 6.2% of the sample used dieting information acquired through television, video, friends or family. Eating smaller portions and avoiding certain foods was reported by 69.7% of the sample and creating own diet ("a diet that you have made up") was reported by 68.7% of the sample. Responses to the remaining dieting methods were as follows: counting grams of fat, 36.3%; skipping meals, 27.4%; counting calories, 20.2%; avoid eating when hungry, 12.9%; using a diet that is part of a commercial weight loss program, 5.0%; using a liquid formula or drink that replaces a meal(s), 4.5%; and fasting, 2.8%. At 6 months, 4.5% of the sample (frequency of 'Often' to 'Always') obtained dieting information from books, magazines or newspapers, 21.3% received dieting information from a health professional and 5.0% of the sample used dieting information heard about on television, video or from friends or family. Responses to the remaining dieting methods were as follows: eating smaller portions and avoiding certain foods, 63.1%; creating own diet, 60.9%; counting grams of 87 Results Table 15: Multiple Regression Analysis of Baseline Duration of Dieting Variables on Change in Symptom Intensity Score in Study Participants Duration of Dieting Variables B Beta t Age at First Diet (y) 0.30 0.14 1.50 Long-Term Dieting (y) -0.98 1.85 -0.53 Short-Term Dieting (d/6 mo.) -0.01 -0.11 -1.35 R2=0.04, F(3,159)=2.14 Constant = -4.32 N o significant difference found 88 Results fat, 36.5%; counting calories, 15.0%; skipping meals, 13.9%; avoid eating when hungry, 7.8%; using a diet that is part of a commercial weight loss program, 2.8%; liquid formula or drink that replaces a meal(s), 2.8%; and fasting, 1.7%. Table 16 presents the mean responses to the dieting methods measured at baseline and six months. There was a significant decrease in the total number of dieting methods used over the 6 month study period. In addition, there were significant reductions in the use of commercial weight-loss programs, use of liquid formulas, skipping meals, counting calories, and avoiding eating when hungry. However, no significant differences in frequency of counting grams of fat, creating own diet, consuming smaller portions or avoiding certain foods, fasting, or in obtaining dieting information from the various sources were observed over the study period. The Dieting Practices Survey provided an opportunity for the participant to list any other dieting method(s) used to lose weight. Table 17 presents a summary of the 'other' methods reported by the subjects at baseline and 6 months. Seventy-six subjects at baseline and 56 subjects at 6 months indicated that they used a dieting method instead of, or in addition to, the dieting methods listed in the Dieting Practices Survey. The most frequently cited methods included using calorie- or fat-reduced foods, filling up on water or low-calorie beverages, and eliminating red meat from the diet. B. Factor Analysis of Dieting Methods Principal Components Factor Analysis was used to identify underlying constructs and group dieting method variables. All dieting methods, except 'consuming smaller portions/avoiding certain foods', showed a significant positive relationship with the Symptom Intensity Score. Bartlett's Test of Sphericity (286.6, p<.000001) was used to confirm that the dieting methods correlation matrix was not an identity matrix. The Kaiser-Meyer-Olkin measure of sampling adequacy was sufficiently high (.603) to proceed with the analysis. Information received from a 89 Results Table 16: Self-Reported Use of Dieting Methods by Study Participants Dieting Methods Baseline 6 Months (n=179) (n=179) Total Number of Dieting 5.1 ± 2 * 4.4 ± 2.0a Methods Used a) Diet from a book, newspaper 0.9 ± l.Ot 0.8 ±1 .0 or magazine b) Information received from a 1.4 ± 1.2 1.4 ± 1.2 health professional c) Diet heard about on TV, 0.8 + 1.1 0.7 ± 1.0 video, or from friends/family d) A diet that is part of a 0.4 ± 0.9 0.2 ± 0.6a commercial weight-loss program e) A diet that you have made up 2.8 ± 1.1 2.5 ± 1.2 f) A liquid formula that replaces 0.5 + 0.9 0.3 ± 0.7a meals g) Smaller portions/ avoids or 2.9 ± 1.0 2.7+1.1 cuts down on certain foods h) Skips meals 1.6 ± 1.2 1.3 ± l . l a i) Counts calories 1.2+1.4 0.9 ± 1.2a j) Avoids eating when hungry 1.2 ± 1.1 1.0+ 1.0a k) Fasts (goes without food for 0.4 ±0.8 0.3 ±0 .6 more than 24 hours) 1) Counts grams of fat 1.8 ± 1.6 1.7 ± 1.5 Mean ± S D | M e a n derived from: 'Never' = 0 to 'Always' = 4. a Significantly different from baseline value: ap<.001 (paired sample t-Test with Bonferroni Correction for multiple comparisons). 90 Results Table 17: 'Other' Dieting Methods Reported by Study Participants 'Other* Reported Dieting Methods* Baseline 6 Months (n = 76) (n = 56) Uses Calorie- or Fat-reduced Foods 18 n 15 Fills up on water or low-calorie beverages 15 9 Eliminates Red Meat from Diet 8 7 Eats "healthy"/increased vegetable intake 6 7 Eats only one food for extended periods 4 3 Avoids eating after a certain time in the evening 4 2 Avoids all animal products/"gone vegetarian" 3 2 Avoids snacking 3 — Food combining 3 — Uses Fibre pil!s/"colonics" 3 — Uses food diary/Prepares eating schedule 2 2 Uses Chinese herbal tea 2 1 Consumes small, frequent meals 1 1 Uses weight-loss gum 1 1 Goes on a week-long "cleanse" 1 1 Eats up to a certain energy level then stops 1 — Avoids being at home 1 — Eats out less frequently — 1 Uses smaller plates or dishes — 1 Follows Overeaters Anonymous (12 Steps) — 1 Uses Prozac to lose Weight — 1 •Responses derived from open-ended question "List any other dieting methods used to lose weight". Results health professional was eliminated from the matrix due to its small values for the measure of sampling adequacy. Factor loadings less than 0.5 were omitted in the sorting. Factor analysis yielded five factors with eigenvalues greater than one, accounting for 69% of the total variance. (Eigenvalues represent the total amount of variance explained by a factor (Munro et al., 1986)). Factor labels were based on the common meanings of the grouping of variables. The factors include: Factor 1, labeled 'Food Avoidance Behaviours', consisted of skipping meals, avoiding eating when hungry, and fasting; Factor 2, labeled 'Fat & Calorie Counting', consisted of counting calories and counting grams of fat; Factor 3, labeled 'Use of Diet Aids', consisted of using a diet plan that is part of a commercial weight-loss programs and using a liquid formula or drink that replaces a meal or several meals; Factor 4, labeled 'Information from Popular Media', consisted of using a diet(s) read about in newspapers, books or magazines and using information or advice heard about on TV, videos or from friends/family; and Factor 5, labeled 'Self-Selected Dieting', consisted of creating own diet and eating smaller portions/avoids certain foods. C. Correlation Between Dieting Method Factors and Symptom Intensity Score Table 18 reports the correlations between baseline dieting method factors and number of dieting methods used and the Symptom Intensity Scores at baseline and 6 months and the change in Symptom Intensity Score. Food Avoidance, Fat & Calorie Counting, Number of Methods, and Use of Diet Aids correlated with Symptom Intensity Score at baseline and 6 months. Only Number of Dieting Methods Used demonstrated a correlation with Change in Symptom Intensity Score. D. Correlation Between Dieting Method Factors and Binge-Eating and Compensatory Behaviours Table 19 presents the correlations between baseline dieting method factors and number of dieting methods used and use of binge eating and compensatory behaviours at baseline and 6 months. Multiple correlations were observed. Loss of control over eating, crash dieting, fasting and 92 Results Table 18: Correlation between Dieting Method Factors and Symptom Intensity Score and Change in Symptom Intensity Score in Study Participants Symptom Intensity Change in Score Symptom Dieting Method Factors* (Baseline) Baseline 6 Month Intensity Scoret (n=179) (n=179) (n=179) Food Avoidance .51§ c .49c -.05 Fat & Calorie Counting .31c .30c -.01 Use of Diet Aids .27c .22b -.07 Information from Popular Media .14 .12 -.04 Self-Selected Dieting .03 .06 -.10 Number of Methods Used .60c .49c -.16a Dieting Method Factors: 1) Food Avoidance Behaviours: includes skipping meals, avoiding eating when hungry, and fasting 2) Fat & Calorie Counting: includes counting calories and counting grams of fat 3) Use of Diet A ids: includes use of a diet plan that is part of a commercial weight-loss programs and use of a liquid formula or drink that replaces a meal or several meals 4) Information from Popular Media: includes using a diet(s) read about in newspapers, books or magazines and using information or advice heard about on T V , videos or from friends/family; 5) Self-Selected Dieting: includes creating own diet and eating smaller portions/avoids certain foods. "("Corrected Symptom Intensity Score = Baseline Symptom Intensity Score - 6 Month Symptom Intensity Score (Score determined from the Health Information Questionnaire). §Pearson Product Moment correlations (r) S ign i f icant correlation: a p<.05, bp<.01, cp<.001 93 Results Table 19: Correlation between Use of Binge Eating and Compensatory Behaviours and Dieting Method Factors in Study Participants Binge Eating and Compensatory Behaviours Food Avoidance Baseline Dieting Method Factors Fat& Use of Informatio Self-Calorie Diet n from Selected Counting Aids Media Dieting (n=174) Number of Dieting Methods BASELINE Binge-eating .10* .12 .07 .02 -.14 .12 Loss of Control .27 c .28 c .09 .13 -.16 a .29 c over Eating Crash dieting .53 c .24b .18 a .07 .10 .47 c Diet Pills .12 .16 a .22 b .16 a .02 .35 c Diuretics .19 a .19 a .10 .08 -.04 .19 a Fasting .64 c -.09 .20 -.07 -.04 .38 c Laxatives .12 .11 .18 a .02 .004 .30 c Vomiting .15 a .15 .11 .03 -.04 .22b 6 MONTHS Binge-eating .19 a .19 a .12 .15 -.16 a .22b Loss of Control .27 c .25 b .18 a .14 -.14 .31 a over Eating Crash dieting .48 c .19 a .02 .11 .04 .34 c Diet Pills .17 a .10 .19 a .10 -.11 .25 b Diuretics .12 .24b -.02 .04 -.14 .10 Fasting .47 c -.07 .12 -.06 -.09 .23b Laxatives .25b .19 a .20 b -.08 -.09 .28 c Vomiting .25b .26b .18 a .05 -.10 .33 c Pearson Product Moment correlations (r) S ign i f icant correlation: ap<.05, b p<.01, cp<.001 94 Results vomiting were correlated with Food Avoidance dieting methods both at baseline and 6 months. Use of diuretics correlated with Food Avoidance dieting methods only at baseline while use of diet pills and laxatives correlated with this factor only at 6 months. Feelings of loss of control over eating, crash dieting, and diuretic use correlated with Fat & Calorie Counting both at baseline and 6 months. At baseline, use of diet pills, and at 6 months, laxative use and vomiting to control weight was significantly correlated with Fat & Calorie Counting. Use of diet pills and laxatives were correlated with the Use of Diet Aids both at baseline and 6 months. Significant correlations were also observed between Use of Diet Aids and crash dieting (baseline only) and loss of control over eating and vomiting to lose weight (6 months only). Diuretic use and obtaining Information from Popular Media was correlated at baseline. Negative correlations were observed between Self-Selected Dieting and loss of control over eating (baseline only) and binge-eating (6 months only). A l l compensatory behaviours, except use of diuretics at 6 months, were positively correlated with number of dieting methods used at baseline. Binge-eating was significantly correlated with the number of dieting methods used only at 6 months. E. Regression of Dieting Method Factors on Symptom Intensity Score Multiple regression analysis of baseline dieting method factors and number of dieting methods used at baseline was used to predict Change in Symptom Intensity Score. Results are presented in Table 20. The number of dieting methods used at baseline was significantly correlated to the Change in Symptom Intensity Score; the overall regression equation, however, did not reach significance, F (6, 166) = 1.42, NS. 6. Relationship Between Macronutrient Intake and Eating Disorder Symptomology A. Macronutrient Intake Macronutrient intake was derived from the Food Frequency Questionnaires. Questionnaires that appeared inaccurate or incomplete (n = 5) were excluded from the analysis. Analyses are based on a sample size of 174. 95 Results Table 20: Multiple Regression Analysis of Baseline Dieting Method Factors on Change in Symptom Intensity Score in Study Participants Dieting Methods Factors B Beta t Food Avoidance 0.95 0.13 1.24 Fat /Calorie Counting 1.07 0.15 1.48 Diet Aids 0.66 0.10 0.92 Information from the Media 0.71 0.10 1.06 Self-Selected Dieting -0.23 -.03 -0.40 Number of Methods -1.36 -0.38 - 2 . 3 2 a R2 = 0.05, F(6 ,166)= 1.42 Constant = 5.63 Significant correlation: p<0.05 96 Results Table 21 presents the mean intake (g) and relative intake (%) of protein, carbohydrate and fat per day over the previous 6 months as estimated by the FFQ. On average, the dieting women in the sample consumed 1900 kcal with. 58% of the energy derived from carbohydrate, 16% from protein, and 26% from fat. B. Correlation Between Macronutrient Intake and Symptom Intensity Score Table 22 reports the correlations between the baseline macronutrient intake, and percent of energy derived from the macronutrients, and the Symptom Intensity Scores at baseline and 6 months, and the Change in Symptom Intensity Score. A significant negative relationship was observed between grams of carbohydrate consumed at baseline and Change in Symptom Intensity Score. This suggests that as carbohydrate intake increased, symptom intensity score likewise increased. No other significant relationships were observed. C. Correlation Between Macronutrient Intake and Binge-Eating and Compensatory Behaviours Table 23 presents the correlations between baseline macronutrient intake and use of binge eating and compensatory behaviours at baseline and 6 months. A lower total energy and carbohydrate intake at 6 months, and a lower fat intake at baseline, was associated with the use of diet pills to control weight. Lower absolute carbohydrate and protein intakes, were associated with fasting behaviour at six months. A higher relative carbohydrate intake was associated with use of vomiting to lose weight at 6 months. In addition, consuming a greater proportion of protein as energy was associated, both at baseline and 6 months, with use of diet pills and laxatives to control weight. A lower fat intake, both absolute and relative, was associated with loss of control over eating and a greater frequency of vomiting. 97 Results Table 21: Baseline Macronutrient Intake of Study Participants Determined by Food Frequency Questionnaire Macronutrient Intake (n=174) Energy (kcal) 1915 ±657* Carbohydrate (g) 285 ±106 % Carbohydrate 5 8 ± 9 t Protein (g) 78 ± 3 0 % Protein 16 ± 3t Fat(g) 56 ± 2 6 % Fat 2 6 ± 7 t *Mean ± S D ; Determined by the Food Processor II Nutrient Analysis Software ( E S H A Research, Salem, O R ) ^Calculated as energy (kcal) from C H O (g/d x 4 kcal/g), protein (g/d x 4 kcal/g), and fat (g/d x 9 kcal/g) divided by the total energy intake. 98 Results Table 22: Correlation between Macronutrient Intake and Symptom Intensity Score in Study Participants Macronutrient Intake* Symptom Intensity Change in (Baseline) Score Symptom Baseline 6 Months Intensity Score! (n=174) (n=174) (n=174) Energy Intake (kcal) -.04§ -.06 -.15 CHO (g) .08 -.04 -.16a % Energy from Carbohydrate^ .11 .08 -.02 Protein (g) .03 -.06 -.13 % Energy from Proteinrj: -.01 .02 .03 Fat(g) -.07 -.15 -.12 % Energy from Fatrj: -.15 -.12 .02 •Macronutrient Intake determined by response to Food Frequency Questionnaire tCorrected Symptom Intensity Score = Baseline Symptom Intensity Score - 6 Month Symptom Intensity Score (Score determined from the Health Information Questionnaire). ^Pearson Product Moment correlations (r) X Calculated as energy (kcal) from C H O (g/d x 4 kcal/g), protein (g/d x 4 kcal/g), and fat (g/d x 9 kcal/g) divided by the total energy intake. S ign i f i can t Correlation: p<.05 99 Results Table 23: Correlation between Use of Binge Eating and Compensatory Behaviours and Macronutrient Intake in Study Participants Binge Eating and Compensatory Behaviours Energy (kcal) CHO (g) Baseline Macronutrient Intake CHO Protein Protein (%)* (g) (%)* (n=174) Fat (g) Fat (%)* BASELINE Binge-eating -13t .13 -.03 .11 -.02 .08 -.21 Loss of Control over Eating .01 .03 .07 .01 .02 -.07 -.11 Crash dieting .02 .05 .11 .01 -.01 -.08 -.13 Diet Pills -.11 -.09 .02 -.002 .23b -.16a -.11 Diuretics -.06 -.06 .02 -.03 .07 -.08 -.04 Fasting -.01 -.07 -.08 -.04 -.06 .04 .13 Laxatives -.03 -.0007 .04 .07 .23° -.09 -.13 Vomiting -.02 .003 .06 .01 .08 -.08 -.10 6 MONTHS Binge-eating .05 .04 .03 .06 .06 -.02 -.05 Loss of Control over Eating -.07 -.03 .13 -.06 .01 -.19a -.17a Crash dieting -.05 -.04 .09 -.05 -.01 -.11 -.08 Diet Pills -.15a -.16a -.07 -.05 .230 -.12 -.01 Diuretics -.01 .05 .13 -.05 -.07 -.10 -.14 Fasting -.14 -.17a -.07 -.18a -.12 -.05 .13 Laxatives -.05 -.06 -.06 .04 .18a -.04 -.01 Vomiting -.05 .01 .15a -.03 .05 -.17a -.19a *Calculated as energy (kcal) from C H O (g/d x 4 kcal/g), protein (g/d x 4 kcal/g), and fat (g/d x 9 kcal/g) divided by the total energy intake. tPearson Product Moment correlations (r) S ign i f i can t correlation: a p<.05, bp<.01 Results D. Regression of Macronutrient Intake on Symptom Intensity Score Multiple regression using baseline Macronutrient Intake was used to predict Change in Symptom Intensity Score. As seen in Table 24, no variable predicted Change in Symptom Intensity Score and the regression equation did not reach significance, F (3, 159) = 2.14, NS. 7. Relationship Between Restrictiveness of the Diet and Eating Disorder Symptomology A. Restrictiveness Variables Table 25 reports the mean estimated energy intake and estimated energy requirement, the restrictiveness of the diet, and rate of weight loss for the subjects. The restrictiveness was calculated as the estimated energy requirement minus the energy intake. As the FFQ was administered only at baseline, only one measure of intake and restrictiveness was calculated. A significant difference was observed in the rate of weight loss per week of dieting between baseline and 6 months. B. Correlation Between Restrictiveness of the Diet and Symptom Intensity Score Table 26 reports the correlations between the baseline restrictiveness of the diet and the rate of weight loss and the Symptom Intensity Scores at baseline and 6 months and the Change in Symptom Intensity Score. Restrictiveness did not correlate with Symptom Intensity Score or Change in Symptom Intensity Score. Rate of weight loss was correlated with Symptom Intensity Score at both baseline and 6 months. No correlation, however, was observed between rate of weight loss and Change in Symptom Intensity Score. 101 Results Table 24: Multiple Regression Analysis of Macronutrient Intake on Change in Symptom Intensity Score in Study Participants Macronutrient Intake Variables B Beta t Energy Intake (kcal) -.0004 -.55 -.84 Carbohydrate Intake (g) .09 .58 .71 % Carbohydrate Intake* -.01 -.21 -.97 Protein Intake (g) -.10 -.36 -.56 % Protein intake* .03 .16 .51 Fat Intake (g) 1.56 .64 1.14 % Fat Intake* -.03 -.44 .45 R 2 = 0.06, F(7,165)= 1.59 Constant = -.83 •Calculated as energy (kcal) from C H O (g/d x 4 kcal/g), protein (g/d x 4 kcal/g), and fat (g/d x 9 kcal/g) divided by the total energy intake. N o significant differences found 102 Results Table 25: Estimated Energy Requirement and Level of Energy Restriction of Study Participants Parameter Baseline 6 Months (n=174) (n=174) Energy Intake (kcal)* 1915 +657t — Estimated Energy Requirement (kcal)§ 2080 ± 300 — Restrictiveness^ 166 ±667 — Weight lost/week of dieting (kg) 0.69 ± .66 0.55 ± .63a •Der ived from the Food Frequency Questionnaire t M e a n ± S D §Calculated from the Mif f l in Equation (Miff l in et al., A m J C l in Nutr V o l . 51: 241-247, 1990) ^Calculated as the Estimated Energy Requirement less the Energy Intake Signi f icant ly different from Baseline: p < .01 (Paired sample t-Test) 103 Results Table 26: Correlation between Restrictiveness Variables and Symptom Intensity Score in Study Participants Symptom Intensity Change in Score Symptom Restrictiveness Variables Baseline 6 Months Intensity Score (n=174) (n=174) (n=174) Restrictivenesst .05§ .06 .03 Weight lost/week Dieting (kg) 3\b .33^ .01 Corrected Symptom Intensity Score = Baseline Symptom Intensity Score - 6 Month Symptom Intensity Score (Score determined from the Health Information Questionnaire). fCalculated as the Estimated Energy Requirement (calculated from the Mif f l in Equation (Miff l in et al., A m J Cl in Nutr V o l . 51: 241-247, 1990)) less the Energy Intake (derived from the Food Frequency Questionnaire) §Pearson Product Moment correlations (r) S ign i f icant Correlation: a p<.05, bp<.01 104 Results C. Correlation Between Restrictiveness of the Diet and Binge-Eating and Compensatory Behaviours The correlations between baseline restrictiveness of the diet, and rate of weight loss, and the use of binge eating and compensatory behaviours are presented in Table 27. No significant relationships were observed between binge-eating and compensatory behaviours and restrictiveness. However, rate of weight loss was found to correlate with binge eating (baseline only) and fasting (baseline and 6 months). D. Regression of Restrictiveness of the Diet on Symptom Intensity Score Multiple regression using baseline Dieting Restrictiveness was used to predict Change in Symptom Intensity Score. As seen in Table 28, no variable predicted Change in Symptom Intensity Score and the regression equation was not significant, F(2, 113) = 0.58, NS. 8. Vitamin/Mineral Supplementation Vitamin/Mineral supplementation practices were determined by the Food Frequency Questionnaire. Eighty-eight (49.2%) subjects used a vitamin/mineral supplement on a weekly or more frequent basis while 91 (50.8%) subjects did not use supplements on a regular basis. No significant differences in subject characteristics, binge-eating and compensatory behaviours, duration of dieting, dieting method factors, macronutrient intake, or restrictiveness was observed when the sample was divided based on use of vitamin/mineral supplements. There was, however, a significant difference in the Symptom Intensity Scores both at baseline (supplement users: 15.8; non-users: 18.8; F=4.7, p<.05) and 6 months (supplement users: 13.9; non-users: 18.2;F=10.8,p<.01). 105 Results Table 27: Correlation between Use of Binge Eating and Compensatory Behaviours and Baseline Restrictiveness Variables in Study Participants Baseline Measures Binge Eating and Restrictiveness Weight lost/week Compensatory Behaviours of the Diet Dieting (kg) (n=174) BASELINE Binge-eating -.081 .19a Loss of Control over Eating .01 .03 Crash dieting .02 .14 Diet Pills .11 -.13 Diuretics .12 -.03 Fasting -.02 .21a Laxatives .05 -.13 Vomiting .08 .07 6 MONTHS Binge-eating -.08 .13 Loss of Control over Eating .08 .03 Crash dieting .06 .06 Diet Pills .12 .01 Diuretics .02 .07 Fasting .09 .24b Laxatives .03 .10 Vomiting .08 -.0003 •Calculated as the Estimated Energy Requirement (calculated from the Mif f l in Equation (Miff l in et al., A m J C l in Nutr V o l . 51: 241-247, 1990)) less the Energy Intake (derived from the Food Frequency Questionnaire) fPearson Product Moment correlations (r) S ign i f icant Correlation: ap<.05, bp<.01 106 Results Table 28: Multiple Regression Analysis of Restrictiveness Variables on Change in Symptom Intensity Score in Study Participants Dieting Restrictiveness Variables 15 Beta t Restrictiveness* 0.001 0.10 1.04 Weight lost/wk Dieting (kg) 1.02 0.04 0.42 R 2 = 0.01,F(2, 113) = 0.58 Constant = -2.05 •Calculated as the Estimated Energy Requirement (calculated from the Mif f l in Equation (Miff l in et al., A m J C l in Nutr V o l . 51: 241-247, 1990)) less the Energy Intake (derived from the Food Frequency Questionnaire) N o significant differences found 107 Results 9. Vegetarianism In total, 22 (12.3%) subjects in the sample followed a vegetarian diet. Vegetarian status was determined at baseline by the Food Frequency Questionnaire. The vegetarian subjects were categorized as follows: lacto-ovo-vegetarian: 10 (5.6%); partial vegetarian: 11 (6.1%); vegan: 1 (0.6%). No significant differences in subject characteristics, Symptom Intensity Scores, binge-eating and compensatory behaviours, duration of dieting, dieting method factors, macronutrient intake, or restrictiveness was observed when the sample was divided based on vegetarian status (means weighted for unequal sample sizes). 10. Comparison of Dieting Practices in Subjects According to Dieting Status at 6 Months At 6 months, 125 (70.2%) subjects identified themselves as dieters while 53 (29.8%) subjects indicated that they were no longer dieting to lose or maintain their weight. One subject did not respond to this question. A. Age, Weight and BMI Table 29 presents the comparison of baseline characteristics of subjects who identified themselves as a dieter at 6 months and subjects who indicated that they were no longer dieting to lose or maintain their weight at 6 months. No significant differences in age, weight and B M I were observed. B. Eating Disorder Symptomology Table 30 presents the comparison of Symptom Intensity Score and Compulsive Eating Score of subjects according to dieting status at 6 months. Subjects who identified themselves as dieter at 6 months had significantly greater Symptom Intensity Scores and Compulsive Eating Scores both at baseline and at 6 months, than those subjects who indicated that they were no longer dieting. 108 Results Table 29: Comparison of Age and Baseline Weight, Height and BMI of Study Participants According to Dieting Status at 6 Months Dieting Status at 6 Months Baseline Characteristics Not Dieting Dieting (n=53) (n=125) Age (y) 22 ± 3.5* 21 ±3 .0 Height (cm) 166.5 ±6.9 164.8 ±7.1 Weight (kg) 64.7 ± 14.3 61.6 ±10.4 BMI (kg/m2) 23.3 ±4.3 22.6 ± 3.0 * M e a n ± S D ; comparisons made with M A N O V A (Hotelling's T Squared Statistic) N o significant differences between groups 109 Results Table 30: Comparison of Symptom Intensity Scores and Compulsive Eating Scores of Study Participants According to Dieting Status at 6 Months Symptom Intensity and Compulsive Eating Score51 Dieting Status at 6 Months Not Dieting Dieting (n=53) (n=125) Baseline Symptom Intensity Score Compulsive Eating Score 6 Months Symptom Intensity Score Compulsive Eating Score 14.7±8.6t 2.6 + 2.0 11.9 ±7.8 2.3 ±2 .0 1 8 . 5 ± 9 . 5 a 3 . 5 ± 2 . 3 a 18 .0±9 .0 C 3.5 ±2 .2° Symptom Intensity Score and Compulsive Eating Score determined from the Health Information Questionnaire ^Mean ± S D ; comparisons made with M A N O V A (Hotelling's T Squared Statistic) Significant difference between groups: F = 4.75, p<0.01 Signi f icant ly different than Not Dieting Group: ap<.05, b p<.01, cp<.001 110 Results Table 31 presents the comparison of baseline binge eating and compensatory behaviours of subjects who indicated they were continuing to diet at 6 months and subjects who discontinued dieting at 6 months. There were no significant differences in behaviours between groups. C. Duration of Dieting Table 32 presents the comparison of baseline duration of dieting of subjects who identified themselves as dieters at 6 months and subjects who indicated they were no longer dieting at 6 months. Subjects who were dieting at 6 months had dieted for more days during the six months prior to the study than the group of subjects that did not identify themselves to be a dieter at 6 months. D. Dieting Method Factors Table 33 presents the comparison of baseline dieting method factors of subjects according to dieting status at 6 months. Those subjects who identified themselves as dieters used a greater number of dieting methods overall, used Fat & Calorie Counting, and Self-Selected Dieting Methods more often than the group subjects who indicated they had discontinued dieting. E. Macronutrient Intake Table 34 presents the comparison of baseline macronutrient intake of subjects according to dieting status at 6 months. Those subjects who indicated they continued to diet at 6 months were differentiated from those who indicated they were not dieting at 6 months by a lower baseline energy, fat and carbohydrate intake, a higher baseline percent protein intake. F. Restrictiveness of the Diet Table 35 presents the comparison of baseline restrictiveness variables of subjects according to dieting status at 6 months. No significant difference in rate of weight loss was seen. Those that identified themselves as dieters, however, had a more calorie-restricted diet at baseline relative to their energy needs than those subjects who stopped dieting. I l l Results Table 31: Comparison of Baseline Binge Eating and Compensatory Behaviours of Study Participants According to Dieting Status at 6 Months Binge Eating And Dieting Status at 6 Months Compensatory Behaviours Not Dieting Dieting (n=53) (n=125) Binge-eating 1.9 ±1.4* 1.8 ± 1.2 Loss of Control over Eating 1.4 ± 1.0 1.6± 1.4 Crash dieting 0.8 ± 1.3 1.3 ± 1.4 Diet Pills 0.1 ±0 .3 0.2 ±0.7 Diuretics 0.1 ±0 .3 0.1 ±0 .4 Fasting 0.2 ±0.5 0.3 ± 0.7 Laxatives 0.0 ± 0.2 0.1 ±0 .5 Vomiting 0.1 ±0.5 0.4 ± 1.0 * M e a n ± S D ; comparisons made with M A N O V A (Hotelling's T Squared Statistic); means derived from: 'Never' = 0 to 'More than Once Everyday' - 5 (Binge-eating, Fasting, Diet Pills, Diuretics, Vomit ing, Laxatives); 'Never' = 0 to 'A l l the time' = 5 (Loss of control over eating); 'Never' = 0 to 'More than Once in the Last Month' = 4 (Crash Dieting). N o significant differences between groups 112 Results Table 32: Comparison of Baseline Duration of Dieting Variables of Study Participants According to Dieting Status at 6 Months Baseline Duration of Dieting Dieting Status at 6 Months Not Dieting Dieting (n=53) (n=125) Age at First Diet (y) 15.4 ±3.2* 15.5 ±3.7 Long-Term Dieting (y) 4.0 ± 3.6 4.0 ± 3.0 Short-Term Dieting (d/6 mo.) 56.7 + 61.2 94.9 + 57.3a * M e a n ± S D ; comparisons made with M A N O V A (Hotelling's T Squared Statistic) Significant differences between groups: F = 5.32, p<0.01 Signi f icant ly different than Not Dieting Group: ap<.001 113 Results Table 33: Comparison of Baseline Dieting Method Factors of Study Participants According to Dieting Status at 6 Months Dieting Status at 6 Months Baseline Dieting Not Dieting Dieting Method Factors* (n=53) (n=125) Food Avoidance -0.2 ± 0.9t 0.1 ± 1.1 Fat /Calorie Counting -0.4 ± 1.0 0.1 ± 1.0° Diet Aids 0.04 ± 1.0 -0.01 ± 1.0 Information from the Media -0.2 + 0.9 0.1 ± 1.0 Self-Selected Dieting -0.3 ± 1.0 0.1 ± 1.0a Number of Methods 4.4±2.1§ 5.4 ± 1.8a Dieting Method Factors: 1) Food Avoidance Behaviours: includes skipping meals, avoiding eating when hungry, and fasting 2) Fat & Calorie Counting: includes counting calories and counting grams of fat 3) Use of Diet Aids: includes use of a diet plan that is part of a commercial weight-loss programs and use of a liquid formula or drink that replaces a meal or several meals 4) Information from Popular Media: includes using a diet(s) read about in newspapers, books or magazines and using information or advice heard about on T V , videos or from friends/family; 5) Self-Selected Dieting: includes creating own diet and eating smaller portions/avoids certain foods. f M e a n ± S D (Factor Scores); comparisons made with M A N O V A (Hotelling's T Squared Statistic) Significant differences between groups: F = 3.41, p< 0.01 Significantly different than Not Dieting Group: a p<.05, bp<.01 § M e a n ± S D 114 Results Table 34: Comparison of Baseline Macronutrient Intake of Study Participants According to Dieting Status at 6 Months Dieting Status at 6 Months Macronutrient Intake* Not Dieting Dieting (n=53) (n=125) Energy Intake (kcal) 2075 ± 632§ 1 8 3 8 ± 6 5 2 a Carbohydrate Intake (g) 315 ±111 270 ± 100a % Carbohydrate Intakef 58 ± 9 57 ± 8 Protein Intake (g) 81 ± 2 9 77 ±31 % Protein Intake! 15 ± 3 17 ± 3 ° Fat Intake (g) 61 ± 2 0 54 ± 28a % Fat Intakef 26 ± 6 26 ± 7 •Der ived from the Food Frequency Questionnaire § M e a n ± S D ; comparisons made with M A N O V A (Hotelling's T Squared Statistic) Significant differences between groups: F = 3.01, p< 0.01 Signi f icant ly different than Not Dieting Group: ap<.05, bp<.01 tCalculated as energy (kcal) from C H O (g/d x 4 kcal/g), protein (g/d x 4 kcal/g), and fat (g/d x 9 kcal/g) divided by the total energy intake. 115 Results Table 35: Comparison of Restrictiveness Variables of Study Participants According to Dieting Status at 6 Months Dieting Status at 6 Months Baseline Restrictiveness Variables Not Dieting Dieting (n=53) (n=125) Restrictiveness* 40 ± 599f 226 ± 685a kg lost/wk Dieting (kg) 0.74 + 0.77 0.68 ± 0.65 •Calculated as the Estimated Energy Requirement (calculated from the Mi f f l in Equation (Miff l in et al., A m J Cl in Nutr V o l . 51: 241-247, 1990)) less the Energy Intake (derived from the Food Frequency Questionnaire) f M e a n ± S D ; comparisons made with M A N O V A (Hotelling's T Squared Statistic) Significant differences between groups: F = 3.96, p<.05 Signi f icant ly different than Not Dieting Group: ap<.01 116 Discussion & Conclusions V Discussion and Conclusions Dieting is a common practice among college-age women. It has been speculated that dieting, and some specific dieting practices, (such as the duration of dieting, dieting method, diet composition and restrictiveness of the diet) may predispose susceptible individuals to unhealthy weight loss behaviours. Such behaviours may mark the early stages of serious health problems including eating disorders. Examination of the role of these dieting characteristics in the development of eating disorders, however, has received little attention in current research. The purpose of the present study was to determine, in female college or university students who were dieting, the long- and short-term duration of dieting, the methods used to lose weight, the macronutrient intake, and the restrictiveness of the diet (the energy intake, the extent to which the energy intake deviates from the energy requirement, and the rate of weight loss). A second purpose was to investigate whether any of these specific dieting practices, measured at the beginning of the university year, could predict eating disorder symptomology at the end of the school year, six months later. 1. Major Findings Subject Characteristics The majority of the dieting women were at a normal weight or were underweight. However, those with a higher weight and Body Mass Index (BMI) were more likely to have a higher Symptom Intensity Score and engage in binge-eating or feel a loss of control over eating. In addition, weight predicted an increase in Symptom Intensity Score. Eating Disorder Symptomology A high frequency of disordered eating was observed in this sample of dieting women. At the end of the study, approximately 12% of the subjects responded in a way to yield a probable diagnosis of Eating Disorders Not Otherwise Specified, while 4% of subjects appeared to progress to more 117 Discussion & Conclusions clinical forms of eating disorders, predominately bulimia nervosa. Over the study time period, however, the mean Symptom Intensity Score decreased and use of crash dieting to lose weight was significantly reduced. Frequency of binge-eating and all other compensatory behaviours did not change. Duration of Dieting Subjects who had a longer dieting history, both in terms of beginning to diet at a younger age and the total number of years dieting, had higher Symptom Intensity Scores and were more likely to use crash dieting to lose weight and feel a loss of control over eating. More frequent dieting over the previous 6 months was associated with the use of vomiting to control weight. Dieting Methods The majority of the subjects reported using moderate dieting methods (eating smaller portions and avoiding certain foods, creating own diet, and counting grams of fat) and received weight loss information from a health professional. Use of a greater number of dieting methods predicted an increase in Symptom Intensity Score and was significantly associated with binge-eating and most compensatory behaviours. Principal Component Factor analysis of the dieting methods resulted in the identification of several dieting factors. Of these, greater use of Food Avoidance dieting methods, rigid control over Fat and Calorie Counting, and the use of Diet Aids, were associated with higher Symptom Intensity Scores. Macronutrient Intake On average, the dieting women in the sample consumed a reasonable diet with regards to energy intake and macronutrient distribution. No significant correlations between Symptom Intensity Score and macronutrient intake were observed. However, carbohydrate intake and Change in Symptom Intensity Score were correlated. In addition, a macronutrient intake that combined a 118 Discussion & Conclusions lower fat and carbohydrate intake with a higher protein intake appeared to be a pattern of intake associated with abnormal eating behaviours. Restrictiveness of the Diet No relationship between the restrictiveness of the diet and Symptom Intensity Score was observed. However, rapid weight loss was associated with a higher Symptom Intensity Score, fasting to lose weight, and binge-eating. Regression of Dieting Variables on Symptom Intensity Score None of the dieting variables (duration of dieting, dieting method, macronutrient intake and restrictiveness of the diet) had a significant influence on the prediction of Symptom Intensity Score. Only between 1-5% of the variance in the Symptom Intensity Score was explained by the variables examined in this study. Comparison of Baseline Dieting Practices in Subjects According to Dieting Status at 6 Months Significant differences were observed in baseline variables between subjects who identified themselves as dieters at follow-up and subjects who indicated they were no longer dieting. Subjects who continued dieting had greater initial Symptom Intensity Scores, dieted for a greater number of days, and used more dieting methods (in particular, Fat and Calorie Counting, and Self-Selected Dieting) over the previous 6 months than those subjects who were not dieting. The composition of the diet was also different. For those subjects who continued to diet, baseline energy, carbohydrate and fat intake was lower, the calorie restriction of the diet was higher, and the percent protein intake was higher. 119 Discussion & Conclusions 2. Sample Characteristics A. Sample Characteristics The anthropometric measurements of the sample compare well with data from other studies in female college students (Davis et al., 1993; Drewnowski et al., 1988; Drewnowski et al., 1994; Frank et al., 1991; Horner & Utermohlen, 1993; Schotte & Stunkard, 1987). Enrollment data available from Langara Community College, Vancouver Community College (including King Edward Campus and City Centre Campus) and Simon Fraser University indicated that the majority of female students attending these institutions in the Fall of 1994 were between the age of 20-24 (Simon Fraser University, Office of Analytical Studies, Personal Communication, Jan. 1995; Ministry of Skills, Training and Labour, Post Secondary Education Division, Personal Communication, Jan. 1995; Office of Institutional Research, Vancouver Community College, Fact File. 1995). This is in close agreement with the average age of participating students in the present study. B. Age Age did not predict a Change in Symptom Intensity Score. There appears, however, to be some transition in the approach to dieting from adolescence to adulthood as evidenced by the negative correlation between age and Symptom Intensity Score. Younger subjects were more likely to use more extreme weight loss methods, such as crash dieting and vomiting, than the older subjects. A similar trend was noted by Frank et al. (1991) in female college students. Frank et al. (1991) reported that the use of abnormal weight loss behaviour, particularly the use of vomiting and diet pills to lose weight, decreased as the students matured from adolescence to young adulthood. C. Weight and BMI The current cultural preoccupation with thinness may promote the weight and body image disturbances that lead to dieting. Weight dissatisfaction and dieting usually accompany subjective impressions of feeling fat even in the absence of realistic reason for such feelings 120 Discussion & Conclusions (Striegel-Moore et al., 1986). This is evident in the present study by the non-concordance between dieting status and body weight. A large percentage of the participants had no weight-related reason to be dieting. Approximately 18% of the sample were underweight (BMI < 20) and 71% were at a normal B M I (20 < 25). At 6 months, 15% of the subjects were classified as underweight, and 73% were normal weight. It must be noted that in 8% of subjects at baseline and 16% of subjects at 6 months, weight was self-reported. Due to the tendency of women to underestimate their weight and overestimate their height, it is possible that the mean B M I in the present study was slightly underestimated. The reliability of self-reported weight in those subjects is unknown. However, when the self-reported body weight data at baseline and 6 months were excluded, there were no significant differences in the distribution of women who were underweight, normal, overweight, or obese. The proportion of dieting women who were underweight was similar to the results of Stephenson et al. (1987) who found that of those women between the ages 18-29 years who were trying to lose weight, 19% were 'lean' (BMI <20.5). The present estimates for dieting in the normal weight women (71%), however, are greater than other published reports (Levy et al., 1993; Serdula, 1993; Stephenson et al., 1987) in which dieting in normal weight women, aged 18 and older, varied from 44% to 62%. The larger estimates obtained in the present study may be due to a number of reasons. Perhaps the larger estimates reflect an increase in 'normative' dieting in this weight group. Conversely, the inclusion of 'weight maintenance' in the definition of dieting may have led to a greater estimate of dieting behaviour than those studies that define dieting only by attempts at weight loss (Levy et al., 1993; Stephenson et al., 1987). Differing procedures for categorizing weight groups (BMI, Metropolitan life tables) make comparison between studies difficult. And finally, use of the B M I may also have affected results. The B M I has been considered the most useful measure of weight status for individuals from the age 20 to 65 years (Health and Welfare Canada, 1988). In the current sample, approximately 42% were under 20 121 Discussion & Conclusions years old. At this age, growth is variable making the B M I an inconsistent measure when used as a comparison tool. In a survey of college women, Megel et al. (1994) reported that female students gained an average 1.1 kg over the academic term. In the present study, the participants maintained their weight over the term and did not gain a significant amount of weight overall. Maintenance of weight may, indeed, be related to the dieting behaviour of the sample. It is likely that some subjects may have lost and regained weight throughout the year. As only pre and post-study measurements were obtained, the true variability may have been obscured. However, when the first year students were examined separately, a weight gain of 0.7 kg over the 6 months was noted. Results from both the present study and that of Megel et al. (1994), do not give credibility to the myth of the "freshman five", the weight gain commonly associated with starting college or university. Body weight was inversely correlated with Change in Symptom Intensity Score. This suggests that those dieting women who had a larger body weight were more likely to experience a worsening of eating disorder symptomology over the academic term. In addition, small but significant correlations between B M I and binge-eating and, at 6 months, loss of control over eating were observed. This is in agreement with previous work that found a correlation between weight and the prevalence and frequency of binge-eating (Telch et al., 1988). As weight is a central component of body image and self-esteem, the more overweight a woman felt she was, the less physically attractive she felt (Megel, 1994). This may precipitate extreme weight loss behaviours. The observed relationship between weight and eating disorder symptomology in the present study, however, was expected to be stronger. As the present study examined both probable diagnoses of anorexia nervosa (characterized by a low body weight) and bulimia nervosa (characterized by a normal or above normal weight), this relationship may have been obscured. 122 Discussion & Conclusions D. Vegetarianism As a vegetarian diet is often adopted by young women as a means of losing or controlling weight, it was expected that a large percentage of subjects may endorse vegetarianism. While it is evident that some women in the sample regarded red meat as 'fattening' (approximately 5% of the sample indicated that they eliminated red meat from the diet and 2% avoided all animal products as a means of controlling their weight), only 12.3 % of the sample were classified as vegetarian. A vegetarian lifestyle was not associated with Symptom Intensity Score. 3. Eating Disorder Symptomology A. Eating Disorder Diagnoses The prevalence of probable anorexia nervosa and bulimia nervosa in this sample of dieting women was 0.5% and 3.4%, respectively. However, 14% of the sample at baseline and 12% at 6 months were classified as probable Eating Disorder Not Otherwise Specified. While this incidence rate of probable anorexia nervosa and bulimia nervosa is comparable to that in the general population (Drewnowski, 1988) it is notably lower than the incidence rate reported in a group of dieters. In a study of high school students (Patton et al., 1990), the majority of students classified as dieters did not develop an eating disorder at one year follow-up. However, 21% of the dieters were diagnosed as having an eating disorder (bulimia nervosa) at follow-up. Indeed, Patton et al. suspected that this was an underestimate as the population had not passed though the age of risk for eating disorders. Some movement between probable eating disorder diagnostic categories was observed over the six month study period. Drewnowski et al. (1994) argued that the frequency and severity of abnormal eating behaviours are distributed along a continuum and demonstrated that shifts in diagnosis on their 'Eating Pathology Scale' occurred primarily between adjacent categories. While the Eating Pathology Scale was not used in this study, some have argued that eating disorder diagnoses exist on a similar continuum of No Diagnosis to Eating Disorders Not 123 Discussion & Conclusions Otherwise Specified (EDNOS) to bulimia nervosa or anorexia nervosa, and an individual may progress through these categories (Adams & Shafer, 1988; Patton, 1992; Shisslak et al., 1995). The results of the present study do not confirm this theory. While there was a larger number of subjects moving from the No Diagnosis category to the probable EDNOS category than subjects moving from the EDNOS category to the probable diagnosis of bulimia nervosa or anorexia nervosa, it was found that new cases of eating disorders were drawn from both the EDNOS category and the No Diagnosis category. Fifty four percent of those subjects with a probable diagnosis of EDNOS at baseline demonstrated an improvement in diagnosis. This improvement in symptoms has been reported in other studies. Patton et al (1990) found that 40% of those individuals diagnosed with an eating disorder spontaneously improved at follow up. Likewise, Zuckerman et al. (1986) noted that as many as 50% of college women who met the diagnostic criteria for bulimia during the fall term of their freshman year no longer met these criteria nine months later. Fairburn and Belgin (1990) have also noted that bulimic symptoms 'wax and wane' in severity over time. B. Previous Treatment Approximately 50% of the subjects (11/21) who reported at baseline they had received treatment in the past for eating disorder symptoms, denied any previous professional treatment at 6 months. No other study has reported this phenomena. This discrepancy could be due to a misunderstanding of the question on the Health Information Questionnaire; respondents may have interpreted the question to mean have they received treatment for eating disorders over the past 6 months or since the last survey was administered. C. Symptom Intensity Score Despite the progression to more serious forms of eating disorders by 4% of the total sample, a significant decrease in the overall mean Symptom Intensity Score was observed. At 6 months, 124 Discussion & Conclusions eating disorder symptomology improved for 55.6% of the sample, worsened for 37.1% of the sample and remained the same for 7.3% of the sample. The drop in Symptom Intensity Score over the course of the study may be a result of the 53 subjects who indicated that they were no longer dieting to lose or maintain their weight at 6 months. When the sample was divided based on dieting status at 6 months, a significant decrease in Symptom Intensity Score was seen in the not dieting group between baseline and 6 months while no significant difference in Symptom Intensity Score was seen in the dieting group. This suggests that when dieting stops, there is an improvement in eating disorder symptom severity. D. Binge Eating and Compensatory Weight Loss Behaviours With the exception of crash dieting, prevalence of binge-eating and compensatory behaviours were consistent over the study period. There was, however, a trend of an overall decreased frequency. This trend was not explained by the 30% reduction in number of women dieting as there were no differences in the frequency of behaviours between dieters and nondieters at 6 months. Comparisons of the prevalence of binge-eating and compensatory behaviours to other studies with similar samples were limited to those specifying a frequency of weekly or more. Some diversity is evident. Use of diuretics and laxatives to lose weight was similar to that reported in other studies of college women (Halmi et al., 1987; Pyle et al., 1991; Pyle et al., 1983; Schlundt & Johnson, 1990). Use of diet pills and fasting to lose weight was less, while vomiting to lose weight was far more prevalent. Weekly or more frequent binge-eating was a prevalent behaviour in this sample and was comparable to that reported by Schotte & Stunkard (1987) in female college students. Comparison of the prevalence of loss of control over eating and crash dieting to other studies with the aforementioned referent period was not possible. However, a large percentage of the sample indicated that they felt a loss of control over eating and used crash dieting to lose weight. 125 Discussion & Conclusions The high prevalence of binge-eating, vomiting, and feeling a loss of control over eating in the present study may be related to the inclusion criteria. The sample consisted solely of dieters at baseline while other studies included both dieting and non-dieting college women. Due to the association between dieting and use of abnormal eating behaviours, it was likely that a higher frequency of bulimic behaviours would be detected. Conversely, these results may be a product of selection bias; individuals with disturbed eating practices may have been more attracted to this type of study. 4. Duration of Dieting A. Age at Onset of Dieting The women in the study began dieting at an average age of 15 years. This is shortly after the onset of puberty and is consistent with the notion that dieting to avoid the weight gain that accompanies normal pubertal development is a major cause of dieting (Adams & Shafer, 1988). Similarly, Frank et al. (1991) reported that the median age for beginning restrictive dieting was 16 years old. Mussell et al. (1995), however, found that the age of onset of significant dieting was 22.0. In Mussell's study the definition of dieting was defined by the age at which the subject first lost 10 pounds by dieting. The negative correlation observed between the 'Age at first diet' and Symptom Intensity Score suggests that those individuals who began dieting at a younger age were more likely to use inappropriate eating practices. Individuals reporting a younger age when they started dieting also had a stronger tendency to endorse the statements that they felt a loss of control over eating and used crash dieting as a method of weight control. Beginning to diet at a younger age, however, did not predict a change in the severity of symptoms. Johnson et al., (1984) reported similar associations between age at first diet and disordered eating. In their survey of high school 126 Discussion & Conclusions students, a significant percentage of students reporting bulimic symptoms had started to diet at an early age (14 years old). B. Long- and Short-term Duration of Dieting On average, the women in the sample had dieted for four years and reported to be on a diet for approximately half of the previous six months. A longer duration of dieting was associated with some abnormal eating behaviours. Specifically, a greater number of days dieting was associated with vomiting to lose weight, while a greater number of years dieting was associated with a higher initial Symptom Intensity Score and loss of control over eating. Johnson et al. (1984) reported similar results in high school students where bulimic symptoms were associated with the duration of dieting. Due to the association between chronic dieting and abnormal eating behaviours, a stronger relationship between the Symptom Intensity Score and long- and short-term dieting was expected. However, large variability in the measures of duration of dieting was apparent which may have obscured the relationship between these variables. 5. Dieting Methods A. Sources of Information Overall, use of the three sources of dieting information among the women surveyed was relatively low (6 to 21%). However, the source of dieting information used most frequently was that obtained from a health professional rather than information obtained from a book or magazine or solicited information from family, friends, TV, or videos. This was observed at both baseline and 6 months. This is in contrast to results from The Weight Loss Practices Survey (Levy & Heaton, 1992) in which respondents were more likely to use information from 'informal' sources. This may be explained by the differences in sample composition and context of the study. In the present study, the sample was obtained from a university and college setting. These women may have had a greater level of education and a greater number of available health 127 Discussion & Conclusions resources (student health services, health education programs) than the women in participating in The Weight Loss Practices Survey. B. Dieting Methods The most frequently reported dieting methods in the study included, in decreasing order, eating smaller portions and avoiding certain foods, creating 'own' diet, and counting grams of fat. However, the use of dieting methods showed some variability over time. There was a trend to a decreased use of all methods except counting grams of fat. When the mean frequency of use of diet methods was analyzed, there were significant reductions in the total number of methods used, use of commercial weight-loss programs, use of liquid formulas, skipping meals, counting calories, and avoiding eating when hungry. It can only be speculated as to why some dieting methods stayed constant while others were used with reduced frequency. There may be different weight control practices over the summer compared to the winter months. In addition, the students may have experienced an increased exposure and access to health and nutrition education leading to an awareness of the health consequences or the ineffectiveness of some weight loss practices for weight control. Perhaps some of these dieting methods were too challenging to continue through the school year or dieting may have become less of a priority during the school years. It could be assumed that the most likely reason for the reduced frequency of use of diet methods over the 6 month study period was due to a 30% reduction in active dieters. However, when use of the individual dieting methods over the study period were analyzed separately for dieters and non-dieter, the reduced use of methods overall was due to a significant drop in frequency of use by those women who indicated that they continued dieting at 6 months. Whether the reduced use of these methods was permanent or temporary could not be answered by the present study. The number of dieting methods used was included in the analysis with the dieting factors. Only the number of dieting methods used, however, was related to the change in symptom severity 128 . Discussion & Conclusions suggesting that participants who used more methods to lose weight may experience more severe eating disorder symptomology. This is in agreement with Drewnowski's survey of college women which found that the use of a greater number of dieting methods was associated with assignment to higher categories of the eating pathology scale. Food Avoidance dieting methods, Fat and Calorie Counting, and Use of Diet Aids were associated with both the Symptom Intensity Score and the use of binge eating and compensatory behaviours. The association between Food Avoidance dieting methods and crash dieting and fasting was expected as these variables essentially measure the same behaviours (fasting, avoiding to eat when hungry, and skipping meals). The Self-Selected Dieting factor consisted of dieting methods (eating smaller portions and avoiding certain foods, and creating own diet) deemed safe in terms of their limited potential for abuse. It was possible, however, that even reasonable dieting methods could be taken to the extreme and be associated with eating disorder symptomology. In the present study, however, no relationship between Self-Selected Dieting and the Symptom Intensity Score or binge eating and compensatory behaviours was observed. This lack of association suggests that, in the present sample of women, these methods of dieting may be considered safe and potentially reasonable elements of weight loss diets. While factor analysis assisted in reducing the data set and identifying underlying relationships between methods in this sample of women, some caution must be used in interpreting the results. The identified dieting factors may display a different factor structure when used in a different sample, for example in predominantly overweight individuals. The consistency of these factors, therefore, needs to be assessed in several populations. 6. Macronutrient Intake According to the Nutrition Recommendations of Canadians, the Canadian diet should include no more that 30% of energy from fat and about 55% of energy as carbohydrate and the remainder (~ 129 Discussion & Conclusions 15%) as protein (Health Canada, 1990). On average, the dieting women in this sample consumed 1900 kcal per day with a macronutrient distribution of 58% of the energy derived from carbohydrate, 16% from protein, and 26% from fat. Like restrained eaters and individuals with eating disorders, these dieting women consumed a diet that had a tendency to be lower in percent fat and higher in percent protein (Kaplan, 1993; Drewnowski, 1995). There were, however, no significant correlations between macronutrient intake (absolute and relative) and Symptom Intensity Score. A weak but significant association between carbohydrate intake and Change in Symptom Intensity Score was observed. This suggests that a higher carbohydrate intake was associated with an increased Symptom Intensity Score. Multiple correlations between the macronutrient intake and use of compensatory behaviours were observed. The pattern of intake that appears to be associated with abnormal eating behaviours included a lower fat and carbohydrate intake and a higher protein intake. This is similar to that described for individuals with bulimia nervosa as discussed by O'Connor et al. (1987). It is not clear from this study whether the macronutrient intake resulted in the compensatory behaviours, or if the compensatory behaviour resulted in the subsequent intake. However, such associations suggests that extremes in macronutrient intake, particularly low intakes, are associated with abnormal eating behaviours. As the macronutrient intake was derived from the FFQ, caution is advised in interpreting the results. A large variation in intake was noted, as has been experienced by others (Kushi, 1994). Reasons for this wide variation are not clear but differential recall among individuals as they attempted to determine their average intake over the preceding six months and differing familiarity with the format of FFQ may influence this variation. 7. Restrictiveness of the Diet 130 Discussion & Conclusions The average age of women in the study was 21 years. According to the Nutrition Recommendations for Canadians, females at this age should consume 2100 kcal per day (Health and Welfare Canada, 1990). The mean estimated energy requirement for this sample, as determined by the Mifflin Equation, was strikingly similar at 2080 ± 300 kcal/d. According to the Food Frequency Questionnaire (FFQ), the average energy intake was 1900 ± 660 kcal/day or 91% of the recommended intake. The energy deficit or restrictiveness of the diet was, therefore, determined to be approximately 170 ± 670 kcal/d. The restrictiveness of the diet did not correlate with any measure of eating disorder symptomology. This result suggests that those individuals who consumed an energy intake below their requirement were no more likely to engage in abnormal eating behaviours than those meeting their estimated energy intake. In addition, the restrictiveness of the diet did not correlate with any of the expected compensatory behaviours, such as binge-eating or crash dieting. Given that severe weight concerns and restrictive dieting practices are incorporated into the definition of anorexia nervosa and bulimia nervosa (French & Jeffrey, 1994), it was unusual that restrictiveness and Symptom Intensity Score were not highly correlated. There may be a number of reasons for this finding. The results may have been obscured by the inclusion of dieting to maintain weight as part of the definition of dieting. The level of energy restriction may be different for those attempting to lose weight and those wishing to maintain their current weight. Thus, the association between restrictiveness and development of eating disorder symptoms may also be different between those attempting to maintain weight and those trying to lose weight. The ability of the FFQ to accurately quantify energy intake may also have contributed to the lack of significant results. While the goal of the FFQ was to characterize usual intake, it is not clear whether the respondents quantified the energy intake from eating binges, if any. In addition, the presence of day-to-day variation in intakes, the differential recall among individuals in determining their average intake, and the validity of self-report, gives rise to the 131 Discussion & Conclusions possibility of both bias and random error. These sources of error may impact upon the estimate of usual dietary intake. Thus, the relationship between the restrictiveness and eating disorder symptomology may not have been evident. It is unlikely that the estimated energy requirements were overestimated as the calculated requirement was close to that proposed by Health and Welfare Canada. However, use of the Mifflin Equation was not without limitations. The Mifflin Equation was originally derived in a large sample of women between the ages of 19 and 78 years. In the present study 24% of the women in the sample were under the age of 19. Estimated energy requirements for the sample, therefore, may not be valid. However, the advantages of the Mifflin Equation (more recently derived, strong predictive abilities) over more popular equations, such as the Harris-Benedict Equation, may be sufficient justification for use despite this limitation. Another factor may account for the lack of association between restrictiveness of the diet and eating disorder symptomology. Perhaps the perceived deprivation and restrictiveness of the diet is more likely to be related to eating disorder symptomology than actual intake (Tuschl, 1990a). In fact, Heatherton et al (1988) demonstrated that despite their chronic dieting efforts, restrained eaters do not differ from unrestrained eaters in terms of caloric intake. Perception of the level of restrictiveness of the diet may be an important dimension of dieting. The perceived level of restriction, however, was not measured in the present study. The rate of weight loss reported by the students agrees with other estimates reported in similar age populations (Williamson et al., 1992). However, the rate of reported weight loss decreased significantly over the study time period. This difference was accounted for by the change in dieting status over the study period. The mean rate of weight loss reported by those individuals who were still dieting at 6 months demonstrated a significant reduction while those who ceased dieting reported no change in the rate of weight loss (see Appendix O). It is of interest that the 132 Discussion & Conclusions rate of weight loss was less in the individuals who indicated they were still dieting than in those who ceased dieting. Reasons for this difference can only be speculated and cannot be directly determined by the present study. The rate of weight loss in dieters may have been influenced by seasonal variation (Rosen et al., 1990; Forman et al., 1986; Levy & Heaton, 1993). While there exists the possibility that some individuals with bulimia nervosa may experience a worsening of symptoms (increase in binge eating and purging, weight gain, and feeling worse) over the fall or winter months (Lam et al., 1996; Lam et al., 1994; Levitan et al., 1996), there are no data, however, on what, if any, relevant seasonal factors might be involved in dieting individuals. It is also possible that those that individuals who indicated that they ceased dieting may be considered "successful" dieters. These individuals may have lost the desired amount of weight and ceased dieting. In addition, repeated dieting may have resulted in a reduced metabolic rate making subsequent weight loss more difficult (Brownell et al., 1986; Blackburn et al., 1989). Other studies have not found a difference in energy expenditure or rate of weight loss over successive diets (McCargar et al., 1992, 1993; Wadden et al., 1991) A small but significant correlation was observed between the rate of weight loss and Symptom Intensity Score. This suggests that students reporting a more rapid weight loss were more likely to use inappropriate weight control behaviours. In particular, these women were more likely to report binge-eating and fasting to lose or control weight. Rate of weight loss did not, however, predict a change in eating disorder symptom severity. In addition to the high variability in rate of weight loss, the higher rate of weight loss reported by those individuals who indicated that they were no longer dieting suggests that this variable may not be valid as a marker of restrictiveness of the diet. Also, as rate of weight loss is usually a function of both dietary energy restriction, physical energy expenditure, and is partially dependent on initial weight, determination of the amount of weight lost specifically from dieting was not feasible. 133 Discussion & Conclusions 8. Dieting Status Attempts at weight control are often transient as evidenced by other surveys in college and high school students (Drewnowski et al., 1994; Patton et al., 1990). Similarly, the sample in the present study were heterogeneous with respect to their degree of commitment to dieting. After 6 months, approximately 30% of the students indicated that they were no longer dieting to lose or maintain their body weight. The Symptom Intensity Score decreased from baseline in those subjects who indicated they were no longer dieting but was maintained in those subjects who identified themselves as dieters. This suggests that dieting is not without some risks. However, the absence of predictive values of the key dieting variables suggests that dieting alone is insufficient to predict the development of eating disorders. Hsu (1990) and Garfinkel et al. (1992), however, have suggested that certain factors may intensify or perpetuate the dieting, indirectly increasing the risk of eating disorders. In the present study, the comparison of baseline features between the dieting and not dieting groups indicated several dieting practices that differentiated the two groups. The dieting group had dieted for a greater number of days and used a greater number of dieting methods (including Fat and Calorie Counting, and Self-Selected Dieting) over the previous 6 months. The composition of the diet was also different. Those subjects who indicated they continued to diet at 6 months consumed less energy, fat and carbohydrate, and consumed a greater proportion of their intake from protein. The restrictiveness of the diet was higher for those still dieting. These differences may be construed as possible factors which may perpetuate dieting. A longer prospective study of these dieting practices, however, would be required to confirm potential risk. 9. Limitations The limitations of the present study include those inherent in a study involving self-report instruments. These limitations may affect the validity and generalizability of the results. 134 f Discussion & Conclusions Important limitations are addressed below. The results of this report should be interpreted with consideration given to these limitations. A. Questionnaires The use of questionnaires in research has some major benefits. In addition to being easy to administer, practical and non-invasive, questionnaires are cost-effective relative to individual interviews (Bright-See et al., 1994; Margetts et al., 1989; Mullen et al., 1984; Sheatsley, 1983). Participants may be more willing to respond honestly to sensitive issues when using an anonymous self-report questionnaire instrument; socially-sanctioned answers may be reduced when confidentiality is assured. In addition, use of a closed-ended response format with standard response categories permits easier coding and statistical analysis of the data (Perkins, 1992; Woodward & Chambers, 1986). No measure, however, is without error and the limitations of questionnaire instruments must be recognized (Willett, 1994). Often questionnaires ignore context or have the potential to be too standardized. Respondents may also feel they are forced into, what appears to them, an unnatural response. A closed-ended response format has the disadvantage of limiting options and risks missing important dimensions of the behaviour. In addition, with closed ended formats, respondents do not have the opportunity to qualify or explain their answers. (Woodward & Chambers, 1986; Perkins, 1992). i . The Dieting Practices Survey A major limitation of the Dieting Practices Survey is that subject responses were not validated. Recall of dates, ages, the specific duration of dieting, and dieting methods used is subject to memory errors, and may depend on the individual's previous dieting history. In addition, dieting methods considered unusual may be under-reported if considered by the respondent as socially undesirable. Furthermore, due to prevalence of dieting behaviours in society, there may be no 135 Discussion & Conclusions known group validity. It may be difficult to determine where 'normal eating' ends and symptoms of eating disorders start. The extent to which the accuracy of self-report varies by factors such as dieting, is unknown. Measurements of these dieting practices and behaviours in population that were either not dieting (little dieting history) and a group with diagnosed eating disorders, would confirm the validity of this instrument. The reliability and internal consistency of the Dieting Practices Survey, however, is supported by several analyses. Test-retest reliability was found to be adequate by the nonsignificant differences between baseline and 6 months measure of Age at First Diet and Long-term Dieting Frequency. Significant difference in these two variables would not be expected over the 6 month period. Internal consistency is supported by significant correlations between items that essentially measure the same concepts. For example, fasting behaviours were significantly correlated (r = 0.7) between the Health Information Questionnaire, the Dieting Practices Surevy, and through the factor analysis of the dieting methods. i i . The Food Frequency Questionnaire The Food Frequency Questionnaire (FFQ) used in this study was extensively used The Ontario Heart Survey (50,000 residents over the age of 12) and the relative validity was assessed in 147 individuals. The limitations of this FFQ, however, must be noted. The best method for validation of any food intake instrument includes the use of some biochemical marker. However, only relative validation with a three-day food record in a small sample of pretest subject (n=9) was used in the present study. In addition, the small number of changes made to the FFQ means that the data supporting the relative validity, as described by Bright-See et al. (1994), may not be extended to this study's population. In general, the FFQ has been shown to estimate intakes higher on a number of nutrients and with a greater variance than other instruments. This was evident in the pretest with the estimates of 136 Discussion & Conclusions protein and fat. Subjects may estimate portion size incorrectly or may feel compelled to indicate a quantity of food even if that item is rarely eaten or they may reply with amounts of food eaten that are in excess of their usual intake (Gibson, 1987). Other sources of error in the FFQ include within-person variability and effect of time delay. In addition, estimation of usual food intake is subject to memory errors. A l l of these sources of error will affect both statistical analyses and derived estimates. Random error reduces the statistical power of the analyses and estimates of the relationship between variables may be attenuated or inflated (Margetts, 1989; Tarasuk & Beaton, 1992). i i i . The Health Information Questionnaire Use of a questionnaire to rate eating disorder severity and divide the sample into probable diagnosis is not without precedence. Similar studies have used simulated DSM-ffl and D S M -BTR questionnaires for eating disorder diagnosis (Johnson et al., 1984; Halmi, 1981; Drewnowski et al., 1994; Schotte & Stunkard, 1987). The validity of the HIQ as a classification instrument has been verified in a previous study (Geller et al., in press). The Eating Disorder Inventory composite score was significantly greater for those undergraduate women receiving a probable eating disorder diagnoses than those receiving no diagnosis (Geller et al., in press). In addition, the HIQ has been shown to differentiate between normal controls and individuals with known clinical eating disorders (Geller, 1996). Use of The Health Information Questionnaire (HIQ) in the current study, however, raises a number of important methodological and interpretive issues. Eating disorder symptoms and diagnoses were not confirmed by direct interview. Validation of subject responses, therefore, was not possible. Clinical interviews would have determined the number of false positives and negatives found in this sample using the DSM-IV criteria in self-report format. As such, the sensitivity (% of cases correctly identified), specificity (% non-cases correctly identified), and the positive predictive (% of HIQ positive responses correctly identifying cases) value of the HIQ in 137 Discussion & Conclusions the current study is unknown. The rate of false-positive diagnoses using a questionnaire based on DSM-UJ and DSM-ITfR criteria has been reported as high as 22% (Schotte & Stunkard, 1987). In addition, purgative behaviours may be considered secretive; prevalence rates may underestimate the actual prevalence of bulimic behaviours. It is likely, therefore, that the reported incidence of eating disorders in the present study may have a margin of error. Furthermore, the HIQ relied on the student to interpret subjective concepts such as binge eating or binge frequency (Gilbert, 1986). It must be acknowledged, however, that the measurement of some criteria for eating disorders, for example, loss of control over eating, relies on self-report even within the context of a clinical interview (Shaw and Garfinkel, 1990). Therefore, even with the inclusion of a concurrent interview, some eating disorder concepts may escape validation. B. Experimental Design The limitations of the study design must also be acknowledged with respect to the generalizability of the results. As the study sample was not randomly selected from the study population and was limited to volunteers, dieting practices and weight control behaviours may differ depending on willingness to participate in research studies. This study may have attracted women concerned with their weight and eating habits. Consequently, there exists the possibility of selection bias; non-responders may have differed in frequency of dieting practices or other risk factors for eating disorders. In addition, subjects were selected from a university or college environment. There is no guarantee that similar findings would result from studies in populations that do not share the characteristics of this setting, for example in women who were younger, or were not attending college or university. For practical purposes, the study was implemented during the fall and winter academic semesters. While this study may have more accurately reflected the subject's natural environment compared to artificial laboratory or hospital settings, the time frame may not have been long enough to observe differences in Symptom Intensity Score. Observed changes in 138 Discussion & Conclusions behaviours may have been the result of the impact of school or seasonal variation in dieting and weight loss behaviours and not dieting to lose or maintain weight (Forman et al., 1986; Rosen et al., 1990). Some of the aforementioned limitations are, however, an inherent aspect of a survey design. A l l efforts were made to minimize the effect of the limitations by ensuring confidentiality, and by providing incentives and clear instructions. 10. Conclusion This thesis hypothesized that specific dieting practices, such as the duration of dieting, dieting methods, macronutrient intake and restrictiveness of the diet, are related to the development of abnormal eating behaviours in women who are already dieting. As such, individuals reporting a longer duration of dieting, use of unsound dieting methods, an unbalanced macronutrient intake, or a greater energy restriction, would experience an increase in Symptom Intensity Score, a measure of eating disorder symptomology, after 6 months. While a high frequency of disordered eating was observed in this sample of dieting women, these data fail to reject the null hypotheses. The dieting variables did not predict an increase in symptomology and most of the variability in Symptom Intensity Score remained unexplained by these dieting variables. The results suggest that development or propagation of eating disorders may not be contingent solely on dieting and specific dieting practices. A number of risk factors, in addition to dieting, may be necessary to create a condition where an eating disorder emerges. The dieting variables examined in this population, however, were associated with disordered eating behaviours. Individuals with a history of dieting from a young age may be at especially high risk of engaging in unhealthy eating behaviours. A longer dieting history, rapid weight loss and a higher B M I was related to the use of inappropriate weight loss behaviours. Similar associations were found with food avoidance dieting methods (such as skipping meals, not eating 139 Discussion & Conclusions when hungry and fasting), counting grams of fat or calories, and use of diet aids (such as structured diet plans and liquid formulas). In addition, a dietary intake combining a lower fat and carbohydrate intake with a higher protein intake represented a pattern associated with prolonged dieting and abnormal eating behaviours. Reliance on a greater number of dieting methods, greater use of fat and calorie counting, and a longer dieting duration, in terms of the number of days dieting, was also associated with subsequent dieting in this sample. These associations do not necessarily imply that use of these dieting practices leads to disordered eating. Instead, these dieting practices may be used more frequently by women susceptible to disordered eating. The results also suggest that not all dieting is the same; there may be safer forms of dieting. Several methods of altering the diet to lose weight, such as obtaining information from a health professional or consuming smaller portions, were not associated with disordered eating in this sample. These methods, therefore, appear to be relatively safe and healthy weight loss methods. The conclusions presented above were mitigated by several limitations. A non-random sample, reliance on self-report, and the lack of documented validation of some the questionnaires may have contributed to the limited generalizability of the results. In addition, the short time frame of the study may have obscured the extent to which these dieting practices predict the development of eating disorders. 11. Clinical Application Dieting has long been considered a putative risk factor in the development of eating disorders. While the dieting variables investigated in the present study left much of the variability in Symptom Intensity Score unexplained, the associations between dieting practices and some abnormal eating behaviours have several implications for action. It is premature to suggest that specific dieting practices, such as a longer duration of dieting or fat and calorie counting, are markers for later abnormal eating behaviours. However, these dieting practices may be used more frequently by women who are susceptible to disordered eating. Early detection and 140 Discussion & Conclusions treatment of disordered eating is thought to lead to more positive outcomes. Therefore, presence of these dieting practices may suggest that further investigation of the individual's eating behaviours is warranted. Sensitivity to an individual's dieting history, including the number and types of dieting methods used, the onset of dieting, usual food choices and rate of weight change, may distinguish those at risk for abnormal eating behaviours. This study emphasizes the need for preventative education programs that address realistic expectations for body shape and size. From the large number of normal weight women who were dieting, it appears that their desired weight was less than what may be defined, by the B M I categories, as 'medically appropriate'. This suggests that the influence of the media and the unreasonable expectations of a women's weight and shape are pervasive and seem to have been internalized by the participating students. In addition, current nutrition education messages promoting healthy weights and weight loss must be carefully targeted so as not to influence women who are already at a healthy weight. The women in this sample began dieting at approximately age 15. Therefore, preventative education programs may be more effective if aimed at the teen or pre-teen age group. It has been speculated that dieting at this age may be in direct response to the body shape changes associated with puberty. Detailed education on normal pubertal development and its impact on body size/shape, body image and self-esteem may be warranted. 12. Future Research In the present study, the duration of dieting, rate of weight loss and certain dieting methods were associated with disturbed eating behaviours. Further consideration of these characteristics over a longer study period is warranted. In addition, extension of the hypotheses to different populations, including overweight populations or individuals with eating disorders, would also be of interest to determine the stability of the observed associations with abnormal eating behaviours. Separation of dieters into groups according to whether they were dieting to maintain 141 Discussion & Conclusions weight or dieting to lose weight may further clarify the relationship between dieting practices and abnormal eating behaviours. A concurrent interview with a trained clinician would address the issue of validity associated with self-report questionnaires. A major challenge in studies of eating disorders is to understand why some individuals develop anorexia nervosa and other individuals develop bulimia nervosa. In the present study, abnormal eating behaviours were assessed with a single score (Symptom Intensity Score). The reported associations between the dieting variables and abnormal eating behaviours, as measured by the Symptom Intensity Score, therefore, suggest that the dieting variables are non-specific risk factors. That is, a behaviour that may be associated with disturbed eating practices may not be specific to a particular category of eating disorder. This raises the question of whether there are eating and dieting behaviours that are associated with the development of anorexia nervosa while others may be linked to bulimia nervosa. The significant relationship observed between the duration of dieting variables and abnormal eating behaviours, combined with the onset of dieting at the age of 15, suggests that focusing on the prevention of dieting among adolescent women may be beneficial. A prospective study examining the impact of preventative education programs for teenagers would determine whether or not these practices are, indeed, amenable to change and the effectiveness of an educational strategy. 142 References References Adams, L.B. , Shafer, M . B . Early manifestations of eating disorders in adolescents: defining those at risk. Journal of Nutrition Education. Vol . 20(6): 307-313, 1988 American Dietetic Association Reports. Position of the American Dietetic Association: Nutrition intervention in the treatment of anorexia nervosa, bulimia nervosa, and binge eating. Journal of the American Dietetic Association. Vol . 94(8): 902-907, 1994 American Psychiatric Association. Diagnostic Criteria For Eating Disorders. American Psychiatric Association: Diagnostic and Statistical Manual, Fourth Edition. Washington, D.C.: American Psychiatric Association, 1994 American Psychiatric Association. Practice Guideline for Eating Disorders. First Edition. Washington, D.C.: American Psychiatric Association, 1993 Ash, J.B., Piazza, E. Changing symptomatology in eating disorders. International Journal of Eating Disorders. Vol . 18(1): 27-38, 1995 Bakan, R., Birmingham, C.L., Aeberhardt, L. , Goldner, E . M . Dietary zinc intake of vegetarian and nonvegetarian patients with anorexia nervosa. International Journal of Eating Disorders. Vol . 13(2): 229-233, 1993 Bailey, S.M., Goldbery, J.P., Chomitz, V.R., Houser Jr., R.F. Relationship between body dissatisfaction and physical measurements. International Journal of Eating Disorders. Vol . 9: 457-461, 1990 Barr, S.I., Janelle, C , Prior, J.C. Vegetarian vs nonvegetarian diets, dietary restraint, and subclinical ovulatory disturbances: prospective 6-mo study. American Journal of Clinical Nutrition. Vol . 60: 887-894, 1994a Barr, S.I., Prior, J .C , Vigna, Y . M . , Restrained eating and ovulatory disturbances: possible implications for bone health. American Journal of Clinical Nutrition. Vol . 59: 92-97, 1994b Beals, K .A . , Manore, M . M . The prevalence and consequences of subclinical eating disorders in female athletes. International Journal of Sport Nutrition. Vol . 4: 175-195, 1994 Beggs, L. , Hendricks, S., Scwartz, N.E., Biro, K. Tracking nutrition trends: Canadians' attitudes, knowledge and behaviours regarding fat, fibre and cholesterol. Journal of the Canadian Dietetic Association. Vol 54(1): 21-25, 1993 Bennett, E . M . Weight loss practices of overweight adults. American Journal of Clinical Nutrition. Vol . 53: 1519S-1521S, 1991 143 References Birmingham, C.L. Anorexia nervosa and bulimia: Medical complications and management. British Columbia Medical Journal. Vol . 31(3): 155-158, 1989 Blackburn, G.L., Wilson, G.T., Kanders, B.S., Stein, L.J. , Lavin, P.T., Adler, J., Brownell, K.D. Weight cycling: the experience of human dieters. American Journal of Clinical Nutrition. Vol. 49: 1105-1109, 1989 Borrelli, R., Cole, T.J., DiBiase, G., Contaldo, F. Some statistical considerations on dietary assessment methods. European Journal of Clinical Nutrition. Vol . 43: 453-463, 1989 Bright-See, E., Catlin, G., Godin, G. Assessment of the relative validity of the Ontario Health Survey Food Frequency Questionnaire. Journal of the Canadian Dietetic Association. Vol . 55(1): 33-38, 1994 Brownell, K.D. , Greenwood, M.R.C. , Stellar, E., Shrager, E.E. The effects of repeated cycles of weight loss and regain in rats. Physiology & Behaviour. Vol . 38: 459-464, 1986 Brownell, K.D. , Rodin, J. The dieting maelstrom. Is it possible and advisable to lose weight? American Psychologist. Vol . 49(9): 781-791, 1994a Brownell, K.D. , Rodin, J. Medical, metabolic and psychological effects of weight cycling. Archives of Internal Medicine. Vol . 154: 1325-1330, 1994b Brownell, K.D. , Rodin, J., Wilmore, J.H. Eating, Body Weight and Performance in Athletes. Philadelphia: Lea & Febiger, 1992 Button, E.J., Whitehouse, A . Subclinical anorexia nervosa. Psychological Medicine. Vol . 11: 509-516, 1981 Cohen, J. Statistical power analysis for behavioural sciences. Second Edition. New York: Academic Press, 1988 Cooper, Z. The development and maintenance of eating disorders. In: Brownell, K.D. , Fairburn, C.G. Eating Disorders and Obesity. A Comprehensive Handbook. New York: The Guilford Press, 1995 Daly, J .M., Heymsfield, S.B., Head, C.A., Harvy, L.P., Nixon, D.W., Katzeff, H. , Grossman, G.D. Human energy requirements: Overestimation by widely used prediction equation. American Journal of Clinical Nutrition. Vol . 42(6): 1170-1174, 1985 Dancyger, I.F., Garfinkel, P.E. The relationship of partial syndrome eating disorders to anorexia nervosa and bulimia nervosa. Psychological Medicine. Vol . 25: 1019-1025, 1995 Davis, E., Furnham, A. Body dissatisfaction in adolescent girls. British Journal of Medical Psychology. Vol . 59: 279-287, 1986 144 References Davis, C , Durnin, J., Gurevich, M . , Le Maire, A. , Dionne, M . Body composition correlates of weight dissatisfaction and dietary restraint in young women. Appetite. Vol. 20: 197-207, 1993 Del Prete, E., Balkowski, G., Scharrer. Meal pattern of rats during hyperphagia induced by long-term food restriction is affected by diet composition. Appetite. Vol . 23: 79-86, 1994 Dickstein, L.J. Current college environments: Do these communities facilitate and foster bulimia in vulnerable students? In: Whitaker, L .C. (ed.) The Bulimic College Student. Evaluation, Treatment and Prevention. New York: The Haworth Press, 1989 Drewnowski, A . Energy intake and sensory properties of food. American Journal of Clinical Nutrition. Vol . 62(suppl): 1081S-5S, 1995 Drewnowski, A. , Hopkins, S.A., Kessler, R.C. The prevalence of bulimia nervosa in the US college student population. American Journal of Public Health. Vol . 78(10): 1322-1324, 1988 Drewnowski, A. , Yee, D.Y., Kurth, C.L., Krahn, D.D. Eating pathology and DSM-IU-R bulimia nervosa: A continuum of behavior. American Journal of Psychiatry. Vol . 151(8): 1217-1219, 1994 Dwyer, J.T. Nutritional consequences of vegetarianism. Annual Review of Nutrition. Vol . 11: 61-91, 1991 Dwyer, J. Sixteen popular diets: Brief nutritional analyses. In: Stunkard, A . J.(ed.) Obesity. Philidelphia: W.B. Saunders & Company, 1980 Eisler, I., Szmukler, G.I. Social class as a confounding variable in the eating attitudes test. Journal of Psychiatric Research. Vol . 19: 171-179, 1985 Emmons, L. Predisposing factors differentiating adolescent dieters and nondieters. Journal of the American Dietetic Association. Vol . 94(7): 725-729, 1994 E S H A Research Incorporated. Food Processor II Nutrient Analysis System. Version 3.24. Enhanced. Salem OR: ESHA Research Incorporated, 1990 Forman, M.R., Trowbridge, F.L., Gentry, E .M. , Marks, J.S., Hogelin, G.C. Overweight adults in the United States: The behavioral risk factor surveys. The American Journal of Clinical Nutrition. Vol . 44: 410-416, 1986 Fairburn, C.G., Beglin, S.J. Studies of the epidemiology of bulimia nervosa. American Journal of Psychiatry. Vol . 147:401-408, 1990 Frank, R.E., Serdula, M.K. , Adame, D. Weight loss and bulimic eating behaviour: Changing patterns within a population of young adult women. Southern Medical Journal. Vol . 84(4): 457-460, 1991 145 References French, S.A., Jeffrey, R.W. Consequences of dieting to lose weight: Effects on physical and mental health. Health Psychology. Vol . 13(3): 195-212, 19940 French, S.A., Story, M . , Downes, B., Resnick, M.D. , Blum, R.W. Frequent dieting among adolescents: Psychosocial and health behaviour correlates. American Journal of Public Health. Vol . 85(5): 695-701, 1995 Garfinkel, P.E., Garner, D . M . Anorexia nervosa: A multidimensional perspective. New York: Brunner/Mazel, 1982 Garfinkel, P.E., Halmi, K., Shaw, B. Applications of current research findings to treatment: What we need for the future. In: Anderson, G.H., Kennedy, S.H. (eds.) The Biology of Feast or Famine: Relevance to Eating Disorders. Toronto: Academic Press, 1992 Garfinkel, P.E., Lin, E., Goering, P., Spegg, C , Goldbloom, D.S., Kennedy, S., Kaplan, A.S., Woodside, D.B. Bulimia nervosa in a Canadian community sample: Prevalence and comparison of subgroups. American Journal of Psychiatry. Vol . 152(7): 1052-1058, 1995 Garner, D .M. , Garfinkel, P.E. Socio-cultural factors in the development of anorexia nervosa. Psychological Medicine. Vol . 10: 647-656, 1980a Garner, D .M. , Garfinkel, P.E., Schwartz, D., Thompson, M . Cultural expectations of thinness in women. Psychological Reports. Vol. 47(2): 483-491, 1980b Garner, D .M. , Olmstead, M.P. The eating disorder inventory manual. Odessa, FL: Psychological Assessment Resources, 1984 Garner, D .M. , Olmstead, M.P., Polivy, J. Development and validation of a multidimensional eating disorder inventory for anorexia nervosa and bulimia. International Journal of Eating Disorders. Vol . 2: 15-33, 1983 Garrow, J.S. The safety of dieting. Proceedings of the Nutrition Society. Vol . 50: 493-499, 1991 Geller, J., Johnston, C , Madsen, K. The role of shape and weight in self-concept: The shape and weight based self-esteem inventory. Cognitive Therapy and Research. (In Press) Geller, J. Body shape and weight as determinants of women's self-esteem. Ph.D. Thesis. The University of British Columbia. 1996 Gibson, R.S. Sources of error and variability in dietary assessment methods: A review. Journal of the Canadian Dietetic Association. Vol . 48(3): 150-155, 1987 Gilbert, S. Pathology of eating: Psychology and treatment. New York, N Y : Routledge & Kegan Paul, 1986 146 ', References Goldbloom, D.S., Garfinkel, P.E. Anorexia nervosa and bulimia nervosa - Diagnostic issues and risk factors. In: Kennedy, S.H. (ed.) The Handbook of Eating Disorders. The University of Toronto, 1993 Goldner E . M . , Birmingham, C.L. Anorexia nervosa: Methods of treatment. In: Alexander-Mott, L. , Lumsden, D.B. (ed.) Understanding Eating Disorders: Anorexia Nervosa, Bulimia Nervosa and Obesity. Washington, D.C.: Taylor & Francis, 1994 Gordon, R., Bulimia: A sociocultural interpretation. In: Whitaker, L .C. (ed.) The Bulimic College Student. Evaluation, Treatment and Prevention. New York, N Y : The Haworth Press, 1989 Greenfeld, D., Quinlan, D .M. , Harding, P., Glass, E., Bliss, A . Eating behaviour in an adolescent population. International Journal of Eating Disorders. Vol . 6: 99-111, 1987 Grocery Products Manufacturers of Canada. Grocery attitudes of Canadians 1994. Don Mills, Ontario: Grocery Products Manufacturers of Canada, 1994 Grodner, M . "Forever Dieting": Chronic dieting syndrome. Journal of Nutrition Education. Vol . 24 (4):207-210, 1992 Gwirtsman, H.E., Curtis, S.R., Lyter, L . M . Energy intake and dietary macronutrient content in women with anorexia nervosa and volunteers. Journal of the American Dietetic Association. Vol . 89(1): 54-57, 1989 Halmi, K . A . , Falk, J.R., Schwartz, E. Binge-eating and vomiting: A survey of a college population. Psychological Medicine. Vol . 11(4): 697-709, 1981 Halmi, K . A . , Casper, R.C., Eckert, E.D., Goldberg, S.C., Davis, J .M. Unique features associated with the age of onset of anorexia nervosa. Psychiatry Research. Vol . 1(2): 209-215, 1979 Harris, J.A., Benedict, F.G. A biometric study of basal metabolism in man. Washington, DC: The Carnegie Institute, 1919 Harvey, J., Wing, R., Mullen, M . Effects on food craving of a very low calorie diet or a balanced, low calorie diet. Appetite. Vol . 21: 105-115, 1993 Health and Welfare Canada. Nutrition Recommendations for Canadians. Ottawa, ON; Minister of Supply and Services, 1990 Health and Welfare Canada. Promoting Healthy Weights: A Discussion paper. Ottawa, ON; Minister of Supply and Services, 1988a Health and Welfare Canada. Rootman, I., Warren, R., Stephens, T., Peters, L . (eds.) Canada's Health Promotion Survey: Technical Report. Ottawa, ON; Minister of Supply and Services, 1988b 147 References Health and Welfare Canada. Stephens, T., Fowler-Graham, D. (eds.) Canada's Health Promotion Survey 1990: Technical Report. Ottawa, ON; Minister of Supply and Services, 1993 Heatherton, T.F., Herman, CP., Polivy, J., King, G.A., McGree, S.T. The (mis)measurement of restraint: An analysis of conceptual and psychometric issues. Journal of Abnormal Psychology. Vol. 97(1): 12-28, 1988 Herman, CP., Mack, D. Restrained and unrestrained eating. Journal of Personality. Vol. 43: 647-660,1975 Herman CP., Polivy, J. Studies of eating in normal dieters. In: Walsh, B.T. (ed.) Eating Behaviors in Eating Disorders. Progress in Psychiatry. Washington, DC: American Psychiatric Press, Inc. 1988 Hill, J.O., Drougas, H., Peters, J.C. Obesity treatment: Can diet composition play a role? Annals of Internal Medicine. Vol. 119 (7 part 2): 694-697, 1993 Horm, J., Anderson, K. Who in America is trying to lose weight? Annals of Internal Medicine. Vol. 119(7): 672-676, 1993 Horner, T.N., Utermohlen, V. A multivariate analysis of psychological factors related to body mass index and eating preoccupation in female college students. Journal of the American College of Nutrition. Vol. 12(4): 459-465, 1993 Howell, D.C Statistical Methods for Psychology. Third Edition. California: Duxbury Press, 1992 Hsu, L.K.G., Eating Disorders. New York: Guilford Press, 1990 Huon, G.F., Wootton, M. Towards the prevention of eating disorders: Some preliminary findings from an empirical investigation. Psychotherapy and Psychosomatics. Vol. 48: 181-188, 1987 Johnson, C.L., Lewis, C , Love, S., Lewis, L., Stuckey, M. Incidence and correlates of bulimic behaviour in a female high school population. Journal of Youth and Adolescence. Vol. 13(1): 15-26,1984 Johnson, C. L., Lewis, C , Love, S., Stuckey, M., Lewis. A descriptive survey of dieting and bulimic behavior in a female high school population. In: Ross Laboratories. Understanding Anorexia Nervosa and Bulimia. Report of the Fourth Ross Conference on Medical Research. Columbus, Ohio: Ross Laboratories, 1983 Johnson-Sabine, E., Wood, K., Patton, G., Mann, A., Wakeling, A. Abnormal eating attitudes in London schoolgirls - A prospective epidemiological study: factors associated with abnormal response on screening questionnaires. Psychological Medicine Vol. 18:615-622,1988 148 ' , References Kagan, D .M. , Squires, R.L.S. Dieting, compulsive eating, and feelings of failure among adolescents. International Journal of Eating Disorders. Vol . 3: 15-26, 1983 Kaplin A.S. Medical and nutritional assessment. In: Kaplin, A.S., Garfinkel, P.E. (eds.) Medical issues and the Eating Disorders: The Interface. New York, N Y : Brunner/Mazel, 1993 Kalucy, R.S. The "new" nutrition. The Medical Journal of Australia. Vol. 147: 529-530, 1987 Kendler, K.S., MacLean, C., Neale, M . , Kessler, R., Heath, A. , Eaves, L. The genetic epidemiology of bulimia nervosa. American Journal of Psychiatry. Vol . 148(12): 1627-1637, 1991 Keys, A. , Brozek, J., Henschel, A. , Micklelsen, O., Taylor, H.L. The Biology of Human Starvation. Minneapolis: University of Minneapolis Press. 1950 Kiernan, M . , Rodin, J., Brownell, K.D. , Wilmore, J.H., Crandeall, C. Relationship of level of exercise, age, and weight cycling history to eating and weight disturbances in male and female runners. Health Psychology. Vol . 11:418-421, 1992 Killen, J.D., Hayward, C , Wilson, D .M. , Taylor, C.B., Hammer, L.D. , Litt, I., Simmonds, F., Haydel, K. Factors associated with eating disorder symptoms in a community sample of 6th and 7th grade girls. International Journal of Eating Disorders. Vol. 15(4): 357-367, 1994 Krahn, D.D. The relationship of eating disorders and substance abuse. Journal of Substance Abuse. Vol . 3: 239-253, 1991 Krenz, J. McCargar, L. , Donald, L. , Canuel, E. Eating disorders and diabetes: Prevalence in a diabetes education centre population. Journal of the Canadian Dietetic Association. Vol . 54 (1): 17-21, 1993 Krey, S.H., Palmer K., Porcelli, K . A . Eating disorders: The clinical dietitian's changing role. Journal of the American Dietetic Association. Vol . 89(l):41-43, 1989 Kristal, A.R., White, E., Shattuck, A.L . , Curry, S., Anderson, G.L., Fowler, A. , Urban, N . Long-term maintenance of a low-fat diet: Durability of fat-related dietary habits in the women's health trial. Journal of the American Dietetic Association. Vol. 92(5):553-559, 1992 Kuehnel, R.H., Wadden, T.A. Binge eating disorder, weight cycling and psychopathology. International Journal of Eating Disorders. Vol. 15(4):321-329, 1994 Kurtzman, F.D., Yager, J., Landsverk, J., Wiesmeir, E., Bodurka, D.C. Eating disorders among selected female student populations at U C L A . Journal of the American Dietetic Association. Vol . 89(1): 45-53, 1989 149 References Kushi, L . H . Gaps in epidemiological research methods: design considerations for studies that use food-frequency questionnaires. American Journal of Clinical Nutrition. Vol . 59 (suppl): 180S-4S,1994 Laessle, R.G., Tuschl, R.J., Kotthaus, B.C., Pirke, K . M . Behavioral and biological correlates of dietary restraint in normal life. Appetite. Vol. 12: 83-94, 1989 Lam, R.W., Goldner, E .M. , Grewal, A . Seasonality of symptoms in anorexia and bulimia nervosa. International Journal of Eating Disorders. Vol . 19(1): 35-44, 1996 Lam, R.W., Goldner, E .M. , Solyom, L. , Remick, R.A. A controlled study of light therapy for bulimia nervosa. American Journal of Psychiatry. Vol . 151(5): 744-750, 1994 Lappalainen, R., Sjoden, P., Hursti, T., and Vese, V . Hunger/craving responses and reactivity to food stimuli during fasting and dieting. International Journal of Obesity. Vol . 14: 679-688, 1990 Leibowitz, S., Shor-Posner, G. Brain serotonin and eating behaviour. Appetite. Vol . 7:(suppl)l-14,1986 Levin, G. V. , Zehner, L.R., Saunders, J.P., Beadle, J.R. Sugar substitutes: Their energy values, bulk characteristics and potential health benefits. American Journal of Clinical Nutrition. Vol . 62(suppl): 1161S-8S, 1995 Levitan, R.D., Kaplan, A.S., Rockert, W. Characterization of the "seasonal" bulimic patient. International Journal of Eating Disorders. Vol . 19(2): 187-192, 1996 Levy, A.S., Heaton, A.W. Weight control practices of US adults trying to lose weight. Annals of Internal Medicine, Vol . 119 (7 part 2): 661-666, 1993 Levy, R.L., Richey C.A. Measurement and research design. In: Blechman, E.A. , Brownell, K.D. (eds.) Handbook of Behavioral Medicine for Women. New York: Pergamon Press, 1988 Louis-Sylvestre, J., Tournier, A. , Chapelot, D., Chabert, M . Effect of a fat-reduced dish in a meal on 24-hour energy and macronutrient intake. Appetite. Vol . 22: 165-172, 1994 Margetts B . M . , Cade, J.E, Osmond, C. Comparison of a food frequency questionnaire with a diet record. International Journal of Epidemiology. Vol . 18: 868-873, 1989 McCargar, L.J. , Crawford, S.M. Metabolic and anthropometric changes with weight cycling in wrestlers. Medicine and Science in Sports and Exercise. Vol . 24: 1270-1275, 1992 McCargar, L.J. , Taunton, J., Birmingham, C.L., Pare, S., Simmons, D. Metabolic and anthropometric changes in female weight cyclers and controls over a 1-year period. Journal of the American Dietetic Association. Vol. 93(9): 1025-30. 1993 150 References McCargar, L.J. , Sale, J., Crawford, S.M. Chronic dieting does not result in a sustained reduction in metabolic rate. Journal of the American Dietetic Association. In Press. McCreary Centre Society. Adolescent Health Survey: Province of British Columbia. Vancouver, B.C.: The McCreary Centre Society, 1993 McCreary Centre Society. The Provincial Task Force Eating Disorder Report. Vol . 1-3. Vancouver, B.C.: The McCreary Centre Society, 1989 McDonald, P.W., A case of the treatment of obesity. Journal of the Canadian Dietetic Association. Vol . 56(3): 131-136, 1995 Megel, M.E. , Hawkins, P., Sandstrom, S., Hoefler, M.A . , Willret, K., Wade, F., Norton., J, Zajic, K., Partusch, M.E. , Tourek, N . Health promotion, self-esteem, and weight among female college freshmen. Health Values. Vol . 18(4): 10-19, 1994 Mifflin, M . D., St Jeor, S.T., Hi l l , L .A. , Scott, B.J., Daugherty, S.A., Koh, Y.O. A new predictive equation for resting energy expenditure in healthy individuals. "American Journal of Clinical Nutrition. Vol . 51: 241-247, 1990 Moore, D.C. Body image and eating behavior in adolescents. Journal of the American College of Nutrition. Vol . 12(5): 505-510, 1993 Mullen, B.J., Krantzler, N.J., Grivetti, L. , Schutz, H.G., Meiselman, L. Validity of a food frequency questionnaire for the determination of individual food intake. American Journal of Clinical Nutrition. Vol . 39: 136-143, 1984 Munro, B .H. , Visintainer, M.A. , Page, E.B. Statistical methods for health care research. Philidelphia: J.B. Lipponcott Company, 1986 Mussels, M.P., Mitchell, J.E., Weller, C.L., Raymond, N.C., Crow, S.J., Crosby, R.D. Onset of binge eating, dieting, obesity, and mood disorders among subjects seeking treatment for binge eating disorder. International Journal of Eating Disorders. Vol . 17(4):395-401, 1995 National Institutes of Health. Technology Assessment Conference Panel. Methods for voluntary weight loss and control. Annals of Internal Medicine. Vol . 116 (11): 942 - 949, 1992 National Task Force on the Prevention and Treatment of Obesity. Weight cycling. Journal of the American Medical Association. Vol. 272 (15): 1196-1202, 1994 Nevo, S. Bulimic symptoms: Prevalence and ethnic differences among college women. International Journal of Eating Disorders. Vol . 4: 151-168, 1985 Nylander, J. The feeling of being fat and dieting in a school population: Epidemiologic interview investigation. Acta Sociomedica Scandinavica. Vol . 3: 17-26, 1971 151 References O'Connor, M.A . , Touyz, S.W., Dunn, S.M., Beumont, J.V. Vegetarianism in anorexia nervosa? A review of 116 consecutive cases. Medical Journal of Australia. Vol . 147: 540-542, 1987 Omichinski, L. Reduction of dieting attitudes and practices after participation in a non-diet lifestyle program. Journal of the Canadian Dietetic Association. Vol . 52(2): 81-85, 1995 Owen, O.E., Holup, J.L., D'Alessio, D.A., Craig, E.S., Polansky, M . , Smalley, K.J . , Kvale, E.C., Bushman, M.C. , Owen, L.R., Mozzoli, M.A. , Kendrick, Z.V., Boden, G. A reappraisal of the caloric requirements of men. American Journal of Clinical Nutrition. Vol . 46(6): 875-885, 1987 Owen, O.E., Kvale, E.C., Owen, R.S, Polansky, M . , Caprio, S., Mozzoli, M.A . , Kendrick, Z.V., Bushman, M.C. , Boden, G. A reappraisal of the caloric requirements of women. American Journal of Clinical Nutrition. Vol. 44(1): 1-19, 1986 Patton, G.C. Eating disorders: Antecedents, evolution and course. Annals of Medicine. Vol . 24: 281-285, 1992 Patton, G.C. Mortality in eating disorders. Psychological Medicine. Vol . 18: 947-951, 1988 Patton, G.C. The spectrum of eating disorders in adolescence. Journal of Psychosomatic Research. Vol . 32: 579-584, 1988b Patton, G . C , Johnson-Sabine, E., Wood, K., Mann, A . H. , Wakeling, A . Abnormal eating attitudes in London schoolgirls - A prospective epidemiological study: Outcome at twelve month follow-up. Psychological Medicine. Vol . 20: 383-394, 1990 Perkins, J. Design and use of questionnaires in research. In: Monsen, E. R. (ed.) Research. Successful Approaches. Chicago, IL: The American Dietetic Association, 1992 Polivy J., Herman, C P . Breaking the diet habit: The natural weight alternative. New York: Basic Books, Inc., 1983 Polivy, J., Herman, C. P. Dieting and bingeing: A causal analysis. American Psychologist. Vol . 40(2): 193-201, 1985 Polivy, J., Thomsen, L . Dieting and other eating disorders. In: Blechman, E. A. , Brownell, K.D. (eds.) Handbook of Behavioral Medicine for Women. New York: Pergamon Press, 1988 Pyle, R.L., Mitchell, J.E., Eckert, E.D., Halvorson, P.A., Neuman, P.A., Goff, G . M . The incidence of bulimia in freshman college students. International Journal of Eating Disorders. Vol . 2: 75-85, 1983 Pyle, R.L., Neuman, P.A., Halvorson, P.A., Mitchell, J.E. An ongoing cross-sectional study of the prevalence of eating disorders in freshman college students. International Journal of Eating Disorders. Vol . 10: 667-677, 1991 152 References Reid, D.J., Hendricks, S.M. Consumer understanding and use of fat and cholesterol information on food labels. Canadian Journal of Public Health. Vol. 85(5): 334-337, 1994 Rock, C.L., Yager, J. Nutrition and eating disorders: A primer for clinicians. International Journal of Eating Disorders. Vol. 6(2): 267-280, 1987 Rodin, J., Silberstein, L., Striegel-Moore, R. Women and weight: A normative discontent. In: Nebraska Symposium on Motivation. Vol. 32: 267-307, 1984 Rolls, B.J., Hammer, V.A. Fat, carbohydrate and the regulation of energy intake. American Journal of Clinical Nutrition. Vol. 62 (suppl): 1086S-95S, 1995 Rolls, B.J., Kim-Harris, S., Fischman, M.W., Foltin, R.W., Moran, T.H., Stoner, S.A. Satiety after preloads with different amounts of fat and carbohydrates: implications for obesity. American Journal of Clinical Nutrition. Vol. 60: 476-87, 1994 Rosen, J.C, Compas, B.E., Tacy, B. The relation among stress, psychological symptoms and eating disorder symptoms: a prospective analysis. International Journal of Eating Disorders. Vol. 14(2): 153-162, 1993 Rosen, J.C, Gross, J., Vara, L. Psychological adjustment of adolescents attempting to lose or gain weight. Journal of Consulting and Clinical Psychology. Vol. 55: 742-747, 1987 Rosen, J.C, Tacy, B., Howell, D. Life stress, psychological symptoms and weight reducing behavior in adolescent girls: A prospective analysis. International Journal of Eating Disorders. Vol. 9(1): 17-26, 1990 Rossiter, E.M., Agras, W.S., Telch, C.F., Bruce, B. The eating patterns of non-purging bulimic subjects. International Journal of Eating Disorders. Vol. 11(2): 111-120. 1992 Schlundt, D.G., Johnson, W.G. Eating disorders: Assessment and treatment. Boston: Allyn & Bacon, 1990 Schotte, D.E., Stunkard, A.J. Bulimia vs bulimic behaviours on a college campus. Journal of the American Medical Association. Vol. 258 (9): 1213-1215, 1987 Schutz, Y. Macronutrients and energy balance in obesity. Metabolism: Clinical and Experimental. Vol. 44 (9 suppl 3): 7-11, 1995 Serdula, K.K., Collins, M.E., Williamson, D.F., Anda, R.F., Pamuk, E., Byers, T.E. Weight control practices of U.S. adolescents and adults. Annals of Internal Medicine. Vol. 119 (7): 667-671,1993 Shaw, B.F., Garfinkel, P.E. Research problems in the eating disorders. International Journal of Eating Disorders. Vol. 9 (5): 545-555, 1990 153 References Sheatsley, P.B. Questionnaire construction and item writing. In: Rossi, P.H., Wright, J.D., Anderson, A . B . (eds.) Handbook of Survey Research. New York, N Y : Academic Press. 1983 Shisslak, C M . , Crago, M . , Estes, L.S. The spectrum of eating disturbances. International Journal of Eating Disorders. Vol . 18 (3): 209-219, 1995 Shisslak, C M . , Crago, M . , Neal, M.N . , Swain, B. Primary prevention of eating disorders. Journal of Consulting and Clinical Psychology. Vol . 55(5): 660-667, 1987 Spitzer, R.L., Devlin, M . , Walsh, B.T., Hasin, D., Wing, R., Marcus, M . , Stunkard, A. , Wadden, T., Yanovski, S., Agras, S., Mitchell, J., Nonas., C. Binge eating disorder: A mulisite field trial of the diagnostic criteria. International Journal of Eating Disorders. Vol . 11(3): 191-203, 1992 Statistical Packages of the Social Sciences (SPSS), Professional Statistics for Windows, (Version 6.0) Chicago, IL, 1993 St Jeor, S.T. The role of weight management in the health of women. Journal of the American Dietetic Association. Vol . 93: 1007-10012, 1993 Stephenson, M.G. , Levy, A.S., Sass, N.L. , McGarvey, W.E. 1985 NHIS Findings: Nutrition knowledge and baseline data for the weight-loss objectives. Public Health Reports. Vol . 102(1): 61-67, 1987 Stoner, S.A., Fedoroff, I .C, Anderson, A.E. , Rolls, B.J. Food preferences and desire to eat in anorexia and bulimia nervosa. International Journal of Eating Disorders. Vol . 19(1): 13-22, 1996 Story, M . , Rosenwinkel, K., Himes, J. H. , Resnick, M . , Harris, L. J., Blum, R. W. Demographic and risk factors associated with chronic dieting in adolescents. American Journal of Diseases in Childhood. Vol . 145: 994-998, 1991 Striegal-Moore, R.H., Silberstein, L.R., Rodin, J. Toward an understanding of risk factors for bulimia. American Psychologist. Vol . 41(3): 246-263, 1986 Sundgot-Borgen, J., Oseid, S., Maehlum, S. Etiology and pathogenesis of eating disorders in female elite athletes. Medicine and Science in Sports and Exercise. Vol . 24:(suppl) 184 (abstract), 1992 Szmukler, G.I. Weight and food preoccupation in a population of English schoolgirls. In: Ross Laboratories. Understanding anorexia nervosa and bulimia. Report of the fourth Ross conference on medical research. Columbus, Ohio: Ross Laboratories, 1983 Szmulker, G.I. The epidemiology of anorexia nervosa and bulimia. Journal of Psychiatric Research. Vol . 19(2/3): 143-153, 1985 154 References Tarasuk, V. , Beaton, G.H. Statistical estimation of dietary parameters: Implications of pattern in within-subject variation - a case study of sampling strategies. American Journal of Clinical Nutrition. Vol . 55: 22-27, 1992 Taub, D.E., Blinde, E . M . Eating disorders among adolescent female athletes: Influence of , athletic participation and sports team membership. Adolescence. 27(108): 833-48, 1992 Telch, C.F., Agras, W.S., Rossiter, E .M. , Binge eating increases with increasing adiposity. International Journal of Eating Disorders. Vol .7 : 115-119, 1988 Telch, C.F., Agras, W.S. The effect of a very low calorie diet on binge eating. Behavior Therapy. Vol . 24: 177-193, 1993 Thompson, J., Manore, M . Predicted and measured resting metabolic rate of male and female endurance athletes. Journal of the American Dietetic Association. Vol . 96(1): 30-34, 1996 Tonkin R.S. Practical approaches to eating disorders in adolescence. Primer for family physicians. Canadian Family Physician. Vol . 40: 299-304, 1994 Tonkin, R.S. Eating disorders: Are we doing enough?. British Columbia Medical Journal. Vol . 31(3): 144-146, 1989a Tonkin, R.S., Wigmore, E.J. Epidemiology, classification, and diagnosis of eating disorders. British Columbia Medical Journal. Vol . 31(3): 147-150, 1989b Tonkin, R., Cox, D., Milner, R. Adolescent Health Survey: Province of British Columbia. Vancouver, B . C . : The McCreary Centre Society, 1993 Tuschl, R.J. From dietary restraint to binge eating: some theoretical considerations. Appetite. Vol . 14: 105-109, 1990a Tuschl, R.J., Laessle, R.G., Platte, P., Pirke, K . M . Differences in food-choice frequency between restrained and unrestrained eaters. Appetite. Vol . 14: 9-13, 1990b Wadden, T.A., Bartlett, S., Letizia, K .A. , Foster, G.D., Stunkard, A.J. , Conill, A . Relationship of dieting history to resting metabolic rate, body composition, eating behaviour, and subsequent weight loss. American Journal of Clinical Nutrition. Vol . 56: 203S-208S, 1991 Wadden, T.A., Foster, G.D., Letizia, K . A . One-year behavioral treatment of obesity: Comparison of moderate and severe caloric restriction and effects of weight maintenance therapy. Journal of Consulting and Clinical Psychology. Vol . 62(1): 165-171, 1994 Weingarten, H.P., Elston, D. The phenomenology of food cravings. Appetite. Vol . 15:231-246, 1990 155 References Westenhoeffer, J., Pudel, V . Pleasure from food: Importance for food choices and consequences of deliberate restriction. Appetite. Vol . 20: 246-249, 1993 Whitaker, A. , Davies, M . , Shaffer, D., Johnson, J., Abrams, S., Walsh, B.T., Kalikow, K. The struggle to be thin: a survey of anorexic and bulimic symptoms in a non-referred adolescent population. Psychological Medicine. Vol. 19: 143-163, 1989 Whitehouse, A . M . , Button, E.J. The prevalence of eating disorders in a U.K. college population: A reclassification of an earlier study. International Journal of Eating Disorders. Vol . 7: 393-397, 1988 Willet, W. Future directions in the development of food-frequency questionnaires. American Journal of Clinical Nutrition. Vol . 59 (suppl): 171S-174S, 1994. Williamson, D.F., Serdula, M.K. , Anda, R.F., Levy, A. , Byers, T. Weight loss attempts in adults: Goals, duration, and rate of weight loss. American Journal of Public Health. Vol . 82 (9): 1251-1257, 1992 Wilson, G.T. Relation of dieting and voluntary weight loss to psychological functioning and binge eating. Annals of Internal Medicine. Vol. 119 (7) Part 2: 727-730, 1993 Wilson, G.T., Nonas, J.A., Rosenblum, G.D. Assessment of binge-eating in obese patients. International Journal of Eating Disorders. Vol. 13: 25-34, 1993 Wilson, G.T., Eldredge, K .L . Pathology and development of eating disorders: Implications for athletes. In: Brownell, K.D. , Rodin, J., Wilmore, J.H. (eds.) Eating and Weight Problems in Athletes. Philadelphia: Lea & Febiger, 1992 Woodward, C.A., Chambers, L.W. Guide to questionnaire construction and question writing. Ottawa, Ont.: The Canadian Public Health Association. 1986 Zeman, F.J., Ney, D . M . Applications in medical nutritional therapy. Second Edition. Columbus, Ohio: Merrill, 1996 Zuckerman, D .M. , Colby, A. , Norma, C , Lazerson, J.S. The prevalence of bulimia among college students. American Journal of Public Health. Vol . 76(9), 1986 156 Appendices 157 Appendix D: Questionnaires Administered to Subjects at Baseline D l : The General Information Sheet General Information Sheet 1. Birth Date: 2. What year of university are you in? . What is your degree? Major? ,— 3. What is your highest level of education? _ _ 4. Please check the situation that best describes the family you grew up in: a) Two parent family b) One parent family c) Step-family d) Other. Please Describe: 5. Mother's occupation: 6. Father's occupation: '. 7. What is your ethnic background? D2: The Dieting Practices Survey Dieting Practices Survey The concept of dieting has many meanings in our society. In our study, we have defined "dieting" to mean an actual change in eating habits for the purpose of losing weight. We also consider regulating your food intake to maintain your current weight as a form of dieting. 1. At what age did you start your first diet? years old 2. Estimate the total number of years in which you have dieted. Include only those years that you were on a diet for a total of 2 weeks or longer. years 3. Over the last 6 months, estimate the total number of days that you were dieting. days 4. On average, how much weight do you lose each week of dieting? pounds 5. Over the last 6 months, how often do you exercise for a period of at least one hour? i • less than once a week 2Q 1-3 times per week 3D more than three times per week 165 6. People often use a variety of methods to lose or control their weight Please indicate how often you have used any of the following dieting methods over the past 6 months. Never Rarely Sometimes Often Aiwa a) A particular diet that you have read about (for example in a book, magazine or newspaper) oQ id 2 d 3 d <• b) Information or advice you have received from a health professional (dietitian, nurse, doctor) od i d 2 d 3 d 4d c) A diet you have heard about on television, video or from friends or family od iQ 2 d 3d 4d d) A diet plan that is part of a commercial weight loss program (ex. Weight Watchers®, od Jenny Craig® or others) iQ id 3 Q <• e) A diet that you have made up yourself oQ i d 2 d 3d <• £) A liquid formula or drink that replaces a meal or several meals oQ I Q 2 d 3d 4d g) You eat the same way you usually eat but you eat smaller portions and avoid or cut down on certain foods oQ id 2 Q 3d 4d h) You skip meals od iQ 2Q 3d 4d i) You count calories oQ iQ 2 Q 3d 4d j) You avoid eating when hungry od id 2 d 3d aQ k) You fast (that is, go without food for more than 24 hours) od id 2 d 3d 4d 1) you count grams of fat od id 2 d 3d 4d I) Please list any other dieting method(s) that you may use to lose weight: 166 7. When on a diet some people eat at very fixed times each day while others have no such routine. For each of the following time periods, please indicate if you usually have a meal or snack. We have separated week days from weekends. (Do not include any water, diet sodas, tea or coffee that you may drink.) Example: When dieting, Jane usually has a glass of juice at 7:30 AM and then a muffin and a glass of milk before class at 820 AM. She normally skips lunch but at 5:30 PM has a salad and bowl of soup. After studying until 11:00 PM, Jane usually has a snack such as a bowl of popcorn. Her answer would look like this: 0 1 2 3 4 5 It varies from dav to dav 5:00-9:00 AM a a a • a • 9:00 -11:00 AM a • • • • • 11:00 AM-2:00 PM • • • • • a 2:00-5:00 PM • • • a • • 5:00 - 7:00 PM a • • • • • 7:00 -9:00 PM • • • • • a 9:00 PM-12.00 AM • • a • • • 12:00 AM - 5:00 AM • • • • • • Monday - Friday Number of meals and/or snacks eaten during these time periods while dieting 0 I 2 3 4 5 It varies from dav 5:00 - 9:00 AM a a • a • a a 9:00- 11:00 AM • • • • • • a 11:00 AM-2:00 PM • a • • • • a 2:00 - 5:00 PM a a a a • a a 5:00 - 7:00 PM • a a a a • a 7:00 - 9:00 PM a a a • • a a 9:00 PM- 12.00 AM • • • • • • • 12:00 AM-5:00 AM • • • • • • a 167 7. (Continued) Saturdays and Sundays Number of meals and/or snacks eaten during these time periods while dieting 0 1 2 3 4 5 It varies from day to day 5:00 - 9:00 AM • • • • • • • 9:00-11:00 AM a a a • • • • 11:00 AM-2:00 PM • • a a a a a 2:00 - 5:00 PM • • • • • • a 5:00 - 7:00 PM a • a a • • a 7:00 - 9:00 PM a • • a • a • 9:00 PM-12.00 AM a • • a • • • 12:00 AM-5:00 AM • a a a a • • 168 D3: Food Frequency Questionnaire Part II Food Intake This part of the survey is designed to determine your usual food intake over the last 6 months. Complete the chart on the next pages including foods and beverages consumed both at home and away from home. Please read the items carefully and take your time filling in the chart. For every food mark Yes or No. If Yes, indicate the number of times and mark day, week or month. Mark one of the serving sizes. Here are some examples showing how to complete the chart. "Sarah chinks 1% milk once a day - about 1 1/2 cups each time" This is how she would show that on the chart Examples 2. 1% milk and beverages made with it Do you have this food or beverage at least once a month ? O No About how many times per day or week or month? About how much do you have each time? ©'Day O Week O Month O 1/2 cup 0 1 cup More than 1 cup "Sarah eats whole wheat bread in a sandwich for lunch about five times a week, two slices each time." She would record her bread this way. 22. Wholewheat or light rye ^ bread and rolls &Yes ONo ODay O Month 1-2 slices O 3-4 slices O 5 or more slices "Sarah only eats roast beef or steak every 3 or 4 months." She would show that on the food chart like this. 27. Beef and steak roasted or stewed O Yes- O Day O Week O Month O 4 ounces O More than 4 ounces O Less than 4 ounces 169 Do you have this food or beverage at least once a month? White or Chocolate Milk To Drink About how many times per day or week or month? About how much do you have each time? 2. 5. Skim milk and beverages O Yes made with it O No 1% milk and beverages O Yes made with it O No 3. 2% milk and beverages O Yes • made with it O No Whole milk and O Yes beverages made with it O No Milkshakes O Yes ONo O Day O Week O Month O Day O Week O Month O Day O Week O Month ODay O Week O Month O Day O Week O Month O 1/2 cup O 1/2 cup O 1/2 cup O 1/2 cup O 1/2 cup 0 1 cup 0 1 cup O 1 cup O 1 cup O 1 cup O more than 1 cup O more than 1 cup O more than 1 cup O more than 1 cup O more than 1 cup Cheese, Yogurt and Eggs 6. Hard cheese such as cheddar O Yes ONo O Day O Week O Month 0 1 inch cube O More than 1 inch cube O Less than 1 inch cube 7. Skim milk cheese O Yes such as low fat O No mozzarella O Day O Week O Month O 1 inch cube O More than 1 inch cube O Less than 1 inch cube Processed cheese slices O Yes (including on O No sandwiches and hamburgers) O Day O Week O Month 0 1 slice O . 2 slices O more than 2 slices Cottage cheese O Yes ONo O Day O Week O Month O 1/2 Cup O More than 1/2 cup O Less than 1/2 cup 10. Low fat cottage cheese O Yes-» O No ODay O Week O Month O 1/2 Cup O More than 1/2 cup O Less than 1/2 cup 170 Do you have this food or beverage at About how many times least once a per day or week or month? month? About how much do you have each time? 11. Any other cheese and O Yes —> O cheese spreads ONo O O 12. Yogurt O Yes-> O ONo O O 13. Low fat yogurt O Yes ~» O ONo O O 14. Eggs OYes-> O ONo O O Day O O O Week 1 in. cube More man Less than Month /1 Tbsp. cube/Tbsp. cube/Tbsp. Day O O O Week small large 1/2 cup Month carton carton Day O O O Week small large 1/2 cup Month carton carton Day O O O Week 1 2 3 or more Month egg eggs eggs Breakfast Cereals 15. Whole grain hot cereals O Yes (rolled oats, red river) O No 16. Instant hot cereals OYes ONo 17. Cold cereals, no sugar O Yes (Shredded Wheat, ONo Com Flakes, Rice Krispies, Cheerios) 18. Bran type cold cereals O Yes (Bran Flakes, All Bran ONo Raisin Bran, etc.) 19. Sweetened cold cereals O Yes (FrostedFlakes,Sugar ONo Smacks) 20. Granola OYes ONo O Day O Week O Month ODay O Week O Month O Day O Week O Month ODay O Week O Month O Day O Week O Month ODay O Week O Month O 3/4 cup O 3/4 cup O 3/4 cup O 3/4 cup O 3/4 cup O 3/4 cup O more than 3/4 cup O more man 3/4 cup O more than 3/4 cup O more than 3/4 cup O more than 3/4 cup O more man 3/4 cup O less than 3/4 cup O less than 3/4 cup O less than 3/4 cup O less than 3/4 cup O less than 3/4 cup O less than 3/4 cup 171 21. If you eat cereal a) Do you usually add sugar? O Yes ONo b) Do you usually add artificial sweetener? O Yes O No c) Which one of the following do you use most often on your cereal? O Cream/Half & Half OWholemilk O 2% milk O l%milk O Skim milk d) How much milk do you add to your cereal? O O O 1/2 cup 1 cup more than 1 cup Do you have this food or beverage at About how many times least once a per day or week or month? month? About how much do you have each time? Breads, Rolls and Muffins 22. Wholewheat or light rye O Yes -> bread and rolls O No ODay OWeek O Month O 1-2 slices O 3-4 slices O 5 or more slices 23. Dark rye, pumpernickel O Yes —> fibre-enriched bread O No and rolls ODay OWeek O Month O 1-2 slices O 3-4 slices O 5 or more slices 24. White, Italian, French O Yes -> egg, raisin bread O No and rolls, bagels, hotdog or hamburger buns ODay OWeek O Month O 1-2 slice O 3-4 slices O 5 or more slices 25. Bran or com muffins O Yes -> ONo ODay OWeek O Month 0 1 muffin O 2 muffins O 3 or more muffins 26. Any other muffins, O Yes -> such as blueberry, plain O No chocolate chip ODay OWeek O Month 0 1 muffin O 2 muffins O 3 or more muffins 27. Pancakes or waffles O Yes —• ONo ODay OWeek 0 1 O 2 O 3 or more O Month 172 28. If you eat bread, do you add Always Usually Sometimes Rarely/Never Butter, margarine or cream cheese O O O O Diet margarine or cream cheese O O O O Mayonnaise or salad dressing O O O O Low calorie mayonnaise and dressing O o O O Peanut butter O o O O Jelly, jam, honey or other sweet spread O o O O If you eat muffins, do you add Always Usually Sometimes Rarely/Never Butter, margarine or cream cheese O O O O Diet margarine or cream cheese O O O O Mayonnaise or salad dressing O O O O Low calorie mayonnaise and dressing O O O O Peanut butter O O O O Jelly, jam, honey or other sweet spread O O O O Do you have this food or beverage at least once a month? About how many times per day or week or month? About how much do you have each time? Meat, Poultry, Fish and Alternatives 29. Beef and steak, O Yes -> roasted or stewed O No 30. Pork and pork chops, O Yes • roasted or stewed O No 31. Fried or breaded O Yes • beef, steak, pork O No pork and pork chops 32. Liver, any type OYes ONo 33. Chicken, turkey or O Yes • other poultry, roasted, O No stewed or barbecued ODay O Week O Month ODay O Week O Month O Day O Week O Month ODay O Week O Month ODay O Week O Month O 4 ounces O 4 ounces O 4 ounces O 4 ounces O 1-2 slices o O More than Less than 4 ounces 4 ounces O O More than Less than 4 ounces 4 ounces O O More than Less than 4 ounces 4 ounces O O More than Less than 4 ounces 4 ounces O O 3-4 slices 5 or more slices 173 Do you have this food or beverage at About how many times least once a per day or week or month? month? About how much do you have each time? 34. Fried chicken, nuggets, O Yes -> O Day O O O chicken sandwiches ONo OWeek 2 pieces 4 pieces More than O Month 6 nuggets 9 nuggets 4 pieces / 1 s'wich 2 s'wich 9 nuggets 35. Fish, canned, fresh, OYes-> O Day O O O frozen (ex. tuna ONo OWeek 4 ounces More than Less than salmon, sushi) . O Month 4 ounces 4 ounces 36. Fried fish, fried O Yes-> O Day O O O fish sandwiches ONo OWeek 2 pieces 4 pieces more than O Month 1 s'wich 2 s'wich 9 nuggets 37. Hamburgers and O Yes-> ODay O O O cheeseburgers ONo OWeek 4 ounces more than less than O Month 4 ounces 4 ounces 3 8. Wieners, hot dogs O Yes-> ODay O O O ONo OWeek regular large/2 regular more than O Month 1 large/2 re 39. bacon O Yes-> O Day O O O ONo OWeek 1-2 3-4 5 or more O Month slices slices slices 40. Sausages O Yes-> O Day O O O ONo OWeek 1-2 3-4 1-2 large O Month links links sausages 41. Coldcuts, luncheon O Yes-> O Day O O O meats (bologna, salami, ONo OWeek 1-2 3-4 5 or more chicken loaf or ham) O Month slices slices slices 42. Tofu, soy bean curd O Yes-> O Day O O O ONo OWeek 1/2 cup more than less than O Month 1/2 cup 1/2 cup 43. Always Usually Sometimes Rarely/Never a) If you eat meat or chicken, do you add gravy ? O O O O b) If you eat meat, do you eat the fat? O O O O c) If you eat chicken do you eat the skin? O O O O d) If you eat fish, do you have tartar sauce or O O O O mayonnaise with it? 174 Do you have this food or beverage at About how many times least once a per day or week or month? month? About how much do you have each time? Mixed meat, fish or chicken dishes 44. Meat and chicken pies O Yes-> ONo O Day OWeek O Month O 1-2 slices O 3-4 slices O 5 or more slices 45. Any other mixed dishes O Yes —> made with ground meat, O No chicken and fish O Day OWeek O Month O 1 cup O more than 1 cup O less than 1 cup 46. Spaghetti, lasagna, other pasta with meat-tomato sauce OYes-> ONo ODay OWeek O Month O 1 cup O more than 1 cup O less than 1 cup 47. Macaroni and cheese, other pasta dishes with cheese O Yes ONo O Day OWeek O Month O 1 cup O more than 1 cup O less than 1 cup 48. Pizza O Yes-> ONo ODay OWeek O Month O 1-2 slices O 3-4 slices O 5 or more slices 49. Any other pasta or noodles O Yes-> ONo ODay OWeek O Month O 1 cup O more than 1 cup O less than 1 cup 50. Rice, any type O Yes-» ONo O Day OWeek O Month O 1 cup O more than 1 cup O less than 1 cup Soups 51. Vegetable or noodle-type soup O Yes -> ODay O O O ONo OWeek 1 cup more than less than O Month 2 cups 2 cups O Yes-» O Day O O O ONo OWeek 1 cup more than less than O Month 2 cups 2 cups 175 Do you have this food or beverage at About how many times least once a per day or week or month? month? About how much do you have each time? Vegetables 53. Broccoli OYes-> ONo ODay OWeek O Month O 1/2 cup O more than 1/2 cup O less than 1/2 cup 54. Carrots O Y e s - » ONo ODay OWeek O Month O 1/2 cup O more than 1/2 cup O less than 1/2 cup 55. Corn Q Yes-> ONo ODay OWeek O Month O 1/2 cup small cob O more than 1/2 cup O less than 1/2 cup 56. Green Peas OYes-> ONo ODay OWeek O Month O 1/2 cup O more than 1/2 cup O less than 1/2 cup 57. Greens (spinach, kale, bok choy, leeks) O Yes-> ONo ODay OWeek O Month O 1/2 cup O more man 1/2 cup O less than 1/2 cup 58. Green beans, string beans, yellow beans O Yes -> ONo O Day OWeek O Month O 1/2 cup O more than 1/2 cup O less than 1/2 cup 59. Any other beans, peas lentils (lima beans, navy, baked, pork and beans, kidney beans) OYes-> ONo ODay OWeek O Month O 1/2 cup O more than 1/2 cup O less than 1/2 cup 60. Potatoes, baked, salad, boiled OYes-). ONo O Day OWeek O Month O 1 cup O more than 1 cup O less than 1 cup 61. French fries, home OYes -> fries, pan fried potatoes ONo hash browns ODay OWeek O Month O 1/2 cup O more than 1/2 cup O less than 1/2 cup 62. Squash, all types O Yes-> ONo O Day OWeek O Month O 1/2 cup O more than 1/2 cup O less than 1/2 cup 63. Salad - combination lettuce and tomato O Yes-> ONo ODay OWeek O Month O 1 cup 0 more man 1 cup 0 less man 1 cup 176 64. Any other salads such as coleslaw, carrot, bean, spinach 65. Any other vegetables such as cabbage, Brussels sprouts Do you have this food or beverage at least once a month? OYes-* ONo OYes ONo About bow many times per day or week or month? About how much do you have each time? ODay OWeek O Month ODay OWeek O Month O 1 cup O 1/2 cup O more man 1 cup O more than 1/2 cup O less than 1 cup O less than 1/2 cup 66. Always a) If you eat potatoes or rice do you add • butter, margarine, gravy or sour cream? O • diet margarine, defatted gravy or diet sour cream? O b) If you eat vegetables, do you add • butter, margarine, cheese or other sauces? O • diet margarine, low fat cheese sauces? O c) If you eat salads, do you add • regular mayonnaise, salad dressing or salad oil? O • diet, low fat, low calorie dressing or mayonnaise? O Usually O O O O O O Sometimes Rarely/Never O O O O O O O O O O O O Fruit 67. Apples, applesauce OYes-* ONo ODay OWeek O Month O 1 apple/ 1/2 cup 0 2 apples/ 1 cup O More than 2 apples/2 cups 68. Bananas O Yes-> ONo ODay OWeek O Month O 1 banana O 2 bananas O 3 or more bananas 69. Oranges, grapefruit OYes-> ONo ODay OWeek O Month O 1 orange 1/2 g'fruit O 2 oranges 1 g'fruit O more than 3 oranyi g'fruit 70. Pears, peaches, nectarines, grapes plums O Yes-> ONo O Day OWeek O Month O 1 fruit 1/2 cup 0 2 fruit 1 cup O more than 3 fruit/2 cups 71. Raisins, prunes other dried fruit OYes-* ONo ODay OWeek O Month O 1/2 cup O 1 cup 0 more than 1 cup 177 Do you have this food or beverage at least once a month? About how many times per day or week or month? About how much do you have each time? 72. Cantaloupe O Yes-* ONo ODay OWeek O Month O less than 1/4 melon O 1/4 melon 73. Any other fruit, including berries, fruit cocktail and salad OYes-* ONo O Day OWeek O Month O 1 fruit 1/2 cup O 2 fruit 1 cup Beverages 74. Orange juice and other citrus juices O Yes -* ONo ODay OWeek O Month O 1/2 cup O 1 cup 75. Apple and other juices O Yes-> ONo ODay OWeek O Month O 1/2 cup O 1 cup 76. Tomato, mixed vegetable juices OYes-* ONo ODay OWeek O Month O 1/2 cup O 1 cup 77. Fruit drinks such as Tang or Kool-Aid O Yes-* ONo O Day OWeek O Month O 1/2 cup O 1 cup 78. Regular soft drinks (TVOrdiet) OYes-* ONo ODay OWeek O Month O small or 1 can O medium 79. Diet soft drinks O Yes-> ONo ODay OWeek O Month O small or 1 can O medium 80. Beer, wine or liquor OYes-* ODay O O ONo OWeek O Month 1 glass wine or 1 can beer or 1 oz liquor more than 1 glass wine or 1 can beer or 1 oz liquor O more than 1/4 melon O more than 2 fruit/2 cups O more than 1 cup O more than 1 cup O more than 1 cup O more than 1 cup O large O large O less than 1 glass wine or 1 can beer or 1 oz liquor 81. Coffee OYes ONo ODay OWeek O Month O 1 cup O 2 cups O 3 or more cups 178 82. Tea O Yes ^  ODay O O O O No O Week 1 cup 2 cups 3 or more O Month cups 83. If you drink coffee a) Do you add sugar? O Yes O No b) Which ONE of the following do you use most often? O No milk or cream O Cream or O 2% milk O l%milk O Skim milk whole milk 84. If you drink tea a) Do you add sugar? O Yes O No b) Which ONE of the following do you use most often? O No milk or cream O Cream or O 2% milk O l%milk O Skim milk whole milk Dessert and Snacks Do you have this food or beverage at least once a month? About how many times per day or week or month? About how much do you have each time? 85. Ice cream, ice milk, sherbet, frozen yogurt 86. Cake OYes-* O No OYes-* ONo ODay OWeek O Month ODay OWeek O Month O 1 scoop O 1 slice O 2 scoops O 2 slices O 3 or more scoops O 3 or more slices 87. Pie O Yes -* ONo ODay OWeek O Month O O 1 slice 2 slices O 3 or more slices Cookies OYes-* ONo O Day OWeek O Month O 1-5 O 5-10 O more than 10 89. Crackers, O Yes Ritz, cheese-type, O No Triscuits O Day OWeek O Month O 1-5 O 5-10 O more than 10 90. Donuts, danish croissant OYes ONo O Day OWeek O Month 0 1 O 2 O 3 or more 179 Potato chips Popcorn Peanuts, other nuts seeds Chocolate Do you have this food or beverage at least once a month? O Yes-> ONo OYes ONo OYes ONo OYes ONo About how many times per day or week or month? ODay OWeek O Month ODay OWeek O Month ODay OWeek O Month ODay OWeek O Month About how much do you have each time? O small bag O 2 cups O 1/2 cup O regular bar o O more than less than small bag small bag O O more than less than 2 cups 2 cups O O more than less than 1/2 cup 1/2 cup O O large bar 2 pieces Do you use at least About how many once a capsules or tablets per month ? day, week or month? Vitanun/mineral O Yes -> O Day supplements O No O Week O Month 180 D4: The Health Information Questionnaire HIQ Please indicate the extent to which you feel you are overweight or underweight. a) Extremely overweight b) Very overweight c) A little overweight d) Weight is just right e) A little underweight f) ^ Very underweight g) Extremely underweight 2) Please indicate the extent to which you feel your shape is attractive. a) Shape is extremely unattractive b) Shape is very unattractive c) Shape is a little unattractive d) Neutral e) Shape is moderately attractive f) Shape is very attractive g) Shape is extremely attractive 3) What Is your present weight? (pounds) 4) What is your present height? (feet/inches) 5) Have you LOST more than 10 pounds in the last year? (please circle) YES NO If YES, how many times has this happened? If YES, was there any special reason? 6) Have you GAINED more than 10 pounds In the last year? (please circle) YES NO If YES, how many times has this happened? If YES, was there any special reason? , 7) Many people at some time feel afraid to eat because they think they will gain weight During the past year, have you ever had this fear? a) Never b) Hardly ever cj Sometimes d) Often e) Very often f) All the time During the past year, has feeling afraid to eat ever led you to do any of the following? (please circle) a) Refuse to eat even though you were hungry YES NO bj Try to get rid of food you have just eaten YES NO c) Pretend to others you have eaten YES NO 8) During the past month, have you been worried about your eating habits? a) NO, not at all worried b) YES, slightly worried c) YES, somewhat worried d) YES, very worried If YES, have you worried for any of the following reasons? a) I dont eat enough YES NO b) I eat too much YES NO c) I get urges to stuff myself YES NO d) I cant seem to stop eating once I start YES NO e) I eat when I am upset YES NO f) I eat when I am not hungry YES NO g) 1 eat too much junk food YES NO 181 9) Do you ever experience episodes of eating a large amount of food in a relatively short amount of time (i.e., less than 2 hours)? a) Never b) j_ Less than once a month in the last year c) About once in the last month d) About once a week e) Between 2 and 6 times per week f) Every day g) ' More than once every day 10) Do you ever feel that you CANNOT STOP [EATING or CONTROL what or how much you are eating? a) All the time b) Very often c) Often d) Sometimes e) Hardly ever f) Never 11) Do your ever try to CONTROL YOUR WEIGHT by EXERCISING (i.eM exercising with the primary Intention of burning calories)? a) Never b) Less than once a month in the last year c) About once In the last month d) Once or twice a week e) Between 3 and 5 times per week f) About every day g) More than once every day 12) Do you ever try to LOSE WEIGHT by going on a 'CRASH DIET" (Le., eating at least some food but much less than you usually eat for at least a few days)? a) Never b) Once in the last year c) More than once in the last year d) Once In the last month e) More than once in the last month 13) Do you ever try to CONTROL YOUR WEIGHT by FASTING (i.e., no solid food for at least 24 hours)? a) Never b) Less than once a month in the last year c) About once in the last month d) About once a week e) Between 2 and 6 times per week f) Every day g) More than once every day 14) Do you ever try to CONTROL YOUR WEIGHT by using OIET PILLS? a) Never b) Less than once a month in the last year c) About once in the last month d) About once a week-e) Between 2 and 6 times per week f) Every day g) More than once every day 15) Do your ever try to CONTROL YOUR WEIGHT by using DIURETICS or WATER Pit IS? a) Never b) Less than once a month in the last year c) About once in the last month d) About once a week e) Between 2 and 6 times per week f) Every day g) More than once every day 16) Do you ever try to CONTROL YOUR WEIGHT by VOMITING? a) Never b) Less than once a month in the last year c) About once in the last month d) About once a week . e) Between 2 and 6 times per week f) Everyday g) More than once every day 17) Do you ever try to CONTROL YOUR WEIGHT by using LAXATIVES? a) Never b) Less than once a month in the last year c) About once in the last month d) About once a week e) Between 2 and 6 times per week f) Everyday g) More than once every day 18) In the past 3 months, have you had the experience of others telling you that you are thin or normal weight whereas you feel fat? a) All the time b) _j Often c) Sometimes d) Hardly ever e) Never 19) Do you feel that being at your current weight presents any significant health risks? a) No risk at all b) Possibly some risk c) Definitely at risk 20) To what extent is the way you feel about yourself (i.e„ positively or negatively) affected by your shape and weight? a) Almost entirely b) To a very large extent c) Somewhat d) _ A « t l e e) Not at all 21) For Females Only. Are your periods regular? (please circle) YES NO 22) For Females Only. Have your periods stopped AT ANY TIME during the past year? YES NO If YES: For how many months? Was this at a time you had lost weight? YES NO If this was due to physical illness, please describe below: 23) Have you ever received treatment for an eating disorder? YES NO If YES, when? 183 D5: The Health Information Questionnaire (Coding Key) Only items 7 through 22 on the Health Information Questionnaire contributed to the Symptom Intensity Score. Total score ranges between 0 and 80. Items were coded as follows: Item 7 Never = 0 to A l l the time = 5 YES = 1 NO = 0 Item 8 NO, not at all worried = 0 to YES, very worried = 3 YES = 1 NO = 0 Items 9, 11, 13-17 Never = 0 to More than once everyday = 6 Item 10 A l l the time = 5 to Never = 0 Item 12 Never = 0 to More than once in the last month = 4 Item 18 A l l the time = 4 to Never = 0 Item 19 Coded only if the criteria for anorexia nervosa are met No risk at all = 2 to Definitely at risk = 0 Item 20 Coded only if the criteria for anorexia nervosa and bulimia nervosa are met Almost entirely = 4 to Not at all = 0 Item 21-22 If response to Item 21 is No and If response to Item 22 is > 3 and YES, then Items 21 & 22 coded together as 1 184 E1: The Dieting Practices Survey (6 Month Version) Dieting Practices Survey The concept of dieting has many meanings in our society. In our study, we have defined "dieting" to mean an actual change in eating habits for the purpose of losing weight. We also consider regulating your food intake to maintain your current weight as a form of dieting. 1. By the definition of dieting given above, are you currently dieting? (Yes or No) Please answer aU of the following questions even if you are not currently dieting. At what age did yon start your first diet? years old 3. Estimate the total number of years in which you have dieted. Include only those years that you were on a diet for a total of 2 weeks or longer. years 4. Over the last 6 months, estimate the total number of days that you were dieting. days 5. On average, how much weight do you lose each week of dieting? pounds 186 6. People often use a variety of methods to lose or control their weight Please indicate how often you have used any of the following dieting methods over the past 6 months. Never Rarely Sometimes Often Always a) A particular diet that you have read about (for example in a book, magazine or newspaper) <>• iQ iQ b) Information or advice you have receivec. from a health professional (dietitian, nurse, doctor) <>• iQ 2 Q c) A diet you have heard about on television, video or from friends or family <>• iQ iQ iU <• d) A diet plan that is part of a commercial weight loss program (ex. Weight Watchers®, oQ Jenny Craig® or others) iQ 2d 3 Q <• e) A diet that you have made up yourself oQ iQ 3 Q <a f) A liquid formula or drink that replaces a meal or several meals oQ iQ zQ 3Q g) You eat the same way you usually eat but you eat smaller portions and avoid or cut down on certain foods oQ iQ JQ 3 Q h) You skip meals oQ iQ iQ 3Q <a i) You count calories oQ iQ 2 Q 3d j) You avoid eating when hungry oQ iQ iQ zU AQ k) You fast (that is, go without food for more than 24 hours) oQ iQ IO AQ 1) you count grams of fat oQ id 3Q AU 1) Please list any other dieting method(s) that you may use to lose weight: 187 7. When on a diet some people eat at very fixed times each day while others have no such routine. For each of the following time periods, please indicate if you usually have a meal or snack. We have separated week days from weekends. (Do not include any water, diet sodas, tea or coffee that you may drink.) Example: When dieting, Jane usually has a glass of juice at 7:30 AM and then a muffin and a glass of milk before class at 8:20 AM. She normally skips lunch but at 5:30 PM has a salad and bowl of soup. After studying until 11:00 PM, Jane usually has a snack such as a bowl of popcorn. Her answer would look like this: 0 1 2 3 4 5 It varies from dav to dav 5:00-9:00 AM • • • • a • 9:00 -11:00 AM • a a a • • 11:00 AM- 2:00 PM a • • • a • 2:00-5:00 PM • a • a • • . 5:00-7:00 PM • • • a • • 7:00-9:00 PM • • • • • • 9:00 PM-12.00 AM • • • • a • 12:00 AM-5:00 AM a • a • a • Monday - Friday Number of meals and/or snacks eaten during these time periods while dieting 0 I 2 3 4 5 It varies from ds 5:00 - 9:00 AM • a • • • • a 9:00- 11:00 AM a • • • • • • 11:00 AM-2:00 PM • • a a • a • 2:00 - 5:00 PM a a a a • • a 5:00 - 7:00 PM a • a a • • a 7:00 - 9:00 PM a a • a • • a 9:00 PM-12.00 AM a • • • • • a 12:00 AM-5:00 AM a a • • • • • 188 7. (Continued) Saturdays and Sundays Number of meals and/or snacks eaten during these time periods while dieting 0 I 2 3 4 5 It varies from day to dav • 5:00 - 9:00 AM • • a • • a • 4 9:00-11:00 AM • • • a a • • • 11:00 AM -2:00 PM a a • • a a • • 2:00 - 5:00 PM • a • • a • • • 5:00 - 7:00 PM a a a • • a • • 7:00 - 9:00 PM a • a • • • a • 9:00 PM-12.00 AM • a • a a a • • 12:00 AM-5:00 AM a • • • a • a 189 Appendix F: Letter to Subjects with 'Not In Service' Telephone Numbers at 6 Months T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A School of Family and Nutritional Sciences Division of Human Nutrition 2205 East Mall Vancouver, B.C. Canada V6T 1Z4 March 30, 1994 Dear , In October of last year, you participated in our Eating Behaviours of College-Age Women Study. We are now beginning the second part of tie study which involves the completion of a second questionnaire. This questionnaire requires only about 20 minutes of your time. Once completed your name will be placed in the draw for one of three prizes: First prize - a weekend for two at Whistler (or $6"00 cash) Second prize-$300 Third prize -$100. Recently, I have been unsuccessful in contacting you as the phone number I have for you is not in service. We were fortunate to have you participate in the first part of the study and would be disappointed not to have you complete the study. It would be greatly appreciated if you would phone the School of Family and Nutritional Sciences at 822-3934 and leave a message that includes your current phone number (this number is a 24-hour answering service). I will contact you soon after and arrange to mail you the questionnaire. Thank you for your time. Your input is essential to the success of the study. Yours sincerely, Ellen Mackay, B. Sc., Graduate Student, Human Nutrition, UBC. 190 Eating Behaviours in College-Age Women Summary Dieting is a common practice among college-age women. It has been speculated that dieting may predispose susceptible individuals to unhealthy weight loss practices and eating disorders. The purpose of this prospective study was to examine the relationship between specific dieting practices and disordered eating in female college and university students who were dieting. In September and October, 1994, students were recruited from six post-secondary campuses in the Greater Vancouver area: The University of British Columbia, Simon Fraser University, Capilano College, Langara Community College, King Edward College, and Vancouver Community College. Several questionnaires were administered to the students, including the Dieting Practices Questionnaire, a food frequency questionnaire and The Health Information Questionnaire. Responses to The Health Information Questionnaire yielded a Symptom Intensity Score, a measure of disturbed eating practices. The higher the Symptom Intensity Score, the more frequently the individual engaged in abnormal eating behaviours such as binge eating, purging and fasting. At the end of the academic year (March and April, 1995), the Dieting Practices Survey and the Health Information Questionnaire were re-administered to the students. One hundred and seventy-nine dieting women completed both parts of the study. Based on responses to the follow-up Health Information Questionnaire, 15 students developed a probable eating disorder. Significant correlationships were found between the Symptom Intensity Score and age at first diet, number of years dieting, counting grams of fat, counting calories and total number of dieting methods used. These results suggest that individuals with a long history of dieting and those individuals using these dieting practices are more likely to have abnormal eating behaviours. Overall, however, dieting practices did not predict the development of disordered eating in this sample. Significant differences were observed between students who continued dieting through the academic year and students who ceased dieting. Students who continued dieting had a greater initial Symptom Intensity Score, reported a higher frequency of calorie/fat counting and food avoidance dieting methods (such as skipping meals and fasting), reported lower carbohydrate, fat and energy intakes, and were dieting more days over the previous 6 months than the group of students who eventually stopped dieting. These results suggest that specific dieting practices may not predict disordered eating, but may have a role in perpetuating dieting. (This study was supported by a grant from the British Columbia Medical Services Foundation.) 193 Appendix J: Serving Size Key used in Coding the Food Frequency Questionnaire White or Chocolate Milk To Drink Item 1-5 more than 1 c —¥ 1.5 c Cheese, Yogurt and Eggs Item 6-7 more than 1 inch cube —» 1.5 x 1.5 inch cubes less than 1 inch cube —> 0.5 x 0.5 inch cube Item 8 more than 2 slices —> 3 slices Item 9 - 10 more than 0.5 c -> 0.75 c less than 0.5 c -> 0.25 c Item 11 more than 1 inch cube/Tbsp —> 1.5 cube/Tbsp less than 1 cube/Tbsp —> 0.5 cube/Tbsp Item 12-13 small carton —> 175 g large carton —> 750 g Item 14 3 or more eggs —> 3 eggs Breakfast Cereals Item 15 - 20 more than 0.75 c -» 1.5 c less than 0.75 c —> 0.5 c Item 21a 1 tsp. sugar per 1 c serving of cereal indicated Item 2 Id more than 1 c —> 1.5 c of the type of milk indicated in Item 21c Breads, Rolls and Muffins Item 22-24 5 or more servings —> 5 servings Item 25 -27 3 or more muffins —» 3 muffins Item 28 always —» 1 tsp./slice (serving) usually —> 0.75 tsp./slice (serving) sometimes —> 0.5 tsp./slice (serving) rarely —> 0 195 Meat, Poultry, Fish and Alternatives Item 29 - 32 Item 33,39,41 Item 34 Item 35 Item 36 Item 37 Item 38 Item 40 Item 42 Item 43 (a & d) Item 43 (b & c) more than 4 oz —» 6 oz less than 4 oz —> 2 oz 1-2 slices —> 1.5 slices 3-4 slices —» 3.5 slices 5 or more slices —» 5 slices 2 pieces —> 3 oz 4 pieces —» 6 oz more than 4 pieces —> 9 oz more than 4 oz —> 6 oz less than 4 oz —> 2 oz more than 9 nuggets —> 9 oz more than 4 oz —» 6 oz less than 4 oz —> 2 oz regular —> 40 g large/2 regular —> 100 g more than 1 large/2 regular —» 140 1-2 links -» 1.5 links 3-4 links -» 3.5 links 1-2 large sausages —» 1.5 sausages more than 0.5 c —» 0.75 c less than 0.5 c -» 0.25 c always -» 1 Tbsp./serving usually —> 0.75 Tbsp./serving sometimes —> 0.5 Tbsp./serving rarely —> 0 always —> 1 tsp. fat/serving usually —> 0.75 tsp. fat/serving sometimes —> 0.5 tsp. fat/serving rarely —> 0 Mixed meat, fish or chicken dishes Item 44 Item 45 - 47 Item 48 Item 49 - 50 Soups Item 51 - 52 Vegetables Item 53-59, 61-62, 65 Item 60, 63, 64 Item 66 Fruit Item 67 1-2 slices —» 1.5 slices 3-4 slices —> 3.5 slices 5 or more slices —> 5 slices more than 1 c —> 1.5 c less than 1 c —> 0.5 c 1-2 slices —» 1.5 slices 3-4 slices —> 3.5 slices 5 or more slices —> 5 slices more than 1 c —» 1.5 c less than 1 c —> 0.5 c more than 2 c —> 2.5 c less than 2 c -> 1.5 c more than 0.5 c —> 0.75 c less than 0.5 c -» 0.25 c more than 1 c —> 1.5 c less than 1 c —» 0.5 c always —> 1 tsp./slice (serving) usually —> 0.75 tsp./slice (serving) sometimes —» 0.5 tsp./slice (serving) rarely —> 0 more than 2 apples —> 3 apples or 2 c Item 68 3 or more bananas —> 3 bananas Item 69 more than 3 oranges/ 1 grapefruit —> 1.5 grapefruit Item 70 more than 3 fruit/2 c - > 2 c Item 71 more than 1 c —> 1.5 c Item 72 less than 0.25 melon -» 0.17 (l/6th) melon more than 0.25 melon —> 0.5 melon Item 73 more than 2 fruits/2 c —> 2 c Beverages Item 74-77 more than 1 c —> 1.5 c Item 78-79 small/ 1 can -> 355 mL medium —> 2 c large —> 2.5 c Item 80 more than 1 glass wine, etc —> 1.5 glass wine, etc. (average of group) less than 1 glass wine, etc. —» 0.5 glass wine, etc. (average of group) Item 81-82 3 or more c - > 3 c Item 83a & 84a 1 tsp. sugar per serving indicated Dessert and Snacks Item 85 1 scoop —> 0.33 c 2 scoops —> 0.66 c 3 scoops -» 1 c Item 86-87, 90 3 or more -» 3 Item 88-89 more than 10 -> 12.5 Item 91 small bag-> 2 oz more than a small bag —» 3 oz less than a small bag —> 1 oz Item 92 more than 2 c —> 3 c less than 2 c —> 1 c 198 Item 93 more than 0.5 c -» 0.75 c less than 0.5 c -> 0.25 c Item 94 2 pieces —> 2 oz Appendix L: Pretest Questionnaires Dieting Practices Survey Please indicate the exact time you begin this questionnaire. © Part I For each question, please mark the appropriate box unless otherwise indicated. The concept of dieting has many meanings in our society. In our study, we have defined "dieting" to mean an actual change in eating habits for the purpose of losing weight We also consider regulating your food intake to maintain your current weight as a form of dieting. Yes No 1. By this definition, are you currently dieting? • • > I JfYes, please answer Questions 2-10 If No please turn to Part II 2. At what age did you start your first diet? years old 3. Not including your current diet, approximately how many attempts at dieting did you make over the last 12 months? • 0 • 1-3 • 4-6 • 7-10 • more than 11 • I am always dieting • do not know 4. In your opinion, are you dieting • more often now than you were 6 months ago? • less often than you were 6 months ago? • about the same amount over the last 6 months? • do not know. S. Is the diet that you are following now different from the way you eat while not dieting? • not at all different • a little different • very different 200 6. People often use a variety of methods to lose or control their weight Please indicate how often you have used any of the following dieting methods over the past 6 months. Never Rarely Sometimes Often Always a) A particular diet that you have read about • (for example in a book, magazine or newspaper) • • • • b) Information or advice you have received from a health professional (dietitian, nurse, doctor) • • • • • c) A diet you have heard about on television, -video or from friends or family • • • • • d) A diet plan that is part of a commercial weight loss program (ex. Weight Watchers®, • Jenny Craig® or others) • • • • ') A diet that you have made up yourself • • • • • 0 A liquid formula or drink that replace a meal or several meals • • • • • g) You eat the same way you usually cat but you eat smaller portions and avoid certain foods • • • • • h) You skip meals • • • • • i) You count calories • • • • • i) You avoid eating when hungry • • • • • k) You fast (that is, go without food for more than 12 hours) • • • • • I) Please list any other method(s) that you may use. On average, how long does your diet last? • hours • 1-3 days • 4-6 days • between 1 and 2 weeks • 3-4 weeks • more than I month • I am always dieting 201 8. When on a diet, some people eat at very fixed time each day while others have no such routine. For each of the following time periods, please indicate if you usually have a meal or snack. We have separated week days from weekends. (Do not include any water, diet sodas, tea or coffee you drink.) Example: On week days, Jane usually has a glass of juice at 7:30 AM and then a muffin and a glass of milk before class at 8:20 AM. She normally skips lunch but at 5:30 PM has a salad and bowl of soup. After studying until 11:00 PM, Jane usually has a snack such as a bowl of popcorn. Her answer would look hke this: 0 1 2 3 4 5 It varies from dav to dav 5:00 - 9:00 AM • • a a a • 9:00-11:00 AM a a • a • • 11:00 AM- 2:00 PM a a a • • • 2:00-5:00 PM • • • • • a 5:00 - 7:00 PM • • a a a • 7:00 -9:00 PM a a • a • a 9:00 PM-12.00 AM • a-" • a • a • 12:00 AM-5:00 AM • • • a • • Monday - Friday Number of meals and/or snacks eaten during this time period 0 I 2 3 4 5 It varies from dav to dav 5:00 - 9:00 AM • a a • a a • 9:00- 11:00 AM a a • a a a • 11:00 AM-2:00 PM a • a • • • a 2:00-5:00 PM a a • a • • a 5.00 - 7:00 PM a • • • • a a 7:00 - 9:00 PM • • a • • a • 9:00 PM - 12.00 AM • • a a • a • 12:00 AM-5:00 AM a • a a a • a 202 8. (continued) Saturdays and Sundays Number of meals and/or snacks eaten during this time period 0 1 2 3 4 5 It varies from day to dav 5:00 - 9:00 AM • • • • • • a 9:00- 11:00 AM • • • a • • • 11:00 AM -2:00 PM • • • • • • • 2:00 - 5:00 PM • • • • • • • 5:00 - 7:00 PM • • • • • • • 7:00 - 9:00 PM • • • • • • • 9:00 PM - 12.00 AM • • • • • • • 12:00 AM - 5:00 AM • • • • • • • 9. With each new diet you start, do you: • eat about the same amount of food each time you diet? • eat more food than you did the last time you dieted? • eat less food than you did the last time you dieted? • do not know. 10. What is the most common cause for ending your diet? • you've lost enough weight • eating a food you consider "bad" or "forbidden strong feeling of hunger • boredom • a depressed mood • being upset or nervous • pressure from family or friends to stop dieting • an uncontrollable urge to eat • you feel you're not losing enough weight • none of the above • other (please specify) 203 Part II Food Frequency Questionnaire This part of the survey is designed to determine your average food intake over a month. Please read the items carefully and take your time filling in the chart. Do not hesitate to ask for assistance if necessary. 204 Three Day Food Record Instructions Please choose two week-days and one weekend day for this record. The days do not have to be consecutive. Please write down all food, snacks and beverages consumed each day. You may use the attached sheets or a note book of your own. Keep your food record with you so that you can record food and snacks soon after consuming them Record the amounts of food consumed. You may use the measuring units you are most familiar with such as cups, litres, tablespoons, millilitres, ounces or grams. Be as specific about the food or drink as possible. This will ensure our nutritional analysis of your diet is accurate. For example: O include milk and sugar used in tea or coffee, if applicable. •=t> record any margarine or butter used on toast indicate if you had orange juice or orange drink •=> write down the brand name if known •=> be sure to indicate if the food or drink is 'diet* or lite' (ex: diet pop, 'lite' beer, diet mayonnaise or margarine, artificial sugar, low fat cheese, etc.) •=> for casserole dishes or meals with multiple ingredients, try to write down all ingredient if possible or provide the recipe •=> indicate how the food was prepared (baked, barbecued, microwaved, fried, stir fiied, roasted, steamed, etc.) Example Day Breakfast: Vi large pink grapefruit with 1 tsp. white sugar % c Kellogg"s Bran Flakes with V4 c 1% milk 2 x 1 c coffee ( 1 Tbsp 1% milk in each cup) iLunch: VAc l°/omilk VAc baked beans in tomato sauce 1 plain bagel (toasted) with 1 Tbsp 'lite' margarine | Snack: lean diet 7-up 1 cinnamon roll (from UBC cafeteria) ) Supper: 1 large (7 inch) com on the cob with salt and 1 tsp butter 4 oz piece of chicken, with skin, barbecued 1 Tbsp barbecue sauce 114 c tossed green salad with 2 Tbsp Thousand Islands 'lite' Dressing 2 bottles beer 1 c fruit salad (strawberries, watermelon, blueberries and apples) If you have any problems or questions, please leave a message for Ellen Mackay at 822-2502. Thank you for your participation in this study! 205 F O O D RECORD Diet Record Day#:__ Day of Week: Date: TIME & PLACE DESCRIPTION OF FOOD & BEVERAGE ITEMS QUANTITY CONSUMED (specify amounts) • • bid you take a vitamin/mineral supplement on this day? (Y/N) If yes, please state the type, brand name, and the number of pills you took. Was this a fairly typical day for you? (Y/N) ~ If not, please give reason(s): 206 Questionnaire Evaluation ~& Please indicate the exact time at which you finished this questionnaire. © Answer Pari I only if you completed this section, otherwise please complete Part II of the evaluation. Part I - Dieting Behaviors Were you able to understand all the instructions on the questionnaire? Were you able to understand all the questions on the questionnaire? Were you able to understand all the options for your answers? Were all the options appropriate for the question? Was the order of the questions logical? Part II - Food. Frequency Questioimaire Were you able to understand all the instructions on the Food Frequency questionnaire? Were all the foods you commonly eat on the questionnaire? Which foods were missing? Overall Please give us your overall impression on the questionnaire. (Was it easy to fill out? Easy to read? Did the definition of dieting make sense? Was it too long? Other general comments?) Please use the back of this sheet it you need more room. Thank you for your participation! 1!! 207 Appendix M : Self-Reported Use of Dieting Methods by Study Participants at Baseline Dieting Methods Never Rarely Sometimes Often Always n(%) (n=174) a) Diet from a book, newspaper 82(46.1) 46 (25.8) 36 (20.2) 11 (6.2) 3(1.7) or magazine b) Information received from a 61 (34.7) 31 (17.6) 48 (27.3) 28 (15.9) 8 (4.5) health professional c) Diet heard about on TV, 99 (55.9) 27 (15.3) 40 (22.6) 8 (4.5) 3 (1.7) video, or from friends/family d) A diet that is part of a 151 (84.4) 6 (3.4) 13 (7.3) 7 (3.9) 2(1.1) commercial weight-loss program e) A diet that you have 14 (7.8) 6 (3.4) 36 (20.1) 76 (42.5) 47 (26.3) made up f) A liquid formula that replaces 125 (69.8) 23 (12.8) 23 (12.8) 7 (3.9) 1 (0.6) meals g) Smaller portions/avoids or 7 (3.9) 9(5.1) 38 (21.3) 67 (37.6) 57 (32.0) cuts down on certain foods h) Skips meals 47 (26.3) 32(17.9) 51 (28.5) 41 (22.9) 8 (4.5) i) Counts calories 78 (43.8) 34(19.1) 30(16.9) 21 (11.8) 15 (8.4) j) Avoids eating when hungry 54 (30.2) 57 (31.8) 45 (25.1) 17 (9.5) 6 (3.4) k) Fasts (goes without food for 134 (74.9) 28 (15.6) 12(6.7) 3(1.7) 2(1.1) more than 24 hours) 1) Counts grams of fat 60 (33.5) 22 (12.3) 32(17.9) 28 (15.6) 37 (20.7) 208 Appendix N: Self-Reported Use of Dieting Methods by Study Participants at 6 Months Dieting Methods Never Rarely Sometimes Often Always n (%) (n=174) a) Diet from a book, newspaper 90 (50.3) 45 (25.1) 36 (20.1) 6 (3.4) 2(1.1) or magazine b) Information received from a 67 (37.6) 24(13.5) 49 (27.5) 33(18.5) 5 (2.8) health professional c) Diet heard about on TV, video, 104 (58.4) 33(18.5) 32(18.0) 7 (3.9) 2(1.1) or from friends/family d) A diet that is part of a 156 (87.6) 11 (6.2) 6 (3.4) 5 (2.8) 0 (0.0) commercial weight-loss program e) A diet that you have made up 21 (11.7) 13 (7.3) 36 (20.1) 72 (40.2) 37 (20.7) f) A liquid formula that replaces 149 (83.2) 19 (10.6) 6 (3.4) 4 (2.2) 1 (0.6) meals g) Smaller portions/ avoids or 10 (5.6) 9(5.0) 47 (26.3) 69 (38.5) 44 (24.6) cuts down on certain foods h) Skips meals 56 (31.3) 45 (25.1) 53 (29.6) 21 (11.7) 4 (2.2) i) Counts calories 94 (52.5) 34 (19.0) 24 (13.4) 23 (12.8) 4 (2.2) j) Avoids eating when hungry 72 (40.2) 54 (30.2) 39 (21.8) 10 (5.6) 4 (2.2) k) Fasts (goes without food for 148 (82.7) 18(10.1) 10 (5.6) 3 (1.7) 0 (0.0) more than 24 hours) 1) Counts grams of fat 65 (36.5) 17 (9.6) 31 (17.4) 37 (20.8) 28 (15.7) 209 Appendix O: Self-reported Weight Loss (kg) per Week Dieting According to Dieting Status at 6 Months in Study Participants Weight loss/week of dieting (kg) Baseline 6 Months Sample Mean (n=174) 0.69 ± .66* 0.55 ± .63a Dieters at 6 months (n=125) 0.68 ± 0.65 0.45 ± 0.56bc Non-dieters at 6 months (n=53) 0.74 + 0.77 0.74 ± 0.75 * M e a n ± S D ' a Signif icantly different from Baseline: p < .01 (Paired sample t-Test) b S i gnificantly different from Baseline: p < .001 (Paired sample t-Test) c Signif icantly different from Non-dieters at 6 months: p<.01 (Independent t-Test) 210 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0087211/manifest

Comment

Related Items