Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A computerized system for instruction in food selection practice Prince, Peter Robert 1979-03-03

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
831-UBC_1979_A6_7 P75.pdf [ 19.89MB ]
Metadata
JSON: 831-1.0094593.json
JSON-LD: 831-1.0094593-ld.json
RDF/XML (Pretty): 831-1.0094593-rdf.xml
RDF/JSON: 831-1.0094593-rdf.json
Turtle: 831-1.0094593-turtle.txt
N-Triples: 831-1.0094593-rdf-ntriples.txt
Original Record: 831-1.0094593-source.json
Full Text
831-1.0094593-fulltext.txt
Citation
831-1.0094593.ris

Full Text

A COMPUTERIZED SYSTEM FOR INSTRUCTION IN FOOD SELECTION PRACTICE by PETER ROBERT PRINCE B. Sc., University of British Columbia, 1969 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES SCHOOL OF HOME ECONOMICS DIVISION OF HUMAN NUTRITION We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA July 1979 (c) Peter Robert Prince, 1979 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 representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Depa rtment The University of British Columbia 2075 Wesbrook Place Vancouver, Canada V6T 1W5 i i ABSTRACT This thesis has developed arprototypical system which provides informa tion on dietary practices for those individuals interested in applying nutritional principles to their eating habits. The system has the poten-tion to provide information which both accurately reflects nutritional guidelines and facilitates adoption of recommendations, by providing a self-explanatory statement of foods to consume and by 1imiting suggested changes in present food pattern. The prototypical computerized system developed has two major functions: (i), diet-assessment to appraise the acceptability of individual's dietary practices; and (ii), diet-planning to recommend modifications in the diets of those individuals not meeting specified limits. The focus of the system is a constrained-optimization algorithm that generates a revised food plan which both satisfies nutrient constraints, and minimizes the deviation of food items rand item groups from the original amount consumed by the client. Testing has been restricted to a descriptive evaluation of some of the algorithm's characteristics -- specifically, the design assumptions which define the acceptability of deviating from an original inventory, and the revised diets developed when these assumptions are modified. The results illustrate that altering these design assumptions produces marked varia tions in the revised diets with respect to observed parameters. Further modifications in the algorithm have been suggested. The explorative evaluation provides'ia foundation for more systematic evaluation of the validity of the algorithm. Recommendations for facili tating the further development and testing of the system are^outlined. 1 ii This thesis has shown that mathematical modeling provides an effective means of collating the vast amount of data required to develop cogent dietary recommendations which are nutritionally accurate, straightforward, and acceptable to the client. iv TABLE OF CONTENTS ABSTRACT 11 TABLE OF CONTENTS 1V. . LIST OF TABLES v1"1"LIST OF FIGURES x. ACKNOWLEDGEMENTS x1. INTRODUCTION 1 1.1 Background and Need for the Study 1 1.1.1 Fundamental Tasks of Nutrition Education Programs... 1 1.1.2 Criticisms of Nutrition Education Programs 2 1.1.2.1 Developing a Scientific-Nutritional Code 4 1.1.2.2 Communicating a Nutritional Code to the Public 5 1.1.2.2.1 Communication With a Variety of Audiences 6 1.1.2.2.2 Complex Dynamics of Communication With the Public 7 1.1.2.2.2.1 Changing Food Market 7 1.1.2.2.2.2 Mass Media and Advertising 9 1.1.3 Summary and Conclusions 11 1.2 Thesis Goal :and Objectives 3 1.2.1 Thesis Goal 11.2.2 Thesis Objectives1.2.2.1 The First Objective of the Thesis 13 1.2.2.2 The Second Objective of the Thesis 4 2. REVIEW OF LITERATURE 15 2.1 Definition of an Adequate Diet 12.2.1 Influences of Diet in Human Populations 15 2.1.1.1 Influences on the Individual 12.1.1.2 Socio-Economic Influences of Diet 17 2.1.2 Criteria of an Adequate Diet 12.1.3 Nutrient Requirements 19 2.2 Relevant Systems 21 2.2.1 Food Guides2.2.1.1 Fallibility in Guiding Food Selection Practices... 22 2.2.1.2 Inapplicability to the Present Food Supply 24 2.2.1.3 Relevance to Contemporary Nutrition Problems 25 2.2.1.4 Ineffectiveness as a Teaching Tool 26 V 2.2.2 Other Relevant Systems 27 2.2.2.1 Pennington's Dietary Nutrient Guide 22.2.2.2 Nutrition Labelling 28 2.2.2.3 Computer Applications in Nutrition and Dietetics.. 29 2.2.3 Critique and Conclusions 32 2.3 Dietary Assessment 35 2.3.1 Data Collection 9 2.3.1.1 Data Collection Methods 32.3.1.2 Evaluation of Data Collection Methods 42 2.3.1.2.1 Validity and Repeatability of Methods... 44 2.3.1.2.2 Time Frame of Methods 46 2.3.1.2.3 Randomness and Size of Population-Samples 48 2.3.1.2.4 Other Metholodigical Considerations 49 2.3.1.3 Summary and Conclusions 50 2.3.2 Data Analysis 52 2.3.3 Data Evaluation 62.3.3.1 Dietary Standards2.3.3.1.1 Formulation of Dietary Standards 64 2.3.3.1.2 Use and Interpretation of Dietary Standards 66 2.3.3.1.3 Limitations of Dietary Standards 72 2.3.3.2 Further Dietary Recommendations 4 2.3.3.2.1 Recommendations for Nutrients Not in the Dietary Standard.; 74 2.3.3.2.2 Recommendations for Maximum Intakes 76 2.3.3.3 Conclusions 77 2.4 Dietary Prescription 80 2.4.1 Diet Planning2.4.1.1 Food Planning Models 82 2.4.1.1.1 Food Planning Without Palatability : ,.: i Considerations2.4.1.1.2 Food Planning Models with Palatability Considerations 87 2.4.1.2 Menu Planning Models 93 2.4.1.2.1 Menu Planning Models - The Random Approach 93 2.4.1.2.2 Mathematically-Programmed, Multistage, Menu-Planning 94 2.4.1.2.3. Mathematically-Programmed, Single-Stage, Menu-Planning 7 vi 2.4.2 Information and Behavior Change 98 2.4.2.1 Development of Food Behavior and Factors in Food Selection 99 2.4.2.2 The KAP Gap 100 2.4.2.3 Increasing the Effectiveness of Communication 102 3. DEVELOPMENT OF THE PROTOTYPICAL SYSTEM 105 3.1 Introduction 103.1.1 Developing Nutritional Guidelines 105 3.1.2 Communicating Information to Individuals 106 3.1.2.1 Comprehensibility of the Information 106 3.1.2.2 Acceptability of Suggested Changes 107 3.2 System Design and Characteristics 109 3.2.1 Data-Collection 110 3.2.1.1 Client's Initial Diet 113.2.1.2 Client's Demographic Data for Defining Nutrient Limits 113 3.2.2 Data-Analysis 113.2.3 Data-Evaluation 7 3.2.4 Diet-Planning 121 4. TESTING OF THE PROTOTYPICAL SYSTEM 128 4.1 Introduction 124.2 Diet-Planning Model's Premises and Assumptions 136 4.2.1 The First Premise and Related Assumptions 136 4.2.2 The Second Premise and Related Assumptions 139 4.3 Observations of the Objective Function's Characteristics 140 4.3.1 Unconstrained Objective Function 144.3.2 Constrained Objective Function 141 4.3.2.1 First Term: Shape of the Curve 144.3.2.2 First Term: Penalty Coefficient wi 145 4.3.2.2.1 Penalty Coefficient, w., Based on Amount Consumed 146 4.3.2.2.2 Penalty Coefficient, , Based on Initial Consumption 158 4.3.2.2.3 Penalty Coefficient, , Further Comments 175 4.3.2.3 First Term: Further Modifications of the Algorithm 176 4.3.2.4 Second Term of the Objective Function 177 vii .4.3.2.5 Second Term; Shape of the Curve 180 4.3.2.6 Second Term: Penalty Coefficients, w.. and Pjk 182 4.3.2.6 Second Term': Further Modifications of the Algorithm 184.4 Summary and Comments on Testing the Algorithm 196 5. SUMMARY, RECOMMENDATIONS, AND CONCLUSIONS 200 5.1 Summary 205.1.1 The Thesis Goal 205.1.2 First Objective of the Thesis 200 5.1.2.1 Diet-Assessment Function5.1.2.2 Diet-Planning Function.... 201 5.1.3 Second Objective of the Thesis 202 5.2 Recommendations 203 5.2.1 First Recommendation 205.2.2 Second Recommendation 5 5.3 Conclusions 20LIST OF REFERENCES 207 APPENDICES. A Food-Item File 224 B Abridged Food-Item File 236 C Food-Composition File 8 D Abridged Food-Composition File 284 E Nutrient-Limits File 286 F Abridged Nutrient-Limits File 301 G Attribute Group Matrix 303 yiii LIST OF TABLES 2.1 Annual subsistence diet for a moderately active adult male, calculated using linear programming methods 85 3.T Excerpt from prototypical intake questionnaire -- dietary intake format Ill 3.2 Excerpt from prototypical intake questionnaire -- client demographic data 114 3.3 Excerpt from the evaluation output used for system testing . 117 3.4 Excerpt from a proposed evaluation output format 118 3.5 Excerpt from the diet-planning output format used for system testing 121 4.1 Upper and lower nutrient constraints for a standard subject . 129 4.2 Standard Initial Diet 1 (SID-1) and Standard Initial Diet 2 2 (SID-2) 131 4.3 Nutrient composition of SID-1 and SID-2, and the upper and lower nutrient constraints for a standard male subject 135 4.4 Revised diets developed from SID-1 and SID-2 using quadratic objective functions (Eqns. 4.8, 4.9) 148 4.5 Nutrient composition of SID-1 and SID-2, and the revised diets developed using quadratic objective functions (Eqns. 4.8, 4.9) 152 4.6 Average absolute deviation from initial consumption levels for items in revised diets developed from SID-1 and SID-2 using quadratic objective functions (Eqns. 4.8, 4.9) 155 4.7 Revised diets developed from SID-1 using quadratic objective functions (Eqns. 4.9, 4.12) 160 4.8 Revised diets developed from SID-2 using quadratic objective functions (Eqns. 4.8 - 4.12) 164 4.9 Nutrient composition of SID-1, and the revised diets developed using quadratic objective functions (Eqns. 4.9.,-4.12) 168 4.10 Nutrient composition of SID-2, and the revised diets developed using quadratic objective functions (Eqns. 4.8 -4.12) 169 4.11 Average absolute deviation from initial consumption levels for items in revised diets developed from SID-1 using quadratic objective functions (Eqns.; 4.9, 4.12) 172 4.12 Average absolute deviation from initial consumption levels for items revised diets developed from SID-2 using quadratic objective functions (Eqns. 4.8 - 4.12) 173 4.13 Revised diets developed from SID-2 using quadratic objective functions (Eqns. 4.8, 4.26 - 4.28) 184 4.14 Nutrient composition of the SID-2, and the revised diets developed using quadratic functions (Eqns. 4.26 - 4.28) ... 188 4.15 Average absolute deviation of item clusters'and of attribute groups in each hierarchical level from initial levels, and the . penalty coefficients assigned for each hierarchical " -level 190 4.16 Effect of penalty assignment on the number of attribute groups containing consumed items 192 LIST OF FIGURES 2.1 Individual variability in nutrient requirements 68 2.2 Proportion of the population having actual requirements above nutrient intake and the probability of deficiency in individuals ingesting a particular level of nutrients .. 69 2.3 Theoretical model of the relationship of the nutrient intake of a population to the prevalence of deficiency 70 3.1 Overview flowchart of prototypical system for assessment and planning of individual's diets 109 4.1 Graph of quadratic function reflecting penalties for deviation from initial consumption of food i 142 4.2 Graph of quadratic and linear functions reflecting penalties for deviation from initial consumption of food i 144 4.3 Graph of quadratic function for deviation from initial consumption of attribute group G- 180 xi ACKNOWLEDGEMENTS For those who have borne with me in my discovery that the shortest distance between two points is not a thesis, my sincerest thanks. John H. Mil sum William T. Ziemba Donald B. Hausch Nancy E. Schwartz -Melvin Lee Joseph Leichter Thomas J. Abernathy CHAPTER 1 INTRODUCTION 1.1 Background and Need for the Study 1.1.1 Fundamental Tasks of Nutrition Education Programs Nutrition education, as a process to promote the public welfare, is universally needed to ensure healthful food selection. Since "there is no instinct that guides man to select those foods which meet the nutritional needs of the body ... each new generation must be taught what foods to select ..." (Tddhunter 1969, p. 9). Furthermore, it is needed because societal forces, which influence food availability, must parallel'iknowledge of nutritional well-being to ensure that appropriate products will be available for selection. The American Dietetic Association (1973), in "its "Position Paper on Nutrition Education for the Public", defines nutrition education as "... the process by which beliefs, attitudes, environmental influences, and understandings about food lead to practices that are scientifically sound, practical, and consistent with,.individual needs and available food resources" (p. 429). As the above definition implies, the fundamental tasks of nutrition education personnel are two-fold. First, nutrition educators provide a code of nutritional practice through the judicious interpretation of scientific studies, and strive to reduce the time lag between the discovery of nutrition knowledge and its application to food practice (Gifft et al. 1972; Leverton 1974). Second, nutrition educators 2 communicate available knowledge to the public sector with awareness of the variety of audiences - individual citizen, "activated" consumer, govern mental bodies, and industry - that may be approached to ultimately effect changes in food practices by their influence on either available food supply or food selection behaviours (American Dietetic Association 1973; Leverton 1974). Also, in communicating with the public sector, the nutrition educator must be aware of the complex dynamics of communication with target audiences that potentiate or deter change in food habits. 1.1.2 Criticisms of Nutrition Education Programs Man's knowledge of food and its implication for well-being has evolved through the course of centuries. Traditional knowledge of food, previously based on "... a long series of trials and errors -- sometimes mortal errors -- ..." (Mayer 1973, p. xxi) has been greatly extended by the systematic investigation of the present day. The new scientific-technologic tradition has produced an explosion in knowledge of dietary consequences for health, and with this a growing suspicion that this knowledge is not being applied to its full potential. In 1941 nutrition-related medical problems were shown to be affecting much of the potential American military manpower. In that year ...John R. Murlin, participating in a White House Conference on Nutrition and Defense, stated: "Knowledge is like a rock set upon a shelf. It does no harm, and it does no good, so long as it rests there, but let somebody jar the shelf and let the rock fall off, and then something happens. That is the state of things today. We know more than we are doing." (Hill 1969, p. 14) Similar sentiments have been echoed to the present day. Briggs (1969) suggests that "advances in nutrition-related research are only useful when extensively applied ... but the link between this knowledge and its 3 application is extremely weak" (p. 8). Concerning the state of knowledge in the science of nutrition, White (1976) states that "obviously much more is known about nutrition and human needs than is manifest by the current practices of our population" (p. 54). Current criticisms on the gap between nutritional knowledge and its apparent application, stem from western population studies (U. S. Depart ment of Agriculture 1968; U. S. Department of Health, Education, and Welfare 1972; Canada 1973) identifying significant'incidences of nutrition-related problems, and in some, a trend for decline in dietary quality as compared to previous decade's. Also, there have been great increases in mortality from diseases with dietary involvement, which appear to be largely preventable. While health and medical expenditures have continued to rise, no strong evidence exists to indicate that any corresponding benefits have occurred in the major health indices. For example in the United States, life expectancy of those over twenty has not improved in the last twenty-five years, a period when expenditures have increased eightfold (Cornely 1974). Mayer (1975) aptly comments, "like Alice in Wonderland, we are running faster and faster to stay in the same place" (p. ID-Briggs (1969) suggests that "... we do have sufficient scientific knowledge to conduct sound programs of nutrition education ... " (p. 8). If so, then what are the reasons for the discrepancy between this knowledge and population food practices? The reasons for this discrepancy may be categorized in two primary areas which coincide with nutrition education's aforementioned responsibilities, namely, the problem of developing a scientific-nutritional code, and the complexity of translating this code to the public sector. 4 1.1.2.1 Developing a Scientific - Nutritional Code As indicated in "A New Perspective on the Health of Canadians" (Lalonde 1974): The spirit of enquiry and skepticism, and particularly the Scientific Method, so essential to research, are, however, a problem in health promotion. The reason for this is that science is full of "ifs',' "buts", and maybes" while messages designed to influence the public must be loud, clear and unequivocal (p. 57). Indeed, uncertainty in scientific circles is often reflected in ambiguous statements to the public on which course of action to follow. As Mayer (1975) indicates, "we have been telling people what they should eat but avoiding strong statements about what they ought not to eat or to eat less of. We 'stay away from controversial subjects' ..." (p. 8). Unresolved scientific issues would not matter provided they stayed in academic circles. However, the environment in which sound nutrition practices are promoted is competitive and often hostile. Misinformation, often with a seed in scientific debate, is part of this environment -- a fact which vacillation on issues by the scientific community does nothing to alleviate. In fact, vacillation may ultimately breed distrust, by a significant proportion of the public, in science and technology generally, and specifically in the quality of the food supply and in the potential of nutritionists to reflect meaningfully on these issues (Leverton 1974). When information becomes public knowledge, regardless of whether scientists believe the evidence is sufficient for a conclusion, the public begins to act through the voices and opinions of media. Mayer (1972) recommends "given this situation, it is better to act on the conjecture of scientists than on the guesses of newspapermen" (p. 240). Otherwise, "as the nutritionally trained people contemplate their disagreements on approaches to nutrition, many persons not trained in nutrition are making 5 decisions on what foods are available in stores and in what form" (Ullrich 1973a, p. 184). As the complexity and scope of nutrition knowledge increases, and therefore the task of interpretation appears more difficult, it becomes particularly important that comprehensive and coherent recommendations on dietary practices be provided. The failure of nutrition education to meet fully its obligations to provide a code of appropriate practice is suggested by criticisms (Robinson 1976) indicating the absence of clear recommendations for action on controversial aspects of diet, and on newer findings in dietary relationships for health. 1.1.2.2 Communicating a Nutritional Code to the Public The ultimate recipient of nutrition information is the public and it is their dietary well-being which indicate the success of nutrition education activities. The delivery of nutrition information to the public is not the exclusive domain of the nutrition educator, nor is the medium of exchange restricted to "bookish ministrations". In order to effectively communicate nutrition information "nutrition and health educators must be concerned with human behaviour and therefore must compete with all internal and external forces that define and control how an individual behaves" (White 1976, p. 54). Criticisms, discussed below, suggest that nutrition educational programs have failed to keep pace with the changing communication mileau of nutrition information in western societies, both in terms of the variety of audiences that must be approached to ultimately influence the dietary practices of the public, and also the complex dynamics of communi cation with the public in an informationally-competitive society. 6 1.1.2.2.1 Communication with a Variety of Audiences Food habits of individuals are determined by a number of factors including the foods available in the marketplace, and information dissem inated:.in the media -- both factors which have been influenced massively by food technologists, manufacturers, advertisers, and legislators. Thus, in addition to communicating a nutritional code directly to the public through schools, hospitals, and media, nutrition educators must recognize audiences in both industry and government through which the public is influenced, and through which nutrition information can indirectly reach the public (Anon 1972). For example, dietary lifestyle can be influenced through food processing and manufacturing regulations, media policies and advertising guidelines. As Ullrich (1974a) indicates, "objective nutrition education should not be the exclusive responsibility of any one group but a balance among government agencies, food industry, and educational institutions" (p. 84). However, in efforts to promote sound nutrition, as Gussow (1972) observes, "... most professionals were much more worried about the excesses of the 'health food' stores than the excesses of what at least one observer has called the 'unhealth food stores' [and] until very recently, this misplaced concern extended even to advertising" (p. 48 and 49). In attempts to combat misinformation, the colourful, but perhaps insignificant antics of the charlatan have attracted the greatest attention of nutritionists. For example, recent popular food movements, the "health foodists" as coined by Wolff (1973), may be allies and not enemies of nutrition education. Professional emphasis on such groups may have resulted in nutritionists ignoring other, perhaps more significant, forces which influence food habits and which may be important factors promoting misinformation. Further, 7 Hall (1975) suggests that nutritional science is outmoded in its ability to deal with the modern world of nutrition. Nutritionists "... practice a science, outmoded by technological reality, [thus] they can not influence ... [the] political process" (Hall 1974 p. 9). 1.1.2.2.2 Complex Dynamics of Communication with the Public Beyond the influence nutrition educators can have on the public's dietary practices via intermediary bodies, such as government and industry, the nutrition educator can affect public consumption patterns by directly providing information to individuals. The food choices of individuals are determined by a complex of internal and external forces that define and control behaviour (Gifft et al. 1972). Among the external or environmental forces influencing food behaviour are the foods available, mass media, cultural tradition, and governmental policy. Thus, in providing information directly to indivi duals, consideration must be given to the many competing forces that influence food practices -- forces which define the complexity of the informational environment in which nutrition educators must function. With respect to two environmental influences -- foods available in the marketplace and diet-related discussions generated by mass media and advertising --criticisms suggest that nutrition education programs have been inadequate competitors. 1.1.2.2.2.1 Changing Food Market Food technology, industrial development, and rapid transportation have greatly increased the number, kinds, and availability of food products (Todhunter 1969). Whereas in 1928 the average supermarket 8 contained around 900 items, large supermarkets now carry in excess of 10,000 items (Gussow 1972). Thus, "today, more than ever, there are more opportunities to make poor food choices because of the broad array of new foods available" (Anon 1972, p. 34). "... The shopper has the difficult problem in properly selecting the best buys in both nutritional and monetary value" (Todhunter 1969, p. 9). As Ullrich (1975) states, "the technology of providing a large variety of foodstuffs in the marketplace has far outstripped the knowledge of the consumer to make wise choices" (p. 48). The nature of the food supply has been virtually transformed as a result of technological development. "In 1941, only 10 percent of our foods were highly processed; today, that amount has risen to 50 percent" (Mayer 1972, p. 239). In addition to the alterations that processing may cause in nutrient characteristics and distribution of nutrients in foods, approximately 1,830 additives are available for routine use in foods (Hall 1973). Many of these substances are either new in the diet or are present in greater proportion than previously. With the change in food supply has come a change in food habits as illustrated in the following statements by Ullrich (1974b): A look at the present state of our national food consumption compared with 25 or 50 years ago shows a decline in the consumption of foods of high nutrient quality in relation to consumption of foods of low nutrient quality. In part, the decline may be due to the over-zealous development and heavy advertising of "fabricated" food products that are not equal to the conventional foods they replace (p. 4). There is concern in nutritional circles that consumption of products of low nutrient quality coupled with current patterns of low energy expenditure may lead to nutritional problems (Harper 1974). Mertz (1972) suggests that increased use of textured vegetable proteins and refined 9 grains may result in depletion of trace elements. However, the Ten State Survey (U. S. Department of Health, Education, and Welfare 11972) indicates in its "Highlights", " ... inadequate information is available on the distribution of nutrients in today's food supply ... " (p. 12). The long term effect of fabricated diet consumption is not known. Additionally, the consequence of long term consumption of many additives and particularly of combinationsrof additives is not known. To be effective, nutrition education programs must keep pace with the changing food supply, and with the consequent influence of food supply on behaviour and nutritional well-being. Regardless of the cause of the present transformation in foods available, or of the difficulty of interpreting the consequences of these changes, it is the responsibility of the nutrition education profession to provide rational and useful guidelines for the public, industry, and government -- guidelines that are relevant to the present type and variety of foods available. 1.1.2.2.2.2 Mass Media and Advertising As Gussow (1972) points out, "... between 1928 and 1968 people had learned to eat thousands of new food items" (p. 48). If people, as nutritionists say, have food habits which once established are difficult to change, then what change agent has been effective? Her implication is that advertising is a significant factor, and particularly advertising on television. Manoff (1973), who has also publicized the importance of media and advertising as a nutrition education (or miseducation) force, indicates that, "of the 6,000 to 8,000 items on sale in American food stores, 50 percent of them did not exist 20 years ago. This would not have been possible without commercial television, which began roughly 10 20 years ago" (p. 126). Presumably the Canadian situation is similar. These factors are of particular significance to the nutritionist since the massive influence of advertising is motivated by marketing forces and not by a concern for nutritional value. As Manoff (1973) suggests, "food manufacturers produce anything they can sell at a profit. This is the elementary law of marketing" (p. 128). Although the food industry claims to act in the consumer's interest by supplying consumer demands, millions of dollars are spent each year on advertising to influence consumer purchase of foods hav.ing high profit, but often low nut ritive value (Gussow 1972; Manoff 1973 ). Mayer (1975) further questions the motives of the food industry, stating, " ... the opposition to the proposed Consumer Protection Agency in [the United States] is being led by the large food companies. (He adds,]] I leave you to draw your own con clusion on what this may represent" (p. 11). Manoff (1975), paraphrasing and quoting Mayer, suggests: ... the reasons that food manufacturers appear reluctant to concentrate on foods of higher nutritional value ... is that there is no insistent demand for such foods. "Until the American mentality changes", he said, "food manufacturers will feel no strong injunction to provide such products", (p. 139). Although mass media are a significant force in molding population food practices, " ... the mass media has been virtually abandoned by nutrition educators to the commercial food marketer and his nutrition education ..." (Manoff 1973, p. 125). If consumer mentality is to change, "legitimate" nutrition education, in addition to the traditional nutrition education vehicles of school, clinic, and hospital, must make effective use of media and advertising techniques to match the efforts of the food industry (Anon 1972; Manoff 1973; Manoff 1975). Additionally, the formidable force of advertising in 11 molding present attitudes and practices, establishing relevant public nutrition issues, and influencing the nature of the food supply, must be recognized when establishing nutrition education programs. 1.1.3 Summary and Conclusions Since the public is the ultimate user of nutritional information, it is the task of the nutrition education profession to provide the consumer with comprehensive food selection guidelines; and to influence which foods and information are available through activity in.government,:the food industry, educational services, and the marketplace. Nutrition education programs in western countries have been criticized for not keeping pace with the complex changes in food available in the marketplace, and the complex discussion within the community on dietary issues influencing health, whether of legitimate or artifactual origin. If the above difficulties plague the nutrition education profession in its efforts to guide nutritional practice within society, then they also plague individual citizens when nutrition education programs are not effective. That is, when these programs cannot properly provide the neces sary information to the public, the individual must become the sole arbiter of nutrition information. Thus, the individual has the confusing tasks of resolving the complex dietary issues; of interpreting the worth of con flicting recommendations and admonitions from nutritional, medical, media, and marketing sources; of coordinating food selection practice in the face of a vast number of dietary provisos; and of selecting items from an over whelmingly complex food supply -- in short, an information overload. As stated by Mayer, and quoted by Ullrich (1973b), concerning the nutritional literacy of the American people: " ... we have a very tired mind subjected 12 day in and day out to a tremendous amount of information which is mostly misinformation by people who have something to sell" (p. 224). Programs providing nutrition education services directly to the public must consider the complex forces influencing food decisions of individuals and families -- forces which nutrition education can not presently buffer through means other than public education. Unfortunately, the tendency has been to believe that individuals cannot handle complex information on foods, and therefore will benefit more with simplified guidelines on food selection practice. In so doing, nutrition educators leave the consumer as prey to the whims of industry and advertising. Oversimplified nutrition information does not properly equip the consumer for the present world of nutrition information. Although it is not necessary that each individual be a nutrition specialist, she or he should be offered the opportunity to confidently eat as one. It is the responsibility of the nutrition education profession to provide the consumer with the resources necessary to collate available information on foods and nutrition, so that rational choices can be made. This includes useful guidelines which are comprehensive in coverage of recognized and proposed dietary issues, and relevant to the present variety and types of food available in the marketplace. Proposals by leading authors in nutrition indicate a trend to more comprehensive and detailed information. For example, Mayer (1975) states, "in order to teach nutrition to a broader audience, we obviously have to embrace a much broader concept of nutrition than talking about nutrients and foods which existed 25 years ago" (p. 8). 13 1.2 Thesis Goal and Objectives 1.2.1 Thesis Goal The overall goal of the thesis was to develop a prototypical system which provides information on dietary practices for individuals who: (i) want to apply nutritional principles to their eating habits; and (ii) have sufficient resources (eg. time, energy, education, money) to use the information which defines healthful dietary practices for them. This development was undertaken bearing in mind: (i) The complexity of developing a scientific-nutritional code due to the many unresolved dietary issues under investigations, and the consequent problem of translating available knowledge for health promotion at the community level, (ii) The complex factors influencing communication with individuals which the nutrition educator must consider -- that is, an environment with an overwhelming array: of foods to choose from, for an equally overwhelming number of reasons -- and the unsatis factory resolution of this condition by individual exposure to contradictory information provided by media sources. 1.2.2 Thesis Objectives 1.2.2.1 The First Objective of the Thesis The first objective, arising from the thesis goal, was to develop a prototypical computerized system with two major functions, namely: 14 (i) diet-assessment in order to appraise the dietary intake of individuals, and (ii) diet-planning in order to recommend modifications in food intake for those individuals with diets which do not meet specified nutrient limits. Specific characteristics of the system are defined in "System Design and Characteristics" (p.109"). 1.2.2.2 The Second Objective of the Thesis The second objective, arising from the thesis goal, was to test the diet-planning component of the prototypical system. This testing explored some of the conceptual assumptions of a model designed for modifying diets which do not meet specified nutrient limits. The results of this work is reported in "Testing of the Prototypical System" (p.128). 15 CHAPTER 2 REVIEW OF LITERATURE 2.1 Definition of an Adequate Diet The criteria used for appraising diets are ultimately derived from a knowledge of the dynamics of these diets in human populations. Presently, the criteria for appraising diets of normal individuals are defined in the dietary standard (Passmore e_t aVi 1974; United States 1974; Canada 1975) -- in a sense, a compendium of known nutrient requirements used in the evaluation and design of diets. Any discussion about providing information on' food selection practice should consider the theoretical foundation of the dietary standard, and more broadly, the basis for defining an adequate diet. Thus, this discussion begins with a cursory examination of the influence of diet on the human individual and on the society in which the individual lives. 2.2'il Influences of Diet in Human Populations 2.1.1.1 Influences on the Individual Diet has many influences on the human organism. Both human physiology and perception are influenced by dieUs nutrient and non-nutrient constit uents, and its characteristics conferred by human imagination and cultural beliefs. With respect to nutrient constituents, more than forty are presently recognized as essential for normal body functions of growth, maintenance, and repair (United States 1975). Deficiencies of one or more of these nutrients results in a plethora of physiologic and psychologic symptoma-16 tology, touching all systems of the body (Goodhart and Shils 1973; Pike and Brown 1975). Similarly, overnutrition is also associated with disease symptomatology. For example, a nutritional component has been suggested for many of the degenerative diseases (Canada 1976a), and for such ubi quitous problems as dental disease (McBean and Speckmann 1974). In -addition to nutrient involvement in the causation of disease, nutrients may alter physical and mental potentials within the vaguely defined limits of normal well-being. For example, a special athletic dietary regimen, called the glycogen loading diet, maximizes available energy for per formance (Astrand and Rodahl 1970; Astrand 1973). As a further example, Davis and Williams (1976) suggest that diet may influence such factors as sleeping pattern and healing time. Diet can also be a vehicle for an immense array of potentially harmful or beneficial non-nutritive substances. Toxic food substances are found o in the environment naturally (United States 1973) -- for example, l.athyrism and aflatoxin -- and as a biproduct of modern technology which has intro duced many new chemicals into the environment and into foods (Hall 1973; Hall 1977; W. H. 0. 1978). Beneficial non-nutritive substances found in food include the active anti-infective property called lactobifidus factor found in human milk (JeTTiffe and Jelliffe 1971), and dietary fiber (Burkitt and Painter 1974; Klevay 1974; Spiller and Amen 1975). Further, food includes substances which provide taste, color, texture, smell, and other sensations. These influence the desirability of items for consumption, and the nature of the eating experience (Kinder 1973).. Diet in some cultures is believed to have properties beyond the physical. For example, the idea that the "heat" or flesh of a brave enemy or animal conferred courage has existed in cultures from as early as the 17 Stone Age (Lowenberg et_ al_. 1974). Additionally, beliefs based on hygenic observations, and taboos not so empirically founded, exist inrmost cultures (Lowenberg ejt a]_. 1974). Even if these beliefs have no basis in fact, their influence may be sufficient to confer either benefits or detriments; in any case they have consequences for the population even if a physiological effect is absent. 2.1.1.2 Socio-Economic Influences of Diet Diet has implications beyond its physiologic and psychologic influences on the individual, extending to those of a social, economic, and political nature. Authors who have addressed these issues include: Berg (1973), who examines the role of nutrition in national development; Correa (1975), who discusses the influence of nutrition on socio-economic development; and Mitchell (1975) with a consideration of the dynamics of food production and supply in context of Canadian economic policy. 2.1.2 Criteria of an Adequate Diet As indicated, nutritive and non-nutritive elements of the diet have many influences on the individual and on the society in which the individual lives. Thus, to properly identify an optimal or ideal diet potentially requires consideration of this multitude of dietary agents and their con sequences. In turn, to identify desirable consequences may require con sideration of the objectives and aspirations of individuals and their society. "A major objective of the Food and Nutrition Board of the NRC continues to be to encourage the development of food use practices by the population of the United States that will allow for maximum dividends in 18 the maintenance and promotion of health" (United States 1974, p. 1 ); where "health is defined according to the World Health Organization, as 'a state of complete physical, mental and social well-being and not merely the absence of disease or infirmity"1 (United States 1974, p. 1 ). However, even when collective recognition of the desirability of healthful food practices exists, as exemplified by the objectives of a recognized body such as the Food and Nurtition Board, defining the optimal diet may still be difficult -- quite apart from the problems of defining health alone. For example, it may be that not all apparently desirable dietary practices are mutually compatible. "It is not at all certain ... that the nutritional requirements for maximum size, early maturity, active sex life, and maximum muscular development are identical with those for maximum longevity" (Goodhart 1973, p 403 ), resistance to chronic diseases, and maximum physical and mental performance in old age (Harper 1974). Furthermore, "diets designed to protect the individual against bacterial infections, such as Tuberculosis, may lower resistance to certain viral infections and predispose the individual to obesity and coronary heart disease in later years" (Goodhart 1973, p.403 ). Thus, as Goodhart (1973) summarizes, " ... statements such as 'An adequate diet is one which meets in full all the nutritional needs of the person have little meaning unless they can be interpreted in terms of either the person's ambitions for him self or the community's designs for or on him" (p.403 ). Another factor that may make definition of the ideal diet difficult is conflict between dietary objectives and other objectives of an individual or a society. For example, bottle feeding of infants is becoming socially fashionable in third world societies, but economic and technologic constraints often make this practice nutritionally unsound (Jelliffe 1973). 19 However, increased well-being may conceivably result in some instances when apparently less than ideal diets are used in favour of meeting other per sonal or social priorities, as for example, that moderate alcohol consump tion increases well-being, with respect to coronary risk (Yano et al. 1977). In summary, the optimal diet -- the best dietary option described by environmental, biological, societal and personal circumstances -- is the diet which most effectively contributes to the achievement of the goals and aspirations of the society and of the individuals in the society. Its definition involves, in addition to identification of the many relevant consequences diet has for societal and individual realization, choosing among a variety of apparently equivalent but mutually incompatible dietary objectives, which in turn conflict with other personal and societal aspirations. 2.1.3 Nutrient Requirements Although complex criteria, such as those discussed above, are relevant for defining an optimal diet, at present a more modest formulation of desirable dietary practices has been defined. This is the "adequate diet", as described by Goodhart (1973), which is based on human metabolic require ments for essential nutrients. Nutrient dynamics in the body can be envisioned as a contiuum from the extremes of depletion to repletion, presumably with a circumscribed optimum between (Arroyave 1971). Ideally, to monitor these changes, criteria are required which differentiate the metabolic continuum from these extremes, and which identify the optimum for an individual in any particular situation for any nutrient. In the absence of suitable methods 20 and criteria to identify a "nutrient optimum", the establishment of a requirement for a nutrient, defined by the Food and Nutrition Board (United States 1974) as " ... the minimum intake that will maintain normal function and health" (p. 8), " ... rests on the production of a deficiency? and on the definition of that daily intake which prevents or cures the deficiency state" (Mertz 1972, p .19). Criteria to determine these requirements are based on sensitive measures of biological response to nutrient depletion, such as, gross clinical symptomatology, and metabolic, biochemical, or other biophysical-functional changes (Arroyave 1971). The known nutrient requirements are largely outlined in the dietary standards of various countries and international agencies, such as: the "Dietary Standards for Canada" (Canada 1975), the "Recommended Dietary Allowances" (United States 1974) in America, and the "Handbook on Human Nutritional Requirements" (Passmore et_ al_. 1974) by the World Health Organization. The dietary standards, will be discussed more fully in "Data Evaluation" (p. 62). As discussed later (p. 74), other dietary goals beyond those presented in the dietary standard have been proposed. 0. Deficiency is defined as "...an habitual intake by the individual below his own true requirement. The manifestations of such 'deficiency' will depend upon the criteria by which the requirement has been defined ..." (Beaton 1972, p.357). 21 2.2 Relevant Systems 2.2.1 Food Guides Although there are many different approaches to teaching nutrition (Ahlstromand Rasanen 1973), food guides have been the standard model for outlining the adequate diet in nutrition education (Chandler and Perloff 1975; Winarski 1976). Examples of food guides include: Canada's Food Guide (Canada 1977) the Basic Four and Basic Seven (U. S. D. A. 1976, U. S. D. A. 1971), and the Type A Pattern for school lunches (Head ejt al_. 1973). Food exchange lists (Caso 1950)are examples of guides developed for use in planning therapeutic diets. The food guide is intended as a simple and reliable nutrition education device for teaching the principles of healthful food selection. The food guide is a translation of nutrient requirements into a guideline of sug gested serving of items from a few basic groups of foods with roughly equivalent nutrient composition (Ahlstrom and Rasanen 1973; Hertzler and Anderson 1974). Foods are classified into basic groups according to both their major nutrient contributions to the diet, and other criteria which assure the guides' applicability for the population and its utility as a teaching tool. These other criteria include: the food's function in the meal system of the population; foods available; educational status, income, and lifestyle of the population; and local nutritional needs. (Hayes et al. 1955; Lachance 1972; Ahlstrom and Rasanen 1973; Hertzler and Anderson 1974; United States 1974). The problem in the design of food guides is to strike the balance between simplicity -- so that the information can be understood and remembered -- and validity -- so that the guide accurately reflects the dietary standards. To this end adaptations and modifications have been 22 made to meet the needs of different populations and for different purposes, but the basic model of a limited number of natural food groups has remained (Hertzler and Anderson 1974). Even though the utility of the food guide is recognized, it has been criticized as a teaching tool for a number of reasons, including: its falliblity in guiding proper food selection practices; its inapplicability to the present food supply; its lack of relevance for contemporary nutrition problems of the propulation; and its ineffectiveness as a teaching tool in an informationally-competitive society. Each of these criticisms will be dealt:; with below. 2.2.1.1 Fallibility in Guiding Food Selection Practices The validity of the simple plan, based on four or five groups, which is used i.hothe United States and Canada for directing food selection practice has been questioned (Anon 1972). Even for populations whose preferred style of eating is reflected by the food guide, it is recognized that the guide is fallible. For example, it is known that the food guide system can be invalidated by consistently making poorer choices in a food group, skimping on serving sizes, or using improper cooking and preparation practices (Bogert et al_. 1973). Also, the required allowance of nutrients can be obtained from a wide variety of food combinations and patterns, besides those of the food guide (United States 1974). A simplified eating plan, as in the food guide, is based on two basic assumptions: first, only a few key nutrients need to be monitored out of forty-plus possible known and unknown nutrients; and second, the nutrients 23 not being monitored inevitably occur in the variety of foods selected for the key or index nutrient (Bogert et al_. 1973). To test these assumptions Pennington (1976) evaluated two diets selected by Page and Phipart (1957) in accordance with the "Basic Four" plan, against the 1973 Recommended Dietary Allowances of the Food and Nutrition Board (United States 1974). Although the four index nutrients (protein, calcium, vitamin A, and vitamin C) were acceptable, other nutrients (thiamin, riboflavin, niacin and iron) and energy were below the recommended levels. Pennington (1976) comments as follows: Whether or not the deficit for these 4 nutrients or any of the remaining essential nutrients (which total 45 and are ignored with this plan) are met depends on foods chosen to round out energy needs. If "empty calorie" foods are selected ... the chance of getting adequate nutrients is lessened (p. 4 ). Further Pennington (1976) states that "the major problem with the Basic Four seems to be the nonuniformity of major and coincidental nutrients ... within the groups" (p. 6 ). Thus, the Basic Four food group concept inten tionally provides for only 4 nutrients out of a possible 50 or so. The observed nonuniformity of the index and other coincident nutrients in the Basic Four plan would indeed indicate potential fallibility of the plan. In a more recent evaluation, King e_t al_. (1978) compared the nutrient content of 20 published menus based on the Basic Four Food Guide with the 1974 Recommended Dietary Allowance (RDA) for an adult reference male. The Basic Four Foods met or exceeded the RDA's for only 8 of the 17 evaluated nutrients. For 5 of these nutrients, menus supplied 60 percent or less of the standard. A modification of the Basic Four Food Guide for adults was suggested, in order to improve the adequacy of menus developed from this guide. Two studies of food guides were also undertaken in Canada. In the first Canadian study, Milne e_t al_. (1963) compared nutrient intakes to 24 patterns of food usage for adolescents. They found teenagers may consume nutrients in recommended amounts without ingesting foods in the amounts suggested in Canada's Food Guide. Similarly, McClinton et al.''(1971), in the second Canadian food use study, indicated a wide divergence between patterns of foods recommended in the food guide, and those patterns of foods selected by the study population. Out of 1,418 people who met the recom mended allowance for nutrient intake, only one person included foods as recommended in the Canada Food Guide. These results agree in concept with the findings^of Pennington (1976). On the other hand, these findings do not necessarily evaluate the educational effect of the food guide. The fact that the population sampled had adequate nutrient intakes may reflect favourably on the educational efficacy of the food guide, if in fact the food guide had served as a basis of instruction. 2.2.1.2 Inapplicability to the Present Food Supply Present food guides have also been criticized as inapplicable to present food supplies because of the trend, in western society, towards the use of highly refined fabricated foods, and of nutrient supplements (Bogert et a_]_. ,1973; Hertzler and Anderson 1974; Fremes and Sabry 1976). The development of an effective food guide requires that the food supply, and consequent pattern of food usage, have a reasonably consistent nutrient distribution which allows translation of the foods into a limited number of food groups. Many of the synthetic and fabricated foods may be concentrated sources of one or more nutrients which reflect those in a food group but do not supply sufficient amounts of other nutrients. Thus, although the food guide is presumed to give reasonably sound information on a diet selected from a limited number of natural food groups, many of the manu-25 factured items presently available which resemble natural foods and may in fact replace items in the food plan, do not compare favourable in nutrient content (Bogert e_t al_.1973; Hertzler and Anderson 1974; Fremes and Sabry 1976). 2.2.1.3 Relevance to Contemporary Nutrition Problems Food guides have been criticized for their lack of relevance to the present nutrition problems of the population. A food guide is a tool to improve eating habits and consequently it should indicate where improvement is needed. Food guides, although revised occasionally, were developed at a time when nutrition knowledge was not as extensive as today, and when emphasis on dietary needs was different (Hertzler and Anderson 1973). For example, Canada's Food Guide was originally established in 1942, at a time when the needs of national defense dictated a policy of food allocation and rationing practices. However, as suggested by Fremes and Sabry (1976), the nutritional problems of today are not ones of shortage, as they were during thai war years, but problems of abundance. Bogert et a]_. (1973) states that in developing a food guide the goal is to " ... add to the scientific basis of the Four Food Groups, making finer distinctions and discriminations among the group alternatives based on newer knowledge of nutrients and the foods in which they occur" (p. 442). Examples of this evolution exist (Hertzler and Anderson 1974). For example, embellishments to meet particular needs have been incorporated in food guides, such as indications of yitamin D requirements, and provisos for the use of iodized salts. Similarly, subclassification of the fruit and vege table group has been used to discriminate foods rich in vitamin A and C. Modern examples of food guides for use by vegetarians (MacMillam.'and Smith 26 1975; Smith 1975) and for coronary prevention (Conner 1967; Jansen et al. 1975) are available on a limited basis. However, these guides must necessarily incorporate additional food groups or qualifying remarks to direct food choices. 2.2.1.4 Ineffectiveness as a Teaching Tool To clarify the concept of healthful food selection and to facilitate learning, the need for maximum simplicity in food guide design has been stressed. Simplicity has been obtained by limiting the number of food groups and by using familiar names of foods (Hertzler and Anderson 1974). However, the trend in North America to attain maximum simplicity may not be as practical today as in earlier years, consequently, some authors have questioned the present food guide approach as the exclusive model for teaching food selection. (Hayes et_ aj_. 1955; Anon 1972; Lachance 1972; Poolton 1972; Manoff 1975). The media and advertising in the current era of food fads and fad diets have made the public aware of the basic nutrients, and a host of other food-associated claims (Ullrich 1971). Simple food-grouping systems may not be effective as teaching tools because the increased impact of media on nutritional awareness has established issues to which nutritional instruc tion by food guides does not adequately respond (Anon 1972; Lachance 1972; Manoff 1975). Another criticism of the food guide as a teaching tool has been the effect on interest in nutrition observed when it has been used extensively and repeatedly through several years of nutrition education (Poolton 1972; Bogert et_ al_. 1973; Leverton 1974). Food guides do not adequately hold Peoples' interest and consequently their ability to learn suffers. As Lachance (1972) stresses, " ... teaching nutrition by food groups is like 27 teaching mathematics by astronomy. It can be done, but it's not exactly the optimal method" (p. 44) 2.2.2 Other Relevant Systems The food guide is not the only method for presenting the principles of appropriate food selection. As indicated there are reasons that a more elaborate teaching tool should be considered to replace or supplement the-present food guide. The sacrifice of the food guide's simplicity may be substantially offset by benefits in both teaching effectiveness and accuracy of nutritional information. Such tools include: food labelling (Anon 1973), the Dietary Nutrient Guide (Pennington 1976), and a variety of computerized dietary assessment programs (Hanson 1969; Hansen 1973; Johnson et al_. 1974; Eddison 1975; Action B. C. 1976; Kugler 1976). These tools provide a .nutrient-emphasis rather than the more traditional food-emphasis. 2.2.2.1 Pennington's Dietary Nutrient Guide Pennington (1976) has developed a "Dietary Nutrient Guide" for evaluating diets which utilizes a limited number of "index" nutrients to monitor the total nutrient contribution of diets. Adequacy of the seven index nutrients ensures adequacy of other essential nutrients provided a few other guidelines are followed. The concurrence of index and other nutrients in foods has been determined by using extensive statistical analysis and computer procedures. This approach does circumvent the inadequacies of the traditional food guide by ensuring adequacy for 45 nutrients rather than four. Pennington has been able to avoid the usual pitfalls encountered when the dietary standard is translated into food guides or other patterns for desirable 28 eating -- that is, to deliberately or inadvertently ignore many essential nutrients in order to simplify the computation required in order to group foods. It is interesting to note that the index nutrients identified by Pennington, as indicative of dietary quality in conventional foods, are not those customarily recognized as key nutrients in food guides and which presently appear as key nutrients in nutrition labelling. Pennington's seven index nutrients are vitamin B-6, magnesium, pantothenic acid, vitamin A, folacin, iron, and calcium. 2.2.2.2 Nutrition Labelling Nutrition labelling, as defined in the 1973- U. S. Code of Federal Regulations for food label information panels (Anon 1973), outlines con ditions for voluntary and mandatory declaration of nutrient composition information on food products. While nutrition labelling is not solely and specifically designed as a nutrition education device, it can be "exploited" for nutrition education purposes (Moore and Wendt 1973). Thus, nutrition labelling provides a reference standard for directly comparing the nutritive value of foods, and illustrates the major nutrient contri butions of products whether or not they correspond to the food group concept. It also acts as a planning guide for balancing meals by providing clues to selecting food combinations which contain adequate amounts of key nutrients. Moore and Wendt 1973 suggest that nutrition labelling uses " ... a techni cally sound vocabulary for describing the nutritive quality of foods ..." (p. 123) -- that is, objective nutrient information on specific food products -- rather than a " ... contrived index of nutritive quality that sacrifices accuracy-for oversimplification" (p. 122). Although Moore and Wendt (1973) support the use of percentage of the 29: U. S. RDA per serving or portion as a more practical and meaningful method of providing nutrient content information than either nutrients per calorie or nutrients per unit weight, other authors (Wittwer et al_. 1977) have promoted the use of nutrients per calorie or nutrient density. Nutrient density is expressed as an "Index of Nutritional Quality" which is the ratio of a food's nutrient contribution as a percentage of the nutrient allowance, to its caloric contribution as a percentage of the energy requirement. An index value of 1.0 for a nutrient is the basic goal. 2.2.2.3 Computer Applications in Nutrition and Dietetics During the previous decade, electronic data processing and system design have been extensively applied in food service systems management, hospital dietetic departments, dietetic education, and nutritional research centers to perform a variety of functions (Hoover 1976). Computers have been variously utilized: in hospital dietetic departments and nutrition-research centers to perform nutrient analysis on a variety of dietary data (Brisbane 1964; Hjortland et_ a}_. 1966; Schaum 1973); in diet- and menu-planning procedures which allow simultaneous satisfaction of nutritive, production, economic, and palatibility constraints (Smith 1963; Gelpi et al_. 1972; Balintfy 1976); in patient-and hospital-information systems to support patient care activities such as menu ordering and production (Schaum and Sharp 1973); in obtaining dietary histories by automated-interactive interviewing (Evans and Gormican 1973); and in various other areas of nutrition research such as the study of eating patterns (Pao and Burke 1974). Computer-assisted :: systems have also been used extensively in food services to provide auto mated control of resources (Andrews et_ al_. 1967; Tuthill and Moore 1974), 30 and in food production (Fleetwood et al.1974; Sager 1974). In addition to the industrial, clinical, and research applications of computers in dietetics and nutrition mentioned above, systems utilizing some of the same functions are available to support nutrition education activities. These systems include Dietronics (Hanson 1969), Nutrimetrics 07 (Eddison 1975), the Nutrient Adequacy Reporting System (NARS) (Johnson et al. 1974), the Nutrition, Health and Activity Profile (Kugler 1976), the Nutrient Quality Index (Hansen 1973; Sorenson and Hansen 1975; Sorenson et al_. 1976; Wyse et al_. 1976; Wittwer et al.. 1977), and the Action B. C. nutrition evaluation program (Action B. C. 1976). In addition to their potential nutrition education function, some systems, in particular Dietronics and NARS, were designed as dietary-screening devices for clinical use. All of these programs are designed primarily to evaluate individual's diets, by providing information on the nutrient characteristics of the diet in relation to the estimated needs of the individual. Nutrient analysis and evaluation are performed on information from self-administered dietary questionnaires. Food composition data is recorded as one (Action B. C. 1976) or two-day recalls (Eddison 1975), as records of one (Johnson et al. 1974) to seven days (Hanson 1969; Action B. C. 1976), or as usual consumption pattern (Hanson 1969; Kugler 1976). Additionally, client demographic data are collected on the questionnaire and used for estimating the client's nutrient needs. The various systems employ computer-generated outputs which provide information on nutrient consumption with a comparison of nutrient intake to anindividualized dietary standard for between one (Eddison 1975) and 25 nutrients (Hanson 1969; Kugler 1976). This information is usually 31 displayed as actual nutrient intake and/or percentage of the nutrient allowance. Additionally some programs include information on nutrient ratios and percentages (Hanson 1969; Kugler 1976); the number of servings in the diet from each of the food groups (Hanson 1969; Johnson et_ a]J 1974) and other dietary measures such as sucrose consumption (Hanson 1969; Kugler 1976) and plaque frequency exposure (Hanson 1969). NARS also provides an overall measure of nutrient adequacy for the client, by averaging the percentage of the recommended allowance obtained over 12 nutrients. Although most systems provide tabular displays for the results of analysis and evaluation, the Nutrient Quality Index utilizes a graphical display. In some systems instructional material is appended to, or forms an integral part of the output, so that the results of the nutrient analysis and evaluation can be retranslated to an eating pattern. Nutrimetrics 07, which analyses the diet solely for caloric intake, prescribes caloric res triction where required to achieve ideal body weight by identifying and suggesting limitation of low-nutri nt density foods, such as alcoholic beverages and sweets. The Action B. C. program output provides summary statements on desirable items to include in the diet when a nutrient is limiting. Information on nutrient functions and a recommended activity program for weight loss is also available. The most elaborate output from the systems available is the Nutrition, Health and Activity Profile which proves a 9 to 13 page computer-generated response incorporating extensive discussion on diet and other aspects of lifestyle. Another form of prescription is utilized in the Dietronics output. Brand-name nutrient supplements are recommended to supply those nutrients falling below the nutrient standards. 32 These examples illustrate some of the present uses of computer systems in dietetics and food services, and particularly in nutrition education. As Hoover (1976) indicates, in her review of computers in dietetics, the next decade may bring more extensive use of computers in present areas, and the exploration of further computer applications in dietetics and nutrition. Present computer technology offers more capabilities than have been presently exploited in nutrition. While the computer is the tool used to perform the operations required in the above systems, the opera tions have nothing to do with the computer per se. However, the solution of problems, and performance of tasks of realistic size cannot be practi cally attempted without the aid of high-speed computers. The computer is a tool which can be effectively utilized in situations where rapid error-free processing is required. For example, the numerical manipulations, report preparation, and other routine decision-making functions found in many areas of dietetics and food services make effective use of computer methods. Within these limits the computer has demonstrated substantial advantages in cost and time-saving over conventional methods. 2.2.3 Critique and Conclusions To develop a teaching tool or guide which is generally useful is not an easy task'(UIlrich 1971). As indicated in Section 1.1, the complexity of present food markets, of dietary issues influencing health, of sources of nutrition and food related information, and of communicating to a public with a diversity of needs and levels of knowledge, must be consid ered in designing any tool providing information on food selection practice. This information provides a guideline for both the assessment of present practices and a goal for directing dietary modifications. The acceptability 33 of any system for food selection instruction depends on the worth of these two functions -- assessment and prescription. The food guide has attempted, as Bogert et_ al_. (1973) suggests, to "... combine scientific knowledge with Western cultural wisdom and, faulted though it is, there is not now an equally simple, equally workable alternative plan for Western food patterns" (p. 422). On the other hand, Pennington (1976) comments; A workable and usable food guide should be reliable and understandable but it need not be overly simplified. The Basic Four, as presently used is oversimplified; and is certainly not fool proof' (p. 7). Unlike the food guide the other systems discussed have potential capabilities of monitoring most of the known nutrients in the diet, and in this sense are theoretically less fallible guides for assessment of food selection practices. The use of nutrients as the basis of assessment provides an objective foundation on which to effectively coordinate nutritional information effectively with available food supplies and consumer habits. But these systems have been criticized for this detailed information allegedly exceeding the average consumer's comprehension (Moore and Wendt 1973). Unfortunately, after elaborate analysis and evaluation of the client's diet, these systems must then utilize less accurate methods to generate a food plan, such as the food guide or a listing of primary sources of nutrients. Neither of these concepts -- the food guide with its traditional food-emphasis nor nutrition labelling with its nutrient-emphasis -- are designed to stand alone (Moore and Wendt 1973; Winarski 1976). They both require the informational skills of a nutritionist/dietician to interpret nutrient information and to develop a dietary plan. Presumably the computer systems discussed also require some supplementary instruction 34 to ensure the development of a viable food plan. However, unless the professional has specific auxiliary tools available, the effectiveness of assessment and prescription will be limited by the capacities of the food guide and nutrition labelling. Although the nutrition educator should play an integral part in programs for instruction in food selection practices, the actual inter pretation and translation of food and nutrient information to human diets may be a mathematical and technologic problem (Balintfy 1973; Balintfy 1976). The logic of mathematical modelling, and the data handling capability of computers may be required to resolve the conflict between the requirements for accuracy and for simplicity in the design of systems for food selection instruction. As indicated, several computerized systems for dietary analysis and evaluation are available for nutrition education. These systems provide useful feedback on the adequacy of present practices through accurate and comprehensive assessment of individual diets. Although no computerized nutrition education tools which provide individualized diet or menu planning are presently available, illustrative models do exist in hospital food services and other industrial applications. These systems allow for simultaneous satisfaction of a large number of variables including nutri tive, economic, and palatibility. These may be adapted to nutrition education purposes to provide recommendations on needed alterations in individual diets, and married to present evaluative programs. 35 2.3 Dietary Assessment The objective of dietary assessment is to provide an estimate of a population's or individual's nutritional risk, by accurately estimating food and nutrient intake parameters of a population, group, or individual, and evaluating these parameters against appropriate standards (ICNND 1963; Mongeau 1974; Pike and Brown 1975). Dietary assessment has three phases: first, the collection of data on food intake, second, analysis of dietary data to determine the nutrient intake, and third, evaluation by comparing estimated nutrient intake with appropriate standards. Dietary assessment is only one of the procedures used in the assess ment of nutritional status. Present methods for assessing the nutritional status of individuals includes dietary studies, clinical studies and laboratory investigations (Jelliffe 1966; Christakis 1973). Each of these techniques is important for investigation of nutritional status, and for defining desirable dietary practices. However, only in comprehensive nutrition surveys would all techniques be simultaneously invoked. Other wise different techniques would be selectively employed depending on the objectives of the study or requirements of the situation. Nutritional status according to Christakis (1973) is "the health condition of an individual as influenced by his intake and utilization of nutrients, determined from the correlation of information obtained from physical, biochemical, clinical, and dietary studies" (p. 80). Consequently, nutritional status by definition cannot be judged from dietary intake data alone for two primary reasons. (i) Since the evaluative standard for nutrient intake -- the dietary standard -- is built on a normal curve which has inherent variability and consequently makes the individual's exact requirement 36 unknown, a certain diagnosis of nutrient inadequacy cannot be made from a knowledge of nutrient intake alone (United States 1974; Hegsted 1974). Although it is highly unlikely that intake is inadequate when the dietary allowance or standard is met or exceeded, there still exists a risk that deficiency will occur. Similarly, an intake below the allowance is not, in itself, evidence of nutritional inadequacy. This criticism applied to other methods of nutritional status assessment: ... data, like dietary data, have been traditionally interpreted by selecting some arbitrary cut off point, tabulating those who fall below that point, and using this value to indicate the number of :'>. "deficient" individuals in the population. Obviously, the significance <: of any measure of nutritional status - whether based on dietary, biochemical, clinical, or anthropometric data = depends on how it relates to the conditions one is attempting to diagnose or the specificity of the measure and the variability of the measure itself (Hegsted 1975, p. 18). Any measure of nutritional status can only indicate a degree of risk which may be attached to that measurement. Thus, to assess the nutritional status of an individual, records of nutrient intake should be considered in conjunction with the results of clinical, biochemical, and anthropometric measures. This procedure simply increases the probability of making a correct diagnosis. (ii) Different methods of assessment are not exactly equivalent. Nutritional assessment tools have sensitivity to different aspects of the deficiency continuum as predictors of risk, and|.,provide different types of information about nutritional status. Jelliffe (1966) classifies the assessment tools as: direct methods, including clinical, biochemical, anthropometric, and biophysical techniques; indirect methods, such as information on vital statistics; and ecological methods, wich consider food consumption, cultural influences, socio economic factors, and infectious diseases. Jelliffe's classification illustrates a basic difference in the type of information provided by the 37 different methods. Whereas, the direct and indirect methods of assessment "ideally" identify the consequence of nutrient influence, ecological measures such as diet assessment are used to establish causation. Diet is only one of many factors that may result in poor nutritional status and consequently diet adequacy does not ensure nutritional health. However, the different utility of dietary assessment methods confers some distinct advantages to its use. Dietary assessment monitors a differ ent aspect of the deficiency risk progression and therefore is an important component of nutritional assessment. The progression from sufficiency to deficiency is illustrated in a model of the development of deficiency diseases as follows. Five stages of interference in metabolism, give rise to progressively more overt clinical symptomatology: first, the preliminary stage with depletion of body stores for a nutrient subsequent to a poor diet, disease, or other conditioning factor; second, the biochemical stage where biochemical defects become apparent due to enzymatic depletion of coenzymes and the like; third, the physiologic stage which presents unspecific clinical findings such as general malaise, and irritability; fourth, the clinical stage where overt clinical signs are evident but tissue pathology exhibits nonspecific syndromes such as skin lesions and anemia; and fifth, the pathologic stage where specific syndromes exist and pathology has progressed to more vital functions (Krehl, W. A. 1964). Although there is actually considerable overlap in the specificity of the methods for monitoring the risk continuum, it would appear that " ... dietary information should be the most sensitive method of anticipating risk of deficiency" (Hegsted 1975, p. 18) from dietary origin. Also, dietary assessment may provide cues to nutritional problems that may not be evident from other methods. 38 In summary, dietary assessment procedures aid in the interpretation of other survey findings by, first, contributing to the diagnostic sensi tivity of nutritional status measurements, and second, by providing etiologic clues to morphologic, biochemical, or functional abnormalities. Assessment by itself is of limited value without knowledge of causation, since this is important for defining corrective measures. Additionally, dietary assessment acts as an indicator of early nutritional risk, and provides mechanisms to detect risk of nutritional problems where clinical and biochemical procedures are either not sensitive or are presently unavailable. Consequently, as a public health tool, dietary data have considerable merit in that this serves as the earliest preventive nutritional health monitor. Thus, while recognizing the limitations of defining nutritional status by dietary status alone, dietary assessment can be an effective public health tool to provide evaluation of food selection practices of individuals -- and in fact may be the only practical method to do so. The following three sections deal more specifically with the three phases of dietary assessment -- data collection, analysis, and evaluation -- and how they can be best incorporated into a system to provide instruction on food selection practice. 39 2.3.1 Data Collection 2.3.1.1 Data Collection Methods "The aim of all dietary surveys, whether made on individuals or on groups, is to discover what the persons under investigation are in the habit of eating. Their diets must be those to which they are accustomed and which they freely choose" (Marr 1971, p. 108, quoting Widdowson). Methods are available to collect data on the intake of populations (ICNND 1963), several hundreds of individuals (Paul et aJL 1963), families or households (ICNND 1963), and individuals (Chappell 1955; Taggart 1962). The present concern is with data collection on indivi duals. Food consumption data on individuals may be collected, according to the schema presented by Marr (1971), in the following ways: (i) Present intakes of food can be recorded by weighing methods. The weighing methods can be divided into two types, the precise weighing method and the weighed inventory method (Marr 1971). Both are applicable to the free-living individual, unlike metabloic balance studies which require controlled laboratory conditions (Wilson e_t al_. 1964). With the precise weighing method, all ingredients used in the preparation of dishes, the inedible wastage, cooked portions of food, and the plate waste are recorded. The weighed inventory method only requires that the prepared food be weighed immediately before consumption and the plate waste be weighed at the end of the meal. In either method, the actual measurements of foods eaten may be under the investigator's control (Pekkarinen and Roine 1964). Analysis of the diet may be chemically determined from aliquot samples or may be calculated using tables of food composition. 40 (ii) Present intakes of food can be recorded in household measures. This method, as described by Marr (1971), requires that the record of the food eaten, over a period of usually three to seven days, be described in terms of household measurements or be compared in size with food models. These descriptive terms are then converted to weights which are used for assessing the nutrient and caloric intake by using tables of food con-position. Although direct supervision is not necessary, where cultural patterns do not predispose to systematization, or where literacy is low supervision may be required. (iii) Present intakes can be recorded as a menu. The problem of collecting dietary data on large groups of individuals for epidemiological study, has prompted the development of methods which require only a minimal amount of data from the subjects (Wiehl and Reed 1960; Marr 1971). This method measures the foods consumed over the study period as a menu without quantities. The assumption, when using unquantified intake instead of a measured or weighed intake, is that the size of helping does not vary to any great extent between individuals, or at least not to such an extent that course measures of dietary quality cannot be determined. In some cases factors corresponding to the amount of food in an average portion, or derived from multiple regression analysis, are applied to the frequencies to obtain quantitative approximations of foods or nutrients consumed (Mongeau 1974). (iv) Past intakes of foods actually consumed can be recalled. This method measures actual past intakes as remembered at an interview (Adelson 1960; Guggenheim et_ aj_. 1960; Huenemann et_ al_. 1961; Bransby et al_. 1964) or on self-completion questionnaires (Keen and Rose 1958). The subject is usually asked to recall all food consumed during the 41 previous day and to estimate quantities in ordinary measures or servings. To facilitate quantitication estimation aids such as food models and measuring utensils may be utilized in interviews, and have also been used in questionnaire methods (Johnson ejt al_. 1974). Customarily, foods eaten in the previous 24 hours are recalled -- and for this reason the technique is usually termed the 24-hour recall --, although recalls covering up to the previous three days (Guggenheim ejt al_. 1960) and for periods of one week (Huenemann e_t al_. 1961) have been reported. (v) Past intakes of foods usually consumed can be recalled. The diet history method obtains data on general dietary pattern, as opposed to current diet or past foods actually eaten (Pekkarinen 1970). The method, developed by Burke (1947), makes use of several different approaches to obtain information from an individual about average food intake over a period of time. The three phases of the diet history are, first, a deter mination of the past pattern of eating by using the 24-hour recall coupled with inquiry about the usual eating pattern. Second, using a detailed predetermined list of foods, the pattern of eating reported in the first section is cross-checked to determine inconsistencies and inaccuracies. Finally, a three day food record of present intake in the form of a non-quantified menu is kept by the subject. The results of the history are recorded in household measures, and are converted by use of food composi tion tables to their nutrient values. The reliability of the data on usual consumption, obtained by the first phase of questioning, is verified by the later two phases, for which reason the method is called the cross check dietary history (Hartog e_t al_. 1965). Burke (1947) indicates the history method is suitable only when a well defined dietary pattern exists, vi) Past intakes of foods can be recalled as an unquantified menu. 42 In addition to short-cut methods of data collection by recording, abbreviated methods for collection of recalled dietary data have been developed. These are radically shortened versions of the Burke method which measure usual frequency of consumption data. Both interview (Stefanik and Trulson 1962) and questionnaire formats (Hankin and Huenemann 1967; Hankin et al_.1967; Balogh et al_. 1968):have been tested. 2.3.1.2 Evaluation of Data Collection Methods Although a variety of techniques for determining dietary intake for free-living individuals are available, none has received general acceptance. Indeed as Marr (1971) points out, controversy has raged over what constit utes the best method. She indicates that perhaps the "extremes of con sidered opinion" (p. 109) are respectively illustrated in the views of Widdowson and of Burke. The former emphasizes "the necessity for [accurate] measurement of current intake and the other the need for assessment of [representative] food habits over a considerable period" (Marr 1971, p. 109). Thus, there are conflicting views on the value of accurate measure ment which arise primarily because procedures to maintain validity of measurement may interfere with the representativeness of the sample popu lation, or with the eating patterns of that population (Thomson 1958; Mann et_ al_. 1962; Marr 1971). Mann et al_. (1962) complains in response to this confusion, that " ... a superficial examination of the technical problems experienced in measuring dietary intake meets such a morass of conflicting opinions that the first inclination is apt to be a decision for abandonment" (p. 212). Pekkarinen (1970) and Marr (1971) indicate, in part, the problem of appraising dietary intake methodologies is that no independent device, 43 or absolutely accurate method of dietary data collection, exists against which the validity of the methods used in dietary assessment can be measured. The validity of each method can only be tested comparatively (Marr 1971). Although no single ideal method exists, each method has its advantages and disadvantages. Consequently, the choice of method, its potential and validity, must be assessed against the objectives and purposes tbf the study; and against the circumstances in which the technique is to be used, for example, funds and personnel available for the study, and the study population's characteristics. In terms of validity of measurement alone the precise weighing technique, with investigator supervision and analysis of aliquot samples, theoretically is the method providing the most valid estimate of actual food and nutrient intake (Marr e_t al_. 1959; Whiting and Leverton 1960; Marr 1971). The weighed inventory method with investigator supervision and analysis of aliquot samples would be expected to be a close second. However, the expense required for performing weighing methods, (Pekkarinen and Roine 1964) and the intrusiive nature of the collection procedure (Marr 1971), limit the utility of such studies to small, often non-randomized populations of cooperative individuals (Pekkarinen and Roine 1964; Hartog e_t al_. 1965), and to relatively short periods of reference -- one or two weeks is generally considered to be the maximum duration (Pekkarinen 1970). Further, the intrusive character of the weighing techniques may interfere with normal behavior and thereby limit the validity of the measurements (Pekkarimen 1970; Marr 1971). "To what extent this [intrusion] alters the food intake is difficult, if not impossible, to determine" (Marr 1971, p. 110). Inconsidering other methods, it must then be determined to what 44 extent measurement validity is influenced by the various departures which are necessary in practice to reduce costs and improve subject cooperation, such as: the use of food composition tables instead of chemical analysis for nutrient analysis (Marr 1971); the use of subject versus investigator measurement for data collection (Paul et al_. 1963; Marr 1971); the use of descriptive measures or no measurement instead of direct weighing (Bransby et_ al_. 1948; Young et_ al_. 1952a; Thomson 1958; Marr 1971); and the reliance on memory in recall methods rather than on direct observation (Huenemann et_ al_. 1961; Marr 1971; Mongeau 1974). The loss in accuracy of measurement can in turn be judged against the increased usefulness of data derived from representative samples of a population living their normal lives, for which the weighing methods are not a practical technique (Marr 1971). The less intrusive methods encour age greater subject cooperation resulting in opportunities for longer term studies of larger sample size with randomized samples; and for reduced interference in eating pattern (Pekkarinen 1970; Marr 1971). Data collection at relatively lower cost also permits increased sample size and longer time frame (Marr 1971). 2.3.1.2.1 Validity and Repeatability of Methods Although absolute validity for the weighing methods cannot be estab lished; relative validity has been demonstrated by comparing weighted surveys with the results of other dietary data collection methods (Marr 1971). The weighed inventory method has also been indirectly validated, Durnin (1961) found that caloric expenditure, as measured by indirect calorimetry, and energy intake determined by the weighed inventory method were balanced over seven days. Tests of repeatability show reasonably consistent results with both the precise weighing method (Marr et_ a]_. 1959) and the weighed inventory method (Heunemann and Turner 1942; Adelson 1960). Compared to data recorded by weighed methods, data recorded in household measures and descriptive estimates are considered less precise largely because of errors in estimation of portion size (Young et al."1952a; Thomson 1958; Marr 1971). Interestingly, in a study conducted by Bransby et al_. (1948), comparisons of weighed and measured intakes for both individuals and groups did not show marked discrepancies, providing the measuring equipment was standardized. Dietary records collected as menu items without an indication of quantity are considered less precise than either weighed methods or those using household measures. However, investigators (Marr 1971) testing the utility of frequency-of-consumption information indicate it was accurate enough to classify populations into broad categories with respect to nutrient intake. According to many investigators, grouped mean intakes from recalls of actual intakes and those from weighed records (Morrison ejt a]_. 1949; Adelson 1960; Combs and Wolfe 1960), estimated records (Payton et al_. 1960), or diet histories (Stevens et al_. 1963) give comparable results and thus can replace each other in group surveys. This is so particularly if samples are large and if daily and seasonal variations in intake are small (Pekkarinen 1970). However, agreement of grouped data between recalls and other methods is not found in every case (Bransby et al_. 1948; Young et al. 1952c; Trulson 1954; Thomson 1958; Pekkarinen et al_. 1967). When individual consumption is sampled, variation in intake data are observed between dietary recalls and other methods by some investigators (Morrison et al. 1949; Young et al_. 1952c; Trulson 1954). This finding is not universal. 46 Some authors find that when individual consumption is sampled over the same time frame (Flores e_t al_. 1965), or even different time frames (Adelson 1960), results of recalls and weighed records are in good agreement. Considering the dietary history method, Balogh et_ al_. (1968) have found comparable results from the history method and a weighed record. However, in most cases (Huenemann and Turner 1942; Young et aJL 1952b; Trulson 1954; Paul et al_. 1963; Hartog et al_. 1965; Hart and Cox 1967), attempts to validate dietary histories against weighed or measured food records have demonstrated discrepancies between the methods for group and individual values, particularly when surveying children (Trulson 1954; Beal 1967). Repeatability of the dietary history method had been demonstrated by repeat histories (McCann et_ al_. 1962; Paul e_t al_. 1963). Burke et al. (1943) provides some evidence that indirectly validates the dietary history method. In this study the results of a diet history was validated against an independently measured variable -- the incidence of preeclampsia. Results of validation studies indicate favourable comparisons between the shortened recall methods and research history interviews, weighed, or measured weekly records for group data (Marr 1971). The applicability of this method to individual dietary analysis is doubtful.(Marr 1971). 2.3.1.2.2 Time Frame of Methods Information on long-term food consumption is the objective of most dietary surveys. The efficacy of extrapolating short-term data to habitual intake is questionable (Cantoni ejt a]_. 1961; Pekkarinen 1970; Marr 1971; Mongeau 1974). Studies using weighed records to explore weekly variations in indivi dual's diets have provided conflicting long term intake. Some authors (Thomson 1958; Morris et_ al_. 1963) believe that seven-day weighed records 47 give a sufficiently accurate estimate of usual intake, since in their study populations considerable individual stability of food intake was found over widely separated weeks. However, others (Yudkin 1951; Keys et al. 1966) indicated that a one week survey was not predictive of long term intake. Adelson (1960) reports that for some individuals weekly diet patterns are relatively stable, while this is not the case for other subjects. Authors using weighing methods (Yudkin 1951; Chalmers et al. 1952; Chappell 1955; Fidanza et al_. 1964; Hartog et al_. 1965; Hankin and Huenemann 1967) have indicated that daily, weekly, seasonal, and yearly variations in dietary intake are evident. The extent of the temporal variations is influenced by factors such as age, gender, and occupation (Chalmers et_ al_. 1952; Adelson 1960). Also the different nutrients show different degrees of variability (Chalmers e_t al_. 1952). A general statement about the appropriate length of time to be surveyed using weighing methods, or in fact any method of measurement, cannot be given (Pekkarinen 1970; Marr 1971). The length of time required to establish a representative picture of intake is obviously dependent on the individual's or population's food intake patterns. If extreme varia tions exist in food intake pattern, no method short of extensive sampling is likely to accurately categorize long term dietary intake. As the variation in the diet decreases the time frame required for-accurate estimates of usual consumption can be reduced accordingly. Providing the duration of sampling corresponds to a consistent dietary cycle or rhythm, the sampling frame should be indicative of usual intake. To define the time frame required to record dietary intakes of usual consumption for an individual or group the precision required must be stated, and intake variation per time must be determined for each nutrient and for each sample of the 48 population to be appraised (Marr 1971). Thus, the efficacy of extrapolating from the sample period depends on the consistency of the dietary pattern (Pekkarinen 1970). Even the diet history method, which is intended to measure long-term food pattern, is suitable only when a well defined dietary pattern exists (Burke 1947). Sampling period can be more effectively lengthened by sampling in a number of separate one week periods over the year (Chappell 1955) or as Balogh et al_. (1971) have shown by repeated 24-hour recalls. 2.3.1.2.3 Randomness and Size of Population Samples A major disadvantage of the weighing methods is that they demand a high degree of cooperation from the individual, consequently response rates in population studies may be low. For this reason, weighing methods may not be suitable in studies carried out on random samples (Pekkarinen and Roine 1964), as has been found in certain countries (Hartog et al. 1965). However, in some other countries (Buzina et_ a]_. 1964; Fidanza et_ al. 1964) the use of weighing methods has not prevented randomization of the survey population. Since records using household measures are less demanding than weighirigmethods a higher degree of cooperation may be expected. However, Marr (1971) indicates that higher cooperation rates are not necessarily achieved by descriptive as compared to weighed studies. For all its faults, the 24-hour recall is the only suitable method for use in large scale surveys of heterogeneous populations. The method should allow randomized surveying of large representative sample popula tions, since cooperation rates would be expected to be high (Pekkarinen 1970; Marr 1971). However, Marr (1971) indicates that little information on cooperation rates in recalled surveys of actual food consumption is 49 available in the literature. The diet history method is also reasonably acceptable for random surveys of large population samples, particularly in its abbreviated forms (Marr 1971). 2.3.1.2.4 Other Methodological Considerations In addition to the generic features of the dietary-data collection methods, which have been presented above and which contribute to discussions of the relative validity of these methods, several other methodological considerations which effect the accuracy of dietary intake data have been identified. These considerations include: the professional skills of the interviewer or technician doing the data collection (Adelson 1960); the use of auxiliary procedures, such as visual aids and models to assist in quantification (Balogh et al_. 1968); the characteristics of the diet under study -- its complexity or monotony (Chalmers e_t al_. 1952); the circumstances of the study, such as season undertaken (Mongeau 1974); the study populations characteristics including their skills (Young et al. 1952c; Stevens e_t al_. 1963), age (Marr 1971), attention span, memory capabilities, education, literacy, and intelligence (Pekkarinen 1970); and perhaps a variety of procedural nuances used to reduce errors of portion size estimation and of omission, for example, explicit inquiry about postprandial consumption patterns when doing recalls (Balogh e_t al_. 1968), and a detailed interview following completion of client records to enable the quantities and size of helpings to be checked (Kitchin et_ al_. 1949). The accuracy of data collection methods may be greatly influenced by situational factors, as those above. For example, in quantitative studies using estimation methods, the ability of the subject to accurately estimate quantities is important for achieving accurate results. Stevens et al. 50 (1963) and Young et_ al_. (1952c) found good agreement between methods when using informed subjects, such as Home Economics graduates, whereas the results from other groups of subjects provided inconsistent answers. Presumably had all the subjects in the experiment been able to estimate quantities accurately, the conclusions reached by these authors on the comparative accuracy of the methods would have been different. Present comparative studies of data collection procedures have not explicitly addressed the impact of these factors on the accuracy of methods or their possible importance for defining the exact method appropriate for every field situation. In this regard, new methods are being proposed and developed (Mongeau 1974). Explorations into short schedule epidemiological methods (Hankin and Huenemann 1967) are being made. 2.3.1.3 Summary and Conclusions Evaluation has identified some of the limitations and potentials of each method. The basic problem is that it is difficult, and perhaps impossible, to measure accurately the dietary intake of a large random sample of free-living individuals for long periods of time (Marr 1971), particularly if extreme''daily variation in intake exists (Christakis 1973). No method presently available ensures, simultaneously, validity of measure ment and relatively unbiased sampling and experimental methodology. That is, attempts to ensure absolutely accurate measurement in most cases do not permit data collection from large samples of people, from random z samples of those populations, for long periods of time, and from free-living individuals. As indicated by Marr (1971) and Pekkarinen (1970) the choice of any method -- its acceptability -- is dependent on the characteristics of the population considered, the purposes of therstudy, and the circumstances of 51 the study. Each situation and study purpose has its own best data collection method. Similarly, each of the methods has its own specific utility. Any consideration of the acceptability of a method must take into account the purpose and situation for which the method was designed. In the present evolving field of collection methodology for dietary data, definition of an appropriate method for any given situation may be dependent on more than simply selection from the available genera of study procedures. Available methods provide guidelines from which to define some aspects of procedure; beyond this, in order to ensure an accurate record of the types of foods eaten and the quantities consumed, the stan dard protocol must be modified and elaborated as dictated by circumstance. Thus, both the application as well as the choice of method are significant in/determining the accuracy and acceptability of any method. 52 2.3.2 Data Analysis Beyond choosing how to collect dietary data representative of an individual's intake, and choosing how a particular intake is to be judged, analysis of the dietary intake data is required to provide indices for evaluation. These indices may take the form of nutrient values (Marr 1971), information on foods and food groups (McClinton ejt al_. 1971), or alter natively, dietary scoring patterns (Trenholme and Milne 1963). For the purposes of this thesis, analysis which provides nutrient composition information reflective of the actual nutrient intake is required. Two methods for determining the nutrient content of diets are available. These are calculation of nutrient values from food tables and nutrient analysis by chemical methods. Whiting and Leverton (1960) indicate that, laboratory analysis of individual foods, or of composites, provides the most reliable estimate of the nutrient values of foods actually eaten. However, the laboratory procedures and data collection practices needed to secure food samples for analysis, require careful and time consuming handling. Consequently, available resources generally limit use of these methods to special studies. More commonly, nutrient values are calculated from food tables -- a practice for which the merits have been debated (Harris 1962; Marr 1971). Reflective of the dichotomy of opinions, Widdowson and McCance (1943), in their discussion of the scope and limitations of food tables, wrote: There are two schools of thought about food tables. One tends to regard the figures in them as having the accuracy of atomic weight determinations; the other dismisses them as valueless on the ground that a foodstuff may be so modified by the soil, the season or its rate of growth that no figure can be a reliable guide to its com position. The truth, of course, lies somewhere between these points of view (p. 230). 53 •Comparisons of nutrient determinations using food tables with those using chemical analysis, on either group mean or individual data, show somewhat contradictory results. Some early studies (Bransby et a]_. 1948; Groover ejt al_. 1967) conclude that substantial differences exist between the two methods. They suggest that calculated values do not represent actual nutrient consumption. Interestingly, a reexamination of the study of Bransby et al_. (1948) by Marr (1971) suggests that although absolute agreement between analyzed and calculated values for every individual was not achieved, the conclusions reached by Bransby et_ al_. were overly harsh. Marr found that the calculated and analyzed values were essentially the same for individual and group means. Other authors (Buzina et a]_. 1966; Pekkarinen 1967) indicate relatively good agreement between calculated and analyzed results. Whiting and Leverton (1960) find that whereas the group mean values for some nutrients agree using the two methods, others do not. Differences between the results of the methods depend primarily on how closely the food composition table value corresponds to the nutrient values of the food consumed. Obviously, if the values assigned by food composition tables are the same as the laboratory results, the calculated and analyzed results for a given portion of food will agree. This is illustrated in studies where food tables composed largely of analytic data of local foods are being used. In such cases the results of calculation and analysis have been found to agree (Pekkarinen 1967). Furthermore, closer correspondence of analyzed and calculated values is achieved with nutrient values that are more stable, such as those for energy and protein (Whiting and Leverton 1960). Where the food compositional values may vary widely, for example with vat values, the differences between calculated and analyzed values may be larger (Whiting and Leverton 1960). It is important, .54 therefore, that the food tables used should be appropriately matched to the foods which are eaten and analyzed, if exact correspondence between laboratory and computational methods is desired. In comparing food values obtained by analytic methods with those by calculation from food tables, it should not be expected they would agree exactly. The two methods measure somewhat different information about foods consumed. The analytic value determines the nutrient intake of foods eaten on those days of the study, the calculated value determines the nutrient intake based on food composition which represents the average of food samples taken over an extended time period; perhaps a year or more. Although small special purpose food tables and simple tables for local studies exist, in a complex marketing economy such as North America, it is not useful to disaggregate compositional values for all the local factors that effect nutrient values. Thus, food tables in common use in North America (Watt and Merrill 1963; Church and Church 1975; Adams 1976) contain food compositional information for the country or area which is averaged for factors such as: genetic differences -- variety or breed; environmental effects, for example soil fertility, fertilizers, diet for animals, light, temperature, precipitation, and other climatic elements influencing con ditions of growth; seasonal and geographic differences; methods of harvesting, handling, and storage; and manufacturing and processing procedures (Asenjo 1962; Watt and Merrill 1963). Averaging across factors which create variability in nutritive values is done in order to develop a final figure as representative of the product available the year around for the area considered (Watt and Merrill 1963). Additionally, the statistical procedures used may incorporate averages weighted according to pertinent factors, such as availability or use; other 55 adjustments in the figures, such as excluding widely discrepant results from the final calculations, may be used so that representative figures are developed (Whiting and Leverton 1960; Watt and Merrill 1963). Food values in the 1945 U. S. Department of Agriculture publication, "Tables of Food Composition in Terms of Eleven Nutrients (United States 1945), the 1950 and 1963 editions of Agriculture Handbook No. 8 (Watt and Merrill 1950; Watt and Merrill 1963) are treated in this way. Although averaging the compositional values in this way adds potential variability to nutrient values calculated from food tables when compared to analyzed figures, a certain advantage is attached to the use of average values. When information on usual consumption is desired, then, providing that the average values are reflective of the actual food consumed, the use of food composition tables with average values provides values which are unbiased by season and other variations in food supply throughout the year. In addition to the questions about the suitability of calculations of nutrient values from food tables raised due to the variation between analyzed and calculated values, there are other limitations when using nutrient analysis from food tables. Some of these problems are common to chemical methods. (i) Standard food composition tables (Watt and Merrill 1963) have not provided information on .all the nutrient values in dietary standards, and the standard itself does not provide requirements for all known essential nutrients (United States 1974), Hertzler and Hoover (1977) suggest that: 56. Although more analyses are needed for every nutrient, new tabulations are especially needed for: vitamin E, molybdenum, phosphorus, iodine, chromium, manganese, selenium, amino acids, and individual carbo hydrates. Analyses of dietary fiber are needed because crude fiber represents only a fraction of total fiber in foods (p. 22). Recent progress in the development of larger composition tables may largely rectify this difficulty (Schaum ejt al_. 1973; Hertzler and Hoover 1977). (ii) Food tables do not include values on compounds known to inter fere with nutrient availability, nor do the chemical analysis' on which the tables are based estimate the true availability of all nutrients. Hence, calculations from food tables may overestimate the amounts actually absorbed (Harris 1962; United States 1974). However, in the 1963 Agriculture Handbook No. 8 (Watt and Merrill 1963) foods with high oxalic acid concen trations are footnoted with respect to availability of calcium. Also, although the total amounts of iron, sodium, potassium, and magnesium in foods are listed, the amounts available to the body are recognized, as variable by this publication. (iii) Although until 1940 food tables produced by the U. S. Department of Agriculture listed maximum, minimum, and average values for nutrients when more than one analysis was available (Herzler and Hoover 1977), present tables of nutrient composition to not indicate the variability of nutrient values used to determine the average values (Whiting and Leverton 1960; United States 1974). Thus, while ranges of values would provide some guidance for reliability of calculations, present calculations can only be done from the mean and not the range. (iv) Nutrient analysis is needed for many items, such as: ethnic foods; restored or supplemented items; products with new formulations; and dehydrated, frozen, and ready-prepared products (Hertzler and Hoover 1977). Regradless of the extent of food table development, difficulties will 57 probably still exist in: mbnitoring constantly changing manufacturing practices; describing all the variations of preparation practice such as cooking time;,".and providing food names which cover all local idioms. Regardless of the problems encountered with the use of food tables, resource limitations and circumstances necessitate that nutrient analysis be performed using food composition table values. This applies especially when the intakes of large numbers of people are monitored for many nutrients. Numerous tables of food composition have been developed in North America since the 1890's. This history has been recently reviewed by Hertzler and Hoover (1977). In one of the first comprehensive tables of nutrient compos ition for American foods, Atwater (1895) included values of calories, protein, fat, and carbohydrate for approximately 235 items. Since that time advances in nutrition knowledge, refinements in laboratory procedures, expansion in the number of foods available, and a trend towards increased specificity for most nutrient classifications have resulted in food com position tables of considerably larger size. A projected development of a data bank by the U. S. Department of Agriculture will consider up to 215 food constituents for an undefined but potentially enormous number of foods (Hertzler and Hoover 1977). Tables of food composition used in North America have as their basis tables developed primarily by the U. S. Department of Agriculture. The most comprehensive of these tables, among the bibliography of the USDA food composition publications provided by Chandler and Perloff (1975), is contained in Agriculture Handbook No. 8. This volume was first published in 1950 (Watt and Merrill 1950) and revised in 1963 (Watt and Merrill 1963). Table 1 of the revised publication provides calories and 16 nutrients for 2,483 foods in 100 gram portions. Additional nutrient values for some 58 foods are also included in supplementary tables in the same volume. Publications, such as Church and Church's (1975) "Food Values of Portions Commonly Used", contain food composition tables with items listed according to household portions. Other publications, including the USDA publications Home and Garden Bulletin No. 72, first published in 1960 (United States 1960) and most recently revised in 1971 (United States 1971); and Agriculture Handbook No. 456 (Adams 1975), provide nutrient data for foods in terms of common household measures and market units. Since the late 1950's computer stored data bases have been compiled by the USDA (Hertzler and Hoover 1977). Presently, USDA food composition data sets areoavailable to the public in a variety of machine readable issues. These include the data contained in the 1963 edition and 1972 revised edition of Handbook No. 8 (United States 1977a-b); the 1971 edition of Home and Garden Bulletin No. 72 (United States 1977c); and Handbook No. 456 (United States 1977d). Additionally the USDA began compiling a Nutrient Data Bank in 1972 to serve as an international repository of nutrient data (Watt et al_. 1974; Chandler and Perloff 1975; Hertzler and Hoover 1977)). This facility will integrate data submitted from a variety of sources including industrial, governmental, public, and private laboratories, as- well as that already stored in the USDA's Nutrient Data Research Center. In order to increase handling capacity, a computer based system has been instituted to facilitate storage, processing, and retrieval of data. The data stored in the bank will be processed in three ways. Data from individual analyses will comprise Data Base I; those from the averaged values of identical items create Data Base II; and the averaged values of similar items will form the present revised equivalent of Agriculture Handbook No. 8, Data Base III (Chandler and Perloff 1975). Provision has been made for storage of 59 up to 215 nutrients, related constituents, and analytic values (Hertzler and Hoover 1977). Data Base II will be made available on tape for computer use. Presently, two issues of Data Base III -- those for dairy and egg products, and for spices and herbs -- are available in loose leaf format or on tape (United States 1977e). Schaum et al_. (1973) have reported on another extensive nutrient data bank. The storage space presently allotted permits expansion to a maximum of 10,000 food items with 63 nutrients each. Initially, the file contained 3,600 items, and subsequent work has increased the food item file to approximately 4,800 items (Hertzler and Hoover 1977). In addition to the extensive compilations mentioned above, compressed tables of food composition have been developed and used by a number of authors (Leichsenring and Wilson 1951; Dawber et al_. 1962; Mann et al. 1962; Browe et al_. 1966) for shorthand determinations of nutrient intake. These investigators compress food item listings by foods of similar nutritive value into food groups, and establishing mean nutrient values for each group by arithmetic averages or by a weighted average adjusted to food consumption or preference practices of the population. However, these methods are better for group determinations than for individual determina tions of nutrient intake, since neither the population-averaged nor arithmetically-averaged nutrient value for a food group will accurately reflect the individual's actual intake of foods in that food group (Whiting and Leverton 1960). "The extent to which one may group food in a table depends entirely on the methodology, the nutrients one is interested in, and the amount of detail one is willing to forego for the sake of simplicity and time" (Browe et al_. 1966, p. 107). In nutrition-education programs providing dietary assessments a shortened list of food items is desirable. However, "60 compressed tables have the disadvantage that they may necessitate severe reductions in the number of nutrients assessed and the number of foods considered. Inaccurate assessment may result both from ignoring some nutrient values in order to effectively group foods, and from substituting nonequivalent food items in the condensed food list. Pennington (1976) has produced a "Miniature Food List" and tables of nutrient composition which circumvent some of the problems usual in com pressed food composition tables, by appraising the coexistence of 45 nutrients in a large number of foods. The miniature food list is composed of 202 index foods, from among those most commonly consumed in the United States, with values of 45 nutrients for each food. The index-food items can be substituted with a variety of other designated items which have been found to be similar in nutrient characteristics. The assumption used in defining substitutable items is " ... that the nutrient variation of any one nutrient in any one food (index item) is greater than the variation of all the means of that same nutrient in all the foods included in the group" (Pennington 1976, p. 9). The validity of this tool is based on the demonstrated fact that errors in dietary nutrient value due to the use of food:;substitution and to established serving sizes are much less than errors due to simple variation of nutrients in foods (Pennington 1976). The list does not cover all foods that are potentially available in the marketplace, but does cover a considerable number relative to the index items presented. Pennington's miniature food list may provide an acceptable compressed table of food composition for a nutrition education system performing assessment of individual's diets. It provides a comprehensive table of food items and their nutrient values, without compromising the accuracy of 61 the compositional values. The table provides a balance which maximizes validity of assessment but minimizes computational and data-collection difficulties. Beyond defining the indices for dietary analysis, in this case nutrient composition, the computational method for processing dietary data should be considered. Traditionally, nutrient analysis was performed by hand cal culation, however, the advent of electronic data processing offers methods for rapid and accurate nutrient analysis (Marr 1971). Computers have been used to perform nutrient analysis on a variety of dietary data, for example, on patient food consumption records (Thomson and Tucker 1962; Brisbane 1964; Eagles et al_. 1966; Beal 1967; Schaum 1973; Tuthill 1974), metabolic diets (Hjortland et_ al_. 1966), epidemiological data (Goodloe et_ al_. 1963; Hayes et aJL 1964), and nutrition survey data (Tie et_ al_. 1967). Calculation of menu nutrients by computer demonstrates advantages in cost (Flook and Alford 1974), accuracy, and response time over manual computation (Hoover 1976). Computer systems offer the advantage of comprehensive analysis of nutrients and the option of complex data manipulation, for example, statis tical interpretation of relationships among nutrients. Mathematical accuracy, of course, will not remove inherent errors in food composition tables. However, the computer will carry out routine calculations with speed and complete accuracy. It therefore seems advantageous to use com puters when both ease and accuracy of nutrient assessment procedures would be improved. 62 2.3.3 Data Evaluation The final phase of dietary assessment -- data evaluation -- requires that estimated nutrient intake be judged against appropriate standards. The dietary standard provides the present formal basis on which nutritional science and the nutrition educator define acceptable nutrient intake. Other recommendations which may supplement the values contained in the dietary standard have been proposed by recognized nutrition agencies. The following section discusses the characteristics of the standards available for appraising dietary well-being, and their suitability for evaluating individual dietary practices. 2.3.3.1 Dietary Standards Available information on human nutrient requirements has been incor porated into dietary standards, which are, in a sense, compendia on the known nutrient requirements of man. The most recent revision of the Canadian dietary standard (Canada 1975) is defined as "a statement of the daily amounts of energy and essential nutrients considered adequate, on the basis of scientific data, to meet the physiologic needs of practically all healthy persons in a population" (p. 5). The Recommended Dietary Allowances of the United States Food and Nutrition Board (United States 1974) embodies the same principles. Although the dietary standard is based on actual or extrapolated human requirements, the figures given are "recommended daily intakes" or "acceptable daily intakes", or as the Food and Agriculture Organization of the World Health Organization states "safe level of intakes", and not a statement of the actual nutrient requirements for each individual in the population (Passmore et al_. 1974; United States 1,974; Canada 1975). This 63 distinction is very important in understanding both the design and use of the dietary standard. Individual differences in nutrient requirements, as a result of normal biologic variability, have been considered in establishing a standard which should meet the nutrient requirements of most healthy people in the population (Beaton 1972; United States 1974; Canada 1975). Con sequently, an individual whose average daily nutrient intake meets the standard for all nutrients, would have little likelihood of nutritional inadequacy. In fact, the requirement of most individuals should be less than the standard. The formulation of standards for energy intake is different. Unlike the standard for other nutrients, the recommended caloric intake approximates the predicted average requirement of the popu lation members, instead of lying substantially above the average require ment. For this reason, there is relatively little relationship to indivi dual requirements which may be above or below the recommended intake. Some qualifications for the use of dietary standards are recognized. First, the dietary standards are not formulated to cover additional require ments of persons depleted by infection or injury, other traumatic stresses, prior dietary inadequacies, genetic and metabolic disorders, and the use of pharmaceutical preparations (eg. oral contraceptives); nor do they consider losses of nutrients that occur during processing and preparation of foods (United States 1974; Canada 1975). Second, although the standards are expressed as a daily average, the ability of the body to adapt to daily fluctuations in nutrient intake implies that recommended intakes do not have to be met on a daily basis (United States 1974; Canada 1975). Therefore, it is considered acceptable when estimating dietary adequacy to average weekly intake of nutrients. 64 Third, when using dietary standards, even accurate knowledge of actual nutrient intakes is not synonymous with evaluation of the nutritional status of eigher the individual or the population surveyed. 2.3.3.1.1 Formulation of Dietary Standards The ideal method, rarely if ever achieved, to develop an allowance would be to (1) determine the average requirement of a healthy and representative segment of each age group for the nutrient under consideration; (2) assess statistically the variability among the individuals within the group; and (3) calculate from this the amount by which the average requirement must be increased to meet the needs of nearly all healthy individuals (United States 1975, p. 5). Insofar as possible, the above methodology is followed for developing a nutrient allowance. Population groupings are selected for many of the major identifiable factors influencing requirements -- that is, age, sex, body size, physiological state, and physical activity (Harper 1974; United States 1974). Utilizing available estimates of average requirements and variability within the populations studies, allowances for nutrients other than energy (which is an estimate of average needs of the group), are developed by increasing the average nutrient requirement two standard deviations (Lorstad 1971; Beaton 1972; Harper 1974; United States 1974). Providing or assuming that individual requirements fit a statistically normal distribution, an allowance set two standard deviations above average should be sufficient to meet or exceed the needs of 97.5% of the individuals in the population (Lorstad 1971; Beaton 1972). Unfortunately, such allowances cannot be established for most nutrients, with the exception of thiamin, riboflavin, niacin, iron, and protein, because there is inadequate information about the variability of individual requirements (Beaton 1972). 65 When information on the requirements of large groups of individuals is not available, average requirement and estimated variability are derived from the following data: one or two controlled feeding trials on a limited number of subjects; critical metabolic studies on animals with values extrapolated to man (the translation is fraught with under-tainty); or dietary surveys on the minimum amount of a nutrient known to be consumed by apparently healthy people (Harper 1974; United States 1974; Canada 1975). Unfortunately, the basic requirement and figures for assumed or actual population variability from which the allowances have been derived are not stated in the dietary standards (Lorstad 1971). Factors that influence the efficiency of nutrient utilization are considered in setting allowances. These include precursor conversion ratios, efficiency of absorption, digestibility, assimilability, and utilization of complex nutrients like protein (Young 1964; Jelliffe 1973; Harper 1974; United States 1974). The importance of each of these factors differs from nutrient to nutrient, so the extent to which the allowance for different nutrients must exceed requirements varies accordingly. Although the methodology for establishing allowances is generally accepted, estimates of nutrient requirements arrived at by committees represent the results of accommodation and judgement. Consequently, recommendations from different committees may differ (Goodhart 1973; Harper 1974; Canada 1975). 66 2.3.3.1.2 Use and Interpretation of Dietary Standards Dietary standards have come to serve as guides in an expanding number of areas. However, "there are two primary uses of dietary standards — to serve as guides for the planning of diets and food supplies and to evaluate nutritional adequacy from food consumption data" (Hegsted 1975, p. 13). With respect to both uses, publications of dietary standards have regularly indicated that the values are intended for use as guides applicable to populations and large groups, and that they are not intended for application to the individual. Additionally, as mentioned, dietary standards cannot be used by themselves for the assessment of nutritional status of either individuals or population groups (Goodhart 1973). Although the limitations on the use of dietary standards have been published repeatedly, Beaton (1972) indicates, " ... the meaning and inter pretation of the figures remains a matter of doubt, confusion, and often argument" (p. 356). For example, there is a " ... dichotomy represented by the Food and Nutrition Board's insistence that they were intended for use only when dealing with groups of people, but at the same time, the Board's willingness to give, in the 1974 edition, individual figures for twenty-four age-sex groups plus those for pregnancy and lactation" (Leverton 1975, p. 9). It appears as suggested by Beaton (1972) " ... the recommendations are based upon a consideration of the individual, not population requirements -- or rather, upon a consideration of individual requirements within a population" (p. 356). Referring to the Recommended Dietary Allowance of the Food and Nutrition Board, Hegsted (1975) has further criticized the utility of one standard for two purposes -- evaluation and planning of diets -- and suggests that neither/purpose is fulfilled with the present standard. 67 The dispute over whether the standards are appropriate for use with populations and not for individuals is of particular importance to nutrition educators. The warnings provided in the dietary standard, about their use for evaluating or planning individual's diets, although instructive, fail to address the needs of nutritional practitioners who must evaluate and plan individual's diets (Hegsted 1975). Beaton (1972; 1975) and Lorstad (1971) have been instrumental in the development of a logical approach to the interpretation of nutrient intake data for both populations and individuals. Their approach, discussed below, recognizes that population data defining dietary standards are in fact data on a population of indivi duals. The dietary standards so derived can be used rightfully to evaluate and plan the diets of individuals. The objective of a dietary study is to determine whether deficiency exists in the individual, or in the case of a population, the proportion of the population that is deficient. The problem posed then is: what inter pretation should be placed on those individuals with intakes below the recommended level, arid for the population as a whole, what prevalence of deficiency might be expected (Beaton 1972; Beaton 1975)? The Canadian (Canada 1975) and American dietary standards (United States 1974) have clearly indicated that the intake of a person habitually consuming nutrients below the recommended levels cannot be interpreted as an inadequate intake for those nutrients. The method proposed by Beaton (1972; 1975) is, in short that, providing data are available to describe the distribution of individual requirements in a population, then, by the application of probability statistics it is possible to determine the risk or probability of deficiency to an individual, or the prevalence of deficiency in a population. As Beaton (1975) suggests, 68 " ... assessment of the adequacy of nutrient intake should be based upon a judgement of the probability or risk of deficiency rather than on an 'adequate' or 'inadequate' basis ..." (p. 31). Whether the individual's diet is adequate will depend on where the individual's true requirement lies in relation to the assigned requirement, and this is not known. Similarly, population averages of nutrient intakes can be compared to the recommended allowance, but a population that fails to meet the standard does not necessarily have inadequate diets (Hegsted 1975). Thus, it is only within the framework of probability that the dietary standard can be used legitimately and meaningfully to interpret the relationship of nutrient requirement and the individual's nutrient intake (Beaton 1972; United States 1974; Beaton 1975) or the population's nutrient intake (Beaton 1972; Beaton 1975). The concept outlined by Beaton (1972) will be summarized below. In order to determine individual risk of deficiency, a knowledge of the distri bution of the .individual requirements in the population is needed. This is illustrated below in Figure 2.1 with the assumption of a normal dis tribution about the mean requirement. The recommended intake lies at two standard deviations above the mean, a level which should meet or exceed the requirements of 97.5% of the individuals in the population. mean - .u _ u + 26 requirement Figure 2.1 Individual variability in nutrient requirements (Beaton 1975). 69 Utilizing the data on the variability of nutrient requirements in the population (Figure 2.1), a cumulative distribution can be developed to (Figure 2.2) describe the proportion of the population with requirements above a certain intake level. This curve corresponds to a distribution of probability or risk of deficiency for an individual at any given intake level, and provides a basis for interpretation of individual nutrient intakes. Probability of deficiency in individual s ~* nutrient intake u Figure 2.2 Proportion of the population having actual requirements above nutrient intake and the probability of deficiency in individuals ingesting a particular level of nutrients (Beaton 1975). Beaton (1972) derives a first principle for use in interpreting individual and population nutrient intake data from the above graph (Figure 2.2). The principle states " ... an individual habitually consuming the recommended intake (or more) of a nutrient must be considered to be at low risk of deficiency. As ... intake falls, the risk of deficiency increases" (Beaton 1972, p. 358), in a manner predictable from a knowledge of the distribution of requirements. Thus, although it is not known whether an individual intake is inadequate, the individual's dietary intake can be interpreted in terms of the risk of deficiency associated with that 70 intake. This concept can also be applied to populations to predict the prevalence of deficiency in a population. For this purpose Beaton (1972) derives a second principle. The second principle states, " ... when population data are considered, it is necessary to consider both the variability of nutrient requirements among individuals and the variability of habitual intake among individuals" (Beaton 1972, p. 358), before the prevalence of deficiency can be meaningfully determined. The variability of both requirement and intake must be considered since the ratios of these variabilities will influence the prevalence of predicted deficiency as illustrated in F„igure 2.3 below. A consideration of the average intake of a population alone will not provide a measure of prevalence of deficiency. In this respect Beaton (1972) derives a third principle which states, " ... an observation that the average intake of a population group is at or above the recommended intake does not mean that all individuals are well nourished" (p. 359). Figure 2.3 Theoretical model of the relationship of the nutrient intake of a population to the prevalence of deficiency (Beaton 1972). 71 Thus, the dietary intake of an individual cannot rightfully be deemed adequate or inadequate on the basis of a comparison with the dietary standard, since it in not known where the individuals true requirement lies in relation::to the standard. However, it has been customary to make such assignments. Similarly, nutritional adequacy is not assured if the average intake of a population meets the standards. The prevalence of deficiency will depend on the range of individual requirements and intakes within the population. Therefore statements about population groups such as, " ... 'The average intake meets RDA standards; therefore, there is no problem of nutritional inadequacy.'are ... invalid" (United States 1974, p. 14). Also statements such as the " ... 'RDA includes a large safely factor; therefore a diet that meets two-thirds of the RDA standard should be adequate', have no validity" (United States 1974, p. 14). A nutrient intake of two-thirds of the dietary standard will be adequate for some but f inadequate for others; there is no way of knowing who falls into which category. However, providing the distribution of requirements is known the risk of deficiency at this level can be computed. Unfortunately, the distribution of requirements is not available for most nutrients for which requirements are established. In these cases, requirements are established by committee judgement to include an estimate of population variability. Whether true or estimated variability is used, it would still seem desirable to utilize Beaton's concept of risk so that an individual habitually consuming the recommended intake, or more, must be considered to be at low risk of deficiency. As intake falls,."the risk of deficiency increases. Since individual requirements are not known, the individual's intake can be interpreted in terms of the risk of deficiency associated with a particular intake. Depending on the available data on population variability, this interpretation can be performed either with 72 high numerical accuracy, or alternatively, simply as an indication of a trend towards greater or lesser risk. 2.3.3.1.3 Limitations of Dietary Standards Even if the nutrient content of the daily food ration meets the dietary standard, there are several reasons why this is insufficient information to formulate an ideal, or adequate, or low risk diet. Dietary values should be applied with recognition of the following limitations. Present knowledge of nutritional needs is incomplete. Nutrient needs as yet undiscovered may still exist. For this reason the Bureau of Nutritional Sciences in their publication "Dietary Standards for Canada " (Canada 1975) suggests, " ... recommended intakes should be achieved by eating a variety of foods because unknown nutrients may be present which are essential for the maintenance of health' (p. 7). Similar qualifications are forwarded in the American standards. However, this is not completely consistent with the indication in the same volumes, that standards are acceptable for use as guidelines in processing and fabrication of foods (Hegsted 1975; Leverton 1975). Requirements for many nutrients recognized as essential have not been established (Harper 1974; United States 1974), and for those that have, the requirements are based on a criteria of deficiency rather than a nutrient optimum (Mertz 1972). Since requirements are based primarily on knowledge of deficiency states, they do not differentiate long-term nutrient con sequences other than some conditions of deficiency. Further, safe maximal levels of nutrients have not been established. Many possible factors which may significantly influence requirements in individuals have not been established. For example, standards do not 73 explicitly consider: drug-nutrient interactions, such as oral contracep tives; nutrient-nutrient interactions, such as the effect of vitamin-D on calcium retention, protein intake and the vitamin-B6 requirement, and amino acid balance with protein requirement; the effect with noxious chemicals; individual sensitivity to nutrients; diurnal rhythms; diet pattern; nutrient history; climatic conditions; and the effect of a poten tially large number of environmental factors. As "... a source of nutrients, food has psychological and social values that are difficult to quantify" (United States 1975, p. 2), these psychological vectors of nutrient requirements have not been considered explicitly in defining the present standard. Other factors important in determining the dietary requirements of individuals include: the organism's ability to adapt to its environment, and the aspirations of the individual and the expectations that society has for the individual. Factors, such as those above, must be considered in defining human nutrient requirements, however, there is still lack of knowledge concerning the factors that effect nutrient requirements in individuals. (Goodhart and Shi 1 s 1973; United States 1974; Canada 1975). Additionally, for many nutrients for which requirements have been identified, there exists considerable conflict in their formulation (Harper 1974). For example, Hegsted (1967) points out that, although most of the people of the world do not consume enough calcium to meet the dietary recommendations, and thus are often said to be calcium deficient, there is no convincing evidence that this is true. We do not even know what calcium deficiency looks like in man" (p. 107). It isluncertain.whether criteria to determine requirements in adults should be set to maintain body weight, or prevent nutrient depletion as indicated by balance studies, tissue concentrations, specific functions, or specific deficiency signs (Harper 1974). Note that "for some nutrients there is 74 a considerable difference between the amount that will prevent the develop ment of specific signs of deficiency and the amount required to maintain maximum body stores" (Canada 1975, p. 6). Additional criteria that may be useful in determining requirements, such as sleeping pattern and wound healing time, have been proposed (Davis and Williams 1976). Further, data used in estimates of requirements are often fragmentary and based on limited experimental data on humans (Canada 1975), or in fact may be extrapolated from animal studies (Hegsted 1975). 2.3.3.2 Further Dietary Recommendations 2.3.3.2.1 Recommendations for Nutrients Not in the Dietary Standard The recent Canadian dietary survey (Canada 1973) indicates that some nutritional problems exist in the Canadian population with respect to nutrients for which the "Dietary Standard for Canada" (Canada 1975) has established requirements. However, vital statistics (Canada 1976b) report negligible mortality in the Canadian population from diseases associated with these nutrients. Although the "classic" nutritional deficiency diseases appear to be of minimal public health significance in Canada, dietary components have been implicated as etiologic factors in diseases which are among the major causes of death in Canada. These diseases include, most notably, cardio vascular disease (ischaemic heart disease and cerebrovascular disease) and cancer of the gastro-intestinal tract (Lalonde 1974). Diabetes mellitus also ranks as a significant cause of mortality (Canada 1973). Additionally, these aforementioned diseases have importance for population morbidity figures, as do the diet related problems of dental caries, periodontal disease, obesity, and alcoholism. 75. Whereas the dietary standard provides figures for the nutrients associated with the diet-related diseases of low mortality, appropriate standards are not included for the dietary components and nutrients associated with the high-mortality diseases. However, recommendations and guidelines for these nutrients and food components, which may be used to supplement the dietary standards, are available from recognized nutritional agencies. The American Heart Association (1973), Canadian Committee on Diet and Cardiovascular Disease (Canada 1976a), and Senate Select Committee on Nutrition and Human Needs (United States 1977f) provide recommendations for the dietary intake of fats and salt. These are proposed primarily because of their implication in the etiology of cardiovascular disease. To reiterate briefly, the suggestion by these agencies is for a reduction in consumption of fat. Reduction in saturated fat with a proportionate increase in poly-and mono-unsaturated fat intake is also recommended. The American recom mendations (American Heart Association 1973; United States 1977f) also suggest limiting cholesterol intake. Lower intake of salt is recommended by these agencies since excess dietary sodium is considered an adverse factor in some people prone to hypertension (Men eel y and Battarbee 1976). However, in accepting these recommendations, it should be appreciated that the diet-heart hypothesis has not stood uncriticized (Werko 1976; Mann 1977). Also, the goals of the Senate Select Committee on Nutrition and Human Needs have been both supported (Latham and Stephenson 1977) and criticized (Harper 1977) in recent publications. The Senate Select Committee on Nutrition and Human Needs also presents guidelines for total and simple carbohydrate composition of the diet. Additionally, the Bureau of Nutritional Science of Canada (Cheney 1976) is 76 developing guidelines for meal replacements which provide values for dietary fibre. Decreased dietary fibre intake has been implicated with increased incidence of gastro-intestinal cancer and other gastro-intestinal disorders (Burkitt and Painter 1974; Spiller and Amen 1975); and with coronary heart disease (Klevay 1974). Although dietary recommendations which are based primarily on epidem iological data, such as those from the American Heart Association (1973), may be less precise than those derived from metabolic studies, the potential public health significance of these recommendations may justify, or in fact necessitate, their inclusion in the formulation of prudent nutrition education programs. The implication is that the contribution to the total risk of mortality and morbidity from the secondary nutritional component of the high incidence diseases in North America is as substantial, if not greater, than the mortality and morbidity due to nutritional diseases where diet is the primary agent. Unfortunately, recommendations in this area are highly controvertial and therefore difficult to interpret. 2.3.3.2.2 Recommendations for Maximum Intakes The Canadian dietary standard (1975) suggests that " ... a consideration of possible effects from intakes far in excess of estimated nutrient require ments is outside the scope of present standards" (p. 7). Similarly the American standards (United States 1974) and those of the World Health Organization (Passmore et_ a]_. 1974) do not explicitly consider upper levels on nutrient intakes. Whereas the recommended values presented by the dietary standards are minimum suggested levels of intake, with the exception of calories, many of those discussed directly above correspond to maximum recommended levels of intake. 77 Although quantities in excess of requirements for individual nutrients have been proposed as beneficial (Pauling 1976), the Food and Nutrition Board (United States 1974) suggests that: "... we are aware of no convincing evidence of unique health benefits accruing from consumption of a large excess of any one nutrient. Large doses of individual nutrients may have some pharmacological action, but such effects are unrelated to nutritional function. Claims that large intakes of individual nutrients will cure non-nutritive diseases should be viewed with skepticism ... (p.3 ). No specific benefits or disadvantages are recognized from the ingestion of excessive quantities of many nutrients. However, toxic ceilings have been identified for some nutrients (Goodhart and Shils 1973). In a program designed to monitor nutrient intakes, if possible, assessment should explicitly articulate an acceptable range of intake --that is, both maximum and minimum nutrient limits should be stated. When using standards it is suggested (United States 1974) that intakes substan tially above requirements are not harmful. However, these amounts are not specified. Ignoring values above the requirements established in dietary standards, in light of the availability of high-potency nutrient supplements and extensive food fortification, may be unwise. The necessity of setting minimum and maximum values for nutrients in new products is advised in the American Recommended Dietary Allowance (United States 1974). 2.3.3.3 Conclusions Standards against which individual nutrient intake can be judged are available both from publications of dietary standards and from the recom mendations proposed by recognized nutrition agencies. The standards are not absolute limits which separate good from poor diets, but rather are guidelines or goals which indicate the likelihood or probability that a diet is adequate for an individual. The standards may be used meaningfully 78 to provide guidelines for individual and population dietary practices. When dealing with total populations, a particular prevalance of deficiency can be set as the public health objective -- that is, a certain risk to the population. When intake variability in the population is high, and therefore per-capita nutrient increases needed to reduce the prevalance of deficiency would.be unreasonable, then the public health effort would be to selectively identify those persons with low intakes and raise their intakes. Since actual requirement and consequently true risk are not known, the usual approach in counselling individuals is to urge consumption•of sufficient nutrients to achieve the dietary standard, which corresponds to a low level of risk. Both the individual counselling and the national planning approach can be defended as proper public health objectives. In using the dietary standard, a public health or statistical concept is used rather than a clinical or therapeutic one (Clements 1975: Latham and Stephen son 1977). In many cases these standards or goals are not well accepted. However, the importance of accepting some uncertainty in developing standards is emphasized by Latham and Stephenson (1977), who, in_discussing the nutrition-education profession's responsibility for establishing goals, state: Worrying to us is the use by the opposition of the argument that we lack information to set goals at all. That surely is an abrogation of responsibility, a "cop-out". There will always be uncertainties and more research that needs to be done. But advice has to be given, Goals need to be set ... (p. 152). Further the results of evaluation should not be erroneously extrapolated into a judgement about the nutritional status of the individual. Although there is little virtue in assigning interpretations to dietary data that cannot possibly be valid, one cannot deprecate the value of dietary data as an indication of potential nutritional status. Evaluation of dietary 79 data can be utilized as an effective public health tool to provide assess ment of food selection practices of individuals -- and in fact may be the only practical method presently available to do so. 80 2.4 Dietary Prescription 2.4.1 Diet Planning Conventional methods (Ohlson 1972) of planning human diets for out patient diet counselling, or for modifying pre-assessed hospital routines, include: generalized and simplified models for assessing dietary adequacy, such as food guides, concepts of dietary variety, the prudent diet, and checks on a limited number of nutrient values; computational aides, such as exchange lists; and a variety of rules of thumb and trial-and-error procedures. The above methods help to simplify the inordinately complex procedure of dietary assessment and prescription. However, these methods do not make explicit use of all the information necessary to provide precise conclusions in the planning of human diets (Gelpi et_ al_. 1972; Balintfy 1973a; Head et_ al_. 1973). This problem becomes particularly evident as the complexity of a diet problem increases, or alternatively, as the constraints on an acceptable solution become more severe. Human diet problems are amenable to mathematical definition, formulation, and solution (Balintfy 1973b; Balintfy 1976a), and in fact may require mathematical modelling for effective resolution (Balintfy 1973a). Balintfy (1973a) indicates that "... the problem of diet planning is not a nutritional, but a mathematical one ... [which] ... defies definition, both conceptually and operationally, unless the problem is cast into some mathematical model ..." (p. 581). Therefore, it would seem important that realistic mathematical formulations be available for human diet problems. Without mathematical techniques, nutritionists have limited use for the potentially extensive information required to ensure that nutrient adequacy, preference, production requirements, and budgetary restraints 81 are met when planning diets -- " ... more data means only that more infor mation will have to be ignored" (Balintfy 1973a, p. 581). Consequently, conventional procedures exhibit suboptimal decision processes as compared to mathematical programs (Balintfy 1976b). For example, menu planning models, utilizing mathematical programming techniques designed to optimize costs, have demonstrated raw food cost savings of 5 to 34 percent and improved nutritive control without sacrificing patient satisfaction, as compared to conventional planning techniques (Gelpi e_t al_. 1972; Balintfy 1975). Additionally, increased technical demands on dietitians necessitate the use of tools which improve efficiency. Computerized mathematical programs for diet planning provide dietitians with data, and techniques for data manipulation, which would otherwise be impracticable. Also, there is some indication that the demand for nutrition services may outstrip avail able manpower committments unless supplemented by technological means (Witschi et al_. 1976). Food selection or diet planning problems can be modelled mathematically as constrained optimization problems (Balintfy 1976a). As such, the planning of human diets is a decision process to assure satisfaction of a number of simultaneous requirements. These requirements are certain attributes of food and food intake, such as food cost, nutritive value, palatability, and production characteristics (Gelpi ejt al_. 1972). Although any suitable set of optimization objectives and constraints may be defined for diet planning, typically coefficients of food cost have been optimized within constraints established by nutritive specifications and palatability (Smith 1963; Balintfy 1976a). Recently, considerable exploration has been undertaken of models which optimize measures of food preference (Balintfy 1976a). 82 Human diet problems, according to Balintfy (1973b), fall into two major categories; food planning and meal planning. Food planning is concerned with decisions of food allotment, over an undifferentiated but specified time period, which characteristically must meet given budgetary, nutritional, and acceptability requirements. Alternatively, meal or menu planning is concerned with a temporally-defined food-allotment decision. Meal planning defines " ... an optimum sequence of meals consisting of combinations of prepared foods, called menu items, ... such that the required structure of meals and given budgetary, nutritional, and food production specifications are met" (Balintfy 1973b, p. 1). As discussed below, a number of approaches to defining, formulating, and solving human diet problems by mathematical means have been undertaken. 2.4.1.1 Food Planning Models 2.4.1.1.1 Food Planning Without Palatability Considerations Early food planning models were designed to find "minimum-cost" combinations of foods which met specified nutritional standards. These purely nutritional models were formulated to find the lowest cost of physiological subsistence and not to consider elements of dietary palata bility or acceptability such as: variety, prestige, or other cultural and personal concerns (Smith 1963). In order to optimize food costs within a prescribed nutrient allowance three types of data are required: a list of nutritional requirements, the nutrient composition of available foods, and price coefficients for each food item. A formal statement, in algebraic notation, of the essential relationships between the optimization objective and the three kinds of 83 information used in the construction of least-cost diet models is as follows: Let z be the total expenditure, n the number of foods, p- the unit J price of commodity j, x- the quantity of commodity j to be consumed, J m the number of nutritional requirements (restraints), b. the quan-th titative requirement set by the i restraint, and a^ the quantity of nutrient i contained in one unit of commodity j. In this notation, the problem is to n minimize z = E p.x. j=l J J subject to (1) x- ^ 0 (j = 1, 2, 3, ..., n) 0 n (2) E a, .x. » b. (i = 1, 2, 3 m) j=l 1J J 1 That is, z, the total expenditure on foods purchased, is the sum of the expenditures on each food, where p.x. is the expenditure on the th j food. The quantity z is to be made as small as possible, subject to the two requirements that (1) no negative quantities of foods may be purchased and (2) the total quantity of the i nutrient (the sum of the quantities, a-j-x., provided by each food) shall equal or exceed the required amount, b^, for each of the m nutrients (Smith 1963, pp. 6-7). As indicated by Smith (1963), the earliest formulations of the minimum-cost diet problem were by Cornfield (1951) in 1941 and by Stigler (1945) in 1945. Although both authors developed solutions, Cornfield's was proposed for the case of two foods with any number of nutritional restraints (Dorfman ejt al_. 1958; Smith 1963). Stigler provided the first solution for a general problem with nine equations of nutrient requirements and seventy-seven unknowns for the amounts of food items in the diet (Danzig 1963; Smith 1963). Stigler's solution was not obtained by linear programming but by a systematic procedure of trial and error. Consequently, he could not be 84 certain that an optimal solution had been obtained, although he believed that his result was close to the optimum (Smith 1963). Linear programming of human diets, and in fact linear programming in general, was not available until Dantzig and Laderman offered a general method of solution to the least cost diet problem, called the Simplex Method, in an unpublished paper written in 1947 (Dorfman et_ al_. 1958; Dantzig 1963). Thus, Dantzig's discovery of the Simplex method of linear programming allowed for realistic solutions of an econometric problem previously only solvable in principle. The solutions obtained by Stigler (1945) and in 1947 by Dantzig and Laderman (Dantzig 1963) appear in Table 1. The recalculated solution to Stigler's problem by Dantzig and Laderman, used Stigler's August 1939 price index data, his food composition data for seventy-seven items, and his nine nutrient allowances. This linear programming solution produced an annual subsistence diet with beef liver instead of evaporated milk, and required that the proportions of other items change. However, just as Stigler believed, his solution was close to the true least-cost diet --within 25 cents per annum. In fact, Stigler's answer was only one iteration from the optimum (Vajda 1958). 85 Table 2.1 Annual subsistence diet for a moderately active adult male, calculated using linear programming methods.* Commodity Stigler solution Dantzig-Laderman.solution amount cost amount cost (lb.) ($) (lb.) ($) Wheat flour (enriched) 370 13.33 299 10.77 Evaporated milk 57 cans 3.84 Cabbage 111 4.11 111 4.11 Spinach 23 1.85 23 1.85 Dried navy beans 285 16.80 380 22.28 Beef liver 2.4 .69 total annual expense... 39.93 39.68 * Calculated by Stigler (1945) and recalculated by Dantzig and Laderman (Dantzig 1963) + Cost determined using Stigler's August 1939 price index data. Food planning models using linear programming have been developed by other authors including Vajda (1958), Beckman (1960), and Smith (1963). These models provide solutions of minimum cost diets which meet from 3 to 13 nutrient restrictions, and consider a commodity list of from 8 to 73 food items. Since linear models usually provide solutions with fewer foods than restrictions, the diets obtained from these models contain a predic tably small number of food items. Beckman's solution contains 4 foods, Vajda's solution 3 foods, and Smith's midget model solution 6 foods. Smith (1963) indicates that neither Vajda's model nor his own model can be consid ered, atrue least-cost subsistence model since the food list used was chosen with consideration of each item's palatability. Lower-cost, unpalatable items may not have been included in the calculations. The food list used obviously effects the character of the solution obtained. 86 A recent reflection on the low-cost diet, called the "three-consid eration diet" (for the three statements of nutrient allowance, food com position, and food prices), has been introduced by Lewis and Peng (1977). Daily diets were computed for each member of a four-person reference family using the 1974 U. S. Recommended Dietary Allowances, current food composition data, and Bureau of Labour retail price statistics for Atlanta, October 1975. The authors indicate that the diets produced are less than satisfactory with respect to variety and palatability, although they are somewhat more varied than earlier models. The additional nutrient constraints, 17 altogether, have resulted in diets of 7 or 8 items which includes vege table oils and grain products in addition to other food groups. In com parison Stigler1s solution contained no animal fat or vegetable oil, and the Beckman diet did not contain grain products. Some authors have focused on the application of linear programmed least-cost diet problems for field conditions rather than for illustrating the mathematics of linear programming. Florencio and Smith (1969; 1970) have utilized least-cost diet methods to measure the efficiency of food purchasing among working-class families in Columbia. Efficiency was determined by an index which compared the cost of actual foods consumed with a mathematically optimized least-cost diet which was developed from commonly consumed foods. Least-cost models have also been used by Chamberlain and Stickney (1973) and Kansra ejt al_. (1974) for development of least-cost nutritionally balanced multimixes suitable for children in developing countries. All of the models presented above, and other diet models to be presented below, utilize linear nutrient constraints which are specified in absolute amounts, or where nutrient interdependences exist, as fixed 87 ratios or exact constants. Smith (1974) has developed a food planning model which modifies the standard formulation of dietary constraints by the intro duction of nonlinear constraints for determining protein allowances. The model considers simultaneously the effect of both protein concentration and protein composition on protein allowance in determining the most economical diet which satisfies protein needs and other nutrient require ments. Utilizing the nonlinear protein constraint provides an opportunity to economize by consuming smaller quantities of higher quality protein or larger quantities of lower quality protein while meeting protein require ments. The problem is solved using separable programming to obtain linear approximations to nonlinear restraints. Smith's (1974) model was designed for national food planning problems where efficient use of protein resources is required, for example in the developing countries, Carmel (1976) has developed a model of similar nature and application using the concept of NDpCal percent to account for the relationships between energy content, protein quantity, protein quality, and protein value of diets. Diet planning with variable nutrient coefficients has been discussed by Armstrong and Balintfy (1975). 2.4.1.1.2 Food Planning Models with Palatability Considerations Comparison of the results of purely nutritional models with low-cost diets prepared by nutritionists illustrates discrepancies in dietary cost and composition due to factors beyond those explicitly considered in the modelled diets (Smith 1963). As Calavan (1976) states: The solution to this prescriptive food selection model typically consists of an unpalatable, boring combination of three to seven food items. Except for populations on the verge of starvation, such solutions are optimum only in an arbitrary mathematical sense (pp. 65-66). 88 Consequently, many later diet models have included explicit statements of palatability in their formulations, in attempts to produce diets which are more congruent with actual consumption patterns. Both linear and non linear models have been developed. In an early example of formulations considering palatability, Brown s (1954) attempted to develop a descriptive programming model which would produce a diet for one person similar to the actual diet of the British working class. As Calavan (1976) indicates, this objective was different from that of prescriptive models which attempt to define nutritionally adequate diets. Three initial models tested by Brown were similar to the nutritional models which have been previously discussed. In these models prediction of consumer behaviour was based on the hypothesis that the consumer's objective was to select the least-cost diet which met specified minima of twelve food composition factors. These factors corresponded to nutrient restraints established by population practices and not to physiological requirements identified in dietary standards. The computed diets were derived from a basic list of 15 food groups averaged for price and nutrient composition. The solutions obtained consisted of from five to eight groups for the fifteen food groups. Not unexpectedly, Brown found that as more nutrient restraints are included, the diet becomes more varied and more expensive. In Brown's fourth model an additional objective of consumer behaviour was included. This was an explicit restraint on the levels of bread and potatoes consumed. Brown's most sophisticated model provides a weekly diet for one person which includes eight of the fifteen groups consumed by the population. He judged the fourth model as adequate because it selected 89 foods in all the major groups used by the population, except fruit, and because the calculated weekly expenditure is only about twelve percent lower than actual population expenditures. The major difference between the formulation of Brown's models and earlier nutrition models is the use of quantity restraints on some items. This concept has been further extended by Smith (1963). Smith has developed three food planning models beyond the midget model which was previously discussed (p. 85). These are; the midget model with cooking aids, the small model, and the large model. These models utilize an objective function of cost minimization, and explicit palatability constraints designed to provide solutions which conform to conventional consumption patterns. Smith's most sophisticated model, the large model, provides an inexpensive, nutritious, and reasonable palatable diet of 62 items for a family of three over a four week period. Unlike the other Smith models (and those of Stigler, Beckman, Vajda, and Brown) which have a limited number of food classes, the large model utilizes an expanded commodity list of 572 widely consumed individual food items and narrow commodity classes. In the large model, item acceptability and dietary variety depends primarily on the use of a large number of constraints on the amounts and associations of foods included in the diet, rather than on a restricted commodity list as in earlier models. In fact many of the items in the expanded commodity list are unpalatable raw materials or ingredients which become acceptable only by their association with other ingredients in the diet. The 85 commodity constraints used in the large model include ten complimentary constraints which define proportionalities among different items, 28 maximum quantity and 41 minimum quantity limits on the consumption 90 of certain foods, and six requirements for specified amounts of some commodities. Only the large model uses all four types of commodity res traints. Additionally, thirteen nutritional restraints are defined. Twelve identify minimum acceptable intakes while the thirteenth provides a maximum total caloric intake for the diet. Thus, a total of 98 restraints have been defined in the large model. The commodity constraints protect against excessive amounts of some items, and against exclusion of important items which are common in the usual diet or which are required to make some ingredients palatable. The constraints are in most cases based on exact or adjusted figures derived from the consumption patterns of populations, although in some cases arbitrary or experimental values are incorporated. Calavan (1976) used a linear programming approach to develop a des criptive model of food selection practice which would be appropriate for research on the epidemiology of malnutrition. Data on the socio-economic variation of food-use practice in a northern Thai village was used to identify indices or goals around which the residents optimized dietary behavior. The identified goals of the population were to satisfy energy requirements, maximize dietary variety, maximize intake of animal foods, and maximize fat intake. A tentative model was constructed utilizing a series of linear programs which, while staying within budgetary limits, selected foods to satisfy a specified sequency of the above householder dietary goals. Comparison of model-generated and household-generated food lists was the intended evaluative format, however, insufficient data were available for field testing. Instead, less conclusive tests of comparison indicated that the goals adopted in the original model were consistent with 91 population data on socio-economic variations in dietary behaviour, provided the sequency of these goals was slightly modified. Various nonlinear food models have been developed. The Consumer and Food Economics Division of the U. S. Department of Agriculture has developed a nonlinear programming model to aid in adjusting their family food plans to coordinate with food price fluctuations, changes in established nutrient allowances, and changes in eating habits of the population (Balintfy 1976a). The model provides an adjusted food plan which is nutritionally adequate and which approximates food-group consumption patterns for each of 22 sex-age groups .and income levels established by the 1965-66 Household Food Consumption Survey. The,precise formulation is as follows; Let q.j denote the past consumption of food quantity i and x^ the corresponding quantity in the new food plan. The optimal food plan is thus formulated as the following quadratic programming model: n 2 2 minimize £ w.(q. - x.) , i=l 1 1 1 subject to Ax > b, Rx > d, where w. are weights to equalize the relative contributions of deviations, and where A is the matrix of food cost and nutrient composition data for up to 17 food groups and 18 nutrients per group. The R matrix represents a set or upper and lower bounds as well as proportionality constraints imposed on the components of the solution vector to assure strictly positive and accep table food quantities Balintfy 1976a, p. 328). In 1958 Wolfe developed a nonlinear food planning model which reduces dietary monotony while ensuring economy (Smith 1963). The model incor porates a quadratic index of disutility or "fatigue" in the ojbective 92 function which is based on the assumption that excessive consumption of any food would cause disutility proportional to the square of the quantity consumed. The quadratic expression ensures a disproportionately large penalty at higher levels of consumption for any item and therefore limits the total intake of any item. Wolfe!smodel is formulated as follows: Let n be the number of foods, p. the unit price of commodity j, x- the quantity of commodity j to be consumed, m the number of restraints in the model, b. the quantitative requirement set by the i restraint, a., the quantity of nutrient i contained in one unit of commodity j, 1J +h and f. the "fatigue" or disutility function for the j food. Minimize z for all of the arbitrary numbers P between 0 and 6.8077, where n n 2 z = P Z-PiX,- + z f.x, (j = 1, 2, 3, n) j=l J J j=l 3 3 subject to (1) x. > 0 n (2) E a. .x. * b. (i = 1, 2, 3, ..., m) (Smith 1963, j=1 p. 30) In the above equation, P.is an arbitrary weighting factor which determines the emphasis to be given to dietary economy. At P = 0 there is no cost consideration, whereas at P = 6.8077 cost considerations are dominant as in the classical cost minimization linear solution, z is the total of the weighted cost- of the diet plus the index of disutility. To test the model Wolfe used data from the Stigler model for the nutrient requirements and for cost and nutrient coefficients. Only twenty items from Stigler food list were used. Wolfe found that as the emphasis on economy was decreased, ;the solution developed from the basic Stigler solution of five items to incorporate all twenty items available in the abridged food list. The cost of the diet rose accordingly. 93 In a further nonlinear example, Balintfy (1976a) discusses a model, called the "weightwatcher.'s quadratic diet model", which contains a non linear objective function in its formulation. The model functions to maximize the total dietary utility or "preference" while meeting cost and caloric constraints. The program uses data on the estimated quadratic utility functions for sixteen food groups, and on the cost and caloric content of each food group. Nutrient considerations are not extended beyond the caloric content of the diet since, it is assumed, protein needs will be satisfied by the strong utility of meat and milk products in the objective function, and other nutrient needs can be satisfied by supple mentation. Presumably the model is not ultimately restricted to this limited nutrient domain. 2.4.1.2 Menu Planning Models Menu or meal planning is a decision process of defining the serving sequence and/or serving frequency of prepared foods, called menu items, such that required nutritive, production, economic, and palatability constraints are satisfied (Balintfy 1973b; Balintfy 1976a). Two approaches to menu planning by computer have been reported -- a non-mathematical method called the "random approach" (Eckstein 1967) and a mathematical modelling approach which has evolved to include a variety of linear and nonlinear programming models (Balintfy 1976a). 2.4.1.2.1 Menu Planning Models - The Random Approach The random approach to menu planning (Eckstein 1967) attempts to simulate the routine decision making processes used by dieticians in planning menus. The menu is compiled by randomly selecting items from a variety of meal component classes and sequentially evaluating these items against predeter-.94 mined acceptability criterion used by dieticians. These criteria include colour, texture, shape, flavour, caloric content and cost of the food item, as well as other acceptability factors. Items from each meal com ponent that satisfy the criteria are included in the solution. The random approach is based on the concept of bounded rationality --a decision process where acceptable solutions are reached without consid ering all the possible alternatives. This approach to menu planning has been largely supplanted by mathematical programming methods which, by contrast, produce optimal solutions based on considering all the possible alternatives (Balintfy and Nebel 1966). 2.4.1.2.2 Mathematically Programmed, Multistage, Menu Planning Menu planning as a mathematical programming problem was first identified and solved by Balintfy (1963; 1964). This first approach used a multistage decision rule which planned an optimal combination of menu items for a sequence of meals by considering the problem of scheduling on a sequential meal-by-meal, day-by-day basis. This model has been formulated in both a linear programming version (Balintfy 1966; Neter and Wasserman 1970) and as an integer programming version which approximates the theoretical solution to the problem (Calintfy 1964; Gue and Liggett 1966). This first model utilized menu-item quantities of fixed-portion size as decision variables to produce a nonselective menu. Such a nonselective menu is obtained by assigning one menu item to each of several menu-item classes or courses (for example, appetizer, entree, cereal, bread, and beverage for breakfast courses) over a defined period of time. Gue and Liggett (1966) have extended the concept of menu-items as decision variables to fixed-choice groups of two or more items, and thereby developed 95 selective menu planning model. Selective menu planning allows a choice of items from each menu item class. This concept has been further extended to selective menu planning from variable-choice groups by Balintfy (1971). A refined model based on the above approach was converted into a self-contained food services information processing and menu planning computer package, called "System/360 Computer Assisted Menu Panning" or "CAMP" (Balintfy 1969), for use in institutional food planning in the public domain. The system includes a systemmatic approach to data collection and data management, as well as to menu planning (Balintfy 1975). The menu-planning objectives, which define optimality in the CAMP system, are to determine the least-cost combination of menu items for a sequence of days which meet nutritional and acceptability requirements Balintfy (1975). Item acceptability and dietary variety are provided by structural, separation, and attribute constraints which ensure compatitiTity of items between meals and within meals. Compatibility within meals is provided by use of attribute codes which restrict the entry of items of similar attributes from appearing more than desired in any meal. Further, formal require ments on the structure of the menu are imposed. The structural require ments partition the menu into a customary array of menu components for each meal, to which only appropriate menu items can be assigned. Compatibility between meals is achieved by separation constraints which indicate the desired serving frequency. The separation constraint defines a minimally elapsed number of days between consecutive rescheduling of the same or similar items. Other constraints on proportionality of items, and on production requirements are also included. The constraints used in CAMP are based upon the expressed preferences of patients and upon institutional policy. 96 CAMP has provisions for selective and nonselective menu planning (Balintfy 1975). Selective menu planning is intended to further enhance menu acceptability. Although structural, attribute, and separation constraints are operative in selective menu planning, nutritional constraints are not. It is reasoned that is nutrient-assured meals are required the nonselective first choice must be accepted. Multistage menu planning has evolved to include procedures for variable portion-size and chance-constrained modelling. Armstrong and Sinha (1974) have developed a quasi-integer programming algorithm to plan nonselective menus in which the portion size of the menu items can vary over a specified positive range. This is an advance over earlier systems in which fixed portion size was a technical necessity. Another advance is demonstrated by Balintfy (1976a) with the introduction of random variables as constraints. This particular application considers the nutrient content of any item in probabilistic terms and the constrained solution as meeting nutrient require ments individually or collectively with a specified probability. Multistage menu planning can also incorporate recently developed non linear preference maximization objectives which utilize some measure of consumer satsifaction (Balintfy 1976a). Typically cost-minimization objectives have been used since adequate quantitative measures for consumer preference were absent, and since earlier modelling objectives were chosen to demonstrate the economic impact of mathematical optimization as opposed to conventional methods (Balintfy 1973b). The use of preference maximization objectives has been demonstrated in single-stage menu-planning models to be di scussed. Multistage menu planning has been implemented in a variety of institu tional food services including hospitals, colleges, penal and mental 97 institutions. Applications of the CAMP system (Bowman and Brennan 1969; McNabb 1971; Gel pi et al_. 1972; Balintfy 1975) have demonstrated savings of 5 to 34 percent in raw food costs, improvements in nutritive control, and equivalent acceptability and variety standards as compared to conven tional techniques. Prototypes of the CAMP system (Balintfy 1964; Balintfy and Nebel 1966) and other multistage systems (Gue and Liggett 1966) demonstrate similar advantages. 2.4.1.2.3 Mathematically-Programmed, Single-Stage, Menu Planning Menu planning can be done in a single stage when the frequency of occurrence of menu items over a finite time period is to be determined and not the specific sequence of items in the menu is to be determined (Balintfy 1973b; Balintfy 1976a). Unlike multistage menu planning which provides a consecutive sequency of integer solutions for each of the smallest periods within the planning horizon, single-stage menu planning offers only one mathematical solution for the entire cycle (Balintfy 1966). The common shortcoming of single-stage solutions is that computer scheduling on a meal-by-meal basis is still required to provide a sequence of daily meals (Balintfy 1976a). The difficulty of scheduling for a fixed-time horizon in a single stage is discussed by Balintfy (1974a). The first models incorporating single-stage menu planning (Balintfy 1966) inherited the cost-minimization objective and linear constraints on nutrients, structure, and attributes. The concept of minimum separation of items used in multistage scheduling, as a safeguard for variety, was used in single-stage models to establish upper bounds:onrjthe frequency of occurrence of items. Although the use of these upper-bound constraints was a crude method of maintaining preference levels, studies in a number of situations indicated single-stage menu planning models produced diets 98 which met nutrition and acceptability levels while reducing costs (Gelpi et al. 1972; Balintfy 1976a). With the discovery of time-related food-preference functions (Balintfy et al.197 4) and time-related food-preference and quantity functions (Balintfy 1973b) it became possible to optimize the quantity and frequency of food intake based on an empirical measure of food preference or utility (Balintfy 1976a). Recent single-stage programming models have incorporated nonlinear objective functions which maximize measures of total food preference subject to given nutrient, cost, structural, and assorted attribute, proportionality, production, and other constraints -- constraints similar to those in CAMP (Balintfy 1974b; Balintfy 1976a) An alternative non-linear programming formulation with the preference quantity functions has been proposed (Balintfy 1976a). This model maintains the linear cost-minimization objective normally used in institutional menu-planning models, and incorporates a nonlinear preference constraint which maintains a given food-preference level. 2.4.2 Information and Behavior Change As indicated at the outset, the task of nutrition educators is both to provide a code of nutritional practice and to communicate this available knowledge to the public sector. One view is that this process is one of planned change, where planned change is "... a conscious effort to alter food-related practices or attitudes when the need exists" (Gifft et al. 1972, p. 255). Thus, the ultimate objective of nutrition education programs, according to this approach, is to modify food behaviour through deliberate intervention. 99 2.4.2.1 Development of Food Behavior and Factors in Food Selection Food selection behavior, as with other human behaviors, is the product of a complex interaction of situational and developmental variables, and of individual and environmental variables. Situationally, food selection is dependent of two factors: food availability -- the environmental variable -- and food acceptability -- the individual-related variable (Gifft et_ al_. 1972). The availability of food in the marketplace and home is of paramount importance in food selection behavior. Availability is governed by climatic and geographic factors, by economic, political and technologic factors, and by public policy and individual decisions (Gifft et al_. 1972). A food's acceptability, in turn, determines which of the available items will be selected and eaten. Item acceptability for the individual is determined psychologically through motivations or needs, such as biogenic (sensory) needs, psychogenic (emotional) needs, and sociogenic (goal) needs; and through cognitions such as ideas, attitudes, and beliefs (Lund and Burke 1969). Developmentally, the psychologic elements of an individual are rooted in the unique interaction of his or her biological nature and socio-cultural factors (Gifft ejt aj_. 1972). The intricate process of this social and emotional acculturation provides the food-related experiences which contri bute to the development of food patterns, whereas the biological heritage determines the physiological needs and capacities, and the potential psychologic and sensory structures which will interact with the environment to create a person's food pattern. Consequently, each individual has a characteristic pattern of eating which has developed over a lifetime as a result of many complex processes and influences, and which has an integral part of that individual's total 100 behavior (Gifft et al_. 1972). This pattern is resistent to change, espec ially complex change, unless immediate benefits are evident, or a change is forced by circumstance (Gifft et_ al_. 1972). Much is vested in a par ticular pattern and the tendency is to move towards familiarity or to reinforce what is already known (Thompson 1969; Gifft et_ al_. 1972). Future gains are obscured by the immediate rewards of not changing. 2.4.2.2 The KAP Gap1 Food practices do not change just because people have accurate facts about nutrition (Leverton 1974). Numerous studies in nutrition education (Hampton et al_. 1967; Baker 1972; Bell and Lamb 1973)and other fields of applied education (Young 1967) have indicated the incongruity of different aspects of behavior. Material taught may not be learned, once learned it need not be believed, and even if a change in attitude did occur '" practices would not necessarily be altered. The reciprocal is also true. Behaviors are not based only on particular types of knowledge. In fact a person may not be able to justify his or her actions and beliefs to the satisfaction of others. For example, individuals may not be able to provide valid nutritional reasons for their nutrition practices (Emmons and Hayes 1973). Although evidence of the effect of nutrition knowledge on food practice is limited and conflicting, some evidence does suggest that acquisition of formal nutrition knowledge is positively related to behavior change in 1. For further discussion of the knowledge -- attitude-practice (KAP) gap see Travers (1963) 101 individuals (Young et a]_. 1956 a+b; Hinton et_ al_. 1963; McKenzie and Mumford 1965; George 1971; Boysen and Ahrens 1972). Admittedly this is not a linear relationship (Gifft et al_. 1972) indicate, studies comparing ;theinfluence of formal nutrition education on food practices are difficult to perform and interpret due, in part, to the conflict of other variables which may simultaneously effect behavior. These include factors such as emotional stability and maturation age (Hinton et_ al_. 1963), informational sources besides the nutrition-intervention technique (Rosenstock et al. 1966) , and early learning and food patterns (Litman et_ al_. 1964; Brown 1967) . That nutrition education is able to influence the eating habits of populations, is illustrated most notably by the successful campaigns of various food companies (Tyler 1962; Gussow 1972; Manoff 1973) and other mass media campaigns (Rosenstock et_ al_. 1966). In these cases, most campaigns were aimed at influencing product choice for reasons other than nutritional value. Therefore, there is some reason to believe, provided similar resources are available, that eating habits can be influenced for nutritional reasons (Manoff 1973). In any case, accurate facts are essential for rational decisions, even though information on proper diet may not prevent, and may even result in contrary, incorrect, or exacerbating behavior. Information or messages are a source of guidance in food pattern development and redevelopment, and thus form a tool through which the nutrition educator can promote planned change (Gifft eteH. 1972) 102 2.4.2.3 Increasing the Effectiveness of Communication Although human behavior is complex and difficult to analyze, predict or manipulate, research has provided some principles and guidelines to increase the efficacy of the educational process, and thereby increase the chance of individual compliance with recommendations. However, precise predictions of the efficacy of using these principles are not possible (Gifft et aJL 1972). There are many approaches to and lists of principles for enhancing the effectiveness of educational programs (Gifft et_ al_. -1972). The underlying theme of these principles is to enhance the learner's receptivity to the message by: (i) increasing the learner's interest through: (a) incentives such as teacher and program credibility, individual ized involvement, and advertized benefits; and (b) appropriate exposure to the message, for example, by enlisting the learner's active involvement in the learning situation, changing process demands, keeping messages short to ensure active attention, and emphasizing thinking versus recall activities (Gifft et a]_. 1972). (ii) increasing the learner's receptivity by constructing a message of maximum potential meaning for the learner (Gifft e_t a]_. 1972)t In eithertcase the focus is on the learner's relationship to the information rather than on the message alone. The present concern is with the construc tion of a message with maximum potential meaning for the learner, and not with the development of other aspects of the teaching-learning process> used to increase the learner's receptivity. 103 Change occurs when the information provided is significant enough to the learner to motivate action (Gifft ejt al_. 1972). It is axiomatic, in the field of communication theory (Berlo 1960), that response to bids for change are governed by the ratio between the anticipated benefit and the energy required to respond. The more reward, the more effort which will be put -out, and the less reward the less the effort. Therefore, the potential response to a message can be increased by increasing the reward or decreasing the effort required, or both. Expected benefits can be increased by choosing messages relevant to the learner's interests, perceived needs, and concerns (Gifft et al_. 1972; Leverton 1974). Effort to respond to recommended alterations in food practices can be decreased by gearing information to the physical and mental skills, cognitive sets, attitudes, resources, and emotional readiness of the receiver. For example, information can be adjusted to the individuali's: physiological needs; psychological barriers such as values, attitudes, beliefs, likes and dislikes, and pursuits such as, family responsibilities, professional demands and leisure-time activities (Gifft et_ al_. 1972). An important consideration in decreasing the effort necessary to respond is complexity of change. Since people's food patterns remain relatively stable and are resistant to change, especially complex change, recommendations for change should avoid unnecessarily complex demands. Complexity of change has been suggested as perhaps " ... the strongest determinant of the speed and extent of adoption" (Gifft e_t al_. 1972, p. 265). Thus, nutrition education programs are more effective when emphasis is placed on the maintenance of present desirable habits and the improvement of current food patterns rather than on radical alterations 104 in diet (Todhunter 1969). For example, persuading a person to eat more of a food he or she already eats may be less complex and therefore easier to accomplish than inducing him or her to use a food which has neven been tasted (Gifft ejt aj_. 1972). Similarly, the individual should not be overloaded with information (Leverton 1974). 105 CHAPTER 3 DEVELOPMENT OF THE PROTOTYPICAL SYSTEM 3.1 Introduction The project goal (p.13) was to develop a prototypical system to provide information on dietary practices for those adults who want to apply nutritional principles to their eating habits, and have sufficient resources (eg. time, energy, education, money) to make use of the information which defines healthful dietary practices for them. This development was under taken bearing in mind two primary difficulties, coincident with nutrition education's aforementioned tasks (p. 1), namely: the problem of developing nutritional guidelines suitable for health promotion in the public sector, and the problem of communicating this information to individuals. 3.1.1 Developing Nutritional Guidelines With respect to the issue of developing a message which incorporates accurate nutritional guidelines, standard dietary assessment procedure of data collection, analysis, and evaluation have been discussed (Section 2.3). These appear to be suitable for the prototypical system's design. Guide lines for evaluating diets can be derived from the dietary standards of various countries and international agencies, and the dietary goals proposed by recognized scientific agencies. This procedure has been considered despite certain problems, namely: considerable conflict exists about the formulation and application of dietary standards; dietary assessment alone does not provide the means to establish nutritional status; and "optimal" nutritional status cannot presently be defined. 106 3.1.2 Communicating Information to Individuals With respect to the second problem, that of communicating this information to individuals with the ultimate objective of changing food practices, the use of educational principles was considered relevant to the system's development. The focus has been the development of a message for motivated individuals -- in this instance a statement about what to consume which facilitates adoption of recommendations. In this context, two features of the message are significant, namely: the comprehensibility of the information, and the acceptability of the suggested changes. 3.1.2.1 Comprehensibility of the Information Education principles indicate that the comprehensibility of information recommending change in eating behavior can be increased by not overloading the client with unnecessary information. In particular, recommendations for change in eating habits should not require extensive alteration in the person's perception of his food environment and eating habits. The type of information typically presented in nutrition education programs is either an explicit statement about foods to consume, for example as a daily menu, or alternatively an implicit statement, for example as a daily nutrient allowance. Although the presentation of nutrient information may provide a clear rationale for food selection, nutrient information becomes functional only when translated into foods and meals. Thus, a statement based on foods to consume should be easier to comprehend than one based on nutrients. In developing such a food-based statement, the emerging problem is to accurately translate nutrient requirement data into a viable food plan. The problem arises because the nutrient contribution of each food with its 107 characteristic nutrient pattern must be considered in developing the diet plan. As the number of nutrients under consideration increases and the restrictions on acceptable solutions become more severe., the accounting problem becomes formidable. Mathematical-programming techniques, based on experimental menu-planning procedures for food service applications and on a variety of programs with industrial applications, are useful in overcoming these problems. They provide for simultaneous satisfaction of many variables, including nutritive, production, economic, and palatability constraints. This method may be used for nutrition education purposes to resolve the problem of accurately translating data on nutrient requirements and food composition into a viable food plan. 3.1.2.2 Acceptability of Suggested Changes The acceptability of suggested changes can be increased by limiting the complexity of change. Complexity of change has been suggested as perhaps the strongest determinant of the speed and extent of adoption. Thus, nutrition education programs are more effective when emphasis is placed on the maintenance of presently desirable habits and the improve ment of current food patterns rather than on radical alterations in diet. Factors which may contribute to the perceived complexity of these changes include: budgetary considerations, socio-cultural patterns, foods available, food habits, taste preference, colour preferences, likes and dislikes, needs and interests, and beliefs about foods. In short, any deviation from the characteristics of a usual or desired food pattern may contribute to the perceived complexity of change. Hence the best approach for compliance with nutritional recommendations should be a diet which deviates as little as possible from a dietary inventory 108 identified as desirable either by past consumption or stated preference. Presumably, this would be similar to the diet which the individual would choose if she or he understood, accepted, and used nutrition knowledge. Using the mathematical-programming techniques previously mentioned, a constrained-optimization algorithm can be formulated to find the combina tion of foods which are similar to an individual's actual or desired food plan while simultaneously considering nutritional, budgetary, and palata bility requirements, or for that matter, any other measurable vector of food or food-selection behavior. 2. Usual food pattern may provide a better monitor of actual preference, being the diet chosen under present budgetary restrictions and long term cultural and personal habits. 109 3.2 System Design and Characteristics The prototypical system, as illustrated below in Figure 3.1, is designed both to assess and plan diets by systematizing the procedures used in dietary analysis and counselling by nutritionists and dietitians. Input data from both a client questionnaire and computer files are processed within the two phases of analysis-evaluation and planning. Within this framework, the nutrient characteristics of the client's initial diet can be analyzed and evaluated, and a diet plan can be produced which should remain close to an individual's actual or desired food plan while simul taneously meeting specified nutrient limits. The outcome of this procedure is an output statement of: the client's initial diet; a recommendation of altered food intakes -- the revised diet;and an analytic and evaluative statement of the original nutrient intake. INPUT DATA Client Questionnaire client's initial diet client's demographic data Computer Files food-item file food-composition file nutrient-limits file attribute-group . matrix DATA PROCESSING Analysis and Evaluation Planning OUTPUT STATEMENTS client's initial diet analysis and evaluation of client's initial diet client's revised diet Figure 3.1 Overview fl planning of owchart of prototypical system for assessment and individual's diets. no 3.2.1 Data-Collection 3.2.1.1 Client's Initial Diet The client's initial diet includes those food items and their quantities, selected from the system food-item file (Appendix A), which the client 3 consumes habitually. This information is used both in diet-assessment and in diet-planning, since it provides the basic data with which food-composi tion analysis and evaluation can be performed in diet-assessment and with which the client's diet and dietary-structure can be defined in the planning phase. The client's initial diet is stored in the system during processing and compared with the revised diet on the output presentation. The client's initial diet is determined by using a multiple-purpose intake questionnaire (Table 3.1) which permits use as a one-day recall, weekly record, long-term food frequency history, or other variants.; Items can be quantified by weighing, by household measures using standard portion-sizes or measured portions, or by estimation. The questionnaire can be self-administered by the client, or used in an interview format for super vised inquiry. The individual and/or institution can select the most appropriate data-col lection^procedure for their purposes. Flexible data collection procedures may be more useful for meeting the variety of client requirements, program capabilities, and other situational factors. 3. Alternatively, the diet can represent the desired regime or a diet previously designed in collaboration with a nutritionist. -i Ill Table 3.1 Excerpt from prototypical intake questionnaire -- dietary intake format:* AMOUNT PER .UNIT SERVING CONSUMPTION FREQUENCY FOOD ITEMS AND DESCRIPTION STANDARD PORTION _SK£ ADJUSTED PORTION SIZE SERVINGS PER DAY WEEK MONTH ENTREE ITEMS -- DAIRY AND EGGS AMERICAN PROCESSED CHEESE. BLUE or ROQUEFORT CHEESE. CHEDDAR, JACK or SWISS CHEESE COTTAGE CHEESE, creamed or uncreamed, any curd. CREAM CHEESE or CHEESE SPREADS. SOUR CREAM or CHIP DIP. YOGHURT, whole milk base. YOGHURT, skim milk base. YOGHURT, part-skim milk base. EGG: raw, boiled, poached, fried (add fat). EGG: scrambled, omlet, souffle, spoon bread. ENTREE ITEMS — CEREALS CORN CEREAL, enriched, ready-to-eat. WHEAT CEREAL, enriched, ready-to-eat. WHEAT CEREALS, more refined, enriched: cooked. OATMEAL, all types: cooked. WHEAT CEREALS, less refined: cooked. PANCAKES, WAFFLES or FRITTERS: made with milk and eggs NOODLES, egg-type, enriched: cooked. SPAGHETTI, MACARONI or NON-EGG PASTAS, enriched: cooked RICE, brown: cooked. RICE, white, enriched, unenriched or parboiled: cooked CORNMEAL or CORN GRITS, enriched: cooked. WHEAT GERM. FRENCH or SOURDOUGH BREAD, enriched: fresh or toasted. RAISIN or RAISIN-NUT BREAD, enriched: fresh or toasted. ETC '.. ETC V'-V'-IV (1 oz.) r-r-iv d oz.) r-r-iv (i oz.) H cup (4 oz.) 2 tbsp. (1 oz.) 1 tbsp. 1 cup 1 cup 1 cup 1 egg 2 eggs 1 cup (1 oz.) 1 cup (1 oz.) H cup h cup k cup 2 at 4" dia. h cup h cup H cup H cup H cup 3 tbsp. 1 slice 1 slice (1 oz.) * A complete list of foods and portion sizes is contained in Appendix A. 112 The questionnaire food items, contained in the system food-item file (Appendix A), were derived from the "Mini Food List with Food Substitutions" developed by Pennington (1976). Pennington's Mini Food List contains a table of 202 commonly-consumed American food-items, called index items, and a comprehensive list of 49 compositional values for each of these index items. Each index item represents one-or-more substitutable items classified according to the coexistence of the 49 nutrient values. In order to develop the system's food-item file, 200 of Pennington's 202 index items were selected. This initial list was elaborated to incorporate most of the substitution items within their pre-assigned groups.in order to fit the system's attribute-group matrix. The food item file (Appendix A) developed contains 221 item clusters, where an item cluster is composed of either a single item or two or more exchangable items of similar nutrient characteristics. An abridged food-item file (Appendix B) of 127 item clusters was developed for testing the system. The item clusters have been categorized within a hierarchical structure of groups and subgroups used for planning purposes, called the attribute-group matrix (Appendix G). Each of the item clusters was assigned an attribute-group code number and an item cluster code number which appears with the item in Appendix A and B. The makeup of this hierarchy is discussed further in the material on diet-planning (p. 121). To aid in quantification of questionnaire items, the description of portion sizes in the system food-item file has been elaborated beyond that available from Pennington's text. Food Portions Commonly Used (Church and Church 1975) and Agriculture Handbook No. 456 (Adams 1975) were used in this extension. This procedure was not in every case free of difficulty as noted 113 in Appendix A, since the descriptions .of standard portions per gram quantity, used by the three authors, were considerably different in some instances. Manual computer input of the questionnaire data was used. Alternative input methods such as interactive terminal interviewing or computer read cards, although perhaps preferable in on-going use required for present purposes. 3.2.1.2 Client's Demographic Data for Defining Nutrient Limits Data on the individual's sex, age, size, activity pattern, and pregnancy status are collected on the questionnaire (Table 3.2). This information is used in the assessment and planning phases of the program to define the nutrient limits appropriate to the individual. n.4 Table 3.2 Excerpt from prototypical intake questionnaire -- client demographic data.* AGE (in years) IDEAL BODY WEIGHT (in pounds) SEX and PREGNANCY STATUS (check one) male female not pregnant female pregnant 1st trimester female pregnant 2nd or 3rd trimester female lactating ACTIVITY PATTERN (select pattern appropriate for sex from table below and check one) A B C D Act. Pat.(men ) Act.Pat. (women;) Types of Activity A •BCD A BCD' No. hours/day No. hours/day 1) Resting metabolism 8 8 8 9 8 8 8 9 2) Sitting or standing still 10 10 12 13 11. 11 12 13 3) Walking slowly Light domestic work (eg. ironing, sweeping floor, cooking, dishwashing, dusting) Light office of industrial work (eg. typing, lab. work, sewing, printing, garage mechanics, machine-tool operation) Sports involving light activity (eg. bowling, golf, sailing) 4 5 4 2 4 5 4 2 4) Walking at moderate speed Moderate domestic work (eg. scrubbing floor, window cleaning, furniture polishing) Moderate industrial work (eg. painting, plastering, brick-laying, modern farming) Hobbies with moderate activity (eg. gardening, woodwork, dancing) Active sports (eg. tennis, cycling, skiing, gymnastics, swimming) 2 1 0 0 1 0 0 0 Total 24 24 24 24 24 24 24 24 * Adapted from the Dietary Standards for Canada (Canada 1975). 115 3.2.2 Data-Analysis Following collection of dietary data, the client's initial diet is analyzed for nutrient values. Nutrient data used for this analysis and for generating the new diet plan are contained in the system's food-composition file (Appendix C). The food-composition file is stored in the system for retrieval during operation. It is derived from Pennington's Mini Food List (Pennington 1976), and includes 41 composition values in nutrients per 100 grams of edible food portions for each item cluster in the food-item file. The nutrient values contained are for total calories; protein and eleven amino acids (tryptophan, threonine, isoleucine, leucine, lysine, methionine, cystine, phenylalanine, tyrosine, valine, histidine); total, saturated, and polyunsaturated fatty acids; cholesterol; total carbohydrate; sucrose; fiber; twelve vitamins (thiamin, riboflavin, niacin, pyridoxine, folate, cobalamin, ascorbate, pantothenate, biotin, retinol, cholecalciferol, tocopherol); and nineminerals (calcium, phosphorus, magnesium, iron, iodine, zinc, sodium, potassium, copper). It was assumed that the values obtained from Pennington's publication are applicable to the Canadian marketplace. The composition values used are specific for processing and prepara tion effects outlines with the item description. Consequently, recipes using raw items cannot necessarily be calculated from baked items due to changes in weight with cooking. Further Pennington (1976) suggests that: when possible, fresh cooked items should be substituted by fresh cooked; canned by canned; and frozen cooked by frozen cooked. This will prevent large errors in water-soluble vitamins, sodium, and vitamin-E (p. 16). The sodium content of items is without either salt added at the table or 116 salt added in preparation. Therefore, any added salt must be included as salt listed in the food-item file. An abridged food-composition file (Appendix D) of 22 nutrients was developed to coordinate with the abridged food-item list of 127 item clusters contained in Appendix B, and to provide a reduced composition format for system testing. The nutrients selected for the file are as follows: total calories; protein; total, saturated, and polyunsaturated fatty acids; total carbohydrate; sucrose; fiber; nine vitamins (thiamin, riboflavin, niacin, pyridoxine, folate, ascorbate, retinol, cholecal-ciferol, tocopherol); and five minerals (calcium, phosphorus, magnesium, iron, potassium). For the nutrient tally, the aggregate available quantity of any nutrient is assumed to be the sum of the quantities contained in each of the foods consumed. No allowance has been made for factors that reduce the availability of nutrients, either in a particular item or in other foods eaten simultaneously, with the exception of cashews, spinach, and spinach substitutes where oxalate concentration has been considered. Therefore, the total nutrient composition of the diet is the summation across all foods for each nutrient, namely: n (3.1) E a . x? (q = 1, 2, m) i=l qi 1 where: x° . is the amount of item cluster i per period time in the original diet. An item cluster is a single item, or two or more substitutable items of a similar nutrient composition. n is the ;total number of item .clustersMn the diet- =n is 22> (Appendix A) or 127 (Appendix B). 117 a : is the amount of nutrient q in a unit of item cluster i. The values for a . are contained in the m x n matrices of food composi-tion (Appendix C or D). m ' is the total number of nutrients considered in dietary analysis --either 41 (Appendix C) or 22 (Appendix D). 3.2.3 Data-Evaluation In the data-evaluation phase the system evaluates the client's initial diet by comparing calculated nutrient-intake with nutrient limits which are individualized for the client. The evaluation output comprises a graphical and/or tabular display of the quality of the diet. The output format utilized is shown in Table 3.3. This provides a functional tabular output for system testing and development. An example of a pro- . posed output format (Table 3.4), which is modelled on the "Nutrient Quality Index" (Wittwer et^ al_. 1977), incorporates a graphical presentation of the evaluation output. Table 3.3 Excerpt from the evaluation output format used for system testing NUTRIENT MINIMUM MAXIMUM INITIAL DIET REVISED DIET (/week) (/week) (/week) (/week) PROT (gm) 392.00 558.60 818.46 558.60 CHO-T (gm) 2735.4 4476.1 1572.2 2735.4 T-FAT (gm) 44.209 663.13 825.87 566.47 KCAL 18899. 20889. 18366. 19301. CHO-F (gm) 79.576 159.15 21.596 79.576 SFA (gm) .0 221.04 363.74 208.08 SUCR (gm) .0 746.02 396.53 746.02 PUFA (gm) 44.209 663.13 445.79 317.61 VIT-A (iii) 35000. .14000E+06 44903. 72916. VIT-D (iu) 700.00 4200.0 1623.6 1244.7 VIT-E (mg) 63.000 700.00 48.232 63.000 118 Table 3.4 Excerpt from a proposed evaluation output format.* The table below contains an estimate of your average daily nutrient intake over a period of (1 day, 1 week, 1 month or longer), and a comparison of your estimated intake with minimum and maximum intake standards for an individua-1 with the following characteristics: (activity, sex, age, size, pregnancy status). The table also indicates the average daily amount of nutrients provided by a diet recommended for you, if your old diet does not meet the standards. Nutrient Recommended Intake /day Min. Max. Estimated Diet Intake /day Old New Energy (kcal) 2700 2984 Protein (gm) 56 80 Carbohydrate (gm) 391 639 Fiber (gm) 11 23 Fat (gm) 6 95 6tC•••«*•••••• •••• 2624 117 225 3 118 2757 80 391 11 80 Nutrient Energy (kcal) Protein (gm) Carbohydrate (gm) Fiber (gm) Fat (gm) etc Estimated old dietary intake as % of minimum recommended intake 100% xxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxx xxxxxxxxxx xxxxxxxxxxxxxxxxxxxxx Estimated new dietary intake as % of minimum recommended intake 100% xxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxx Nutrient Estimated old dietary intake as % of maximum recommended intake Energy (kcal) xxxxxxxxxxxxxxxxxx^^ Protein (gm) xxxxxxxxxxxxxxxxxxxxx Carbohydrate (gm) xxxxxxxx Fiber (gm) xxxx Fat (gm) xxxxxxxxxxxxxxxxxxxxx etc The following nutrients, although essential for the maintenance of health, have not been included in assessing your diet: (water, chromium, etc....). The following factors are not considered to be nutrients and consequently are not included in assessing your diet: (nucleic acid, popsicle magic  factor, etc ). Estimated new dietary intake as % of maximum recommended intake 100% xxxxxxxxxxxxxxxxxx xxxxxxx xxxx xxxxxx xxxxxxxxxxxx Nutrient values presented in the evaluation output could correspond to the total evaluation file, to selected values as requested by the client or as needed by the counsellor for illustrating diet problems, or to composites of groups of nutrients 119 The nutrient-limits file (Appendix E) provides the basic data, used in both the assessment and planning phases of the program, to define the range of appropriate nutrient intake for a given period of time. This file is stored in the system for retrieval during operation. Nutrient limits for the client are generated from" the values contained in the nutrient-limits file by applying the rules outlined in Appendix E. These rules are used to translate values from the nutrient-limits file to measurement units which are common to those of the food-composition file, and to individualize these for the client based on the client's demographic data collected on the questionnaire (p.TT.3). Nutrient limits expressed as ratios of two nutrients do not require this modification. The nutrient-limits file contains maximum and minimum nutrient limits expressed as nutrient ratios, as quantities per day, and as quantities per kilogram per day. These limits are disaggregated by age, sex, activity level and pregnancy status. Standards from Health and Welfare Canada, and recommendations from other established nutrition sources, have been used to designate minimum nutrient limits. As indicated, these values with their respective sources have been outlined in Appendix E. It should be noted that the philosophy of the "Dietary Standard" has not been represented in the development of every lower intake limit used in the system. This is due, in part, to the inclusion of "nonessential" nutrients in the evaluation phase -- for example sucrose, saturated fatty acids, and fiber. For these components a lower limit of zero was used, unless benefits from some intake were documented. Where dietary standards were not avail able for essential nutrients included in the system, such as polyunsaturated fatty acids, sodium, and potassium, a measure of minimum maintenance was used as outlined in Appendix E. Other suggested indices of dietary well-120 being not specifically included in the Dietary Standard, such as poly unsaturate and saturate ratios, calcium/phosphorus were also included in the nutrient-limits file. Similarly, maximum limits have been empirically based where possible, and otherwise are arbitrarily set at twice;"! the minimum value for present test purposes. Maximum and minimum energy intakes have been arbitrarily set at plus or minus five percent of the intake for any age, sex, and activity category. In total, 82 nutrient limits -- 41 maximum and 41 minimum -- have been assigned. These values correspond to the 41 food compositional values in Appendix C.with a few exceptions. First, in addition to specific limits on calcium, phosphorus, polyunsaturated and saturated fatty acid intake, restrictions have been included on the ratios of polyunsaturated to saturated fatty acids and of calcium to phosphorus.. Second, instead of maximum and minimum limits for each of the eleven amino acids, only 18 limits have been imposed to cover 7 single amino acids and 2 pairs of amino acids -- phenylalanine and tyrosine; methionine and cystine. An abridged nutrient-limits file (Appendix F) has been designed for system testing. This table has 24 minimum and 24 maximum standards to coordinate with the abridged food-composition file (Appendix D) and food-item list (Appendix B). The nutrients considered are: total calories; protein; total, saturated, and polyunsaturated fatty acids; polyunsaturate to saturate ratio; total carbohydrate; sucrose, fiber; nine vitamins (thiamin, riboflavin, niacin, pryidoxine, folate, ascorbate, retinol, cholcalciferol, tocopherol); five minerals (calcium, phosphorus, magnesium, iron, potassium); and calcium to phosphorus ratio. 121 3.2.4 Diet-Planning The system described thus far, has analyzed the nutrient composition of the client's diet and evaluated the client's diet by comparing the diet to nutrient limits. If the client's initial diet does not meet these limits, then the system generates a revised diet for the client. This computer prints out this revised diet for comparison with the initial diet. Table 3.5 illustrates the numerical listing of the item clusters with quantities in grams per week for the initial and revised diets. Table3.5 Excerpt from the diet-planning output format .used for system testing. Item Cluster Initial Diet Revised Diet Code Number (100 gram/week) (100 grams/week) 003 2.2400 3.9887 004 0.0 0.0 005 0.0 0.0 006 0.0 0.0 008 1.1000 0.0 010 0.0 3.5409 012 0.0 1:5189 013 7.2000 4.4315 015 3.6000 0.0 016 0.0 0.0 017 0.0 0.0 018 0.0 0.4904 019 3.0000 3.0426 020 0.0 0.0 021 0.0 0.0 023 0.6900 0.0 024 1.3800 3.0740 027 0.0 0.0 028 1.3800 0.6316 031 0.0 0.0 The revised diet is developed by a constrained-optimization formulated to find the combination of item clusters that minimizes the sum of the squared differences between the amount of specific item clusters and of 122 g attribute groups in the initial and revised diet, while satisfying nutrient constraints. The formal mathematical statement of the model is: I 2_. _t_ K. j i. 2 (3.2) minimize I w.(x. - x?) t E P.. (E (x- - x?)) i=l 1 1 1 k=lj=l Jk UGjk 1 1 I (3.3) subject to m >, E a . x- > nn (q = 1, 2, ..., Q) q ^ _i q1 1 q I E aui xi (3.4) r ^ ifl > t (u,v = any specified set I of nutrient pairs) 1-! ^ ^ (3.5) Xi > 0 (i = 1, 2, I) where: x^ is the amount, in grams, of item cluster i per time period in the revised and independent d diet. The values of x^ are assumed to be additive x*? is the amount, in grams, of item cluster i per time period, in the client's initial, desired, or nutritionist-prescribed diet, as determined from the client questionnaire. I is the total number of item clusters in the diet. The I vector equals 221 in the complete food-item file (Appendix A), or 127 when the abridged food-item file (Appendix B) is considered. 5. As indicated, an item cluster is one item, or two or more food items, varieties of an item, or prepared variations of an item which are considered nutritionally equivalent. 6. An attribute group is a group of item clusters with similar charac teristics or attributes (see Appendix G). 123 K is the total number of hierarchical levels in the model, namely seven. Each hierarchical level corresponds roughly to a criterion for classifying item clusters into attribute groups on the basis of item similarities. The designated attribute groups form a hierarchical pyramid of groups and subgroups. The matrix of the hierarchical levels and attribute-groups is given in Appendix G. The matrix developed is based, in part, on the examples of food classification schemes by other authors (FAO/WHO 1949; Davenport 1964; Gue and Liggett 1966; Chandler and Perloff 1975; Canada 1977). It represents an attempt to direct appropriate food substitutions among items of the food list by, first, defining general food and diet attributes that may important for indicating an item's or a diet's similarity to another item or diet. Second, foods are classified into groups according to these attributes. By attempting to provide the nearest acceptable substitute for any change in the quantity of an item or class, the compatibility relationships among items and the general characteristics of the altered diet should be most successfully maintained. For further discussion refer to Section 4.2 (p.136). J is the total number of attribute groups j in the diet. J equals 278 in the large model and 178 in the abridged model, is the number of attribute groups j in the k hierarchical level. In the large model J-j = 67, ^ = 55, = 52, = 38, = 35, Jg = 27, and = 4. In the abridged model matrix J-, = 50, J9 = 39, = 34, J. = 23, Jr = 20, Jfi =12, and J7 = 4. 124 Gj^ is the set of item clusters i within the j attribute group of the level. The total amount of food within any attribute group is the sum of the quantities of item clusters contained in that group. The item quantities are assumed to be additive and independent; that is, it is assumed that the common characteristic of items within any group can be measured by the aggregated quantity of the contained item clusters, and that an item cluster can hold simultaneous membership in any number of groups, w- is the weighted penalty associated with deviations of item cluster i from the desired amount. Values of w^ have been assumed in the absence of empirical data. These values are discussed in Chapter 4 (p.145). Pj^ is the weighted penalty associated with deviations of the j^*1 attribute group in the k level from the desired amount. Values of have been assumed in the absence of empirical data. These values are discussed in Chapter 4 (p.182)-a • is the amount of nutrient q in a unit portion of item cluster i. The values of a^ are obtained from the I x Q matrix of food composition (Appendix C or D) which contain nutrient coefficients for edible 100 gram portions of foods. The aggregate available quantity of any nutrient is assumed to be the sum of the quantities contained in each of the foods consumed. The elements of the food composition table are assumed to be additive and independent, and to be constants. Allowance has not been made for factors that reduce the availa bility of nutrients in either a particular item or in other 125 foods eaten simultaneously, with the exception of cashews, spinach, and spinach substitutes where the effect of oxalate concentration has been considered. Q is the total number of nutrients q considered in the diet. For the large model, Q is 41; for the abridged model, Q is 22. The nutritional adequacy of solutions for mathematical diet-models can only be assumed for those nutrients specifically included in the model. For this reason, the large-model formu lation incorporated as many nutrient values as was reasonably possible. m and n are Q vectors of respectively, the maximum and minimum amount of nutrient q allowed in the individual's diet over a defined time period according to the client's age, sex, size, activity and pregnancy status. Values for m^ and-TV are derived from the nutrient-limits file (Appendix E or F). aul- and av1- are the amounts, respectively, of specified nutrients u and v in a unit of item cluster i. u corresponds to either calcium or polyunsaturated fatty acids, v corresponds to either phosphorus or saturated fatty acids. Values of aul- and a^ are contained in Appendix C and D. ruvand t are, respectively, the maximum and minimum allowed ratio for the nutrient pair (u,v). The nutrient pairs considered are calcium and phosphorus, and saturated and polyunsaturated fatty acids. Values of ruv and t are derived from Appendix C or D. 126 The objective function (Eqn. 3.2) defines a list of item clusters by minimizing the aggregate squared difference between specified characteris tics^ the initial and revised diets. The first term of the objective function sums the weighted squares of the difference between the amounts of item clusters in the initial and revised diets. The second term sums The weighted squares of the difference between the initial and revised amounts of a hierarchical sequence of item cluster groups, called attri bute groups. The quadratic term introduces disproportionately larger penalties as the revised diet deviates more widely from the initial diet. This characteristic of the objective function tends to spread deviations uniformly over all item clusters and all attribute' groups. The modelling constraints comprise: nutrient constraints (Eqns. 3.3, 3.4) which designate maximum and minimum limits for the nutrients provided by the diet, and a non-negativity constraint (Eqn. 3.5) which prevents entry of non-negative quantities of variables in the solution. With respect to the nutrient constraints, only those which can be stated either as a fixed amount (Eqn. 3.3), or as linear ratios of two nutrients (Eqn. 3.4) are incorporated in the model. As written the second constraint (Eqn. 3.4) is a nonlinear equation. However, it is easily transformed into linear form. Apart from Equation 3.4, nonlinear constraints were not considered in the model. The model was formulated into matrices suitable for input into a preprogrammed quadratic package based on Lemke's Complementary Slackness Algorithm (Cottle and Dantzig 1968; Lemke 1968). The matrix algebra formulation can be briefly outlines as follows: Minimize: c' x + % x1 D x Subject to: A x >, b Where: x ^ o is the vector of foods. 127 c' is a vector containing the coefficients for linear terms in the objective function. D is a matrix containing the coefficients of quadratic terms in the objective function. A is the food composition matrix, b is the vector of nutrient constraints. 128 CHAPTER 4 TESTING OF THE PROTOTYPICAL SYSTEM 4.1 Introduction The second project objective (p.14) involved testing of the diet-planning phase of the system, since this phase was considered to be the most signi ficant obstacle to overall system development. This testing was restricted to a descriptive evaluation of some of the objective function's character istics. Specifically, assumptions defining the concept of minimum deviation between diets, which are implicit in the objective function, were arti culated and then explored by altering some of these assumptions and observing the consequences for revised diets developed for hypothetical individuals. The impact of altering nutrient constraints was not considered. This evaluation was undertaken to explore the conceptual and technical feasibility of using a mathematical model to provide an effective dietary recommendation -- the individual's nutrient-constrained food-choide -- from a dietary inventory identified as desirable for the individual by his or her past consumption or stated preference. Also, it provides a basis for more definitive evaluation and development of the model. Explorative evaluation does not constitute the means to validate^ the diet-planning model's design. In order to reduce computational costs and clerical work required for evaluation of the diet-planning phase, all testing was done using the abridged data bases outlined in; Appendices B (abridged food-item file), 7. Validation tests that the model is a reasonable representation of reality. 129 D, (abridged food-composition file), F (abridged nutrient-limits file), and G (attribute-group matrix). The nutrient constraints for test runs were determined from Appendix F for a standard male subject of age 19 to 35, activity level code B, and 70 kilograms body weight. The nutrient assigned for this standard subject for a weekly period are outlined below in Table 4.1. The daily intake equivalent is also provided for comparison. Table 4.1 Upper and lower nutrient constraints* for a standard subject4: expressed as weekly amounts (and daily equivalents) for 24 nutrients. Nutr Lower Upper Lower Upper Nutr Lower Upper Lower Upper ient Limit Limit Limit Limit ient Limit Limit Limit Limit (/wk) (/wk) (/day) (/day) (/wk) (/wk) (/day) (/day) ENERGY 18894 20889 2699 2984 PYR 14000 28000 2000 4000 (kcal) (ug) PROT 392.0 558.6 56.0 79.8 F0L 1400 2800 200 400 (gm) (ug) FAT-T 44.21 663.13 6.32 94.73 VIT-C 210 3500 30 500 (gm) (mg) SFA 0.00 221.04 0.00 31.58 VIT-A 35000 140000 5000 20000 (gm) (1u) PUFA 44.21 663.13 6.32 94.73 VIT-D 700 4200 100 600 (gm) (iu) P/S 1 2 1 2 VIT-E 63 700 9 100 (mg) CHO-T 2735.4 4476.1 390.8 639.4 CAL 5600 11200 800 1600 (gm) (gm) SUCR 0.0 746.02 0.0 106.57 PH0SP 5600 11200 800 1600 (gm) (mg) CHO-F 79.58 159.15 11.37 22.74 CA/P 0.8 1.2 0.8 1.2 (gm) THIA 9.94 19.89 1.42 2.84 MAG 2205 4410 315 630 (mg) (mg) NIAC 131.30 262.60 18.76 37.51 IRON 70 140 10 20 (mg) (mg) RIBO 11.94 23.87 1.71 3.41 POT 9800 19600 1400 2800 (mg) (mg* Values for these constraints are derived from Appendix F. + Standard subject was age 19 to 35, activity level code B, and 70 kg. body weight. 130 Two seven-day food-intake records were defined to represent the initial diet obtained from the standard male subject (Table 4.2). One record, called Standard Initial Diet 1 (SID-1), contained 3 items; the other, called Standard Initial Diet 2 (SID-2), contained 83 items. Table 4.3 compares the amounts of nutrients supplied by each of these diets with the upper and lower nutrient constraints. 131 Table 4.2 Standard Initial Diet 1 (SID-1) atid'.Standard Initial Diet 2 (SID-2): Two seven-day food-intake records for a standard male subject, age 19 to 35, activity level code B, and 70 kilograms body weight. GROUP-ITEM FOOD ITEM SID-1 SID-2 CODE # (grams/week) (grams/week) ENTREE-DAIRY 001-003 CHEDDER CHEESE... 182 001-004 COTTAGE CHEESE... 001-005 CREAM CHEESE 002-006 SOUR CREAM 002-008 YOGHURT 003-010 EGG 495 ENTREE-CEREALS 004-012 CORN CEREAL 28 004-013 WHEAT CEREAL 005-015 OATMEAL 270 005-016 WHEAT CEREAL 006-017 PANCAKES 135 007-018 NOODLES 007-019 SPAGHETTI 008-020 RICE, brown 9375 008-021 RICE, white 008-023 WHEAT GERM 280 009-024 FRENCH BREAD 40 009-027 WHITE BREAD 414 009-028 WHOLE WHEAT BREAD 010-031 BISCUITS 70 010-032 HAMBURGER BUN 46 010-033 MUFFIN 40 010-035 ENGLISH MUFFIN... 46 011-037 SALTINES 24 011-040 RYE KRISP ENTREE-MEATS 012-041 BEEF, 30% fat.... 170 012-042 BEEF, 20% fat.... 85 012-043 BEEF, 15% fat.... 170 012-046 PORK, lean cuts.. 85 012-047 PORK, all hams... 43 012-048 BACON 64 013-049 CHICKEN, steamed. 013-051 CHICKEN, fried... 340 JJ 132 Table 4.2 (Continued) 014-055 FRIED FISH 1 014-056 BROILED FISH • 014-058 OYSTERS 01.4-062 SARDINES 014-063 SHRIMP. .• 43 014-065 TUNA 60 015-067 LIVER 017-069 FRANKFURTERS 017-070 FRESH SAUSAGES 40 017-071 LIVERWURST 018-073 BEANS, dried 018-075 SOYBEANS 019-076 ALMONDS 019-077 CASHEW NUTS 15 019-079 PEANUT BUTTER "28 019-080 PEANUTS 120 019-081 PECANS ENTREE-VEGETABLES 020-082 POTATOE, baked.... 020-083 POTATOE, fried.... 020-084 POTATOE, mashed... 020-085 SWEET POTATOE 100 100 200 180 021-089 BEANS, green 65 021-091 BROCOLLI 63 021-092 CABBAGE 65 021-095 LETTUCE 180 021-098 PEAS 170 021-099 PEPPERS 021-101 SPINACH 85 022-103 BEETS 80 022-104 CARROTS, cooked... 1 022-105 CARROTS, raw 76 50 022-106 CORN 022-111 TOMATOE 518 023-113 CUCUMBER 275 023-114 MUSHROOMS 023-115 ONIONS 8 024-116 SUCCOTASH 025-117 OLIVES 025-118 PICKLES, sweet 025- 119 PICKLES, sour ENTREE-FATS 026- 121 LARD 20 34 5 026-124 SOYBEAN OIL 027-125 BUTTER 125 Table 4.2 (Continued) 028-127 CHEESE SAUCE 028-128 GRAVY 029-132 MAYONNAISE 105 029-133 SALAD DRESSING.... 90 030-136 BEVERAGES-DAIRY WHOLE MILK 732 030-137 SKIM MILK 5658 031-140 TABLE CREAM... 195 031-141 WHIPPED CREAM 8 032-143 BEVERAGES-FRUIT APPLE JUICE 032-144 GRAPEFRUIT JUICE.. 120 032-145 LEMON JUICE 5 032-148 ORANGE JUICE 360 033-150 TOMATOE JUICE 034-151 BEVERAGES-MISC. COLA-TYPE 339 035-153 COFFEE 2800 035-154 TEA 1600 036-155 BEER 3240 036-156 DISTILLED SPIRITS. 43 036-158 DRY WINES 400 037-160 SOUPS CREAMED SOUPS 198 037-161 PEA SOUPS 200 037-163 MEAT + VEGIE SOUPS 038-166 DESSERTS-CEREALS COFFEE CAKE 75 038-168 FRUITCAKE 038-170 ICED CAKES 60 039-172 FRUIT PIES 160 039-172 PUMPKIN PIES 150 040-173 COOKIES 60 040-174 FRUIT COOKIES 041-175 CAKE DOUGHNUTS.... 60 041-177 DANISH PASTRY 38 042-178 DESSERTS-DAIRY ICE-CREAM 270 042-179 SHERBERT 043-181 PUDDINGS 185 1,34 Table 4.2 (Continued)'' /:.: DESSERTS-FRUIT 044-183 APPLE 044-184 APPLESAUCE 044-185 BANANA 044-186 CANTALOUPE 044-187 GRAPEFRUIT 044-188 ORANGE 044-192 PINEAPPLE 045-193 DRIED FRUIT DESSERTS-SWEETS 047-195 HONEY 047-197 SUGAR 048-198 JAMS. 048-199 SYRUP 049-201 CHOCOLATE CANDY... 049-202 MARSHMALLOW MISCELLANEOUS 051-205 POT PIES 063-217 SPAGHETTI + MEAT.. 067-221 COCOA MIX Total Grams/Week.. Total # of Items.. 300 185 100 100 200 150 lUb 100 40 227 330 1 15313 18873 83 135 Table 4.3 Nutrient composition of SID-1 and SID-2 and the upper and lower nutrient constraints for a standard male subject. Nutrient Nutrient Constraints SID-1 SID-2 Minimum Maximum Initial Initial (/week) (/week) (/week) , (/week) PROTEIN (gm) 392.00 713.82 512.54 558.60 CHO-T (gm) 2735.4 1872.4 2809.9 4476.1 T-FAT (gm.) 44.209 981.39 92.428 663.13 KCAL 188899. 20195. 14210. 20889. CHO-F (gm) 79.576 32.250 35.125 159.15 SFA (gm) .0 367.46 5.6000 221.04 SUCR (gm) .0 549.60 30.925 746.02 PUFA (gm) 44.209 543.42 22.400 663.13 VIT-A (iu) 35000. 69952. 0.0 .140E+06 VIT-D (iu) 700.00 1111.3 2319.8 4200.0 VIT-E (mg) 63.000 75.273 56.550 700.00 VIT-C (mg) 9.9470 10.292 16.329 19.894 RIBO (mg) 11.936 13.803 13.963 23.873 NIAC (mg) 131.30 178.25 148.67 262.60 VIT-B6 (ug) 15000. 12942. 20777. 28000L FOLATE (mg) 1400.0 1460.9 2019.5 2800.0 POTAS (mg) 9800.0 22490. 17082. 19600. CAL (mg) 5600.0 6137.3 8172.8 11200. PHOSP (mg) 5600.0 11258. 15349. 11200. IRON (mg) 70.000 120.08 73.195 140.00 MAG (mg) 2205.0 2386.5 4471.8 4410.0 CA/P .8 .54514 .53245 1.2 P/S 1.0 1.4789 4.0000 2.0 136 4,2 Diet-Planning Model's Premises and Assumptions A satisfactory procedure for dietary evaluation and suitable data on food composition and nutrient requirements were available for developing constraints to define dietary adequacy. However, no obvious procedure was available to define the relative acceptability of modifications in an initial dietary inventory. As previously mentioned (p.80), an extensive tradition does exist for hospital menu planning and Food Guide generated diets, but precise guidelines are not available for systematically devel oping individualized dietary prescriptions which are both nutritionally adequate and maximally acceptable to the client. Presumably the nature of this process is not fully understood. In the absence of a detailed precedent which defines the relationship between an individual's nutrient-constrained food-choice and his or her initial dietary inventory, two major premises and related assumptions were used to define and operationalize the objective function. 4.2.1 The First Premise and Related Assumptions The first major premise, for developing the model's objective function is that the client's most acceptable diet can be defined either by past consumption patterns -- namely, previous choice tends to define future choice -- or by a food plan identified as desirable by the client. In practice, past consumption may provide the better monitor of actual preference since it is the diet chosen under presently operating budgetary restrictions and long-term cultural and personal habits. It is assumed that an individual identifies a diet's character by a complex of attributes. These attributes provide cues to distinguish one dietary pattern from another and an index to estimate the relative 137 significance of dietary change. In this context, attributes are defined as measureable aspects of food or of behavior applied to foods, such as food taste, colour, ethnocultural pattern of food intake, nutrient content, menu function, food group, cost, serving frequency, preparation procedure, and a variety of food item relationships. These attributes partition or categorize food substance into an array of parts and subparts; for example, into such typical categories as ingredients, food items, food groups, menu items, as well as less formal or personal dietary distinctions such as an individual's preferred items. Thus for each individual, an attribute map of considerable complex ity can be envisioned. Depending on the attributes utilized by the indivi dual, the map may correspond, more or less, to ethno-cultural patterns, familial patterns, and so on. Further, it should be noted that depending on the attributes considered, two diets may differ by one set of criteria but be alike by another. The attribute map developed for the prototypical model is based, in part, on the examples of other authors (FAO/WHO 1949; Davenport 1964; Gue and Liggett 1966; Chandler and Perloff 1975; Canada 1977) who developed food classification schemes for various purposes. Only a limited number of illustrative attributes is considered for the prototypical model --attributes presumably used by the average Canadian for identifying the similarities between food substances. This provisional assignment of attributes includes a definition of a standard item. These are mono-ingredient menu items which are low in complementarity requirements --that is, items which do not require the coexistence of other items to ensure their palatability and acceptability -- and which can be assigned nutrient values from available literature. The other attributes 13a considered are: -generic terms of common usage such as cheeses, poultry, bread, and cultured milks; -meal classification such as breakfast or lunch; food group assignments such as dairy products or fruit products; and -some general physical characteristics including taste, colour, and physical state. On this basis dietary items were partitioned into a hierarchical, attribute-group matrix based on the apparent similarities between defined items. The matrix elements created are discussed in Section 3.2.4 and described in Appendix G. Attributes not explicitly considered in developing the attribute map, such as pricing, are considered independent parameters for present purposes, even though their influence could affect the diets acceptability. It is assumed that under conditions of change the map elements main tain their integrity. However, in some extreme situations, such as where radical dietary alterations are required, new attribute specifications may become apparent as food substance shifts from initial levels. A vast number of attribute maps can be defined, each of which has a particular item list with its inherent interactional properties. The attribute map chosen, and the items and item groups thereby defined, are central to the eventual utility of such a model. The items and item groups selected will effect the accuracy and simplicity of the diet assessment procedure, the acceptability of the diet produced by the planning phase, and the outcome of communicating dietary modifications. For example, using a commodity list of palatable items for diet planning largely guarantees palatibility of the diet developed. 139 4.2.2 The Second Premise and Related Assumptions The second major premise is that acceptable dietary modifications --namely, the best approximation for compliance to nutritional prescriptions -- can be provided by determining the least-altered diet which meets nutrient constraints. The least-altered diet is presumably the best alternative, as perceived by the individual, to his or her presently chosen, or preferred diet. Thus, the mathematical equation for determining the best alternative diet should reflect this concept. The individual is assumed to perceive dietary change as a shift of food substance in his or her personal attribute map. The acceptability or perceived extent of this change then depends on the significance of changing each element of the map from its initial condition with respect to itself or to other elements. For the diet-planning model,.the acceptability of this change is assumed to decrease as the square of the deviation from initial amounts for any element. Where unbounded, this produces a symmetrical quadratic curve centered around the initial amount of the element. This deviation is weighted by a penalty coefficient assigned to represent the relative significance of each attribute element deviating from initial amounts. The value assigned for this coefficient can be adjusted for a variety of hypotheses including client acceptability or preference, initial consumption levels, average serving size of items, or some other provisional pattern such as Canadian average consumption or nutritionist-recommended consumption levels. In the diet-planning model penalties are assigned to reflect initial consumption levels, consumed versus non-consumed status, and for hierarchical membership. Thus, the acceptability of a diet deviating from initial levels is defined as the summation of weighted squares of the difference between the amounts of the attribute elements in the initial and revised diets. 140 4.3 Observations of the Objective Function's Characteristics 4.3.1 Unconstrained Objective Function The unconstrained objective function of the mathematical programming model (viz. no constraints on the nutrients) corresponds to the solution of a problem with an input diet which already satisfies nutrient constraints in all respects. The unconstrained solution allows the sum of weighted squared differences between items and item groups in the initial and revised diets to go to zero. This should result in a recommended diet identical to the initial diet -- the presumed ideal. To verify this, a number of solutions for initial diets were obtained using the diet-planning model. These solutions (Table 4.4, 4.13) were developed from the SID-2 as the input diet and with a variety of objec tive functions (Eqns. 4.8, 4.9, 4.28) which incorporated different penalty coefficients, as described later in this text. The solutions for these diets were rerun using the same programs that generated them. As expected, in each case the original solution was identical to the rerun solution, allowing for rounding errors. These observations illustrate, first, that each solution provided by the model is an optimal one, for the given input diet and nutrient con straints. Second, providing the diet satisfied the nutrient constraints, altering the penalty coefficients in the objective function for positive values of w- and P.. will not affect the outcome. The same phenomena would I J K be expected with other objective functions incorporating the same theme of minimizing the deviation from the original diet, such as a linear objective function. 141 4.3.2 Constrained Objective Function If the most acceptable diet for the client occurs when the sum of differences is allowed to go to zero, as with the unconstrained solutions, less acceptable diets would be those where, because of nutrient constraints, minimization towards zero is less successful. Presumably the larger the deviation the less acceptable the recommended diet will be for the client. The nature of this deviation and perhaps the relative acceptability of the outcome can be altered by modifying the characteristics of the objective function. The following material explores alterations of the objective function which produce different nutrient-constrained output profiles. 4.3.2.1 First Term; Shape of the Curve The mathematical expression used in the model's objective function to minimize deviations of any item cluster from initial dietary levels, irrespective of specific concurrent deviations in other items, is the summed weighted square of differences between the amounts of item clusters in the initial and recommended diet, as follows: I 2 (4.1) minimize .E w. (x- - x.) i=l 1 1 1 This expression, corresponding to the first term of the objective function (Eqn. 3.2), describes a symmetrical quadratic curve centered at the initial consumption level, as shown in figure 4.1. As the difference between the initial and recommended values increases the quadratic objective intensifies the presumed unacceptability of this divergence. The choice of a quadratic curve is arbitrary, but it is the simplest nonlinear function which captures the essence of this phenomenon. 142 Figure 4.1 Graph of quadratic function reflecting penalties for deviation from initial consumption of food i. A simpler alternative, but perhaps less realistic formulation, uses an objective that minimizes the absolute linear difference between the initial and recommended diet, as follows: (4.2) minimize E w. (|x• - x? i=l 1 1 1 In order to program this linear objective function (Eqn. 4.2), the formulation of the algorithm previously described (Eons. 3.2-3.5) can be rewritten in a piece-wise linear form, as follows: (4.3) subject to 1 , + + -minimize £ (w- x- + w- x-) i=l 11 11 (4.4) (4.5) (4.6) + - o xi " xi = xi " xi (i = 1, 2, mq * aqi Xi *',nq (q = ls 2' i=l I) Q) I E aui xi r,,w ^ i=l ^ t (u,v = any specified set uv I uv of nutrient pairs) E' a . x. 1=1 V1 1 143 (4'7) ><i' xt, XT » o (i = 1, 2, ..., I) Where the newly defined terms are as follows: xj, xj are,.respectively, the positive and negative deviation of x^ from x°. + ( (x, - x°) if Xi - x° > 0 xi = ( ( 0 otherwise ( - (Xi - x?) if x. - x° < 0 xi = ( ( 0 otherwise Thatiis, for each i at most one of x| and x!j can be positive. + w^ are the weighted penalty associated with the positive or negative deviation, respectively, of item cluster i from the original amount. In more detail, the linear objective minimized the weighted positive and weighted negative deviation of item clusters from initial levels. Unlike the quadratic formulation the penalty associated with this deviation varies in direct proportion to the difference. The modelling constraints include, in addition to the nutrient constraints previously discussed, an equality constraint (Eqn. 4.4) which defines the positive or negative deviation of the revised diet to the initial diet, and a non-negativity constraint (Eqn. 4.7) which restricts entry of non-negative quantities of variables in the solution. Although linearized solutions were not developed, with which the quadratic results could be directly compared, the two models would be expected to generate different solutions. Compared to the linear version the quadratic term introduces larger penalties as the prescribed diet 144 deviates more widely from the initial diet, but smaller penalties for very small changes, as illustrated in Figure 4,2. Thus, the quadratic objective should moderate extreme fluctuations in any particular item by suppressing single large deviations between initial and revised values in favour of numerous smaller changes. The linear model should be less sensitive to large changes in specific items. Consequently, one can speculate that the linear model may provide dietary solutions with undesirably large changes in some single items, unless these changes are restricted by specific constraints on the amounts of these items. Correspondingly while the quadratic model moderates these extreme fluctuations, its solutions are more likely to include a large number of small decrements or increments. These small deviations may be considered as cosmetic problems which can be remedied by rounding values to the nearest usable portion, but con sequently losing some accuracy of the minimization in the process. Alter natively, either constraints on minimum entry levels to prevent unusably small increments from zero, or integer programming, could be used to over come this difficulty with more accuracy. However, the possibility of many items entering at minimum levels still exists unless specific restrictions are applied to the number of items entering the solution set. Figure 4.2 Graph of quadratic and linear functions reflecting penalties for deviation from initial consumption of food i. 145 4.3.2.2 First Term: Penalty Coefficient, The relative acceptability of change for different items should also be considered, in that alterations in the quantities of different dietary items may not be equally acceptable to the client. Consequently, in developing a recommended diet some items should be preferentially main tained at the initial amount whereas other items may be favourably altered to a greater or lesser extent. By altering the penalty-coefficient terms, w^, of the equation which affect the relative slope of the exponential minimization curve, the value of increments for an item can be exaggerated or depressed. The numerical value assigned to the penalty coefficient can correspond to one of, or the product of, a number of componenti:values used to represent perceived acceptability, (i) The weighting coefficient can be normalized for initial comsumption level, that is, the inverse of the client's initial consumption of an item as determined from questionnaire data, as fol1ows: wn. <* l/(x? + e) ' Here a small increment, e > 0, has been included in the denominator so that percentage change can still be represented when x° = 0. Values of e = 1 and e = .01 were chosen for test purposes, and are not intended to represent established values. Although the value chosen for e is in some sense arbitrary, the values should be small enough so that solutions will not be appreciably altered at typical portion-sized values of x°. At zero values of x°, the small epsilon value of e = .01 weights heavily against the entry of the item into the diet since percentage increase of x° would necessarily be large. Alternatively, initial consumption could be given a 146 lesser weighting by using for example the inverse of the square root of initial consumption, as follows: Thus, as consumption level increases the weighting penalty does not decrease proportionately. (ii) Deviation could be normalized for standard serving size by weighting deviation with the inverse of serving size, as follows: Standard serving sizes are specified in Appendix A. (iii) Further information which may be used in formulating a weighting coefficient to represent client willingness to deviate from original consumption includes: data on Canadian average consumption for normalizing deviations to Canadian norms; weighting coefficients obtained from nutritionists to massage solutions towards specific ends; or some other measure which may possibly represent an individual's acceptability dynamics, such as penalties to inhibit the entry of initially-zero item quantities. Two of the possible variations of the weighting coefficient, mentioned above, have been explicitly considered in the following material. These involve weighting deviation on the basis of initial consumption levels, and preferential weighting of consumed versus non-consumed items in the initial inventory. 4.3.2.2.1 Penalty Coefficient, w^ Based on Amount Consumed The objective function already considered (Eqn. 4.1) can be defined to weight the squared deviation with a coefficient of 1 (Eqn. 4.9). A second approach is to weight deviations on the basis of initial consumption wi oc 1/grams per serving of items cluster i 147 level, or to produce percentage squared deviations (Eqn. 4.9). The cor responding summed expressions for the objective functions are, as follows: I 2 (4.8) minimize z 1 (xi - x°) i=l 1 1 1 1 2 (4.9) minimize z — T (x. - x?) 1=1 (x°+D 1 1 The value e = 1 has been included in the denominator so that percentage change can be represented when x-j = 0, as previously discussed (p.145). Using the above objectives (Eqns. 4^8, 4.9), revised diets were developed from the two standard initial diets, SID-1 and SID-2, as shown in Table 4.4. The corresponding nutrient compositions are tabulated in Table 4.5. 148 Table 4.4 Revised diets developed from SID-1 and SID-2 using quadratic objective functions (Eqns. 4.8, 4.9) GROUP-ITEM CODE # FOOD ITEM SID-1 (grams/week)* Initial Eqn.4.8 Eqn.4.9 SID-2 (grams/week) Initial. Eqn.4.8 Eqn.4.9 ENTREE-DAIRY 001-003 CHEDDER CHEESE. 354 182 519 618 001-004 COTTAGE CHEESE. 001-005 CREAM CHEESE... 002-006 SOUR CREAM .13 002-008 YOGHURT 003-010 EGG 495 117 ENTREE-CEREALS 004-012 CORN CEREAL 203 203 28 210 108 004-013 WHEAT CEREAL... 50 221 88 005-015 OATMEAL 270 187 207 005-016 WHEAT CEREAL... 14 24 21 006-017 PANCAKES 135 47 007-018 NOODLES 007-019 SPAGHETTI. 008-020 RICE, brown.... 9375 7773 3407 75 . 29 008-021 RICE, white 25 39 11 008-023 WHEAT GERM 280 009-024 FRENCH BREAD... 49 40 63 62 009-027 WHITE BREAD.... L 82 414 451 752 009-028 WHOLE WHEAT BREAD 182 138 66 010-031 BISCUITS 70 2 010-032 HAMBURGER BUN.. 82 46 83 73 010-033 MUFFIN 40 28 010-035 ENGLISH MUFFIN. 182 46 184 186 011-037 SALTINES 34 24 6 54 011-040 RYE KRISP 50 221 88 ENTREE-MEATS 012-041 BEEF, 30% fat.. 170 012-042 BEEF, 20% fat.. 85 012-043 BEEF, 15% fat.. 170 012-046 PORK, lean cuts 85 012-047 PORK, all hams. 43 012-048 BACON 64 013-049 CHICKEN, steame d 013-051 CHICKEN, fried. 340 014-055 FRIED FISH 014-056 BROILED FISH... Values rounded to the nearest grams/week, except values below 0.5 grams/ week which are rounded to three decimal places. .149 Table 4.4 . (Continued) 014-058 OYSTERS .1 014-062 SARDINES 014-063 SHRIMP 43 014-065 TUNA'. 60 22 37 015-067 LIVER 017-069 FRANKFURTERS 017-070 FRESH SAUSAGES.... 017-071 LIVERWURST 40 98 61 018-073 BEANS, dried 126 25 018-075 SOYBEANS 33 019-076 ALMONDS 73 019-077 CASHEW NUTS 15 15 019-079 PEANUT BUTTER 88 28 019-080 PEANUTS 79 120 10 43 019-081 PECANS 85 112 ENTREE-VEGETABLES 020-082 POTATOE, baked.... 020-083 POTATOE, fried.... 020-084 POTATOE, mashed... 020-085 SWEET POTATOE 100 100 29 200 82 125 152 180 218 443 021-089 BEANS, green 324 235 65 253 199 021-091 BROCOLLI. 438 352 63 363 277 021-092 CABBAGE 199 178 65 180 175 021-095 LETTUCE 83 180 178 271 021-098 PEAS 355 170 572 924 021-099 PEPPERS 599 304 374 95 021-101 SPINACH 663 231 85 182 214 022-103 BEETS 163 80 159 178 022-104 CARROTS, cooked... 022-105 CARROTS, raw 20 205 76 146 166 181 50 70 113 022-106 CORN 25 109 49 022-111 TOMATOE 59 518 448 577 023-113 CUCUMBER 39 275 210 230 023-114 MUSHROOMS 37 1 023-115 ONIONS 117 8 119 48 024-116 SUCCOTASH 210 158 52 025-117 OLIVES 1443 374 380 84 025- 118 PICKLES, sweet.... 025.119 PICKLES, sour ENTREE-FATS 026- 121 LARD 93 20 20 82 34 30 152 70 5 026-124 SOYBEAN OIL 67 027-125 BUTTER. 104 52 125 22 T 028-127 CHEESE SAUCE 59 .003 028-128 GRAVY 72 28 150 Table 4.4 029-132 029-133 030-135 030-137 031-031-140 141 032-143 032-144 032-145 032- 148 033- 150 034- 151 035- 153 035- 154 036- 155 036-156 036-158 037-160 037-161 037-163 038-166 038-168 038- 170 039- 171 039-172 040-040-041-041-173 174 175 177 042-178 042- 179 043- 181 044-183 044-184 (Continued) MAYONNAISE SALAD DRESSING... BEVERAGES-DAIRY WHOLE MILK SKIM MILK TABLE CREAM WHIPPED CREAM BEVERAGES-FRUIT APPLE JUICE GRAPEFRUIT JUICE. LEMON JUICE ORANGE JUICE TOMATOE JUICE BEVERAGES-MISC. COLA-TYPE. COFFEE TEA BEER. DISTILLED SPIRITS DRY WINES SOUPS CREAMED SOUPS..._ PEA SOUPS MEAT + VEGIE SOUP DESSERTS-CEREALS COFFEE CAKE FRUITCAKE ICED CAKES FRUIT PIES PUMPKIN PIES COOKIES FRUIT COOKIES.... CAKE DOUGHNUTS... DANISH PASTRY.... DESSERTS-DAIRY ICE-CREAM SHERBERT PUDDINGS DESSERTS-FRUIT APPLE APPLESAUCE 27'. 105 21 44 24 49 90 70 74 732 632 5658 4197 1977 13 195 93 60 23 8 6 120 5 360 17:4 44 - 339 328 332 2800 2771 1600 1572 306 3240 3229 4588 39 43 34 41 400 324 159 17 198 145 155 200 99 128 > 20 3 75 11 22 60 28 97 160 219 307 82 150 123 202 60 27 320 447 400 106 60 38 21 270 195 137 20 15 6 9 185 107 143 328 222 300 485 1103 118 120 185 284 419 151 Table 4.4 (Continued) 004-185 BANANA 38 100 44 174 044-186 CANTALOUPE 29 100 55 044-187 GRAPEFRUIT 27 200 135 159 044-188 ORANGE 206 122 150 159 226 044-192 PINEAPPLE 113 90 37 18 045-193 DRIED FRUIT 33 DESSERTS-SWEETS 047-195 HONEY 47 33 18 047-197 SUGAR 8 68 105 109 151 048-198 JAMS 107 57 100 99 123 048-199 SYRUP 332 94 40 129 95 049-201 CHOCOLATE CANDY.. 049-202 MARSHMALLOW 33 8 68 44 11 MISCELLANEOUS 051-205 POT PIES 227 199 261 063-217 SPAGHETTI + MEAT. 067-221 COCOA MIX 330 179 107 52 7 20 Total Grams/Week. 15313 14679 Total # of Items. 3 22 #..of Initial Items 3 2 12891 75 2 18871 83 83 19463 70 53 16831 82 61 152 Table 4.5 Nutrient composition of SID-1 and SID-2, and the revised diets developed using quadratic objective functions .: (Eqns. 4.8, 4.9) 153 Nutrient Nutrient Constraints SID-1. SID-1 SID-1 Minimum Maximum Initial Eqn. 4.8 Eqn. 4.9 (/week) (/week) (/week) (/week) (/week) PROTEIN(gm) 392.00 558.60 512.54 445.72 520.27 CHO-T(gm) 2735.4 4476.1 2809.9 3186.9 2808.5 T-FAT(gm) 44.209 663.13 92.428 511.45 663.13 KCAL 18899. 20889. 14210. 18899. 18899. CHO-F(gm) 79.376 199.15 35.125 79.576 79.576 SFA(gm) .0 221.04 9.6000 139.79 201.04 SUCR(gm) .0 746.02 30.925 291.22 591.58 PUFA(gm) 44.209 663.13 22.400 279.59 372.48 VIT-A(iu) 35000. 140000. 0.0 78874. 92705. VIT-D(iu) 700.00 4200.0 2319.8 1765.4 1612.1 VIT-E(mg) 63.000 700.00 56.550 63.000 97.001 VIT-c(mg) 210.00 3500.0 56.580 1440.5 1105.0 THIA(mg) 9.9470 19.894 16.329 11.669 11.831 RIBO(mg) 11.936 23.873 13.963 12.413 12.062 NIAC(mg) 131.30 262.60 148.67 131.30 131.30 VIT-B6(ug) 14000. 28000. 20777. 19814. 14000. FOLATE(mg) 1400.0 2800.0 2019.5 1682.1 1649.3 POTAS(mg) 9800.0 19600. 17082. 19600. 19600. CAL(mg) 5600.0 11200. 8172.8 8960.0 8960.0 PHOSP(mg) 5600.0 11200. 15349. 11200. 11200. IRON(mg) 70.000 140.00 73.195 110.24 110.56 MAG(mg) 2205.0 4410.0 4471.8 3777.1 3401.8 CA/P .8 1.2 0.53245 0.8000 0.8000 P/S 1.8 2.0 4.0000 2.0000 1.8528 Nutrient Nutrient Constraints SID-1: SID-2 SID-2 Minimum Maximum Initial Eqn. 4.8 Eqn. 4.9 (/week) (/week) (/week) (/week) (/week) PROTEIN(gm) 392.00 558.60 713.82 540.24 550.68 CHO-T(gm) 2735.4 4476.1 1872.4 2735.4 2735.4 T-FAT(gm) 44.209 663.13 981.39 612.94 638.72 KCAL 18899. 20889. 20195. 19239. 19754. CHO-F(gm) 79.576 159.15 32.250 79.576 79.576 SFA(gm) .0 221.04 367.46 221.04 221.04 SUCR(gm) .0 746.02 549.60 746.02 746.02 PUFA(gm) 44.209 663.13 543.42 332.18 359.57 VIT-A(iu) 35000. 140000. 69952. 103460. 120730. VIT-D(iu) 700.00 4200.0 1111.3 705.06 700.00 VIT-E(mg) 63.000 700.00 75.273 93.305 110.41 VIT-C(mg) 210.00 3500.0 886.80 1436.5 1224.0 THIA(mg) 9.9470 19.894 10.292 12.694 12.408 RIBO(mg) 11.936 23.873 13.803 13.931 13.479 NIAC(mg) 131.30 262.60 178.25 131.30 133.04 VIT-B6(ug) 14000. 28000. 12942. 14000. 14000. FOLATE(mg) 1400.0 2800.0 1460.9 1745.2 1977.3 POTAS(mg) 9800.0 19600. 22490. 19600. 19600. CAL(mg) 5600.0 11200. 6137.3 8960.0 8960.0 PHOSP(mg) 5600.0 11200. 11258. 11200. 11200. I RON (nig) 70.000 140.00 120.08 140.00 136.93 MAG(mg) 2205.0 4410.0 2386.5 2875.1) 2800.3 CA/P .8 1.2 .54514 .80000 .80000 P/S 1.0 2.0 1.4789 1.5028 1.6267 154 Although each of these formulations produces revised diets which are optimal solutions in the sense that they minimize total squared deviation, the revised diets are expected to be different because of the different penalty coefficients, w^. Using Equation 4.9, the size of the absolute deviation of items from initial levels should increase as initial consumption level rises. In contrast, the absolute deviation associated with different initial amounts should not vary with the equation using direct squared deviation (Eqn. 4.8), since the penalty coefficient does not incorporate a term for initial consumption levels. This phenomenon is illustrated in Table 4.6 which compares the average absolute deviation of items in the revised diets from their initial con sumption levels* Where the penalty coefficient includes a term for initial consumption (Eqn. 4.9), the average absolute deviation rises from 61 grams at a consumption level of 0 grams/week to 5968 grams at a consumption of 9351-9400 grams/week for the SID-1 revised diet, and from 22 grams at a consumption level of 0 grams/week to 2800 grams at a consumption of 2751-2800 grams/week for the SID-2 revised diet. This extent of rise is not evident with the other equation (Eqn. 4.8)/ The average absolute deviation of items at each initial consumption level is not exactly proportional to the penalty weighting applied. If they were, for example, the objective using direct squared deviation (Eqn.. 4.8) would be expected to produce the same average absolute devia tion for items at each of the initial consumption levels, but it does not. This discrepancy is caused by the unequal distribution of nutrients across items. Those which are efficient sources of nutrients for a given situation deviate more widely. If all items made the same nutrient contribution, the quadratic objectives would spread deviations over all items in exact proportions to the penalty coefficient's relative weighting, 155 in order to satisfy nutrient constraints. On the other hand, if each item contributed only one unique nutrient, the solution outcome would be insensitive to different weighting coefficients. Each item would be a most efficient source of one particular nutrient. Table 4.6 Average absolute deviation from initial consumption levels for items in revised diets developed from SID-1 and SID-2 using quadratic objective functions (Eqns. 4.8, 4.9) Initial Consumption (grams/item) Average Absolute Deviation From SID-1 # of Eqn. 4.8 Items (grams) Average Absolute Deviation From SID-2 Eqn. 4.9 # of Eqn. 4.8 Eqn.4.9 (grams) Items (grams) (grams) 0 124 46 61 44 56 22 0-50 22 43 32 51-100 22 82 77 101-150 8 74 81 151-200 14 129 191 201-250 1 28 34 251-300 1 280 280 4 102 261 301-350 3 167 190 351-400 2 131 303 401-450 1 37 338 451-500 1 378 495 501-550 1 7 7JO 59 701-750 1 100 732 1551-1600 1 28 1294 2751-2800 1 29 2800 3201-3250 1 11 1348 5651-5700 1 1461 3681 9351-9400 1 1602 5968 The different treatment- of deviations by these two objective functions results in marked differences in the solution profiles. With respect to the number of items contained in the diets developed from SID-1 (foot of Table 4.4), both objectives have retained two of the three original items. However, the objective function using percentage squared deviation (Eqn. 4.9) 156 has added 73 new items to the diet for a total of 75 items in the revised diet, whereas the other objective (Eqn. 4.8) has only added 20 new items for a total of 22 items in the revised diet. Thus, each objective has enlisted new items, but the percentage formulation has enlisted a much greater number. This difference is to be expected since the percentage formulation applies a heavy penalty load for deviation of small-quantity items in the initial diet compared to an equivalent deviation in a large-quantity item, while the other formulation does not. Thus, the efficiency of any small-quantity item to supply nutrients to satisfy the constraints is quickly exhausted with the percentage formulation. The consequence is that a longer sequence of alternate efficient sources must be incorporated into the revised diet based on minimum percentage deviation, if these cannot be claimed from items already included in the diet. With respect to the number of items in the diets developed from SID-2. the percentage formulation has retained 61 of the original 83 items and added 21 new items for a total revised solution of 82 items. The solution using direct squared deviation has retained 53 of the original 83 items and added 17 items for a total revised solution of 70 items. Unlike the revised diets developed from the SID-1, those developed from SID-2 do not demonstrate such a large difference in the number of new items entering the solution. In fact, the revised diet developed from SID-2 using the percentage formulation enlists only a few more new items than the revised diet based on direct squared deviation -- 17 and 21, res pectively, as compared, to 20 and 73, for the SID-1 "solutions. The change: in relative proportion of new-item entry for the two formulations is due to the availability of additional consumed items in the SID-2. In this instance, 157 additional items have provided the percentage formulation with a relatively greater number of efficient nutrient sources, and consequently relatively fewer new items have entered the solution. Because larger-quantity items can deviate widely compared to smal1-quantity items with the percentage formulation, less concentrated sources of nutrients can be used efficiently to meet the constraints provided these items are present in the diet in appropriate amounts. Although, in terms of the number of items, there is not a great disparity between the revised diets developed from SID-2 with these two formulations, as compared to those from SID-1, differences are apparent. The revised diet developed from SID-2 by the percentage formulation has more liberally included both original items -- 61 as compared with 53 from the direct formulation arid new items -- 21 as compared with 17 from the direct formulation. It is not clear that the percentage formulation should necessarily incorporate more items in revised solutions developed with every diet. Both formulations, direct and percentage squared deviation are quadratic and hence tend to exhaust the efficiency of any particular item to act as an efficient nutrient source. This spreads deviation over items. In this instance, however, the nutrient distribution over items and the quantities of the items consumed has resulted in the percentage formulation incorporating a greater number of both non-consumed and consumed items in the revised diet. The percentage formulation does not necessarily have a conceptual advantage over the dilrect formulation. The formulation incorporating percentage squared deviation provides the apparent advantage of adjusting the unit value of deviations relative to their initial consumption level. Thus, regardless of the initial amount consumed, the penalty associated 158 with decreasing to zero or doubling in amount is the same.' Viewed one way, with the percentage formulation small-quantity items are not sac rificed to meet nutrient constraints while large-quantity items are pre- „ ferentially maintained, as is the case with the formulation using direct squared deviation. On the other hand, the formulation with percentage squared deviation implies that the penalty for seemingly large deviations in the quantities of large-quantity items should be penalized the same as an equal-percentage, but relatively small, absolute change in small-quantity items. 4.3.2.2.2 Penalty Coefficient, w^, Based on Initial Consumption Penalty.coefficients were chosen to differentially penalize initially consumed versus non-consumed items, as follows: I 2 (100 if x° = 0 (4.10) minimize E w-(x-) ( • - x-) where w.(x-) = ( i=l 1 1 (.1 if x? > 0 I 0 2 (4.11) minimize E w.(x.) (x. - x.) where w-(x-) i=l 1 1 1 1 11 = ( (10,000 if x? = 0 ( o (1 if x° > 0 1 1 o 2 (4.12) minimize £ ——! T (x. - xV) i=l (x? + .01) 1 1 For the formulations using direct squared deviation (Eqns. 4.10, 4.11), penalty weightings of 100 and 10,000. respectively, were applied to items not consumed in the client's original diet. For the formulation1 using percentage squareddeviation (Eqn. 4.12), a penalty value of 10.000 was applied to non-consumed items by defining epsilon as .01 over all items. Although this is not strictly equivalent to the expression: for originally consumed items, the error introduced for portion-sized itemsrs'hould be small. Revised diets (Table 4.7, 4.8) were developed from SID-1 and SID-2 using these objectives (Eqns. 4.10-4.12). The nutrient composition of the initial and revised diets are tabulated in Tables 4.9 and 4.10. 160 Table 4.7 Revised diets developed from SID-1 using quadratic objective functions (Eqns. 4.9, 4.12). SID-1 GROUP-ITEM FOOD ITEM (grams/week)* CODE # [nitial Eqn; 4.9 Eqn. 4.12 ENTREE-DAIRY 001-003 CHEDDER CHEESE... 354 358 001-004 COTTAGE CHEESE... 001-005 CREAM CHEESE 002-006 SOUR CREAM 13 14 002-008 YOGHURT 003-010 EGG ENTREE-CEREALS 004-012 CORN CEREAL 203 202 004-013 WHEAT CEREAL 50 51 005-015 OATMEAL CEREAL... 005-016 WHEAT CEREAL 14 15 006-017 PANCAKES 007-018 NOODLES 007-019 SPAGHETTI 008-020 RICE, brown 9375 3407 3403 008-021 RICE, white 25 25 008-023 WHEAT GERM 280 009-024 FRENCH BREAD 49 49 009-027 WHITE BREAD 82 82 009-028 WHOLE WHEAT BREAD 1 182 182 010-031 BUSCUITS 010-032 HAMBURGER BUN.... 1 82 82 010-033 MUFFIN '-" 010-035 ENGLISH MUFFIN... 182 182 011-037 SALTINES 34 34 011-040 RYE KRISP 50 51 ENTREE-MEATS 012-041 BEEF, 30% fat.... 012-042 BEEF, 20% fat.... 012-043 BEEF, 15% fat 012-046 PORK, lean cuts.. 012-047 PORK, all hams... 012-048 BACON 013-049 CHICKEN, steamed. 013-051 CHICKEN, fried... 014-055 FRIED FISH 014-056 BROILED FISH 014-058 OYSTERS Values rounded to the nearest gram/week. Table 4.7 (Continued) 014-062 SARDINES 1 014-063 SHRIMP 014-065 TUNA 015-067 LIVER-.-*...i'uY 017-069 FRANKFURTERS 017-070 FRESH SAUSAGES... 017-071 LIVERWURST 018-073 BEANS, dri-ed 126 126 018-075 SOYBEANS 33 33 019-076 ALMONDS 73 74 019-077 CASHEW NUTS 019-079 PEANUT BUTTER.... 019-080 PEANUTS 88 88 79 79 019-081 PECANS ENTREE-VEGETABLES 020-082 POTATOE, baked... 020-083 POTATOE, fried... 020-084 POTATOE, mashed... 020- 085 SWEET POTATOE.... 021- 089 BEANS, green 29 30 152 152 235 235 021-091 BROCOLLI 352 352 021-092 CABBAGE 173 173 021-095 LETTUCE 83 84 021-098 PEAS 355 354 021-099 PEPPERS 304 304 021-101 SPINACH 231 232 022-103 BEETS 163 162 022-104 CARROTS, cooked.. 022-105 CARROTS, raw 205 204 181 181 022-106 CORN 25 26 022-111 TOMATOE 59 59 023-113 CUCUMBER 39 39 023-114 MUSHROOMS 37 37 023-115 ONIONS 117 117 024-116 SUCCOTASH. 210 210 025-117 OLIVES 374 373 025-118 PICKLES, sweet... 025- 119 PICKLES, sour.... ENTREE-FATS 026- 121 LARD 93 93 82 82 70 69 026-124 SOYBEAN OIL 67 66 027-125 BUTTER 52 52 028-127 CHEESE SAUCE 59 60 028-128 GRAVY 162 Table 4.7 (Continued) 029-132 029-133 030-136 030- 137 031- 140 031-141 032-143 032-144 032-145 032- 148 033- 150 034- 151 035- 153 035- 154 036- 155 036-156 036-158 037-160 037-161 037-163 038-166 038-168 038- 170 039- 171 039- 172 040- 173 040- 174 041- 175 041-177 042-178 042- 179 043- 181 044-183 044-184 MAYONNAISE 27 27 SALAD DRESSING BEVERAGES-DAIRY WHOLE MILK 49 49 1 SKIM MILK 5658 1977 1936 TABLE CREAM 13 14 WHIPPED CREAM 23 23 BEVERAGES-FRUIT APPLE JUICE 6 6 GRAPEFRUIT JUICE.. LEMON JUICE ORANGE JUICE TOMATOE JUICE 44 44 BEVERAGES-MISC. COLA-TYPE COFFEE TEA BEER DISTILLED SPIRITS. DRY WINES 39 39 SOUPS CREAMED SOUPS 17 18 PEA SOURS MEAT + VEGIE SOUPS DESSERTS-CEREALS COFFEE CAKE 20 20 FRUITCAKE 22 22 ICED CAKES FRUIT PIES 97 97 PUMPKIN PIES 82 82 COOKIES FRUIT COOKIES,,, , 447 445 CAKE DOUGHNUTS DANISH PASTRY J. DESSERTS-DAIRY ICE-CREAM 21 22 SHERBERT 20 20 PUDDINGS 9 10 DESSERTS-FRUIT APPLE 222 221 APPLESAUCE 120 119 163 Table 4.7 (Continued) 044-185 044-186 044-187 044-188 044- 192 045- 193 047-195 047- 197 048- 198 048- 199 049- 201 049-202 051-205 063-217 067-221 BANANA 38 39 CANTALOUPE ?29 '30 GRAPEFRUIT 27 27 ORANGE 122 122 PINEAPPLE 90 91 DRIED FRUIT 33 35 DESSERTS-SWEETS HONEY 47 47 SUGAR 68 68 JAMS 57 57 SYRUP 94 94 CHOCOLATE CANDY... MARSHMALLOW 33 34 68 68 MISCELLANEOUS POT PIES SPAGHETTI + MEAT.. COCOA MIX -V 52 54 Total Grams/Week.. Total # of Items.. # of Initial Items 15313 3 3 12891 75 2 12862 76 2 164 Table 4.8 Revised diets developed from SID-2 using quadratic objective functions(Eqns. 4.8-4.12) SID-2 GROUP-ITEM FOOD ITEM (grams/week)* CODE # Initial Eqn.7.8 Eqn.4.1C Eqn.4.11 Eqn.4.S Eqn.4.12 ENTREE-DAIRY 001-003 CHEDDER CHEESE 182 519 553 546 • 618 696 001-004 COTTAGE CHEESE 001-005 CREAM CHEESE... 002-006 SOUR CREAM 002-008 YOGHURT 003-010 EGG + 495 117 6 ENTREE-CEREALS 004-012 CORN CEREAL.... 28 210 421 431 108 42 004-013 WHEAT CEREAL... 221 13 0.14 38 0.03 005-015 OATMEAL 270 187 185 174 207 125 005-016 WHEAT CEREAL... 24 3 0.03 21 0.01 006-017 PANCAKES 135 47 007-018 NOODLES 007-919 SPAGHETTI 008-020 RICE, brown.... 75 6 0.06 29 0.01 008-021 RICE, white 39 1 0.02 11 0.01 008-021 WHEAT GERM 3 0.04 009-024 FRENCH BREAD... 40 63 50 44 62 44 009-027 WHITE BREAD.... 414 451 370 357 752 981 009-028 WHOLE WHEAT BREAD 138 5 0.05 66 0.02 010-031 BISCUITS 70 2 16 010-032 HAMBURGER BUN.. 46 83 2 73 53 010-033 MUFFIN 40 28 34 010-035 ENGLISH MUFFIN. 56 184 467 478 186 95 011-037 SALTIMES 24 6 146 147 54 27 011-049 RYE KRISP 221 13 0.14 88 0.01 ENTREE-MEATS 012-041 BEEF, 30% fat.. 170 012-042 BEEF, 20% fat.. 85 012-043 BEEF, 15% fat.. 170 012-046 PORK, lean cuts '85 012-047 PORK, all hams. 43 21 012-048 BACON. 64 * Values rounded to the nearest gram/week, except values below 0.5 grams/ week which are rounded to two decimal places. 165 Table 4.8 (Continued) 013-049 CHICKEN, steamec il. 0. Oil. 013-051 CHICKEN, fried. 350 014-055 FRIED FISH 014-056 BROILED FISH... 014-058 OYSTERS 014-062 SARDINES 014-063 SHRIMP 43 17 014-065 TUNA 60 22 167 211 37 67 015-067 LIVER 2 0. 03 017-069 FRANKFURTERS... 017-070 FRESH SAUSAGES.. . 40 5 017-071 LIVERWURST 98 22 0. 03 61 0. 03 019-076 ALMONDS 019-077 CASHEW NUTS 15 115 133 15 15 019-079 PEANUT BUTTER.. 28 20 019-080 PEANUTS 120 10 43 36 019-081 PECANS 85 19 0. 21 112 0. 04 ENTREE-VEGETABLE S 020-082 POTATOE, baked.. 100 7 020-083 POTATOE, fried. 100 020-084 POTATOE, mashed 200 82 125 84 020-085 SWEET POTATOE.. 180 218 324 387 449 539 021-089 BEANS, green... 65 253 368 368 199 141 021-091 BROCOLLI +• 63 363 683 708 277 181 021-092 CABBAGE 65 180 377 401 175 130 021-095 LETTUCE 190 178 97 81 271 321 021-098 PEAS 178 572 1134 1169 924 1247 021-099 PEPPERS 374 11 0. 11 95 0. 04 021-101 SPINACH 85 182 202 212 214 187 022-103 BEETS 80 158 165 153 178 151 022-104 CARROTS, cooked 76 146 11 166 137 022-105 CARROTS, raw 50 70 44 34 113 77 022-106 CORN 109 9 0. 10 49 0. 02 022-111 TOMATOE 518 448 327 310 577 724 023-113 CUCUMBER 275 210 52 29 230 198 023-114 MUSHROOMS 1 0. 00 023-115 ONIONS 8 119 357 381 48 9 024-116 SUCCOTASH 159 4 0 04 52 0. 02 025-117 OLIVES 380 7 0 07 84 0. 03 025-118 PICKLES, sweet. 20 20 21 025-119 PICKLES, sour.. 34 30 33 ENTREE-FATS 026-121 LARD 5 63 78 5 026-124 SOYBEAN OIL.... 027-125 BUTTER 125 22 29 29 1 16 166 Table 4.8 (Continued) 028-127 CHEESE SAUCE... 028-128 GRAVY 0.2? 72 28 46 029-132 MAYONNAISE 105 21 44 50 029-133 SALAD DRESSING. 30 70 65 65 74 78 BEVERAGES-DAIRY 030-136 WHOLE MILS 732 632 446 421 030-137 SKIM MILK 031-140 TABLE CREAM.v.. 031- 141 WHIPPED CREAM.. BEVERAGES-FRUIT 032- 143 APPLE JUICE.... 032-144 GRAPEFRUIT JUICE 032-145 LEMON JUICE 185 93 60 8 8 120 5 5 032- 148 ORANGE JUICE... 033- 150 TOMATOE JUICE.. 360 174 BEVERAGES-MISC. 034-151 COLA-TYPE 339 328 350 350 332 347 035-153 COFFEE ?800 2771 2711 2705 035-154 TEA 1600 1572 1516 1508 306 036-155 BEER 3240 3229 3363 3384 4588 7878 036-156 DISTILLED SPIRITS 036-158 DRY WINES 43 34 28 28 41 42 400 324 241 235 154 SOUPS 037-160 CREAMED SOUPS... 037-161 PEA SOUPS 198 145 25 7 155 112 200 99 37 22 128 81 037- 163 MEAT + VEGIE SOUP DESSERTS-CEREALS 038- 166 COFFEE CAKE 3 0.00 75 11 27 038-168 FRUITCAKE 038-170 ICED CAKES 60 28 43 039-171 FRUIT PIES 160 219 346 352 307 361 039- 172 PUMPKIN PIES.... 040- 173 COOKIES 150 123 45 18 202 224 60 60 58 27 42 040- 174 FRUIT COOKIES.... 041- 175 CAKE COUGHNUTS... 051-177 DANISH PASTRY.... DESSERTS-DAIRY 042- 178 ICE-CREAM 400 9 0.09 106 0.04 60 22 38 26 270 195 85 66 137 9 042-179 SHERBERT 15 1 0.01 6 0.00 043-181 PUDDINGS 185 107 35 19 143 116 167 Table 4.8 (Continued) DESSERTS-FRUIT 044-183 APPLE 300 485 644 643 1105 1920 044-184 APPLESAUSE 185 284 423 430 419 552 044-185 BANANA 100 44 370 430 174 184 044-186 CANTALOUPE 100 55 66 044-187 GRAPEFRUIT 200 135 159 137 044-188 ORANGE 150 159 43 24 226 253 044-192 PINEAPPLE 37 1 0.01 18 0.01 045-193 DRIED FRUIT DESSERTS-SWEETS 047-195 HONEY 33 2 0.02 18 0.01 047-197 SUGAR 105 109 190 197 151 185 048-198 JAMS 100 99 157 162 123 136 048-199 SYRUP 40 129 283 312 95 54 049-201 CHOCOLATE CANDY.. 049-202 MARSHMALLOW 44 2 0.03 11 0.01 MISCELLANEOUS 051-205 POT PIES 227 199 299 312 261 286 063-217 SPAGHETTI + MEAT. 067-221 COCOA MIX 330 179 45 28 107 7 20 8 Total Grams/Week.18871 18463 18748 18748 16831 19850 Total # of Items. 83 70 70 67 82 86 # of Initial Items 83 53 50 47 61 66 # of New Items... 17 20 20 21 20 168 Table 4.9 Nutrient composition of SID-1, and the revised diets developed using quadratic objective functions (Eqns. 4.9, 4.12). Nutrient Nutrient Constraints SID-1 SID-1 SID-1 Minimum Maximum Initial Eqn. 4.9 Eqn. 4.1 (/week) (/week) (/week) (/week) (/week) PROTEIN(gm) 392.00 558.60 512.54 520.27 520.46 CHO-T (gm) 2735.4 4476.1 2809.9 2808.5 2808.9 T-FAT(gm) 44.209 663.13 92.428 663.13 663.13 KCAL 18899. 20889. 14210. 18899. 18899. CHO-F(gm) 79.576 159.15 35.125 79.576 79.576 SFA(gm) .0 221.04 5.6000 201.04 20.179 SUCR(gm) .0 746.02 30.925 591.58 593.21 PUFA(gm) 44.209 663.13 22.400 372.48 371.84 VIT-A(iu) 35000. .14000E+06 0.0 92705. 92840. VIT-D(iu) 700.00 4200.0 2319.8 1612.1 1617.1 VIT-E(mg) 63.000 700.00 56.550 97.001 97.018 VIT-C(mg) 210.00 3500.0 56.580 1105.0 1105.3 THIA(mg) 9J9470 19.894 16.329 11.831 11.824 RIBO(mg) 11.936 23.873 13.963 12.062 12.039 NIAC(mg) 131.30 262.60 148.67 131.30 131.30 VIT-B6(ug) 14000. 28000. 20777. 14000. 14000. FOLATE(mg) 1400.0 2800.0 2019.5 1649.3 1649.6 POTAS(mg) 9800.0 19600. 17082. 19600. 19600. CAL(mg) 5600.0 11200. 8172.8 8960.0 8960.0 PHOSP (mg) 5600.0 11200. 15349. 11200. 11200. IRON(mg) 70.000 140.00 73.195 110.56 110.72 MAG(mg) 2205.0 4410.0 4471.8 3401.8 3408.1 CA/P .8 1.2 0.53245 0.8000 0.80000 P/S 1.0 2.0 4.0000 1.8528 1.8427 169 Table 4.10 Nutrient composition of SID-2, and the revised diets developed using quadratic objective functions (Eqns. 4.8-4.12) 170 Nutrient Nutrient Constraints SID-2 SID-2 SID-2 Minimum Maximum Initial Eqn. 4.8 Eqn. 4.10 Eqn.4.1 (/week) (/week) ((/week) (/week) (/week) (/week) PROTEIN(gm) 392.00 558.60 713.82 540.24 558U60 558.60 CHO-T(gm) 2735.4 4476.1 1872.4 2735.4 2735.4 2735.4 T-FAT (gin) 44.209 663.13 981.39 612.94 575.66 572.11 KCAL 18899. 20889. 20195. 19239. 19026. 18999. CHO-F(Gm) 79.576 159.15 32.250 79.576 79.576 79.576 SFA(gm) .0 221.04 367.46 221.04 221.04 221.04 SUCR(gm) .0 746.02 549.60 746.02 746.02 746.02 PUFA(gm) 44.209 663.13 543.42 332.18 308.42 305.72 VIT-A(iu) 35000. 140000. 69952. 103460. 94146. 91999. VIT-D(iu) 700.00 4200.0 1111.3 705.06 896.00 980.41 VIT-E(mg) 63.000 700.00 75.273 93.305 102.51 101.04 VIT-C(mg) 210.00 3500.0 886.80 1436.5 1361.2 1382.6 THIA(mg) 9.9470 19.894 10.292 12.694 11.699 11.452 RIBO(mg) 11.936 23.873 13.803 13.931 12.421 11.936 NIAC(mg) 131.30 262.60 178.25 131.30 137.02 139.70 BIT-B6(ug) 14000. 28000. 12942. 14000. 14000. 14000. FOLATE(mg) 1400.0 2800.0 1460.9 1745.2 1755.0 1718.1 POTAS(mg) 9800.0 19600. 22490. 19600. 19600. 19600. CAL(mg) 5600.0 11200. 6137.3 8960.0 8419.0 8266.1 PHOSP(mg) 5600.0 11200. 11258. 11200. 10524. 10333, IRON(mg) 70.000 140.00 120.08 140.00 140.00 140.00 MAG(mg) 2205.0 4410.0 2386.5 2875.1 2772.8 2766.9 CA/P .8 1.2 .54514 .80000 .80000 .80000 P/S 1.0 2.0 1.4789 1.5028 1.3953 1.383 Nutrient Nutrient Constraints SID-2 SID-2 SID-2 'Minimun Maximum Initial Eqn. 4.9 Eqn. 4.12 (/week) (/week) (/week) (/week) (/week) PROTEIN(gm) 392.00 558.60 713.82 550.68 538.24 CHO-T(gm) 2735.4 4476.1 1872.4 2735.4 2735.4 T-FAT(gm) 44.209 663.13 981.39 636.72 586.78 KCAL 1.8899. 20889. 20195. 19754. 20075. CHO-F(gm) 79.576 159.15 32.250 79.576 79.576 SFA(gm) .0 221.05 367.46 221.04 221.04 SUCR(gm) .0 746.02 549.60 746.02 746.02 PUFA(gm) 44.209 663.13 543.42 359.57 310.87 VIT-A(iu) 35000. 140000. 69952. 120730. 12T670. VIT-D(iu) 700.00 4200.0 1111.3 700.0 700.00 VIT-E(mg) 63.000 700.00 75.273 110.41 88.360 VIT-C(mq) 210.00 3500.0 886.80 1224.0 1145.9 THIA(mg) 9.9470 19.894 10.292 12.408 11.339 RIBO(mg) 11.936 23.873 13.803 13.479 13.373 NIAC(mg) 131.30 262.60 178.25 133.05 148.24 VIT-B6(ug) 14000. 28000. 12942. 14000. 14000. FOLATE(mg) 1400.0 2800.0 1460.9 1977.3 1939.2 POTAS(mg) 9800.0 19600. 22490. 19600. 19600. CAL(mg) 5600.0 11200. 6137.3 8960.0 8960.0 PHOSP(mg) 5600.0 11200. 11258. 11200. 11200. IRON(mg) 70.000 140.00 120.08 136.93 128.56 MAG(mg) 2205.0 4410.0 2386.5 2800.3 2702.6 CA/P .8 1.2 .54514 .80000 .80000 P/S 1.0 2.0 1.4789 .80000 1.4064 171 The solution dynamics for formulations which selectively penalize change in non-consumed versus consumed items are analogous to those pre viously discussed which penalize on the basis of percentage or absolute deviation. In this instance, however, the penalty curves slope is further altered as a function of the penalty weighting applied to selected items. If a change in a particular item is heavily or lightly penalized, the slope of the penalty curve is proportionately increased or decreased, respectively. Consequently, the size of the average absolute deviation for an item tends to decrease as the penalty weighting increases, and increase as the penalty decreases. This phenomenon is illustrated in Table 4.12 which compares the average absolute deviation from initial consumption levels for each of the revised diets developed from SID-2. As the penalty applied to non-consumed items increases relative to that for consumed items, the average absolute deviation of non-consumed items decreases. For example, consider the objective functions using direct squared deviation (Eqns. 4.8, 4.10. 4.11). With ehese the absolute deviation from zero for non-consumed items has decreased from 56 to 3 to 0.03 with penalty ratios for non-consumed/consumed items of 1/1, 100/1, and 10,000/1, respectively. Interestingly, the effect is roughly proportional to the increase in penalty assignment. Similarly, for the percentage formulations (Eqns. 4.9, 4.12) an increase in relative penalty on non-con-u sumed items of 1 to 10,000 has reduced the average absolute deviation from 22 to 0.01. The decreased deviation in non-consumed items has been reflec ted by a reasonably consistent increase in the deviation of the categories of consumed items, as shown in Table 4.12. Unlike the revised diets developed from SID-2, those developed from SID-1 (Table 4.11) using the percentage formulation (Eqns. 4.9, 4.12) are 172 virtually insensitive to increased penalty weightings on non-consumed items. This different outcome can be explained by the large number of non-consumed items in the initial diet relative to consumed items. The uniform increase of the penalty over all non-consumed items does not appreciably alter the solution dynamics among the vast majority of items in the diet. Also, it is apparent that the utility of the three initially-consumed items to supply nutrients has not changed much, even with the 10,000 fold increase in penalty on the non-consumed items. Table 4.11 Average absolute deviation from initial consumption levels for items in revised diets developed from SID-1 using quadratic objective functions (Eqns. 4.9, 4.12). Initial Average Absolute Deviation From SID-1 Consumption # f 4 4 12 (grams/week) Ums ^grams) Jgrams) 0 124 61 61 >0-5O 51-100 101-150 151-200 201-250 251-300 1 280 280 301-350 351-400 401-450 451-500 501-550 701-750 1551-1600 2751-2800 3201-3250 5651-5700 1 3681 3722 9351-9400 1 5968 5972 173 Table 4.12 Average absolute deviation from initial consumption levels for items in revised diets developed from SID-2 using quad ratic objective functions (Eqns. 4.8-4.12). Initial Average Absolute Deviation From SID-2 Consumption # Qf Ean_ 4>g Eqn> 4/m £qn> 4>11 Eqn> 4g Eqn.4>i2 ^grams/week; Items (grams) (grams) (grams) (grams) (grams) 0 44 56 3 0.03 22 0 >0-50 22 43 102 104 32 11 51-100 22 82 129 136 77 59 101-150 8 74 109 116 81 95 151-200 14 129 248 256 191 :260 201-250 1 28 72 85 34 59 251-300 4 102 209 222 261 526 301-350 3 167 212 218 190 226 351-400 2 131 260 263 303 380 501-450. 1 37 44 57 338 567 451-500 ] 378 489 495 495 495 501-550 1 70 191 208 59 256 701-751 1 100 286 311 732 732 1551-1600 1 28 84 92 1294 1600 2751-2800 1 29 89 95 2800 2800 3201-3250 . 1 11 123 144 1348 4638 174 Penalizing deviation of non-consumed items relative to consumed items has resulted in only minor changes in the solution profiles. With respect to the number of items contained in the diets developed from SID-1 (Table 4.7), both objectives (Eqns. 4.9, 4.12) have retained two of the three original items. The percentage objective weighting against non-consumed items (Eqn.. 4.12) has added 74 new items; one greater than the other per centage objective (Eqn. 4.9). The greater penalty load on non-consumed items has altered the relative utility of one additional item to provide nutrients to meet the constraints. This item has subsequently entered the solution at 1 gram per week. With respect to the number of items in diets developed from SID-2 (Table 4.8), some of the differences between the revised diets developed with the percentage formulation (Eqn. 4.9) arid with the direct formulation (Eqn. 4.8), observed previously, have been amplified with the formulations (Eqn, 4.10-4.12) penalizing non-consumed item deviations. In particular, the number of consumed items retained in the direct formulation solutions have decreased further, from 53 without additional penalty on non-consumed items (Eqn. 4.9), to 50 and 47 with penalties of 100 (Eqn. 4.10) and 10,000 (Eqn. 4.11) on non-consumed items, respectively. The percentage formulation with additional penalties on non-consumed items (Eqn. 4.12) has increased the number of consumed items retained -- 66 as compared to 61 from the formulation with no additional penalty applied to zero items (Eqn. 4.9). Again, it is not clear that the outcome results from other than the cir cumstances associated with this one particular diet. The number of initially non-consumed items in the revised diets developed with the direct formulation have increased from 17 with no penalty on non-consumed items (Eqn. 4.8) to 20 with penalties on non-175 consumed items (Eqns. 4.10, 4.11). Conversely, the revised diet developed with the percentage formulation which penalizes non-consumed item deviation (Eqn. 4.12) has 20 new items as compared to 21 from its coun terpart without the additional penalty (Eqn. 4.9). Although all the above outcomes meet the objective of minimizing the summed deviation of items, the slight reduction in new items entering the solution developed with the percentage formulation is more consistent with the intent of penalizing the deviation of non-consumed items; that is, the penalty should reduce the number of new items entering the solution rather than increase them, as is the case with the revised diets from the formulation using direct squared deviation. This discussion of phenomena associatediwith differently weighting the deviations of specific items has been restricted to a few examples where non-consumed versus consumed items have been differentially penalized. However, these observations should be applicable to other specific situations. 4.3.2.2.3 Penalty Coefficient, w., Further Comments As discussed above, penalty coefficients can be used to influence the deviation of items from initial levels. How a particular penalty coefficient will affect the solution is a product of a number of factors, namely: the nutrient distribution in the items considered, the quantities of the items consumed in the initial diet, and the nutrient constraints applied. The complex interactions inherent in these factors underlines the real problem in trying to establish penalty coefficients which function to provide useful revisions for many different situations. 176 4.3.2.3 First Term: Further Modifications of the Algorithm As presently formulated the objective function (Eqn. 4.1) does not control the direction or extent of change. It would appear realistic that the desirability of an increase or decrease in a particular item may be different, and hence that the penalty coefficient applied should re flect this difference. In this regard, a quadratic objective can be formulated to consider, separately, positive and negative deviations in item clusters. In more detail, this algorithm finds the combination of item clusters which minimizes the total, squared, weighted-positive and weighted-negative deviations in the amounts of each item cluster from initial consumption levels, while satisfying nutrient constraints. The linear form of this objective function has been given already. (Eqn. 4.3). The formulation is as follows: (4.13) ... 1 , + + - .-.2 minimize E (w. x- + w. x.) i=l 11 11 (4.14) (4.15) (4.16) subject to xj - xT = x^ - x° I mq ^ 3qi Xi * "q I .\ aui xi t ruv * 111 > tuv ^avi*i (i = 1, 2, I) (q = 1, 2, Q) (u,v = any specified set of nutrient pairs) xn-, x., x. > 0 (i = 1, 2, I) (4.17 where the terms are as previously defined (p. 142 ). Further elaboration of this algorithm could include separable programming to assign different penalty weighting to different portions of the curve. To ensure that changes fall within specified limits, constraints could 177 be appropriately included to provide minimum and maximum limits on the deviation in the amounts of specific item clusters. This expression is as follows: (4.18) d. » (Xi - x?) » e, where the newly defined terms are: d.j and e^ are, respectively, the upper and lower bound on the deviation in specific item clusters. If d^ = 0 then the revised consumption of the item cluster cannot be greater than the original consumption level. If = 0 then the revised consumption of the item cluster cannot be less than the original consumption level. The constraints used should be carefully applied so as not to interfere with solution feasibility. The formulations just described (Eqns. 4.13-4.18) have not been tested, but provide the basis for desirable future work. 4.3.2.4 Second Term of the Objective Function The first term of the objective function represents the acceptability of an item deviating from its initial dietary levels. In this first term the acceptability of this deviation is described as independent of spec ific concurrent changes in other items. Further, the first term does not consider changes in other dietary attributes, only in terms of the item clusters themselves. Presumably, however, the acceptability of an item deviating from initial dietary levels depends not only upon the nature of the deviation of that food itself, but also upon the nature of concurrent deviations of the other foods in the consumption pattern, and of changes in other characteristics of the diet and in the diet as a whole. In any case, 178 such an assumption would appear more realistic than the assumption that food preferences are not modified by changes in the consumption levels of the different foods. Two basic types of compensatory relationships between foods have been described (Smith 1963), namely, incompatabilities and complementari ties. Incompatible foods are those which clash or disagree, absolutely or in some proportion. They typically vary in inverse proportion. On the other hand, complementary foods either enhance one another, or their coexistence may be essential to palatability or acceptability. Comple mentary foods tend to vary in direct proportion. These relationships are applicable to item clusters and attribute groups. The second term of the model's objective function contains the mathematical expression used to minimize the deviation of attribute groups from their initial levels, and to establish the relationships of concurrent change between item clusters and attribute groups. At present, only incompatibility between item clusters and attribute classes has been considered in establishing the relationships of concurrent change. This term minimizes the summed weighted square of differences between the amounts of item cluster groups -- called attribute groups -- in the initial and recommended diets, as follows: K Jk o 2 (4.19) minimize z L P.. ( £ (x- - xV) ) k-lj-! Jk KGJk The complete formulation incorporating this additional term given in Equations 3.2 to 3.5. The addition of this second term is expected to modify the output profiles as compared to those producedwithout it. . First, if the deviations of attribute groups are increasingly penalized, then the quantity of these 179 attribute groups in the revised diet should approach initial levels. Thus, this term allows for monitoring general food and diet characteristics considered important to the diet's acceptability. Second, if emphasis on maintaining initial levels of attribute groups increases, the attribute structure encourages the item clusters'within each attribute group to vary in inverse proportion. Thus, the second term tends to cause substitution between the presumed nearest acceptable alternative for an item cluster or attribute group. This relationship of concurrent change between item clusters and attribute groups is, in a sense, an artifact of assigning those elements with like characteristics into the hierarchy of successive subgroups, groups, and supergroups, outlined in Appendix G. The number of item clusters contained in any attribute group and the number of attribute groups to which any item is assigned will influence whether substitution between items will be apparent. Where only two items are contained in a group, the shift between these grouped partners should be more apparent than when a large number of items is contained in a group. Where an item cluster has simultaneous membership in many attribute groups, as is the case above, (Eqn. 4.19), the exact outcome is dependent on the relative emphasis on the deviation of each attribute group considered. '.It should be noted that the dynamics between the attribute groups on each hierarchical level, k=l to k=7, should be similar to those previously discussed for item clusters in the zero^ hierarchical level. 180 4.3.2.5 Second Term: Shape of the Curve The second term, as expressed in Equation 4.19, also describes a quadratic curve centered at the initial consumption level of each attribute group defined, as shown in Figure 4.3. Thus the relative unacceptabi1ity of differences between the initial and revised values for any attribute element is strongly intensified by this quadratic objective. Figure 4.3 Graph of quadratic function for deviation from initial consumption of attribute group Gj^. Other non-linear or linear functions may be appropriately used to describe the acceptability of these attribute elements deviating from initial levels, and to describe the interactional properties of individual items and of attribute groups. The second term of the objective function (Eqn. 4.19) can be readily transformed to the linear equivalent. This minimizes the total, weighted-positive and weighted-negative deviation in the amount of specified groups from initial consumption levels.; With the addition of this linear second term the linear formulation (Eqns. 4.3-4.7) becomes: I + + - - ^^k + + (4.20) minimize z {vi. x. + xn- v^) + E E (p-k z-k + p-k + z-k) 181 (4.21) subject to z*k - zTk = z (xj - xT) (j = 1, 2, ..., Jk) i€GJk (k = 1, 2, ..., K) (4.22) xj - xT = x. - x° (i . 1, 1, I) I (4.23) mq ^ i aqi x- nq (q = 1, 2, ..., Q) I (4.24) r ^ i=1 U1 1 >y t (u,v = any specif-uv I uv ied set of nutr-z a . x. ient pairs) (4.25) x., xt, xT, ztu, zT, >, 0 xi' xi' xi' zjk' zjk > Where the newly defined terms are as follows: Zjk' zjk are t'1e Pos1tlve anc' negative deviation, respectively, t h of attribute group j in the k hierarchical level. _ fi G (xjk - xTk) if z (xt - xT) > 0 zik " ^ 1febik JK (0 J otherwise (i€G.k (XJk - Xjk) if * (4 " Xi) < 0 . J (0 otherwise Pjk, pjk are the weighted penalty associated with the positive or negative deviation, respectively, of attribute group j in the k^*1 hierarchical level. Since the linear and quadratic formulations of the objective functions first term are expected to generate different solutions, we can corres pondingly expect changes with these formulations incorporating both the first and second terms. The quadratic objective would be expected to moderate extreme fluctuations in any particular item cluster or attribute group in favour of numerous smaller changes. The linear model, however, 182 should be insensitive to large deviations in specific items, provided these changes are consistent with the overall objective of reducing aggregate deviation of item clusters and attribute groups. For example, if all foods provided the same nutrient contributions, the quadratic objective (Eqn. 3.2) would spread deviations over all items and attribute groups in exact proportion to the penalty coefficients applied. The linear objective (Eqn. 4.20) would be insensitive to how the total devia tion was apportioned over item clusters and attribute groups. These expectations are not illustrated here since the linear solutions were not developed for comparison with the quadratic results. 4.3.2.6 Second Term: Penalty Coefficients, w. and P.. 1 J K The impact on the solution of altering the relative weighting of the penalty coefficients, w- and P.. , should also be considered. By altering these penalty coefficients, the value of increments for an item cluster or attribute group can be exaggerated or depressed, and with it the emphasis on concurrent change of items. As previously discussed, the numerical value of the penalty coefficients can be based on various rationales for perceived acceptability of change. The coefficient chosen can represent: differences in initial consumption levels for each item or group; whether an item or group is initially consumed or not consumed; differences in standard serving sizes; the average consumption levels observed in a population; or some other measure of personal preference. In this instance, all penalty coefficients were arbitrarily assigned as "1" for deviation of any item cluster, and "1" or "0" for deviation of attribute groups depending on whether the deviation of a particular group was to be considered in the computation. In order to observe the impact of including the second term of the 183 objective function (Eqn. 4.19), three different objective functions (Eqns. 4.26-4.28) were defined. Each of these objective functions minimizes the summed, squared deviation of both item clusters and specified attribute groups from initial levels. A numerical value of 1 has been assigned for weighting penalty coefficients, \n. and P^. The first objective function (Eqn. 4.26) sums the deviations of all item clusters and of attribute groups on the first hierarchical level, whereas the second objective function (Eqn. 4.27) sums the deviations of all item clusters and of attribute groups on both the first and second hierarchical levels. The third objective function (Eqn. 4.28) sums the deviations over all item clusters and attribute groups on levels, k=l to k=7, and the total diet.'s amount, k=8. The algebraic expressions corresponding to these objective functions are as follows: 1 n 2 1 ^ 1 n2 (4.26) minimize z 1(x. - xV) + •& z 1( z ( x- - x.) ) i=l k=l j=l i£Gjk 1 I 2 2 ^2 (4.27) minimize z l(x. - x?) + z z 1( z (x, - x°.) )2 i=l 1 1 k=l j=l i€Gjk 1 I 8 J8 (4.28) minimize z Kx- - xV)2 + z z l(i • (x- - xu.) ) i=l k=l j=l iCG 1 1 2 jk Using these objectives (Eqns. 4.26-4.28) revised diets were developed from Standard Initial Diet 2, and compared with the solution obtained from an objective function (Eqn. 4.8) previously described, which minimizes only the summed, squared deviation of item clusters from initial quantities (Table 4.13). Table 4.14 gives nutrient composition of the initial and revised diets for each of these objective functions (Eqns. 4.26-4.28). T84 Table 4.13 Revised diets developed from SID-2 using quadratic objective functions (Eqns. 4.8, 4.26-4.28). GROUP-ITEM FOOD ITEM SID-2 CODE (grams/week)* # Initial Eqn. 4.8 Eqn. 4.26 Eqn.4.27 Eqn.4.28 ENTREE-DAIRY 1001-003 CHEDDER CHEESE. 182 519 538 508 499 001-004 COTTAGE CHEESE. 001-005 CREAM CHEESE... 002-006 SOUR CREAM 002-008 YOGHURT 003-010 EGG 495- 117 96 7 ENTREE CEREALS 004-012 CORN CEREAL... 23 210 196 167 401 004-013 WHEAT CEREAL.. 221 87 112 17 005-015 OATMEAL 270 187 115 35 005-016 WHEAT CEREAL.. 24 96 103 006-017 PANCAKES 135 007-018 NOODLES 007-019 SPAGHETTI 998-020 RICE, brown... 75 69 81 15 008-021 RICE, white... 39 9 008-023 WHEAT GERM.... 009-024 FRENCH BREAD.. 40 63 009-027 WHITE BREAD... 414 451 409 263 104 009-028 WHOLE WHEAT BREAD 138 198 278 327 010-031 BUSCUITS 70 010-032 HAMBURGER BUN. 46 83 64 010-033 MUFFIN 40 010-035 ENGLISH MUFFIN 46 184 267 306 276 011-037 SALTINES 24 6 011-040 RYE KRISP 221 183 176 133 ENTREE-MEATS 012-041 BEEF, 30% fat. 170 74 67 012-042 BEEF, 20% fat. 85 012-043 BEEF, 15% fat. 170 012-046 PORK, lean cut 85 012-047 PORK, all hams 43 100 145 012-048 BACON 64 013-049 CHICKEN- steame j 013-051 CHICKEN, fried. 340 5 014-055 FRIED FISH.... 014-056 BROILED FISH.. Values rounded to the nearest grams/week. Table 4.13 (Continued) 014-058 OYSTERS 267 014-062 SARDINES 014-063 SHRIMP 43 014-065 TUNA 60 22 69 127 102 015-067 LIVER 017-069 FRANKFURTERS... 017-070 FRESH SAUSAGES. 40 017-071 LIVERWURST 98 84 3b 99 018-073 BEANS, dried... 83 018-075 SOYBEANS 019-076 ALMONDS 019-077 CASHEW NUTS 15 019-079 PEANUT BUTTER.. 28 019-080 PEANUTS 120 10 16 019-081 PECANS 85 147 155 184 ENTREE-VEGETABLES 020-082 POTATOE, baked. 100 020-083 POTATOE, fried. 100 020-084 POTATOE, mashed 200 82 88 020-085 SWEET POTATOE.. 180 218 446 411 393 021-089 BEANS, green... 65 253 021-091 BROCOLLI....... 63 363 330 385 481 021-092 CABBAGE 65 180 021-095 LETTUCE 180 178 23 34 90 021-098 PEAS 170 572 525 471 427 021-099 PEPPERS 374 316 329 388 021-101 SPINACH 85 182 3 022-103 BEETS 80 158 193 165 150 022-104 CARROTS, cooked 76 146 172 160 168 022-105 CARROTS, raw... 50 70 112 162 194 022-106 CORN 109 70 022-111 TOMATOE 518 - 448 341 216 -.023-113 CUCUMBER 275 210 137 023-115 ONIONS 8 119 238 201 97; 024-116 SUCCOTASH 149 215 127 50 025-117 OLIVES 380 475 475 507 025-118 PICKLES, sweet. 20 025-119 PICKLES, sour.. 34 ENTREE-FATS 026-121 LARD 5 026-124 SOYBEAN OIL 027-125 BUTTER 125 22 028-127 CHEESE SAUCE... 21 51 028-128 • GRAVY 72 186 Table 4. 029-132 029-133 030-136 030- 137 031- 140 031-141 032-032-032-032-143 144 145 148 033-150 034- 151 035- 153 035- 154 036- 155 036-156 036-158 037-160 037-161 037-163 038-166 038-168 038- 170 039- 171 039- 172 040- 173 040- 174 041- 175 041-177 042-178 042- 179 043- 181 044-183 044-184 13 (Continued) MAYONNAISE.-..'...:. SALAD DRESSING.... BEVERAGES-DAIRY WHOLE MILK SKIM MILK. TABLE CREAM.. WHIPPED CREAM BEVERAGES-FRUIT APPLE JUICE GRAPEFRUIT JUICE.. LEMON JUICE ORANGE JUICE TOMATOE JUICE BEVERAGES-MISC. COLA-TYPE COFFEE TEA BEER DISTILLED SPIRITS. DRY WINES SOUPS 1 CREAMED SOUPS' PEA SOUPS MEAT + VEGIE SOUPS DESSERTS-CEREALS COFFEE CAKE FRUITCAKE ICED CAKES... FRUIT PIES PUMPKIN PIES COOKIES FRUIT COOKIES..... CAKE DOUGHNUTS.... DANISH PASTRY DESSERTS-DAIRY ICE-CREAM SHERBERT PUDDINGS DESSERTS-FRUIT APPLE APPLESAUCE 105 21 90 70 79 76 42 732 632 652 702 711 2 195 93 81 150 191 8 86 151 267 120 5 10 360 174 219 212 158 34 53 339 328 323 328 352 2800 2771 2788 2799 2814 1600 1572 1581 1581 1581 3240 3229 3251 3264 3304 43 34 68 72 71 400 324 321 324 302 198 145 184 233 248 200 99 59 74 106 121 75 60 160 219 248 306 237 150 123 86 88 60 400 496 554 585 60 38 270 185 135 145 6 15 85 102 218 185 107 116 147 104 300 485 608 661 741 185 284 288 225 165 187 Table 4.13 (Continued) 044-185 BANANA .,. 100 44 40 83 044-186 CANTALOUPE 100 044-187 GRAPEFRUIT 200 135 044-188 ORANGE 150 159 217 242 189 044-192 PINEAPPLE 37 18 045-193 DRIED FRUIT DESSERTS-SWEETS 047-195 HONEY 33 19 51 123 047-197 SUGAR 105 109 125 110 25 048-198 JAMS 100 99 49 18 23 048-199 SYRUP 40 129 172 223 179 049-201 CHOCOLATE CANDY... 049-202 MARSHMALLOW 44 30 38 106 MISCELLANEOUS 051-205 POT PIES 227 199 195 155 168 063-217 SPAGHETTI + MEAT.. 330 179 208 213 240 067-221 COCOA MIX 7 Total Grams/Week.. 18889 19461 19288 19117 18933 Total # of Items.. 83 70 68 61 55 # of Initial Items 83 53 48 43 36 # of New Items.... 17 20 18 19 # of Items Dropped 13 15 22 28 188 Table 4.14 Nutrient composition of the SID-2, and the revised diets developed using quadratic objective functions (Eqns. 4.26-4.28). Nutrient Nutrient Constraints SID-2 SID-2 SID-2 SID-2 Minimum Maximum Initial Eqn.4.26 Eqri.4.27 Eqn.4.28 (/week) (/week) (/week) (/week) (/week) (/week) PROTEIN (igm ) 392.00 558.60 713.82 558.60 558.60 558.60 CHO-T (gm) 2735.4 4476.1 1872.4 2735.4 2735.4 2735.4 T-FAT(gm) 44.209 663.13 981.39 658.14 663113 663:13 KCAL 18899. 20889. 20195. 19810. 19832. 19881. CHO-F(gm) 79.576 159.15 32.250 79.576 79.576 79.576 SFA(gm) .0 221.04 367.46 221.04 221.04 210.90 SUCR(gm) .0 746.02 549.60 746.02 746.02 738.58 PUFA(gm) 44.209 663.13 543.42 364.07 360.24 362.29 VIT-A(iu) 35000. .140E+06 69952. 0.107E+06 95816. 94650. VIT-E(mg) 63.000 700.00 75.273 90.631 91.075 94.316 VIT-C(mg) 210.00 3500.0 886.80 1231.6 1248.2 1375.4 THIA(mg) 9.9470 19.894 10.292 11.496 11.190 11.448 RIBO(mg) 11.936 23.873 13.803 13.038 12.289 13.194 NIAC(mg) 131.30 262.60 178.25 131.30 131.30 131.30 VIT-B6(ug) 14000. 28000. 12942. 14000. 14000. 14000. FOLATE(mg) 1400.0 2800.0 1460.9 1400.0 1400.0 1400.0 POTAS(mg) 9800.0 19600. 22490. 19600. 19600. 19600. CAL(mg) 5600.0 11200. 6137.3 8960.0 8960.0 8960.0 PHOSP(mg) 5600.0 11200. 11258. 11200. 11200. 11200. I RON (trig) 70.000 140.00 120.08 135.86 130.87 140.00 MAG(mg) 2205.0 4410.0 2386.5 2751.3 2717.7 2562.4 CA/P .8 1.2 0.54514 0.80000 0.80000 0.80000 P/S 1.0 2.0 1.4789 1.6471 1.6297 1.717 J 189 As noted, the second term in the objective function is expected to reduce the deviation of attribute groups from their initial levels. This deviation should be roughly proportional to the penalty coefficient applied -- in this instance 1 or 0. In order to illustrate this phenomenon, the average absolute deviations of item clusters and of attribute groups in each hierarchical level have been determined for the revised diets of Table 4.13 and compared in Table 4.15. As expected, where the deviation of an attribute group is penalized -- = 1 -- average absolute deviation is consistently lower relative to the condition where the deviation of that levels groups is not penalized -- P^^ = 0. For example, the average absolute deviation of groups in each hierarchical level, k=3 to k=8, is lower when the deviations jof attribute groups are penalized by the objective function (Eqn. 4.28) as compared with the case when the deviations are not penalized (Eqns. 4.8, 4.26, 4.27). Where the objective functions have equally penalized the deviations of groups on any parti cular level, the outcome is roughly dependent on the amount of emphasis on other group deviations. Thus, as emphasis shifts towards attribute groups in higher numbered hierarchical levels, there is a coincident alteration in deviation of item clusters and attribute groups at lower hierchical levels. 190 Table 4.15 Average absolute deviation of item clusters and of attribute groups in each hierarchical level from initial levels, and the penalty coefficients assigned for each hierarchical level. Hierarchical # of Items Average Absolute Deviation From SID-2* and Levels of Groups/ values of w- and P... Levei Eqn.4J8 Eqn.4.26 Eqn.4.27 Eqn.4.28 (gm)(wi/Pjk) (gm)(w./Pjk) (gm)(wi/Pjk) (gm)(w./Pjk) 127 75 1 86 1 95 1 106 50 167 0 136 1 130 1 149 39 190 0 163 0 136 1 124 34 207 0 176 0 149 o: 132 23 297 0 252 0 214 0 176 20 308 0 253 0 210 0 155 12 490 0 402 0 340 0 235 4 763 0 571 0 394 0 244 1 572 0 399 0 288 0 44 k=l k=2 k=3 k=4 k=5 k=6 k=7 k=8 values rounded to nearest gram The different treatment of deviation by these objective functions (Eqns. 4.26 - 4.28) shows in the character of the solution profiles. With respect to the number of items contained in the SID-2 revised diets, as more attribute groups are considered, the total number of items in the revised diet and the number of items included from the initial diet has consis tently decreased (foot of Table 4.13). Although this phenomenon need not occur in every instance, it is not an unreasonable outcome. In order to meet nutrient constraints the objective seeks to change the limited set of most efficient nutrient sources. As indicated previously, the quadratic penalty function restricts the extent to which any particular item can act as an efficient nutrient source. Consequently, the limited set changed may correspond to a larger number of items than would be predicted from comparing the nutrient concentration of items alone. As the deviation in attribute groups in penalized, the deviation of item clusters which act as efficient nutrient sources is amplified. The result 191 is in this instance, that as emphasis on attribute group deviation has increased, progressively more initially-consumed items have declined to zero and thus dropped from the revised diets; that is, thirteen initially-consumed items have deviated to zero when no penalty was applied on attribute group deviations (Eqn. 4.8), whereas 15, 22, and 28 items have deviated to zero as the penalty was increased (Eqns. 4.26-4.28). In these revised diets the number of new items which have entered the solution varies only slightly, between 17 and 20. Another outcome, of particular interest, is the impact of the attribute structure on concurrent deviation of item clusters. That is, does the attribute structure influence compatibility relationships between item clusters by causing these items to vary in inverse proportions, or to substitute? It is apparent from previous results that when attribute group deviations are penalized, item clusters within these attribute groups tend to substitute. That this occurs is, ipso facto, a condition of changes in the item profile of the revised diet and the simultaneous reduction in the attribute group deviation in these revised diets. It is obvious that these shifts have not occured item for item, since there has been a general reduction from the initial condition in the total number of items in the revised diet. However, if the revised diet from the objective function (Eqn. 4.8), which does not penalize deviation of attribute groups, is compared with the objective function (Eqn. 4.26) that penalizes deviation of all attribute groups on hierarchical level "1", one slight substitution effect can be illustrated (Table 4.16); that is, 41 attribute groups represented in SID-2 hierarchical level "1" are more often represented, 38 to 35, in the revised diet when those attribute groups are penalized (Eqn. 4.26) as opposed to when the attribute groups 192 are not penalized (Eqn. 4.8). Table 4.16 Effect of penalty assignment on the number of attribute groups containing consumed item clusters. # of Groups/Level Containing Consumed Item Clusters SID-2 SID-1 SID-2 Initial . Eqn. 4.8 Eqn, 4.26 k=l 50 41 35 38 Although the addition of the second term to the objective function has altered the solution in expected ways, the specific details of these changes are not easily monitored.! If such an approach for modelling concurrent changes between items is realistic, its success depends on appropriately weighting the relative acceptability of each attribute group's deviation from initial levels. 4.3.2.7 Second Term: Further Modifications of the Algorithm Apart from affecting the solutions by altering the penalty weighting for attribute group deviation, it may be useful to consider: selectively penalizing positive and negative deviation of attribute groups; and transforming the objective's second term into a constraint which restricts attribute group deviation. (i) As presently formulated, the objective function's second term (Eqn. 4.19) does not provide an opportunity to specify different penalties for positive, as opposed to negative, deviation of attribute groups. A quadratic second term can be formulated which minimizes the total, weighted-positive and weighted-negative deviation in the amounts of Hierarchy # of Groups/ Levels Level 193 attribute groups from initial consumption levels, as follows: K J (4.29) minimize ^ £ (pjk zjk > p"k z~.^ (4.30) subject to zjk - zTk = z (xn- - x?) (j = 1, 2, ..., Jk) itGjk (k = 15 2, ..., K) (4.31) ^k-'j'k*0 Where the terms are as previously defined (p. 181 ). (ii) To ensure that changes in attribute groups fall within speci fied limits, a constraint can be appropriately included in the algorithm (Eqns. 3.2-3.5), or substituted for the second term of the obj>ective function (Eqn. 4.19). This constraint provides upper and lower bounds on the deviation in the amounts of attribute groups, as follows: (4.32) bjk * z (x. - x?) * cjk (j = 1, 2, Jk) l f G jk (k = 1, 2, ... , K) Where the newly-defined terms are: bjk and Cjk are respectively, the upper and lower bounds on the deviation of attribute group j in the k^ hierarchy level. If b.. = 0 then the revised consumption of the attribute J K group j cannot be more than the original consumption of group j. If Cjk = 0 then the revised consumption of attribute group j cannot be less than the original consumption of group j. (iii) The second term of the objective function (Eqn. 4.19) only considers compatibility relationships among item clusters. Complimentarity between items has been largely neglected in the algorithm's design, 1-9.4 with the exception of choosing items with low complementary characteristics where possible. For example, bread has been chosen for the food item list rather than its ingredients -- flour, water, yeast. Therefore, it may be worthwhile to add a term to the objective function (Eqn. 3.2) which con siders complementary relationships by minimizing the total, weighted squared deviation in the ratio of the amounts of specified item clusters from initial consumption levels. The formulation is: (xn xh )2 (4.33) minimize E f., )— { (h,l = any specified (h,l)fishl ^1 x? ; pair of item clusters) Where the newly-defined terms are: x^, x° are the amounts, in grams, of specified item clusters h and 1, respectively, contained in the client's initial diet. x^, x-| are the amounts, in grams, of specified item clusters hnand 1, respectively, contained in the revised diet. s^i is the set of item cluster pairs (h.l) for which the x° deviation from the initial ratio _h is being penalized. o xl f^l is the penalty associated with the deviation of item cluster pairs (h,l) from the initial ratio _h . *° (iv) The above term (Eqn. 4.33) can be transformed to a constraint which restricts the deviation in the ratios of specific item clusters. The expression is: x. x° (4.34) 3. , > - * \i (h,l = any specified m xl x° pair of item clusters) Where the newly-defined terms are as follows: 3^1 and are respectively, the upper and lower bounds on how much 195 the ratio of item clusters h and 1 will be allowed to x° change from the initial ratio _h . o xl 196 4.4 Summary and Comments on Testing the Algorithm As indicated (p.? 1:28), testing the diet planning phase has been limited to a descriptive evaluation of some of the algorithm's characteristics --specifically, the premises and assumptions which define the acceptability of altering a diet, and the revised diets developed when these assumptions are modified. To reiterate, these assumptions include: (i) A concept of attributes which form the basic practical and conceptual entities that the individual is conditioned to respond to and uses to identify a diet's character. These attributes are used to categorize foods into a matrix of attribute groups -- an attribute map. (ii) The acceptability or perceived extent of change is considered dependent on the significance of maintaining each element ofuthe map at its initial amount with respect to itself and to other elements, and on how this significance is altered when the relative or absolute proportions of the map elements change. Within the context of the premises established for the algorithm's desi (p. 136), modifications in the assumptions about acceptability of dietary change can be proposed, and incorporated in the design of the model's objective function or constraints. Here, only some features of the objective function were explicitly modified -- specifically, the attribute elements defined and the penalty coefficients for deviation of these defined-items. The results illustrate that altering these features pro duces marked variations in the revised diets with respect to the observed parameters -- that is, deviations in the amounts of items and changes in the number of items. It is expected that the other modifications in the algorithm's objective function and constraints, described above, should also have an impact on the revised diets developed. 197 The descriptive evaluation was undertaken to explore the conceptual and technical feasibility of using this type of mathematical model for approximating an individual's nutrient-constrained food choice, given his or her diet inventory. Thus, the evaluation provides insight into the ability of the model to adapt to a variety of output demands. Although this explorative evaluation was not intended to verify the diet-planning model's operation, the implicit question throughout is which, if any, of these approaches provides the closest approximation to the nutrient-allowed food choice an individual would make given that the indi vidual could understand, accept, and utilize nutrient information. It would seem unlikely that all variations of the revised diets developed from a given dietary inventory using different objectives would be equiva lent in acceptability to the client. However, any judgement about the most appropriate formulation must be purely subjective at present. Hence, manipulation of the algorithm and observation of the solution's sensitivity to these modifications, do not immediately provide information on the perceived acceptability of the dietary revision. It may appear that the most appropriate model can be easily determined by selecting the most limiting solution -- that is, the one that deviates least from the original diet. However, as the results demonstrate, this is not a mathematical concept which can be objectively appraised without choosing one of the many possible definitions of "minimum deviation". Given any such definition the algorithm provides a mathematically-optimal solution. Therefore, evaluations to determine which algorithm provides the least-changed solution, can only serve to corroborate or refute whether the algorithm and test procedure use the same optimality criterion. In order to determine the validity of any particular algorithm for 198 approximating the client's nutrient-constrained food choice, an empirical basis is required. The mathematically-optimal deviation of the revised diet from the initial diet must be equated to a behavioral or perceived optimum; or at least an explicitly-stated tradition or professional guide-lineron which to justify revisions. In this instance, as indicated in pagel36no appropriate tradition or guidelines exists for model development or testing. Appropriate empirical methods for evaluating the efficacy of the diet-planning model would include, first, comparison of the system output with client-generated or nutritionist-generated solutions, including observation of the client's behavioral response. Second, a less rigorous evaluation of the algorithm canubecundertaken by considering the response of a client -- or a nutritionist acting as the client's agent -- to the system outputs. Such definitive evaluation was judged premature, pending further explor ative evaluation and system development, because of two major difficulties. First, establishing a test situation where the client or nutritionist could effectively use the same nutritional objectives as the computational method was not presently possible, even if the nutritional objectives could be agreed upon. This equivalent ability is important in order that solutions generated from different sources be comparable; or so that comments by client's and nutritionist's on computationally-produced diets be tempered with recognition of available nutrient-restricted options. Computer-assisted diet planning has been specifically employed in this project to ensure data handling capabilities unavailable to either clients or nutritionists without these means. Second, even if the difficulty of providing a framework for comparison of computational and human methods can be overcome, the computationally-199 generated solutions would be difficult to appraise constructively, either by comparison with human-generated solutions or by consideration of the respon ses of client's or nutritionist's. Although such evaluation could confirm or deny the acceptability of solutions, the relative acceptability of different computational solutions cannot be determined, except for exact or near-exact solution replicates, or certain acceptance of computational recommendations. While difficulties are as relevant now as they were previously, the inquiry and observation undertaken have provided a basis on which to explore more usefully the validity of the algorithm's assumptions. The observations illustrate that the algorithm can be modified to accommodate a variety of different solution requirements. Although iterating the options available may be a lengthy process, these options can be systematically evaluated to establish which, if any, of the algorithms approaches the nutrient-constrained food choices of individuals. Also, such testing may provide for further understanding of the factors which influence the mathematical formulation in producing a solution judged to be useful. Perhaps of more importance, it may provide an opportunity to explore measurably the trade offs between the nutritional goals of presenting accurate nutrient infor mation, and the educational objectives of presenting simple, straightforward information. 200 CHAPTER 5 SUMMARY, RECOMMENDATIONS, AND CONCLUSIONS 5.1 Summary 5.1.1 The Thesi s Goal The overall thesis goal was to develop a system to provide information on dietary practices for adults who are motivated to apply nutritional principles to their eating habits, and who have sufficient resources to make use of the information required to define healthful dietary practices for them. This development was undertaken bearing in mind two primary problems, namely: the developing of nutritional guidelines suitable for health promotion, and the communication of these guidelines to individuals. 5.1.2 First Objective of the Thesis The prototypical computerized system developed has two major functions: first, to assess the dietary intake of individuals (diet-assessment), and second, to recommend changes in food intake for those individuals with nutrient intake which do not meet specified limits (diet-planning). 5.1.2.1 Diet-Assessment Function Standard dietary assessment procedures of data collection, analysis, and evaluation, were adopted for the system's design. In the data collection phase, the initial diet of an individual is estimated by using an intake questionnaire. This information is then translated into nutrient data, in the data analysis phase, and evaluated in the evaluation phase by 201 comparing calculated nutrient-intake with nutrient limits which represent an "acceptable" range of intake for the individual. The list of food consumed, and the results of data analysis and evaluation are displayed in the computer output. The nutritional guidelines used for evaluating diets are largely based on the dietary standards of various countries and international agencies, and on dietary goals proposed by recognized scientific agencies. These guidelines are adopted for the system's design because they provide a technically accurate rationale for healthful food selection practice. However, it is recognized that this complex information on nutrients cannot easily be used to recommend dietary practices for individuals, and therefore would require translation into a statement of food and meals. 5.1.2.2 Diet-Planning Function A revised diet is developed by a constrained-optimization model which determines the combination of foods, subject to nutrient constraints, that minimizes the total squared deviation of food items and item groups from their original amounts as specified on the client questionnaire. The revised diet is provided on the computer output. This model has been developed to provide a recommended dietifor/motivated individuals which: (i) accurately reflects nutritional guidelines, and (ii) facilitates adoption of recommendations by providing a self-explanatory statement of foods to consume and by limiting the changes from present or desired food patterns. Presumably, this would be similar to the diet which the individual would choose if she or he understood, accepted, and used nutritional knowledge 202 -- that is, the individual's nutrient-constrained food choice. The use of mathematical modelling provides an effective means of collating the vast amount of data required to develop dietary recommendations which are both nutritionally accurate, straightforward, and hopefully, acceptable to the client. 5.1.3 Second Objective of the Thesis Following formulation of the diet-planning algorithm, the major task was to test this algorithm in order to explore the conceptual details of this diet-planning approach. Testing was restricted to a descriptive evaluation of some of the algorithm's characteristics -- specifically, the assumptions which define the acceptability of altering a diet and the revised diets developed when these assumptions are modified. The assumptions defined included: first, a concept of attribute elements which form the basic practical and conceptual entities that the individual is conditioned to respond to and uses to identify a diet's character; second, the acceptability or perceived extent of change is considered to be dependent on both the significance of maintaining each element at its original amount with respect to itself and to other elements, and on the extent to which this significance is altered when the relative or absolute proportions of the map elements change. A variety of modifications in the assumptions about acceptability of dietary change can be incorporated in the algorithm's design. In this instance, only certain features of the objective function -- specifically, the attribute elements defined and the penalty coefficients weighting deviation of these attribute elements -- were explicitly modified. The results illustrate that altering these features produced marked variations 203 in the revised diets with respect to the observed parameters, that is, deviation in the amounts of items and changes in the number of items. Other modifications in the algorithm's objective function and constraints would also be expected to have an impact on the revised diets developed. Although the explorative evaluation was not intended to validate whether the diet planning model can provide acceptable revisions of diets for clients, it does provide a foundation for more systematic evaluation of the validity of different approaches. As noted the algorithm can be modified to accommodate a variety of different solution requirements. These options can by systematically tested, by considering client or nutritionist-generated solutions or their responses, to establish which, if any, of the possible algorithms approximate the nutrient-constrained food choice of individuals. 5.2 Recommendations Since iterating the available approaches may be a lengthy process, some recommendations which may facilitate exploration and development of a "commercially" viable model are now given. 5.2.1 First Recommendation Emphasis should be given to reducing the absolute number of changes in revised diets rather than just minimizing deviation of item amounts. It became apparent while analyzing the results, that even though the revised diets.deyiate minimally and in fact could represent the real adjustments that individuals would make, these revisions were difficult to appreciate because so many changes had occured. In a nutrition education setting, realistic solutions may need to maintain many items at initial levels. 204 Some alterations in the algorithm that may reduce changes in the revised diet to a manageable level include: (i) The use of a linear rather than a quadratic formulation. Although the quadratic penalty function may seem more realistic than the linear function, the quadratic function tends to spread deviation over all items. A linear function may provide fewer changes and thereby a more easily understood solution, (ii) A formulation which restricts the items in the solution set to the originally consumed items and to a limited number of alter natives for guaranteeing solution feasibility may simplify out comes; as might one which concentrates changes to specified types of items, such as low nutrient foods. Although the total deviation of item quantities may be greater using these means, the alter ations may be easier for the client to monitor and appreciate, (iii) Constraints on the deviation of items could be included to control the solutions, and would be particularly useful for a linear format. Some items may usefully be retained at initial levels. (iv) In place of an integer formulation which would be difficult to develop, a routine to average item quantities in the revised diets to their nearest serving or half-serving should be considered. The loss of accuracy may be considerably offset by the increase in utility from removing inordinately small increments and decrements, and from 'simplifying the' serving sizes reported;.) (v) Solutions based on a more abbreviated set of item classes or food groups may be easier to comprehend. Aggregating foods into groups and averaging over nutrient values would necessarily reduce the nutritional accuracy of the solutions. However, the increased 205 fulness of the information for illustrating dietary change may substantially offset this disadvantage. 5.2.2 Second Recommendation Only information on the client's initial diet inventory, or alternatively a food plan identified as desirable was used to define modifications of that inventory. This was done to limit the amount of information required from the client on the questionnaire. However, the client (or the client's counsellor) could usefully provide information on the relative acceptability of deviating from initial levels in order to more acceptably direct revisions of the diet. For example, the client could indicate whether the relative acceptability of either an increase or decrease in the amount presently con sumed (or not consumed) is: high, indeterminate, or low. These responses can then be incorporated into the algorithm's penalty coefficients and/or constraints to adjust the solution outcomes. The initially-assigned values for the penalty coefficients and/or constraints will be arbitrary until the relationship of individual response to coefficient assignment is appropriately indexed. 5.3 Conclusions The use of mathematical modelling and computer technology has provided an effective means of collating the vast amount of data required to develop cogent dietary recommendations which are nutritionally accurate, straight forward, and potentially acceptable to the individual. The concept of minimizing deviation from an initial dietary inventory seems a feasible approach to establish the link between these mathematical programming 206 techniques and the objectives of diet planning -- that is, to approximate the individual's nutrient-constrained food choice. 207 LIST OF REFERENCES Action B.C. Nutrition Evaluation Program. Vancouver, B.C., 1976. Adams, C.F. Nutritive Value of American Foods in Common Units. United States Department of Agriculture, Agriculture Handbook No. 456, 1975. Adelson, S.F. Some problems in collecting dietary data from individuals. J. Amer. Dietet. Assoc. 36: 453, 1960. Ahlstrom, A. and Rasanen, L. Review of food grouping systems in nutrition education. J. Nutr. Educ. 5: 13, 1973. American Dietetic Association. Position paper on nutrition education for the public. J. Amer. Dietet. Assoc. 62: 429, 1973. American Heart Association. Diet and coronary heart disease. 1973. Andrews, J.T., Moore, A.N. and Tuthill, B.H. Electronic data processing in intra-departmental food cost accounting. J. Amer. Dietet. Assoc. 51: 332, 1967. Anon. Nutrition education for youth. J. Home Econ. 64: 34, 1972. Anon. Food labeling, Federal Register 38: 2124, 1973; Food labeling, Federal Register 38: 6950, 1973; Food nutrition labeling, Federal  Register 38: 6959, 1973; Labeling of foods in relation to fat, fatty acid, and cholesterol contant, Federal Register 38: 6963, 1973. Armstrong, R.D. and Balintfy, J.L. A chance-constrained multiple-choice programming algorithm. Operations Research 23: 494, 1975. Armstrong, R.D.'and Sinha, P. Application of quasi-integer programming to the solution of menu planning problems with variable portion size. Management Science 21: 474, 1974. Arroyave, G. Standards for the diagnosis of vitamin deficiency in man. In: Proceedings of the Special Session, Ninth Meeting of PAHO Advisory Committee on Medical Research, 16 June, 1970. Pan American Health Organization, Washington, D.C., 1971, p. 88. Asenjo, C.F. Variation in nutritive value of foods. Amer. J. Clin. Nutr. 11: 368, 1962. Astrand, P-0. Nutrition and physical performance. World Rev. Nutr. Diet. 16: 59, 1973. Astrand, P-0. and Rodahl, K. Textbook of Work Physiology. McGraw-Hill Book Company, Toronto, 1970. Atwater, W.0. Methods and Results of Investigations on the Chemistry and  Economy of Foods. United States Department of Agriculture, Bulletin No. 21, 1895. 208 Baker, M.J. Influence of nutrition education on fourth and fifth graders. J. Nutr. Educ. 4: 55, 1972. Balintfy, J.L. Mathematical Programming of Menu Planning in Hospitals. Tulane University, School of Business, New Orleans, 1963. Balintfy, J.L. Menu planning by computer. Communications of the ACM 7: 255,1964. Balintfy, J.L. Linear programming models for menu planning. In_: Hospital Industrial Engineering. Edited by H.E. Smalley and J.R. Freeman. Reinhold, New York, 1966, p. 402. Balintfy, J.L. System/360 Computer Assisted Menu Planning. IBM Contributed Program Library, 360D-15.2.013, IBM Corporation, PID, Hawthorne, New York, 1969. Balintfy, J.L. Computer-assisted menu planning and food service. In: Handbook of Biomedical Information Systems. Edited by W. Spector Encyclopedia Britannica, Inc., Chicago, 1971, p. 85. Balintfy, J.L. Mathematical modeling and human nutrition. Science 181: 581 , 1973a. Balintfy, J.L. Nonlinear Programming and the Food Price Index. Technical Report No. 11, School of Business Administration, University of Massachusetts, Amherst, Massachusets, 1973b. Balintfy, J.L. Large-scale programming properties of menu planning and sched uling./ In: Optimization Methods for Resource Allocation. Edited by R.W. Cottle and J. Krarup. The English Universities Press Ltd., London, 1974a, p. 81. Balintfy, J.L. Mathematical models for menus. Naval Research Reviews 27: 1 , 1974b. Balintfy, J.L. A mathematical programming system for food management applications. Interfaces 6: 13, 1975. Balintfy, J.L. Applications of mathematical programming to optimize human diets. Istituto Nationale di Matematica. Symposia Mathematica. XIX: 313, 1976a. Balintfy, J.L. A scientific and democratic approach. The Cornell H.R.A. Quarterly 17: 33, May 1976b. Balintfy, J.L., Duffy, W.J. and Sinha, P. Modelling food preferences over time. Operations Research 22: 711, 1974. Balintfy, J.L. and Nebel, E.C. Experiments with computer-assisted menu planning. Hospitals 40: 88, June 16, 1966. Balogh, M., Kahn, H.A. and Medalie, J.H. Random repeat 24-hour recalls. Amer. J. Clin. Nutr. 24: 304, 1971. 2Q9 Balogh, M., Medalie, J.H., Smith,'H. and Groen, J.J, The development of a dietary questionnaire for an ischaemic heart disease survey. Israel  J. Med. Sci. 4: .195; 1968. Beal, V.A. The nutritional history in longitudinal research. J. Amer. Dietet. Assoc. 51: 426, 1967 Beaton, G.H. The use of nutritional requirements and allowances. J^n: Western Hemisphere Nutrition Congress III. Edited by P.L. White. Futura Publishing Co., New York, 1972, p. 356. Beaton, G.H. Evaluation of nutrient intake: New statistical approaches. Arch. Latino, de Nutricon 25: 31, 1975. Beckman, M.J. On the determination of an adequate diet at minimum cost. Trabajos de Estadisica 11: 139 1960. (Misprint corrected in Smith, V.E. Electronic Computation of Human Diets. MSU Business Studies, Bureau of Business and Economic Research, Graduate School of Business Administration, Michigan State University, East Lansing, Michigan, 1963.) Bell, C.G. and Lamb, M.W. Nutrition education and dietary behavior of fifth graders. J. Nutr. Educ. 5: 196, 1973. Berg, A. The Nutrition Factor. The Brookings Institution, Washington, D.C., 1973. Berlo, D.K. The Process of Communication. Rinehart and Winston, New York, 1960. Bogert, L.T., Briggs, G.M. and Calloway, D.H. Nutrition and Physical Fitness, 9th edition. W.B. Saunders Company, Toronto, 1973. Bowman, J.L. and Brennan. Computer-assisted menu planning provides control of food service. Hospitals 43: 107, August 16, 1969. Boysen, S.C. and Ahrens, R.A. Nutrition education and lunch surveys with second graders. J. Nutr. Educ. 4: 172, 1972. Bransby, E.R., Daubney, C.G. and King, J. Comparison of nutrient values of individual diets found by calculation from food tables and by chemical analysis. Brit. J. Nutr. 2: 232, 1948. Bransby, E.R., Daubney, C.G. and King, J. Comparison of results obtained by different methods of individual dietary surveys. Brit. J. Nutr. 2: 89, 1948. Bransby, E.R., Berry, W.I.C. and Taylor, D.M. Study of the Vitamin-D intakes of infants in 1960. Brit Med. J. i: 1661, 1964. Briggs, G.M. The need for nutrition education. J. Nutr. Educ. 1: 7, 1969. Brisbane, H.N. Computing menu nutrients by data processing. J. Amer. Dietet. Assoc. 44: 453, 1964. 210 Browe, J.H., Morlley, D.M., McCarthy, M.C. and Gofstein, R.M. Diet and heart disease study in the cardiovascular health center. II. Construction of a food composition table and its use. J. Amer. Dietet. Assoc. 48: 101, 1966. Brown, E.L. College students look at the basis for their food habits. J. Home Econ. 59: 784, 1967. Brown, J.A.C. An experiment in demand analysis: The computation of a diet problem on the Manchester computer. J_n_: Conference on Linear Programming, May 1954. Edited by Anonymous. Ferranti, Ltd., London, 1954, p. 41. Burke, B.S. The dietary history as a tool in research. J. Amer. Dietet. Assoc. 23: 1041, 1947. Burke, B.S., Beal, V.A., Kirkwood, S.B. and Stuart, H.G. Nutrition studies during pregnancy. Amer. J. Qbstet. Gynec. 46: 38, 1943. Burkitt, D.P. and Painter, N.S. Dietary fiber and disease. J. Amer. Med. Assoc. 299: 1068, 1974. Buzina, R., Ferber, E., Keys, A., Brodarec, A., Agneletto, B. and Horvat, A. Diets of rural families and heads of families in two regions of Yugoslavia. Voeding 25: 629,1964. Buzina, R., Keys, A., Brodarec, A. and Fidanza, F. Dietary surveys in rural Yugoslavia. III. Comparison of three methods. Voeding 27: 99, 1966. Calavan, M.M. A model of food selection in northern Thailand. Ecol. Food  Nutr. 5: 63, 1976. Canada. Nutrition Canada, Department of National Health and Welfare. Nutrition: A National Priority. Information Canada, Ottawa, 1973. Canada. Bureau of Nutritional Sciences, Department of National Health and Welfare. Dietary Standards for Canada. Information Canada, Ottawa, 1975. Canada. Department of National Health and Welfare. Report of the Committee  on Diet and Cardiovascular Disease. Ottawa, 1976a. Canada. Statistics Canada. Vital Statistics 1974. Catalogue 84-206, Ottawa, 1976b. Canada. Bureau of Nutritional Sciences, Department of National Health and Welfare. Food Consumption Pattern Report. Information Canada, Ottawa, 1977. Canada. Bureau of Nutritional Sciences, Department of National Health and Welfare. Canada's Food Guide Handbook. Ottawa, 1977. 211 Cantoni, M., Paffenbarger, R.S. and Krueger, D.E. Methods of dietary assessment in current epidemiological studies of cardiovascular diseases. Amer. J. Pub!. HIth. 51: 70,1961. Carmel, J. The prediction of diets of high energy and protein value by linear programming. Ecol. Food Nutr. 5: 161, 1976. Caso, E.K. Calculation of diabetic diets. J. Amer. Dietet. Assoc. 26: 575, 1950. Chalmers, F.W., Clayton, M.M., Gates, L.O., Tucker, R.E., Wertz, A.W., Young, CM. and Foster, W.D. The dietary record - How many and which days? J. Amer. Dietet. Assoc. 28: 711, 1952. Chamber!in, J.G. and Stickney, R.E. Improvement of children's diets in developing countries: an analytic approach to evaluation of alterna tive strategies. Nutr. Rept. Intern. 7: 71, 1973. Chandler, C.A. and Perloff, B.P. Updating food composition values - A progress report. U.S. Department of Agriculture, Nutrition Program  News, September - December, 1975, p. 1. Chappell, G.M. Long term individual dietary studies. Brit. J. Nutr. 9: 323, 1955. Cheney, M.C. Food Enrichment: Nutritional Standards for Synthetic and  Modified Foods. Paper presented at Current Topics in Food and Nutri tion - Workshop. University of British Columbia, Vancouver, B.C., July 27, 1976. Christakis, G. Nutritional assessment in health programs. Amer. J. Publ. Hlth. 63 Part II (Suppl.): 1, 1973. Church, C.F. and Church, H.N. Food Portions Commonly Used, 1.2th edition. J.B. Lippincott Company, New York, 1975. Clements, F.W. Nutrition 11 - Recommended dietary allowances. Med. J. Austr. 2: 24, 1975. Combs, G.F. and Wolfe, A.C. Methods used in dietary survey of civilians in Ecuador. Publ. Hlth. Rep., Wash. 75: 707, 1960. Conner, M.M. A guiding star for modern nutrition. J. Home Econ. 59: 734, 1967. Comely, P.B. There is more to health than just paying bills. Am. J. Public Health 64: 845,1974. Cornfield, J. Letter to Tore Dalenius. Nov. 29, 1951. Unpublished (See Dorfman, R., Samuelson, P. and Solow, R.M. Linear Programming and Economic Analysis. McGraw-Hill Book Co., Inc., New York, 1958; and Smith, V.E. Electronic Computation of Human Diets. MSU Business Studies, Bureau of Business and Economic Research, Graduate School of Business Administration, Michigan State University, East Lansing, Michigan, 1963.) 212 Correa, H. Measured influence of nutrition on socio-economic development. World Rev. Nutr. Diet. 20: 2,1975. Cottle, R.W. and Dantzig G.B. Complementary pivot theory of mathematical programming. Linear Algebra and Applications. 1: 103, 1968. Dantzig, G.B. Linear Programming and Extensions. Princeton University Press, Princeton, New Jersey, 1963. Davenport, E. Calculating the Nutritive-Value-of Diets. ARS-62-10-1. Agriculture Research Service. USDA, Washington, D.C, 1964. Davis, D.R. and Williams, R.J. Potentially useful criteria for judging nutritional adequacy. Amer. J. Clin. Nutr. 29: 710, 1976. Dawber, T.R., Pearson, G., Anderson, P., Mann, G.V., Kannel, W.B., Shirtleff, D. and McNamara, P. Dietary assessment in the epidemiologic study of coronary heart disease: The Framingham study. II. Reliabil ity of measurement. Amer. J. Clin. Nutr. 11: 226, 1962. Dorfman, R., Samuel son, P. and Solow, R.M. Linear Programming and  Economic Analysis. McGraw-Hill Book Co. Inc., New York, 1958. Durnin, J.V.G.A. Appetite and the relationship between expenditure and intake of calories in man. J. Physiol., Lond. 156: 294, 1961. Eagles, J.A., Whiting, M.G. and Olson, R.E. Dietary appraisal - Problems in processing dietary data. Amer. J. Clin. Nutr. 19: 1, 1966. Eckstein, E.F. Menu planning by computer: The random approach. J. Amer. Dietet. Assoc. 51: 529, 1967. Eddison, P. Nutrimetrics 07. Cybus Systems Limited, Suite 604, 350 Sparks Street, Ottawa, Canada, 1975. Emmons, L. and Hayes, M. Nutrition knowledge of mothers and children. J. Nutr. Educ. 5: 134, 1973. Evans, S.N. and Gormican, A. The computer in retrieving dietary history data. I. Designing and evaluating a computerized diabetic dietary history. J. Amer. Dietet. Assoc. 63: 397, 1973. Fidanza, F., Fidanza-Alberti, A., Ferro-Luzzi, G. and Proja, M. Dietary surveys in connection with the epidemiology of heart disease: Results in Italy. Voeding 25: 502, 1964. Fleetwood, M., Moore, A.N., Andrews, J,T. and Tuthill, B.H. Food production and control system at the medical center, UMC. Computer-assisted Food Management Systems. Edited by A.N. Moore and B.H. Tuthill. University of Missouri, Columbia, Missouri, 1974, p. 47. Flook, M.G. and Alfred, B.B. Cost comparison of systems for nutrient analy sis - Computer vs. manual calculations. J. Amer. Dietet. Assoc. 64: 62, 1974. 213 Florencio, C.A. and Smith, V.E. Efficiency of food purchasing among working-class families. J. Amer- Dietat. Assoc. 55: 239, 1969. Florencio, C.A. and Smith, V.E. Towards more efficient expenditure for food among Columbian families. Nutr. Rept. Intern. 4: 207, 1970. Flores, M., Flores, Z. and Lara, M.Y. Estimation of family and mother's dietary intake comparing two methods. Trop. Geogr. Med. 17: 135, 1965. Food and Agricultural Organization, World Health Organization (FA0/WH0). Dietary Surveys - Their Techniques and Interpretation. FAO Nutr. Studies. Washington, D.C., 1949. Fremes, R. and Sabry, Z. Eating Right: New food guide needed. Vancouver  Sun Newspaper, October 13, 1976, p. 90. Gelpi, M.J., Balintfy, J.L., Dennis, L.C. and Findorff, I.K. Integrated nutrition and food cost control by computer. J. Amer. Dietet. Assoc. 61: 637, 1972. George, J.D. A nutrition education unit for a sixth grade. J. Nutr. Educ. 2: 111, 1971. Gifft, H.H., Washbon, M.B. and Harrison, G.G. Nutrition, Behavior and Change. Prentice-Hall, Inc., Englewood Cliffs, New Jersey, 1972. Goodhart, R.S. Criteria of an adequate diet. In: Modern Nutrition in Health and Disease, 5th edition. Edited by R.S. Goodhart and M.E. Shils. Lea and Febiger, Philadelphia, 1973, p. 403. Goodhart, R.S. and Shils, M.E. editors. Modern Nutrition in Health and  Disease, 5th edition. Lea and Febiger, Philadelphia, 1973. Goodloe, M.H.R., Waters, G., Quinlan, C.B and Barrows, J.G. Studies in atherosclerosis. 4. Electronic computers in the processing of dietary data. Amer. J.. Clin. Nutr. 13: 304, 1963. Groover, M.E., Boone, L., Houk, P.C. and Wolf, S. Problems in the quantita tion of dietary surveys. J. Amer. Med. Assoc. 201: 8, 1967. Gue, R.L. and Liggett, J.C. Mathematical programming models for hospital menu planning. J. Industrial Engineering 17: 395, 1966. Guggenheim, K., IIan, J., Vago, T. and Mundel, G. Nutritional status and food consumption of pregnant women in a development area of Israel. Brit. J. Nutr. 14: 347, 1960. Gussow, J. Counter-nutritional messages of TV ads aimed at children. J. Nutr. Educ. 4: 48, 1972. Hall, R.L. Food additives. Nutr. Today 8: 20, 1973. 214 Hall, R.L. Is nutrition a stagnating science? New Scientist 65: 7, 1975. Hall, R.L. Safe at the plate. Nutr; Today. 12: 6, 1977. Hampton, M.C., Huenemann, R.L., Shapiro, L.R. and Mitchell, B.W. Caloric and nutrient intakes of teenagers. J. Amer. Dietet; Assoc. 50: 385, 1967. Hankin, J.H. and Huenemann, R.L. A short dietary method for epidemiologic studies. I. Developing standard methods for interpreting seven-day measured food records. J. Amer. Dietet. Assoc. 50: 487, 1967. Hankin, J.H., Reynolds, W.E. and Margen, S. A short dietary method for epidemiologic studies. Amer. J. Clin. Nutr. 20: 935, 1967. Hansen, R.G. An index of food quality. Nutr. Rev. 31: 1 , 1973. Hanson, W.A. Medical computerized nutritional analysis. J. Applied Nutr. 21: 74, 1969. Harper, A.E. Recommended Dietary Allownaces: Are they what we think they are? J. Amer. Dietet. Assoc. 64: 151, 1974. Harper, A.E. U.S. Dietary Goals - Against. J. Nutr. Educ. 9: 154, 1977. Harris, R.S. Reliability of nutrient analysis and food tables. Amer. J. Clin. Nutr. 11: 377, 1962. Hart, M.L. and Cox, A.G. A comparison of dietary analysis methods during a computer. Nutrition, Lond. 21: 146, 1967. Hartog, C. Den, Van Schaik, T.F.S.M., Dalderup, L.M., Drion, E.F. and Mulder, T. The diet of volunteers participating in a long term epidemiological field study on coronary heart disease at Zutphen, The Netherlands. Voeding 26: 184, 1965. Hayes, O.B., Abraham, S. and Caceres, CA. Computers in epidemiologic dietary studies. J. Amer. Dietet. Assoc. 44: 456, 1964. Hayes, O.B., Trulson, M.F. and Stare, F.J. Suggested revisions of the basic 7. J. Amer. Dietet. Assoc. 31: 1103, 1955. Head, M.K., Weeks, R.J. and Gibb, E. Major nutrients in type-a lunch. J. Amer. Dietet. Assoc. 63: 620, 1973. Hegsted, D.M. Nutrition, bone, and calcified tissue. J. Amer. Dietet. Assoc. 50: 105, 1967. Hegsted, D.M. Dietary standards. J. Amer. Dietet. Assoc. 66: 13, 1975. Hertzler, A.A..and Anderson, H.L. Food guides in the United States. J. Amer: Dietet. Assoc. 64: 19, 1974. 215 Hertzler, A.A. and- Hoover, L.W. Development of food tables and use with computers. J. Amer. Dietet. Assoc. 70: 20, 1977. Hill, M.M. Nutrition committees and nutrition education. J. Nutr. Educ. 1: 14, 1969. Hinton, M.A., Eppright, E.S., Chadderton, H. and Wolins, L. Eating behavior and dietary intake of girls 12 to 14 years old. J. Amer. Dietet. Assoc. 43: 223, 1963. Hjortland, M.C., Duddleson, W.G., Porter, C. and French, A.B. Using the computer to calculate nutrients in metabolic diets. J. Amer. Dietet. Assoc. 49: 316, 1966. Hoover, L.W. Computers in dietetics: State-of-the-art, 1976. J;. Amer. Dietet. Assoc. 68: 39, 1976. Huenemann, R.L., French, F.E. and Bierman, J.M. Diets of pregnant women in Kauai, Hawaii. J. Amer. Dietet. Assoc. 39: 569, 1961. Huenemann, R.L. and Turner, D. Methods of dietary investigation. J. Amer. Dietet. Assoc. 18: 562, 1942. Interdepartmental Committee on Nutrition for National Defense (ICNND). Manual.for Nutrition Surveys, 2nd edition. ICNND, National Institute of Health,"-Bethesda, Maryland, 1963. Jansen, C, Dupont, J. and Blaker, G.G. A tool for individualized management of fat-controlled diets. J. Amer. Dietet. Assoc. 67: 28, 1975. Jelliffe, D.B. The Assessment of the Nutritional Status of the Community. World Health Organization, Geneva, 1966. Jelliffe, D.B. Nutrition in early childhood. World Rev. Nutr. Diet. 16: 1 , 1973. Jelliffe, D.B. and Jelliffe, E.F.P., editors. Symposium: The uniqueness of human milk. Amer. J. Clin. Nutr. 24: 968, 1971. Johnson, N.E., Nitzke, S. and VandeBerg, D.L. A reporting system for nutrient.adequacy. Home Economics Research J. 2: 211, 1974. Kansra, A.C. Srinivasan, N. and Gopalda, T. Least-cost multi-mix supplements: Development of formula to fill dietary deficits of pre-school children. Ind. J. Nutr. Dietet. 11: 276, 1974. Keen, H. and Rose, G.A. Diet and arterial disease in a population sample. Brit. Med. J. i: 1508, 1958. Keys, A., Aravenis, D. and Sdrin, H. The diets of middle-aged men in two rural areas of Greece. Voeding 27: 575, 1966. Kinder, F. Meal Management, 4th edition. The Macmillan Company, N.Y., 1973. 216 Kitchen, A.H., Passmore, R., Pyke, M. and Warnock, G.M. Studies of the diet of students at Edinburgh University. Brit, J. Soc. Med. 3: 10, 1949. Klevay, L.M. Coronary heart disease and dietary fiber. Amer. J. Clin. Nutr. 27: 1202, 1974. Krehl, W.A. The evaluation of nutritional status. Med.Clin. N. Amer. 48: 1129, 1964. Kugler., H.J. Nutrition, Health and Activity Profile. Pacific Research Systems, 2222 Corinth Avenue, Los Angeles, California, 1976. Lachance, P.A. Food guides. J. Nutr. Educ. 4: 44, 1972. Lalonde, M. A New Perspective on the Health of Canadians. Information Canada, Ottawa, 1974. Latham, M.C. and Stephenson, L.S. U.S. Dietary Goals - For. J. Nutr. Educ. 9: 152, 1977. Leichsenring, J.M. and Wilson, E.D. Food composition table for short method of dietary analysis (2nd revision). J. Amer. Dietet. Assoc. 27: 386, 1951. : Lemke, C.E. On complementary pivot theory. In^: Mathematics of the Decision Sciences. Edited by G.B. Dantzig and A.F. Veinott. American.Mathe matical Society, Providence, Rhode Island, 1968, p. 95. Leverton, R.M. What is nutrition education? J. Amer. Dietet. Assoc. 64: T.7, 1974. Leverton, R.M. The RDA's are not for amateurs. J. Amer. Dietet. Assoc. 66: 9, 1975. Lewis, H.C. and Peng, P.H. The three-consideration diet revisited. J_. Amer. Dietet. Assoc. 70: 270, 1977. Litman, T.J., Cooney, J.P. and Stief, R. Views of Minnesota school children on food, J. Amer. Dietet. Assoc. 45: 433, 1964. Lorstad, M.H. Recommended intake and its relation to nutrient deficiency. F.A.O. Nutr. Newsletter 9: 18, 1971. Lowenberg, M.E., Todhunter, E.D., Wilson, E.O., Savage, J.R. and Lubawski, J.L. Food and Man, 2nd edition. John Wiley and Sons, Toronto, 1974. Lund, L.A. and Burk, M.C. A Multidisciplinary Analysis of Children's Food  Consumption Behavior. Tech. Bull. 265. Agriculture Experimental " Station, University of Minnesota, St. Paul, Minnesota, 1969. MacMillan, J.B. and Smith., E.-B. Development of a lacto-ovo vegetarian food guide. J. Can. Dietet. Assoc. 36: 110, 1975. 217 Mann, G.V. Diet-heart: End of an era. New Engl. J. Med. 297: 644, 1977. Mann, G.V., Pearson, G., Gordon, T. and Dawber, T. Diet and cardiovascular • disease in the Framingham study. I. Measurement of dietary intake. Amer. J; Clin. Nutr. 11: 200,1962. Manoff, R.K. Potential uses of mass media in nutrition programs. J. Nutr. Educ. 5: 125, 1973. Manoff, R.K, Towards a new "American Mentality". J. Nutr. Educ. 7: 139, 1975. Marr, J.W. Individual dietary surveys: Purposes and methods. World Rev. Nutr. Diet. 13: 105, 1971. Marr, J.W., Heady, J.A. and Morris, J.N. Repeat individual weighed dietary surveys. Proc. Nutr. Soc. 18: xii, 1959. Mayer, J. Towards a national nutrition policy. Science 176: 237, 1972. Mayer, J. Forward. World Rev. Nutr. Diet. 16: xxi, 1973. Mayer, J. Time for reappraisal. J. Nutr. Educ. 7: 8, 1975. McBean, L.O. and Speckmann, E.W. A Review: The importance of nutrition in oral health. J. Amer. Dent. Assoc. 89: 109, 1974. McCann, M.B., Trulson, M.F. and Stare, F.J. Follow-up of serum cholesterol, diet and physical findings of Italian-American factory workers. Amer. J. Clin. Nutr. 9: 351 , 1961.' McClinton, P., Milne, H. and Beaton, G. An evaluation of food habits and nutrient intakes in Canada: Design of effective food guides. Can. J. Public Health 62: 139,1971. McKenzie, J.L. and Mumford, P. Evaluation of nutrition education programs: A review of the present situation. World Rev. Nutr. Diet. 5: 21, 1965. McNabb, M.E. 90-day nonselective menus by computer. Hospitals 45: 88, May 16, 1971. Meneely, G.R. and- Battarbee, H.D. Sodium and Potassium. In: Present Knowledge of Nutrition, 4th edition. Nutrition Foundation, New York, 1976, p. 259. Mertz, W. Human requirements: Basic and optimal. Ann. N.Y. Acad. Sci. 199: 191 , 1972. Milne, H., Kerr, C. Trenholme, M. and Beaton, G.H. Studies of teenage eating in Ontario. II. Evaluation of a diet scoring method. Can. J. Public  Health 54: 463, 1963. Mitchell, D. The Politics of Food. James Lorimer and Company, Toronto, 1975. 218 Mongeau, E. Dietary assessment of nutritional status. r_n: Assessment of Nutritional Status, Miles Symposium '73. Miles Laboratories, Limited, 1974, p. 13. Moore, J.L. and Wendt, P.F. Nutrition labeling - A summary and evaluation. J. Nutr. Educ. 5: 121, 1973. Morris, J.N., Marr, J.W., Heady, J.A., Mills, G.L. and Pilkington, T.R.E. Diet and plasma cholesterol in 99 bankmen. Brit. Med. J. i: 571, 1963. Morrison, S.D., Russel, F.C. and Stevenson, J. Estimating food intake by questioning and weighing: A one-day survey of eight subjects. Brit. J. Nutr. 3: v, 1949. Neter, J. and Wasserman, W. An experimental comparison between fixed weight and linear programming food price indexes. J. Amer. Stat. Assoc. 65: 49, 1970. Ohlson, M.A. Experimental and Therapeutic Dietetics, 2nd edition. Burgess Publishing Co., Minneapolis, Minnesota, 1972. Page, L. and Phipard, E.F. Essentials of and adequate diet. USDA Agriculture Research Service, Home Economics Research Report 3, 1957. Pao, E.M. and Burke, M.C. A computer-assisted approach to meal patterning. J. Amer. Dietet. Assoc. 65: 144, 1974. Passmore, R., Nicol, B.M., Narayana Rao, M., Beaton, G.H. and Demayer, E.M. Handbook on Human Nutritional Requirements. World Health Organization, Geneva, 1974. Paul, 0., Lepper, M.H., Phelan, W.H., Dupertuis, G.W., MacMillan, A., McKeen, H. and Park, H. .A longitudinal study of coronary heart disease. Circulation 28: 20, 1963. Pauling, L. Vitamin C and the common cold and the flu. W.H. Freeman and Company, San Francisco, 1976. Payton, E., Crump, E.P. and Horton, CP. Growth and development. VII. Dietary habits of 571 pregnant Southern Negro women. J. Amer. Dietet. Assoc. 37: 129, 1960. Pekkarinen, M. Chemical analysis in connection with dietary surveys in Finland. Voeding 28: 1,1967. Pekkarinen, M. Methodology in the collection of food consumption data. World Rev. Nutr. Diet. 12: 145, 1970. Pekkarinen, M. andRoine, P. Studies on the food consumption of the rural population in east and west Finland. Ann. Med. Exp. Fenn. 42: 93, 1964. Pekkarinen, M., Kivioja, S. and Jortikka, L. A comparison of the food intake of rural families estimated by one-day recall and precise weighing methods. Voeding 28: 470, 1967. 219 Pennington, J.A. Dietary Nutrient Guide. The AVI Publishing Company, Incorporated, Westport, Connecticut, 1976. Pike, R.L. and Brown, M.L. Nutrition -An Integrated Approach, 2nd edition, John Wiley and Sons Incorporated, New York, 1975. Poolton, M.A. Predicting application of nutrition education. J. Nutr. Educ. 4: 110, 1972. Robinson, CH. Nutrition education - what comes next? J. Amer. Dietet. Assoc. 69: 126, 1976. Rosenstock, I.M., Hoefner, D.P., Kegeles, S.S.. and Kirscht, J.P. Public knowledge, opinion and action concerning three public health issues: radioactive fallout, insect and plant sprays, and fatty foods. J. Health and Human Behavior. 7: 91, 1966. Sager, J.F. The United States Army Hospital food service data processing system. J_n: Computer-assisted Food Management Systems. Edited by A.N. Moore and B.H. Tuthill. University of Missouri, Columbia, Missouri, 1974, p. 75. Schaum, K.D. Patient-oriented dietetic information system. III. A computer-assisted program for calculating nutrient intake. J. Amer. Dietet. Assoc. 63: 154, 1973. Schaum, K.D. and Sharp, J.L. Patient-oriented dietetic information system. I. A computerized menu print and diet order system. J. Amer. Dietet. Assoc. 63: 35, 1973. Smith, E.B. A guide to good eating the vegetarian way. J. Nutr. Educ. 7: 109, 1975. Smith, P.E. The diet problem revisited: A linear programming model for convex economists. J. Farm Econ. 43: 706, 1961. Smith, V.E. Electronic Computation of Human Diets. MSU Business Studies, Bureau of Business and Economic Research, Graduate School of Business Administration, Michigan State University, East Lansing, Michigan, 1963. Smith, V.E. A. diet model with protein quality variable. Management  Science 20: 971, 1974. Sorenson, A.W. and Hansen, R.G. Index of Food Quality. J. Nutr. Educ. 7: 53, 1975. ' ' Sorenson, A.W., Wyse, B.W.., Wittwer, A.J. and Hansen, R.G. An index of nutritional quality for a balanced diet. J. Amer. Dietet. Assoc. 68: 236, 1976. Spiller, G.A. and Amen, R.J. Dietary fiber in human nutrition. Critical Reviews in Food Science and Nutrition. 6: 39, 1975. 220 Stevens, H.A., Bleiler, R.E. and Ohlson, M.A. Dietary intake of five groups, of subjects: 24-hour recall diets vs. dietary patterns. J. 'Amer. Dietet. Assoc. 42: 387, 1963. Stigler, G.J. The cost of subsistence. J. Farm Econ. 27: 303, 1945. Taggart, N. Diet, activity and body-weight. A study of variation in a woman. Brit. J. Nutr. 16: 223,1962. Thompson, E.M. and Tucker, H. Computers in dietary studies. J. Amer. Dietet. Assoc. 40: 308, 1962. Thompson, J.J. Instructional Communication. Van Nostrand, Toronto, 1969. Thomson, A.M. Diet in pregnancy. I. Dietary survey technique and the nutritive value of diets taken by primigravidae. Brit. J. Nutr. 12: 446, 1958. Tie, L.T., Lian, O.K., Liong-Ong, T.W. and Rose, C.S. Health, development, and nutrition survey of preschool children in Central Java. Amer. J. Clin. Nutr. 20: 1260, 1967. Todhunter, E.N. Approaches to nutrition education. J. Nutr. Educ. 1: 8, 1969. Travers, R.M. Essentials of Learning. The MacMillan Co., New York, 1963. Trenholme, M. and Milne, H. Studies of teenage eating in Ontario. I. Evaluation of teenage food selection. Canadian J. Publ. HIth. 54: 455, 1963. Trulson, M.F. Assessment of dietary study methods. I. Comparison of methods for obtaining data for clinical work. J. Amer. Dietet. Assoc. 30: 991, 1954. Tuthill, B.H. Computer-assisted food management systems at the medical center, UMC. In: Computer-assisted Food Management Systems. Edited by A.N. Moore and B.H. Tuthill. University of Missouri, Columbia, Missouri, 1974, p. 9. Tuthill, B.H. and Moore, A.N. Inventory control system, Department of Nutrition and Dietetics, Medical Center, UMC. Irr. Computer-assisted Food Management Systems. Edited by A.N. Moore and B.H. Tuthill. University of Missouri, Columbia, Missouri, 1974, p. 19. Tyler, K. The influence of mass media on children's food habits. lrr. Proceedings of the Nutrition Education Conference. USDA Misc. Publ. No. 913, Washington, D.C., 1962. .221 Ullrich, H.D. Teaching tools. J. Nutr. Educ. 2: 80, 1971. Ullrich, H.D. The right way: In whose judgement? J. Nutr. Educ. 5: 184, 1973a. Ullrich, H.D. Needed: A definition, legislation, action. J. Nutr. Educ. 5: 224, 1973b. Ullrich, H.D. Nutrition education is survival education. J. Nutr. Educ. 6: 84, 1974a. Ullrich, H.D. Energy crisis today, food crisis tomorrow? J. Nutr. Educ. 6: 4, 1974b. Ullrich, H.D. What's nutritious? J. Nutr. Educ. 7: 48, 1975. United States. Bureau of Human Nutrition and Home Economics, in cooperation with the National Research Council. Tables of Food Composition in  Terms of Eleven Nutrients. USDA Misc. Publ. No. 572, 1945. United States. Nutritive.Value of Foods. USDA Home and Garden Bulletin No. 72, 1960. United States. Food and Nutrition Board. National Research Council. Toxicants Occuring Naturally.in Foods, 2nd Edition. National Academy of Sciences, Washington, D.C., 1973. United States. Food and Nutrition Board, National Research Council. Recommended Dietary Allowances, 8th edition. National Academy of Sciences, Washington, D.C., 1974. United States. Consumer and Food Economics Institute, Agriculture Research Service, USDA. Data published in Composition of Foods - Raw, Processed, Prepared. Rev. USDA Agriculture Handbook No. 8, 1963. Data sets 8-1-0, 2-2-0, and 8-3-0, CFE Form 52, August 1977a. United States. Expansion 1 (March 1972) of data published in Composition of Foods - Raw, Processed, Prepared. Rev. USDA Agriculture Handbook No. 8, 1963. Data sets 8-1-1 and 8-2-1, CFE Form 52, August 1977b. United States. Data published in Home and Garden Bulletin No. 72. Rev., 1971. Data sets 72-4-0, CFE Form 52, August 1977c. United States. Data published in Nutritive Value of American Foods in Common Units, USDA Agriculture Handbook No. 8, 1975. Data sets 456-1, CFE Form 52, August 1977d. United States. Data published in Agriculture Handbook No. 8-1, Rev. 1976. Data sets 8-1 and 8-2, CFE Form 52, 1977e. United States. Select Committee on Nutrition and Human Needs, U.S. Senate. Dietary Goals for the United States. Government Printing Office, Washington, D.C., 1977f. 222 United States Department of Agriculture (U.S.D.A.). The Basic 7 FOod Groups. USDA Leaflet No. 288, 1946. United States Department of Agriculture-(U.S.D.A.). Dietary Levels in the  United. States. Spring 1965. U.S. Govt. Print Office, Washington, D.C, 1968. United States Department of Agriculture (U.S.D.A.). A Daily Food Guide. USDA Leaflet No. 424, 1971. United States Department of Health, Education, and Welfare. Ten-State  Nutrition Survey, 1968-1970. DHEW Publ. No. (HSM) 72-8133, 1972. Vajda, S. Readings in Linear Programming. John Wiley and Sons, Inc., New York, 1958. Watt, B.K. and Merrill, A.L. Composition of Foods - Raw, Processed, Prepared. USDA Agriculture Handbook No. 8, 1950. Watt, B.K. and Merrill, A.L. Composition of Foods - Raw, Processed, Prepared. Rev. USDA Agriculture Handbook No. 8, 1963. Watt, B.K., Gebhardt, S.E., Murphy, E.W. and Buntrum, R.R. Food composition tables for the 70's. J. Amer. Dietet. Assoc. 64: 257, 1974. Werko, L. Risk factors and coronary heart disease - facts or fancy? Amer. Heart J. 91: 84, 1976. White, P.L. Why all the fuss over nutrition education? J. Nutr. Educ. 8: 54, 1976. Whiting, M.G. and Leverton, R.M. Reliability of dietary appraisal: Com parisons between laboratory analysis and calculation from tables of food values. Amer. J. Publ... Hlth. 50: 815, 1960. Widdowson, E.M. and McCance, R.A. Food tables. Their scope and limitations. Lancet i: 230, 1943. Wiehl, D.G. and Reed, R. Development of new or improved dietary methods for epidemiological investigations. Amer. J. Publ. Hlth. 50: 824, 1960. Wilson, C.S., Schaefer, A.E., Darby, W.J., Bridgforth, E.B., Pearson, W.N., Combs, G.F., Leatherwood, E.C, Greene, J.C, Teply, L.J., Plough, I.C, McGanity, W.J., Hand, D.B., Kertesz, Z.I. and Woodruff, C.W. A review of methods used in nutrition surveys conducted by the Interdepartmental Committee on Nutrition for National Defense (ICNND). Amer. J. Clin. Nutr. 15: 29, 1964. Winarski, M. Food guides. Abused and misused. Illinois Teacher of Home  Econ. 19: 225, 1976. Witschi, J., Porter, D., Vogel, S., Buxbaum, R., Stare, F.J. and Slack, W. A computer-based dietary counselling system. J. Amer. Dietet. Assoc. 69: 385, 1976. 223 Wittwer, A.J., Sorenson, A.W., Wyse, B.W. and Hansen, R.G. Nutrient density-Evaluation of nutritional attributes of food.' J. Nutr. Educ. 9: 26, 1977. Wolff, R.J. Who eats for health? Am. J. Clin. Nutr. 26: 438, 1973. World Health Organization (W.H.O.). Chemicals and health. WHO Chronicle. 32: 377, 1978. Wyse, B.W., Sorenson, A.W., Wittwer, A.J. and. Hansen, R.G. Nutritional quality index identifies consumer nutrient needs. Food Technology. 30: 22, 1976. Yano, K., Rhoads, G.G. and Kagan, A. Coffee, alcohol and risk of coronary heart disease among Japanese men living in Hawaii. New Engl. J. Med. 297: 405, 1977. Young, CM., Berresford, K. and Waldner, B.G. What the homemaker knows about nutrition. Ill: Relation of knowledge to practice. J. Amer. Dietet. Assoc. 32: 321, 1956a. Young, CM., Berresford, K. and Waldner, B.G. Nutrition knowledge and practices. Pub. Health Rept. 71: 487, 1956b. Young, CM., Chalmers, F.W., Church, H.N., Clayton, M.M., Hagan, G.C, Steele, B.F., Tucker, R.E. and Foster, W.D. Subject's ability to estimated food portions. Univ. of Mass.,.Agric. Exp. Stat. Bull. 469: 63, 1952a. Young, CM., Chalmers, F.W., Church, H.N., Clayton, M.M., Tucker, R.E., Wertz, A.W. and Foster, W.D. A comparison of dietary study methods. I. Dietary history vs. seven-day record. J. Amer. Dietet. Assoc. 28: 124, 1952b. Young, CM., Hagan, G., Tucker, R.E. and Foster, W.D. A comparison of dietary study methods. II. Dietary history vs. seven-day records vs. 24-hour recall. J. Amer. Dietet. Assoc, 28: 218, 1952c. Young, E.G. Dietary standards. J_n: Nutrition - A Comprehensive Treatise. Edited by G.H. Beaton and E..W. McHenry. Academic Press, New York, 1964, Volume 2, p. 299. Young, M.A.C Reviews of research and studies related to health education practice. Health Education Monographs 23: 12, 1967. Yudkin, J. Dietary surveys: Variations in weekly intake of nutrients. Brit. J. Nutr. 5: 177, 1951. . 224 APPENDIX A FOOD-ITEM FILE Two hundred and twenty-one item clusters with attribute group code number, item cluster code number, standard portion sizes, and gram equivalents per portion. ATTRIB. ITEM GRAMS GROUP CLUSTER PER STANDARD PORTION SIZE CODE # CODE # PORTION 001 001 28 001 002 28 001 003 28 001 004 114 001 005 28 002 006 15 002 007 244 002 008 246 002 009 246 003 010 55 003 on 140 004 012 28 004 013 28 005 014 120 005 015 120 005 016 120 006 017 90 007 018 80 007 019 75 008 020 75 008 021 75 008 022 120 008 023 28 r-r-iy d oz.)^ V-V-lh" (1 oz.)" r-r-1%" (i oz.r h cup (4 oz.)d . 2 tbsp. (1 oz.)Q 1 tbsp.d 1 cupd 1 cupd 1 cup" d 1 egg d 2 eggs0 1 cup (1 oz.)d 1 cup (1 oz.) % cupd h cupd h cupd 2 at 4" dia.d h cupd h cupd h cupd h cupV h cup 3 tbsp. (1 oz.) FOOD ITEMS AND DESCRIPTION ENTREE.ITEMS - DAIRY AND EGGS AMERICAN PROCESSED CHEESE. BLUE or ROQUEFORT CHEESE. CHEDDAR, JACK or SWISS CHEESE. COTTAGE CHEESE, creamed or uncreamed, any curd. CREAM CHEESE or CHEESE SPREADS. SOUR CREAM or CHIP DIP. YOGHURT, base of whole milk (3.5% b.f.) YOGHURT, base of skim milk. YOGHURT, base of part skimmed and 2% nonfat milk solids EGG: raw, boiled, poached, fried (add fat). EGG: scrambled, omelets, souffles, spoon bread. ENTREE ITEMS - CEREALS CORN CEREAL, enriched, eg. corn flakes: ready-to-eat. WHEAT CEREAL, enriched, eg. rice krispies: ready-to-eat. WHEAT CEREAL, more refined, eg. cream of wheat: cooked. OATMEAL, all types: cooked. WHEAT CEREAL, less refined, eg. rolled wheat: cooked. PANCAKES, WAFFLES or FRITTERS, made with milk and eggs. NOODLES, EGG TYPE, enriched: cooked. SPAGHETTI, MACARONI or NON-EGG PASTAS, enriched: cooked. M ro RICE, BROWN: cooked. RICE, WHITE, enriched, unenriched or parboiled: cooked. CORNMEAL or CORN GRITS, enriched: cooked. WHEAT GERM. 009 024 20 1 slice (9/16")J I 009 025 23 1 slice (9/16")° I 009 026 23 1 slice (9/16")° I 009 027 23 1 slice (9/16"ft I 009 028 23 1 slice (9/16")u . 009 029 40 1 square (2"-2"-lV) 009 030 30 1 at 6" dia.d 010 031 35 1 at 2" dia.e 010 032 46 id 010 033 40 ld 010 034 40 ld 010 035 46 ld 011 036 14 1 at 5"-2%"-3/16"£ Oil 037 6 2 at lV'-14"-l/8"t 10 at 3V-1/8" dia. Oil 038 21 1 cups . Oil 039 30 15 chips (1 oz.)J 011 040 14 2 at 3V-14"-V 3^ 8e 012 041 85 1/3 cup (3 oz.jjj 012 042 85 1/3 cup (3 oz.)d 012 043 85 1/3 cup (3 oz.)d 012 044 85 1/3 cup (3 oz.)d 012 045 85 1/3 cup (3 oz.f 012 046 85 1/3 cup (3 oz.)d 012 047 85 1/3 cup (3 oz.)a 012 048 16 2 slicesd 013 049 85 1/3 cup (3 oz.)d 013 050 200 1 cupd . 013 051 85 1/3 cup (3 oz.n 013 052 85 1/3 cup (3 oz.)a FRENCH or SOURDOUGH BREAD, enriched: fresh or toasted. RAISIN or RAISIN-NUT BREAD, enriched: fresh or toasted. RYE BREAD, light or dark: fresh or toasted. WHITE or CRACKED WHEAT BREAD, enriched: fresh or toasted. WHOLE WHEAT BREAD: fresh or toasted. CORN BREAD or JOHNNY CAKE. CORN TORTILLAS or FLOUR TORTILLAS. BISCUITS, ROLLS or POPOVERS. HAMBURGER BUN, HOT DOG BUN, or BAGEL, enriched. MUFFIN, white, bran, blueberry, brownbread. SOURDOUGH ENGLISH MUFFIN. WHOLE WHEAT ENGLISH MUFFIN. GRAHAM CRACKERS, honey-coated or whole wheat. SODA CRACKERS, saltines, holland rusk, matzoth. PRETZELS. POPCORN: popped (add salt and fat is used). POTATOE CHIPS, FRITOS, CORN PUFFS or TORTILLA CHIPS. RYE KRISP. TRISCUITS. WHEAT THINS. ENTREE ITEMS - MEATS AND PLANT PROTEIN BEEF, 30% fat cut, eg. chuck rib, veal: cooked. BEEF, 20% fat cut, eg. regular ground hamburger: cooked. BEEF, 15% fat cut, eg. round: cooked. CORNED BEEF, fresh or canned. LAMB, all muscle cuts, eg. leg, shoulder, chops: cooked. PORK, lean cuts, eg. chops, shoulder, roast: cooked. PORK, all hams, fresh or canned. BACON, thick cut, thin cut or slab: cooked, drained. CHICKEN: steamed, stewed, broiled, baked or canned. CHICKEN: prepared with sauce, eg. fricassee, cacciatore. CHICKEN: fried flesh and skin. TURKEY, DUCK, RABBIT or SQUAB GOOSE: roasted. 014 053 100 014 054 100 014 055 100 014 056 85 014 057 100 014 058 100 014 059 85 014 060 56 014 061 85 014 062 28 014 063 85 014 064 100 014 065 60 014 066 85 015 067 85 016 068 85 017 069 45 017 070 40 017 071 30 018 072 125 018 073 100 018 074 85 018 075 75 019 076 15 019 077 15 5 small (3%oz.)d . CLAMS: canned. 4 to 5 sticks (3%oz.) FISH STICKS, FISH CAKES, FISH LOAFS: cooked. (3%oz.r FRIED FISH, eg. fried haddock: breaded, fried. (3 oz.)a , BROILED FISH, eg. broiled halibut. 5 to 8 medium (3%oz.)Jj OYSTERS: canned. 5 to 8 medium (3*5 oz.) OYSTERS: breaded, fried. (3 oz.)d SALMON, all types: fresh or frozen, cooked. (2 oz.)d SALMON, canned. (3 oz.)d . SMOKED FISH, all types, eg. smoked salmon. 2 medium (1 oz.)d SARDINES or KIPPERS: fresh or canned. (3 oz.)d SHRIMP, LOBSTER, CRAB or ABALONE: fresh or canned. (3% oz. )d . SHRIMP: fried. 1/3 can.(2 oz.) TUNA: canned, oil or water-pack. (3 oz.) RAW FISH, eg. raw tuna. (3 oz.)d LIVER, beef, calf, hog, chicken or lamb: fried. (3 oz.)d KIDNEY. f ALL LUNCH MEATS EXCEPT LIVERWURST. 1 at 5"-^"T f FRANKFURTER: cooked, xl .4(1 link at 4"-^8")T p* KNOCKWURST. xl.l(l slice at 4"-3"-fc")e P HEAD CHEESE, xl .6(1 slice at 4%"-4V'-#')e BOLOGNA. 3 at 2"-$' dia.e VIENNA SAUSAGE. 1 at 3%'-3%'-k"e f SALAMI. x2.8(l slice at 4V'-4V'-&') LOAF MEAT, eg. ham loaf, olive loaf. 3 tbsp.f MEAT SPREADS, eg. deviled ham. 2 links at 3"-h" dia. FRESH COOKED SAUSAGES, eg. pork sausages. 1 slice at 3" dia.-V'e LIVERWURST or PATE DE FOIS GRAS. h cupd . BEANS, WHITE, RED, PINTO, KIDNEY: canned. h cup (3% oz.) BEANS, white, red, pinto or kidney: boiled and drained. h cup ,(3 oz.)d COWPEAS or BLACK-EYED PEAS: boiled and drained. h cupd SOYBEANS: boiled and drained. 12 to 15° e ALMONDS. 10 to 12e FILBERTS or HAZELNUTS. 2 tbsp.; SESAME SEEDS, t tbsp.1" CASHEW BUTTER. ro 019 078 30 019 079 28 019 080 15 019 081 15 020 082 100 020 083 50 020 084 100 020 085 180 020 086 100 020 087 100 021 088 100 021 089 65 021 090 65 021 091 63 021 092 43 021 093 62 021 094 85 021 095 37 021 096 100 021 097 85 021 098 85 021 099 62 021 100 85 021 102 27 022 103 80 022 104 76 022 105 50 6 to 8d CASHEW NUTS. 2 pieces at l"-l"-%" COCONUT: fresh, shredded or dried, cup shreddede . 2 tbsp. (1 oz.)° PEANUT BUTTER and OTHER NUT BUTTERS EXCEPT CASHEWS. 1 tbsp.d , PEANUTS or SPANISH PEANUTS. 2 tbsp., 8 to 10 halves0 WALNUTS, persian, black or english: PECANS. ENTREE ITEMS - VEGETABLES round: 1 at 2%" to 2%" dia.e POTATOE, all white types: baked or boiled, round: xl.5(1 at 3k" dia.)e f long: x2.5(l at 2%' dia-4^')T 10 pes. at h"-k"-2:, cupe POTATOE, all white types: fried, eg. french fried. h cup (3% oz.)" f POTATOE, all white types: mashed with fat and milk. 1 at 2" dia.-5:, or ^cup . SWEET POTATOE: baked in skin. 1 at 2" dia.-4"e, or h cupa SWEET POTATOE: canned in syrup. 2/5 cupe YAM: baked or boiled. 3 f h of 3V dia.-4" , or%cup AVACADO, raw or avacado dip. h cupd BEANS, snap green or wax: fresh or frozen, boiled. h cup0 BEANS, snap green or wax: canned, boiled and drained. h cupd . BROCCOLI: boiled and drained. h cupd CABBAGE, CAULIFLOWER or SAVOY: raw. h cupd CABBAGE, BRUSSEL SPROUTS or CAULIFLOWER: boiled, drained. \ cupd COLLARDS or KALE: boiled, drained. h cup chunksd LETTUCE, crisp head, romaine, iceberg or endive: raw. h cupd GREENS, mustard or turnip: fresh or frozen, boiled, drained. h cupd PEAS, LIMA BEANS or SNOW PEAS: canned, boiled, drained. h cupd PEAS, LIMA BEANS or SNOW PEAS: fresh or frozen, boiled. h cup. PEPPERS, sweet green: raw, canned or boiled. h cup0 SPINACH, BOK CHOY, or BEET GREENS: canned, drained. h cup0 SPINACH: raw. h cupd BEETS or ARTICHOKE: raw or canned, cooked, drained. h cup0 CARROTS or WINTERSQUASH: boiled and drained. £ f MATURE RED PEPPERS or HOT CHILI PEPPERS: canned or fresh. islg.(9"-lV)»|med.(7¥,-lV)T CARROTS: raw. 1 sml. (5" ) >4cup di ced^ h cup shredded' CO 022 106 85 022 107 126 022 108 85 022 109 100 022 no 100 022 111 148 022 112 77 023 113 50 023 114 100 023 115 45 024 116 85 025 117 13 025 118 20 025 119 34 026 120 15 026 121 15 026 122 15 026 123 15 026 124 15 027 125 5 027 126 5 028 127 38 % cup (3 oz.) CORN: fresh or frozen, boiled and drained. % cupd . CORN, cream-style: canned % cup (3 oz.) , CORN, kernal: canned, boiled and drained. h cup (3% oz.)d SUMMER SQUASH, ASPARAGUS or ZUCCHINI: boiled, drained. % cup (3% oz.) f TOMATO: canned solid and liquid. 1 med.(2^'), x^sml.(2^')T TOMATO: raw x 1% lrg.(3")f h cupd TURNIP ROOT or TUTABAGA: boiled and drained, piece (24" dia.-l%") or CUCUMBER: raw. (14" dia.-2")f 1 Irg. stalk(8"-lV) or CELERY: raw 3 sml. stalks(5"-34')f , 10 W to 1") or 6(1" to 1V')TRADISHES: raw h cup (3% oz.)d MUSHROOMS or WATER CHESTNUT: fresh or canned, cooked. k cupd ONIONS, LEEKS, GARLIC or GREEN ONIONS: raw or boiled. h cup (3 oz.)d MIXED VEGETABLES or SUCCOTASH: frozen, boiled. 4(%" dia.9fc"),-3(V dia.-^), OLIVES, green, black or stuffed. 2(%" dia.-l^")1" -1(1" dia.-3V'), 1(1" dia.- PICKLES, sweet, bread and butter or pickle relish. 1^"), 3(1%" dia.-V), overful tbsp. k lrg.(l%' dia.-4"), %cup, PICKLES, dill or sour. h med.(lV dia.-3V)\ 5(1%" dia.-%")f ENTREE ITEMS - FATS AND OILS 1 tbsp.j VEGETABLE FAT, eg. crisco or hardened shortening. 1 tbsp ° LARD, SUET or SALT PORK. 1 tbsp d COTTONSEED OIL. 1 tbsp.d OLIVE OIL. 1 tbsp. SOYBEAN or CORN OIL. 1 tsp.^ BUTTER, sweet or salted butter. 1 tsp. MARGARINE, all brands, whipped or diet. 2 tbsp.d CHEESE SAUCE or FONDUE. ro ro NO 028 128 72 4 tbsp., or h cup 028 129 33 1 tbsp.d 028 130 17 1 tbsp.d 028 131 100 h cupd 029 132 15 1 tbsp d 029 133 15 1 tbsp ° 029 134 15 1 tbsp.0, 029 135 14 1 tbsp. 030 136 244 1 cup (8 fl. oz.)d 030 137 246 1 cup (8 f1. oz.)a 030 138 246 1 cup (8 f1. oz.)a 031 139 121 1 cupd . 031 140 15 1 tbsp. 031 141 8 1 tbsp. whipped, 2 tbsp. whippedd 031 142 15 1 tbsp.d 032 143 120 h cup (4 f1. oz.)d 032 144 120 h cup (4 f 1. oz.)a 032 145 15 1 tbsp, or '6 lemon 032 146 226 1 cup (8 f1. oz.)° 032 147 120 hcup (4 f 1. oz.)a 032 148 120 h cup (4 fl. oz.)a 032 149 120 h cup (4 f1. oz.)a 033 150 100 ^cup (4 f 1. oz.)d 034 151 339 12 f 1. oz.j! 034 152 339 12 f 1. oz. GRAVY, all types. WHITE SAUCE, thick or thin, or HOLLANDAISE SAUCE. TOMATO CATSUP or BARBECUE SAUCE. TOMATO SAUCE. MAYONNAISE, SANDWICH SPREAD or TARTAR SAUCE. SALAD DRESSING, oil types, eg. Italian. SALAD DRESSING, mayonnaise type, eg. thousand island. ROQUEFORT or BLUE CHEESE SALAD DRESSING. BEVERAGES - DAIRY WHOLE MILK, 3.5% b.f. SKIM MILK, BUTTERMILK or NONFAT INSTANT MILK. PART SKIM and 2% NONFAT MILK SOLIDS. HALF AND HALF. LIGHT TABLE CREAM and NONDAIRY COFFEE WHITENERS. WHIPPED CREAM, COOL WHIP or IMITATION TOPPINGS. CANNED EVAPORATED or CONDENSED MILK, sweet or unsweet. BEVERAGES - FRUIT and VEGETABLE APPLE, GRAPE, PRUNE, or PEAR JUICE: fresh, frozen, canned GRAPEFRUIT JUICE: canned and unsweetened. LEMON, LIME or ACEROLA JUICE: fresh, frozen or canned. FRESH LEMONADE, ALL FRUIT ADES and ARTIFICIAL LEMONADES. ORANGE JUICE: canned. ORANGE JUICE: frozen. ORANGE JUICE: fresh. TOMATOE JUICE or VEGETABLE JUICE: canned. BEVERAGES - MISCELLANEOUS COLA-TYPE BEVERAGES, caffeine-containing, carbonated. GINGER ALE, non-caffeine containing, carbonated. 035 153 200 1 cup. 035 154 200 1 cup 036 155 360 12 fl. oz.d . 036 156 43 1.5 oz. jigger 036 157 60 2 fl. oz., sherry glass 036 158 100 3.5 f1. oz., wine glass 037 159 198 8/10 cup, can 037 160 198 8/10 cup, can 037 161 200 8/10 cup, can 037 162 198 8/10 cup, can 037 163 203 8/10 cup, can 037 164 200 8/10 cupd 038 165 45 IV arc of cake 9V dia.-4 or xl .35 of 2Hl arcf 038 166 75 1 piece 3"-2J2"-l%llf 038 167 55 1 " arc of cake 8" dia.-3" xl%(2V dia.), xl.65(2V dia.)f 3 f square 2"-2"-l%,T 038 168 40 1 piece 3"-3"-h"e* 038 169 30 1 piece 3"-3"-h"f f 038 170 60 IV arc cake 8" dia-3" 039 171 160 4V arc 9" dia., l/6tht 039 172 150 4V arc 9" dia., l/6th 040 173 40 2 cookies0 040 174 29 2 barsd 041 175 30 1 average^ 041 176 30 1 average0 COFFEE, all types. TEA, all types. BEER, ALE, or STOUT. DISTILLED SPIRITS, eg. gin, rum, whiskey or vodka. DESSERT or SWEET WINE TABLE or DRY WINE. SOUPS NOODLE or RICE SOUPS WITH MEAT: canned, dry or instant. CREAMED SOUPS MADE WITH MILK: canned, dry or instant. PEA or BEAN SOUP: canned, dry or instant. TOMATOE SOUP WITH OR WITHOUT RICE. MEAT AND VEGETABLE AND VEGETARIAN VEGETABLE SOUPS. ONION SOUP or CONSOMME. DESSERTS AND SWEETS - CEREALS ANGEL FOOD CAKE, SPONGE CAKE or TWINKES. COFFEE CAKE, BANANA BREAD, DATE-NUT BREAD or SNAKIN CAKE CHOCOLATE CAKE WITH ICING. CUPCAKE WITH FROSTING. BROWNIE. FRUITCAKE, light or dark. POUND CAKE. ALL OTHER COMMERCIAL, HOMEMADE or FROZEN CAKE WITH ICING ALL FRUIT PIES, eg. apple, pecan, lemon meringue. PUMPKIN or SWEET POTATOE PIE. ALL COOKIES WITHOUT FRUIT FILLING, eg. sandwich-type. ICE CREAM CONE or COOKIE OF ICE-CREAM SANDWICH. ALL FRUIT-FILLED COOKIES, eg. figbars or raisin cookies. CAKE DOUGHNUT, iced or uniced. YEAST-RAISED DOUGHNUT, iced or uniced. 041 177 38 042 178 90 042 179 97 042 180 339 043 181 130 043 182 246 044 183 150 044 184 130 044 185 100 044 186 100 044 187 100 044 188 100 044 189 126 044 190 100 044 191 139 044 192 80 045 193 28 x 1.6(1)( d % cupd V cup ld % cupd 1 cup 1(2*" dia.), xl^3V dia.), or x*(2V dia.)f . 1(3%" long-2%" diaJe T 1 cup, 20(%" dia.)1" 1 cup, 3(2£" dia.), 4(1%" dia.)f 2of 1/3 (10" dia.-l" slice)f %(6" dia.-lV slice)e h: rund e -2 cup x- ~..nd e s cup1 1 sml.(6"), %cupe f h of 5" dia. % of 4" dia, 1 sml. (2%" x l%med.(8") dia.)( x 1% med.(3" dia.)c ( x 2.35 lrg.(3%" dia.) 10 lrg., 2/3 cupe % cupe 1 at 2%'' dia., 2 ™ f 1* (2%'Y or 3 % cupd % cupd 2 tbsp. (1 oz.)' DANISH PASTRY or HOT CROSS BUN. CINNAMON BUN or SWEET ROLL. DESSERTS AND SWEETS - DAIRY ICE CREAM, ICE MILK or ICE CREAM BARS, all flavors. SHERBERT, all flavors. POPSICLE. TAPIOCA and RICE PUDDING, JUNKET, CUSTARD or PIE FILLING. YOGHURT, add fruit and sugar if included. DESSERTS AND SWEETS - FRUIT APPLE, raw. PEAR, raw. GRAPES, raw. PLUMS, raw. CHERRIES, raw. WATERMELON, raw. HONEYDEW MELON or CASABA MELON, raw. APPLESAUCE: fresh or canned. PEARS, GRAPES, PLUMS or CHERRIES: canned. BANANA or PLANTAIN. CANTALOUPE. GRAPEFRUIT, fresh or unsweetened. ORANGE, raw. STRAWBERRIES, raw. PEACHES, APRICOTS or FRUIT COCKTAIL, heavy syrup pack. PEACHES, PERSIMMONS or NECTARINES, raw. APRICOTS, raw. PINEAPPLE, MANDARIN ORANGE or BOYSENBERRIES, heavy syrup. PINEAPPLE, TANGERINES or BOYSENBERRIES, water or juice pack. RAISINS, PRUNES or OTHER DRIED FRUIT, not peaches or apric. 046 194 80 047 195 21 047 196 20 047 197 5 048 198 20 048 199 20 049 200 28 049 201 28 049 202 15 049 203 20 050 204 125 051 205 227 052 206 238 053 207 220 054 208 220 055 209 224 h cupd GELATIN DESSERT, with or without fruit. DESSERTS AND SWEETS - SWEETS 1 tbsp.J HONEY, all type. 1 tbsp,a MOLASSES, all types. 1 tsp.d SUGAR. 1 tbsp.j! JELLIES, PRESERVES, JAM or MARMALADES. 1 tbsp. CORN SYRUP or MAPLE SYRUP. 1 piece (1"-1"-1V), (loz.)f CARAMEL CANDY, TAFFY or VANILLA-COATED CARAMELS. 1 piece (2"-2"-V)f 3 caramel 1 piece (2"-2"-V), (1 oz.)f CHOCOLATE CANDY, MILK CHOCOLATE or FUDGE. 1 piece (r-l"-lV)f 2(1%" dia.-V), 10(V dia.-% CHOCOLATE DISK. ")f (1 oz.)d CHOCOLATE SYRUP or FUDGE TOPPING. 2 lrg. W dia.) MARSHMALLOW. 1 piece (1"-1"-V), 2 hard6 HARD CANDY. 3 lrg., 6 sml.e MARASHINO CHERRIES. CHEWING GUM 7f f JELLY BEANS. 2 lrg. (%" dia.), 16 sml.T GUM DROPS or JELLY CANDIES. MISCELLANEOUS ITEMS h cupd BEANS WITH PORK AND TOMATOE SAUCE, canned or homemade. 1(4%" dia.), or 1/3(9" dia.) POT PIES, chicken or tuna: commercial or homemade. 1 cupd MEAT AND VEGETABLE STEWS. 1 cupd CHOW MEIN or CHOP SUEY: canned or frozen. £ 1 cupd CHOW MEIN or CHOP SUEY: homemade. 1 cupd CHILI CON CARNE WITH OR WITHOUT BEANS, canned. •GO 056 210 312 1 dinner, (11 oz.)d 057 211 312 1 dinner, (11 oz.)0' 058 212 300 1 dinner, (11 oz.)d 059 213 224 h cupd 060 214 225 h cup 061 215 200 3/8 of 14" pizzaf 3 (5V dia.) sector 062 216 200 1 cupd 063 217 220 d 1 cup 064 218 155 2 from can of 6d 065 219 4 1 cube 066 220 1 1 tsp.d 067 221 7 1 tbsp.d FROZEN DINNER: fried chicken, mashed potatoes and peas. FROZEN DINNER: meatloaf, mashed potatoes and peas. FROZEN DINNER: roast turkey, mashed potatoes and peas. HASHES, CANNED C0RNBEEF or ANY HOMEMADE HASH. MACARONI AND CHEESE: homemade, packaged or frozen. .PIZZA, any kind. )f SPAGHETTI IN TOMATOE SAUCE, CANNED RAVIOLI or NOODLE-0'S. SPAGHETTI WITH MEAT BALLS: homemade or packaged. TAMALES: homemade or canned. BOULLION CUBE. TABLE SALT or MONO-SODIUM GLUTAMATE. COCOA MIX. ro oo For the legend of group codes, see Appendix G of this text. A food item cluster is one item or two or more specific items, item varieties, or prepared variations of an item with common nutrient characteristics. Item clusters are numbered in the sequence they appear in the food item file. The food items list is adapted from Mini List Foods and Food Substitutions in Pennington, J.A., Dietary Nutrient Guide, AVI Publishing Co., Westport, Connecticut, 1976. Gram equivalents per portion are from Mini List Foods, Appendix C, Pennington, J.A., Dietary Nutrient Guide, AVI Publishing Co., Westport, Connecticut, 1976. Standard portion size derived from Mini List Foods, Appendix C, in Pennington, J.A., Dietary Nutrient Guide, AVI Publishing Co., Westport, Connecticut, 1976. Standard portion size derived from Church, C.F. and Church, H.N., Food Values of Portions Commonly Used, 12th edition, J.B. Lippincott Co., New York, 1975. Standard portion size derived from Adams, G.F., Nutritive Value of American Foods in Common Units, Agriculture Handbook No. 456, U.S.D.A., 1975. Description of standard portions per gram quantity differ between e and ro Co ui 236 APPENDIX B ABRIDGED FOOD-ITEM FILE Listing, by group code and item cluster code, of 127 item clusters selected from Appendix A. 237 ATTRIBUTEa ITEM ATTRIBUTE ITEM ATTRIBUTE ITEM GROUP CLUSTER CLUSTER CLUSTER GROUP CLUSTER CODE # CODE # CODE # CODE # CODE # CODE # 001 003 020 082 038 170 001 004 020 083 039 171 001 005 020 084 039 172 002 006 020 085 040 173 002 008 021 089 040 174 003 010 021 091 041 175 004 012 021 092 041 177 004 013 021 095 042 178 005 015 021 098 042 179 005 016 021 099 043 181 006 017 021 101 044 183 007 018 022 103 044 184 007 019 022 104 044 185 008 020 022 105 044 186 008 021 022 106 044 187 008 023 022 111 044 188 009 024 023 113 044 192 009 027 023 114 045 193 009 028 023 115 047 195 010 031 024 116 047 197 010 032 025 117 048 198 010 033 025 118 048 199 010 035 025 119 049 201 on 037 026 121 049 202 on 040 026 124 051 205 012 041 027 125 063 217 012 042 028 127 067 221 012 043 028 128 012 046 029 132 012 047 029 133 012 048 030 136 013 049 030 137 013 051 031 140 014 055 031 141 014 056 032 143 014 058 032 144 014 062 032 145 014 063 032 148 014 065 033 150 015 067 034 151 017 069 035 153 017 070 035 154 017 071 036 155 018 073 036 156 018 075 036 158 019 076 037 160 019 077 037 161 019 079 037 163 019 080 038 166 019 081 038 168 Refer to Appendix A for description of item clusters, standard portion sizes, and gram equivalents per portion associated with each of the coded items. 2.38 APPENDIX C FOOD-COMPOSITION FILE Table of 41 nutrient values, in nutrients per 100 grams of edible portions of food, for 221 item clusters indexed by attribute group code and item cluster code number. 239 ITEM CLUSTER 1 2 3 4 5 ATTRIBUTE GROUP 1 1 1 1 1 NUTRIENTS:b KCAL 370.000 368.000 398.000 106.000 374.000 PROT (GM) 23.20C 21.500 25.000 13.600 8.000 TRY (MG) 320.000 290.000 320.000 150.000 70.000 THR (MG) 840.OOC 810.000 860.000 640.000 350.000 ISO (MG) 1540.CCC 1480.000 1560.000 790.000 460.000 LEU (MG) 2230 .OCO 2140.000 2260 .000 1470.000 ' 800.000 LYS (MG 1 1670.OCC 1550.000 1700.000 1140.000 630.000 MET (MG) 600.000 580.000 600.000 380.000 200.000 CYS (MG) 130.OCC 120.000 130.000 120.000 80.000 PHE (MG J 1200 .000 1200.000 1300.000 760.000 500.000 TYR ( MG) 1110.OOC 1030.000 1110.000 730.000 360.000 VAL (MG) 1640.000 1580.000 1670.000 780.000 470.000 HIS (MG) 760.000 700.000 760.000 440 .000 250.000 FAT-T (GM) 30. OCC 30. 500 32.200 4.200 37.700 SFA (GM) 15.000 17 .000 18.000 2.000 21.OOC PUFA (GM ) 10.OCC 11.000 12.000 1 .000 13.000 CHGLE (GM) 0.150 0.150 0.120 0.015 0. 120 CHO-T (GM) 1.9 00 2. 000 2.100 2.900 2.100 SUCR (GM) 0 .0 0.0 0.0 0.0 0.0 CHO-F (GM) 0.0 0.0 0.0 0.0 0.0 THIA (MG) 0.02C 0. 030 0.030 0. 030 0.020 RIBO (MG) 0.410 0.610 0.460 0.250 0.240 NIACIN (MG) 0.0 1. 200 0. 100 0. 100 0. 100 VIT-B6 (MG) 80.OCC 170.000 80.000 40.000 60.000 FOLIC (UG) 11.000 11.000 6.000 27.000 16.000 VIT-B12(UG) 1.000 1.400 1.000 1.000 0. 220 VIT-C (MG) 0.0 0.0 0.0 0.0 0.0 PANTO (UG) 500.000 1800.000 500.000 200.000 300.000 BIOTIN (MG) 5.OOC 3.000 2.000 2.000 1.000 VIT-A (IU ) 1220.OCC 1240.OCO 1310.000 170.000 1540.000 VIT-D (IU) 30.OOC 30.000 30.000 4.000 30.000 VIT-E (MG ) l.OOC 0. 800 1.300 0.100 1.000 CA (MG » 697.000 315.000 750.000 94.000 62.000 P (MG) 771.000 339.000 478.000 L52.000 95.000 MG (MG) 48.000 20.000 37.000 8.000 5.000 FE (MG) 0.90C 0.500 1.000 0.300 0.200 I (MG) 11.OCC 11.OCO 11.000 6. 000 4.000 ZN (MG) 4.IOC 2.200 0.900 1.400 0. 800 NA (MG ) 1136.OCC 666.OCO 700.000 229.000 250.000 K (MG ) 80.000 78.000 82.000 85.000 74.000 CU (MG) 0. 17C 0. 160 0.130 0.020 0.040 See Appendix A of this text for definition of item cluster and the items associated with the attribute group code and item code numbers. Values of nutrient composition obtained from Mini List Foods, Appendix B, Pennington, J.A., Dietary Nutrient Guide, The AVI Publishing Co., Westport, Connecticut, 1976. 24Q ITEM CLUSTER 6 7 8 9 10 ATTRIBUTE GROUP 2 2 2 2 3 NUTRIENTS: KCAL 211.000 65.000 36.000 59.000 163.000 PROT (GM) 3. OOC 3. 500 3 .600 4.200 12.900 TRY (MG) 40.OCC 50.000 50.000 60.000 210.000 THR (MG) 140.000 160.000 160.000 190.000 640.000 ISO (MG) 190. OCC 230.000 190.000 270.000 850.000 LEU (MG) 300.OCC 350.000 360.000 420.000 1100.000 LYS (MG ) 230.CCC 280.000 280.000 330.000 820.000 MET (MG) 70.000 80.000 90.000 100.000 400.000 CYS ( MG) 30.OCC 30.000 30 .000 40 .000 300.000 PHE (MG) 150 .000 170.000 170.000 200.000 740.000 TYR ( MG) 150.OCC 180.000 180.000 200.000 550.000 VAL (MG) 210.OCC 240.000 250.000 290.000 950.000 HI S (MG) 80.000 90.000 100.000 110.000 310.000 FAT-T (GM) 20.6CC 3. 500 0. 100 2 .000 11 .500 SFA (GM) 11.000 2.000 0.0 1.000 4.000 PUFA (GM) 8. OOC 1.000 0.0 1 .000 6.000 CHOLE (GM ) 0 .070 0.014 0.002 0.002 0. 550 CHO-T (GM) 4. 30C 4.900 5.100 6.000 0.900 SUCR (GM) O.C 0. 0 0.0 0. 0 0.0 CHO-F (GM) 0 .0 0.0 0 .0 0.0 0.0 THIA (MG) 0.030 0. 030 0.040 0.040 0.090 RIBC (MG) 0.150 0. 170 0.180 0.210 0.280 NIACIN (MG) 0.100 0. 100 0. 100 0.100 0.100 VIT-B6 (MG) 30 .000 40.000 40.000 40.000 90.000 FOLIC (UG) 20.000 9.000 9.000 9 .000 30.000 VIT-B12(UG) 0.250 0.400 0.400 0.400 2.000 VIT-C (MG ) 1.000 1.000 I .000 1 .000 0.0 PANTO (UG) 300.OOC 300.000 400.000 400.COO 2000.000 BIOTIN (MG) 4.00C 4.000 2.000 3.COO 23.000 VIT-A (I U) 840.000 140.000 0.0 80.000 1180.000 VIT-D ( IU) 15.000 41.000 41.000 41.000 50.000 VIT-E ( MG) 0.70C 0. 100 0.0 0 .100 1.000 CA (MG) 102.000 118.000 121.000 143.000 54.000 P ( MG) 80.OCC 93.000 95 .000 112 .000 205.000 MG (MG) 10.OOC 13.000 15.000 17.000 12.000 FE (MG) 0.0 0.0 0.0 0. 100 2. 300 I (MG) 6.OOC 7. 000 7.000 80.000 14.000 ZN (MG) 0 .300 0.400 0.400 0.400 1.400 NA (MG) 43.OCC 50.000 52.000 61 .000 122 .000 K (MG) 122.OOC 144.000 145.000 175.000 129.000 CU (MG) 0.17C 0. 150 0.020 0 .020 0.070 .241 ITEM CLUSTER 11 12 ATTRIBUTE GROUP 3 4 NUTRIENTS: KCAL 173.000 386.000 PR07 (GM) 11.200 7.900 TRY (MG) 180.OCC 60.000 THR (MG) 550.000 290.000 ISO (MG) 740. OCC 320.000 LEU (MG) 970.000 1090.000 LYS (MG) 710.000 160.000 MET (MG) 350.000 140.000 CYS (MG) 260.OOC 150.000 PHE (MG) 690.000 360. 000 TYR (MG) 480.0CC 280.000 VAL (MG) 820.OOC 400.OCO HIS (MG) 270.OCC 220.000 FAT-T (GM ) 12.900 0.400 SFA (GM) 5.OOC 0.0 PUFA (GM) 7.OOC 0.0 CHOLE (GM) 0 .380 0.0 CHO-T (GM) 2.40C 85.300 SUCR (GM) 0.0 23.6C0 CHO-F (GM) 0.0 0.700 THI A (MG) 0.08C 0.430 RIBO (MG) 0 .280 0.080 NIACIN (MG) 0.100 2. 100 VIT-B6 (MG) 90.000 70.000 FOLIC (UG.) 19.0C0 5.000 VIT-B12(UG > 2. OCC 0. 0 VI T-C (MG) 0.0 0.0 PANTO (UG) 2300.OCC 200.000 BIOTIN (MG) 17.OOC 1.000 VIT-A (IU) 1C80.00C 0. 0 VIT-D ( IU) 50.000 0. 0 VIT-E (MG) l.OOC 0. 100 CA (MG) 80.000 17.000 P (MG) 189.OOC 45.000 MG (MG) 12 .000 14.000 FE (MG) 1.70C 1.400 I (MG) 13.OOC 14.000 ZN (MG) 1 .20C 0.400 NA (MG) 257.OCC 1005.OCO K (MG) 146.OOC 120.000 CU (MG) 0.05C 0.130 13 14 15 4 5 5 354.000 42.000 55.000 10.200 1 .300 2.000 120.000 20.000 30.000 340.000 20 .000 70.000 470.000 70.000 130.000 840.000 100.000 150.000 340.000 20.000 70.000 120.000 20.000 30.000 180.000 20.000 40.000 500.000 60.000 110.000 300.000 50.000 70.000 540.000 20.000 120.000 220.000 30.000 40.000 1.600 0.100 1 .000 0.0 0. 0 3.000 1 .000 0.0 6.000 0.0 0. 0 0.0 80.500 8 .700 9.700 20.100 0.0 0.0 1.600 0.0 0.200 0.640 0.040 0.080 0. 140 0.030 0.020 4.900 0.400 0.100 290.000 10.000 20.000 18.000 0.0 11.000 0.0 0. 0 0.0 0 .0 0.0 0.0 500.000 100.000 200.000 1.000 16.000 24.000 0.0 0.0 0.0 0.0 0.0 0.0 0.500 0 .0 0.200 41.000 4.000 9.000 309.000 12.000 57.000 96.000 3. 000 24.000 4.400 0 .300 0.600 14.000 1. 000 1.000 2.400 0.100 0.900 1032.000 144.000 218.000 120.000 9.000 61.000 0.450 0 .030 0.03 0 242 ITEM CLUSTER 16 17 18 19 20 ATTRIBUTE GROUP 5 6 7 7 8 NUTRIENTS: KCAL 75.000 225.000 125.000 111.000 119.000 PROT (GM) 2.200 7.200 4.100 3.400 2.500 TRY (MG) 30.000 100.000 50.000 40.000 20.000 THR ( MG) 70.OCC 280.000 170.000 130 .000 100.000 ISO (MG) 140.OCC 400.OCC 200.000 160.000 120.000 LEU (MG) 160.OOC 610.000 270.000 210.000 220.000 LYS (MG) 40.OOC 340.000 140.000 100.000 100.000 MET (MG) 30 .000 150.000 70.000 50.000 50.000 CYS (MG ) 40.OCC 190.000 80.000 70.000 20.000 PHE (MG ) 10 .000 390.000 200.000 180.000 130.000 TYR (MG) 90.000 310.000 100.000 110.000 100.000 VAL (MG) 130.OCC 410.000 240.000 180.000 180.000 HI S (MG) 500.000 170.000 100 .000 80.000 30.000 FAT-T (GM) 0.4CC 7.300 1.500 0.400 0.600 SFA (GM) 0.0 3.000 0.0 0.0 0.0 PUFA (GM) 0.0 5. 000 I.000 0.0 0.0 CHGLE (GM ) 0.0 0.070 0.040 0.0 0.0 CHO-T (GM) 16.900 32.400 23 .300 23.000 25.500 SUCR (GM ) 0 .0 0. 100 0.100 0.100 0.300 CHO-F (GM) 0.50C 0. 100 0.100 0.100 0.300 THIA (MG) 0.C7C 0. 150 0. 140 0. 140 0.090 RI BC (MG) 0.03C 0.240 0.080 0.080 0.02C NIACIN (MG I 0.90C 0. 800 1.200 1 . 100 1 .400 VIT-B6 (MG) 90.000 40.000 20.000 20.000 170.000 FOLIC (UG) 7.OOC 8. 000 2.000 2.000 7.000 VIT-B12(UG) 0 .0 0.0 0.0 0.0 0.0 VIT-C ( MG) 0.0 0.0 0.0 0 .0 0.0 PANTO (UG) 200.OCC 700.OCO 200.000 100.000 400.000 BIOTIN (MG) 16.000 5.000 2.000 0.0 12.000 VIT-A (IU) 0.0 250.OCO 70.000 0. 0 0.0 VIT-D (IU) 0 .0 7.000 1.000 0.0 0.0 VIT-E (MG) 0.300 0.900 0.400 0.400 0.200 CA (MG) 8.000 215.000 10.000 8.000 12.000 P (MG) 76.OCC 260.000 59.000 50.000 73.000 MG (MG) 31.000 14.000 23.000 13.000 29.000 FE ( MG) 0.70C 1. 200 0.900 0.900 0.500 I (MG ) 3.OOC 6. OCO 1.000 1.000 2.000 ZN (MG) 0.90C 0.600 0.600 0. 100 0.300 NA (MG) 0.0 564.000 2.000 1 .000 282 .000 K (MG) 84.OOC 154.000 44.000 61.000 70.000 CU (MG ) 0.28C 0.050 0.010 0.050 0.110 243 ITEM CLUSTER 21 22 ATTRIBUTE GROUP 8 8 NUTRIENTS: KC AL 109.OOC 50.000 PROT (GM) 2.OCC 1. 100 TRY (MG) 20.000 10.000 THR (MG) 80.OCC 40.000 ISO (MG) 90.OCC 30.OCO LEU (MG) 170.OCC 140.000 LYS (MG) 80.OCC 30.OCO MET (MG) 40.000 20.000 CYS (MG) 20.000 10.000 PHE (MG) 100 .000 50.000 TYR ( MG) SO.000 70.000 VAL (MG) 140.OCC 60.000 HI S (MG) 30.OCC 20.000 FAT-T (GM) 0 .IOC 0.2C0 SFA . (GM) 0.0 0.0 PUFA (GM ) 0 .0 0. 0 CHCLE (GM ) 0.0 0.0 CHO-T (GM ) 24.200 10.700 SUCR (GM ) 0.100 0. 200 CHO-F (GM) 0. ICO 0. 100 THIA (MG) 0. 11C 0. 06C RI BO (MG) 0.01C 0.040 NIACIN (MG) l.OCC 0. 500 VIT-B6 (MG) 40.OCC 30.000 FOLIC (UG) 1 .000 1.000 VIT-B12(UG) 0 .0 0.0 VIT-C (MG) 0.0 0.0 PANTO (UG) 200.OCC 100.000 BIOTIN (MG) 5.000 7.000 VIT-A (IU) 0.0 60.000 VIT-D (IU) 0.0 0.0 VIT-E (MG) O.iOC 0. 100 CA (MG) 10.000 1.000 P ( MG) 28.OCC 14.000 MG (MG) 6.000 8.000 FE (MG) 0.9CC 0.400 I (MG ) l.OCC 2. OCO ZN (MG) 0.20C 0.200 NA (MG ) 374.OOC 0. 0 K (MG) 28 .000 16.000 CU (MG) 0.05C 0.030 23 24 25 8 9 9 363.OOC 290.000 262.000 26.600 9.100 6.600 270.000 110.000 80.000 1410.000 260.000 130.000 1250.000 410.000 280.000 1810 .000 700.COO 440.000 1620.000 210.000 170.000 430.000 120.000 100.000 310.000 200.000 150.000 1000.000 500.000 360.000 940.000 260 .000 190.000 1440.000 390.000 300.000 730.000 340.000 250.000 10.900 3. 000 2.800 2.000 1.000 1.000 8.000 1 .000 1.000 0.0 0.004 0.004 46.700 55.400 53 .600 1.000 1-000 4. 100 2 .500 0 .200 0.900 2.010 0.280 0.050 0.680 0.220 0.090 4.200 2.500 0.700 920 .000 50.000 40.000 305.000 9.000 17.000 0.0 0.0 0.0 0.0 0.0 0.0 2200.000 400.000 400.000 20.000 1 .000 1.000 0.0 0. 0 0.0 0.0 o.c 0.0 13.500 0.100 0.100 72.000 43.000 71.000 1118.000 85.000 87.000 323.000 22.000 24.000 9 .400 2 .200 1.300 2.000 9.000 9.000 13.200 1.4C0 1.200 3.000 580.000 365.000 827.000 90.000 233.000 2.910 0.230 0.230 -244 ITEM CLUSTER 26 27 28 29 30 ATTRIBUTE GROUP 9 9 9 9 9 NUTRIENTS: KCAL 243.OCC 270.000 243 .000 207.000 210.000 PROT (GM) 9. IOC 8. 7CO 10.500 7.400 5.000 TRY (MG) 100.OCC 100.000 140.000 70.OCO 30.000 THR (MG) 290.CCC 270.OCC 320.000 270.000 200.000 ISO (MG) 390.000 420.000 480.000 350.000 290.000 LEU (MG) 620.000 690.000 760.000 660.000 810.000 LYS (MG) 290.000 260.000 290.000 360.000 130.000 MET (MG) 140.OCC 130.000 160.000 130.000 70.000 CYS (MG) 2CO.0CG 2 CO.OOC 240.000 70.000 50.000 PHE (MG) 500 .OOC 480.000 570.000 400.000 210.000 TYR (MG) 260.CCC 250.000 400.000 330.000 90.000 VAL (MG) 480.OCC 400.000 500.000 380.000 260.OOC HIS (MG ) 350.COG 330.000 220.000 150.000 70.000 FAT-T (GM ) 1 .100 3.200 3.000 7.200 2.000 SFA (GM) 0.0 1.000 1 .000 2 .000 1.000 PUFA (GM) 1.000 2.000 1.000 4. 000 1.000 CHOLE (GM) 0.0C4 0.005 0.005 0.006 0.0 CHO-T (GM) 52. ICC 50.5 CO 47.700 29.100 45.000 SUCR (GM) 1 .000 1.000 1.000 l.OOC 0.700 CHO-F (GM ) 0.4CC 0. 200 1.6C0 0.500 1 .000 TH I £ (MG) 0.180 0.250 0. 260 0. 130 0. 13C RIBO (MG) 0.07C 0. 210 0.120 0.190 0.050 NIACIN (MG) 1.400 2.400 2.800 0.60C 1.000 VIT-B6 (MG) ICO.OOC 40.000 180.000 110 .000 70.000 FOLIC (UG) 38.CCC 17.000 38.000 4.000 1.000 VI T-E12(UGI 0 .0 0.0 0 .0 O.C 0.0 VIT-C (MG) O.C 0. 0 0.0 1 .000 0.0 PANTO (UG) 5C0.0C0 400.000 800 .000 300.000 100.000 BIOTIN (MG) l.OCC 1.000 2.000 1.000 2.000 VIT-A (IU) 0*0 0.0 0.0 150.000 20.000 VTT-D ( IU) 0.0 15.000 8 .000 5 .000 0.0 VIT-E (MG) 0.300 0. 100 0.400 0.200 0. 100 CA (MG) 75.OCC 84.000 99.000 120.000 200.000 P (MG) 147.OCC 97.OCC 228.000 211.000 140.000 MG (MG) 42.OOC 26.000 45.000 15.000 107.000 FE (MG) 1.6CC 2. 500 2.300 1 .100 3 .000 I (MG ) 9.000 9.000 11. OCC 5.000 6.OOC ZN (MG) 1.60C 1.300 2.800 0.700 0.100 NA (MG) 557.000 507.000 527.000 628.000 110.000 K ( MG) 145.OCC 105.000 273 .000 157.000 16.000 CU (MG) 0.22C 0. 23 C 0.220 0. 080 0.190 .245 ITEM CLUSTER ATTRIBUTE GROUP NUTRIENTS: KC AL PROT (GM) TRY (MG) THR ( MG) ISO (MG) LEU (MG) LYS (MG) MET (MG) CYS (MG ) PHE (MG) TYR ( MG) VAL (MG) HIS (MG) FAT-T (GM) SFA (GM) PUFA (GM) C HOLE (GM ) CHG-T (GM) SUCR (GM) CHO-F (GM) THIA (MG) RI BO (MG) NIACIN (MG) VIT-B6 (MG) FOLIC (UG) VIT-B12(UG) VIT-C (MG) PANTO (UG) BIOTIN (MG) VIT-A (IU ) VIT-D ( IU ) VIT-E (MG) CA (MG ) P ( MG) MG (MG ) FE (MG) I (MG) ZN (MG) NA (MG) K (MG) CU (MG) 31 32 10 10 325.000 270.000 7.100 8.700 90.000 100.000 210.OOC 270.000 330.OCC 420.OCO 550.OCC 690.000 160•OOC 260.000 90.OOC 130.000 140.OCC 200.000 390.000 480.000 230.OCO 250.000 310.000 400.000 130.OOC 330.000 9.300 3. 200 2.OCC 1.000 7.OOC 2. OCO 0.006 0.005 52.3C0 50.500 1 .000 1.000 0.200 0. 200 0.27C C.250 0.250 0.210 2. OCC 2. 400 40.000 40.000 8.OOC 17.000 0 .0 0.0 0.0 0.0 400.OOC 400.OCO 1 .000 1.000 0.0 0. 0 0.0 15.000 0.200 0. 100 68 .000 84.000 232.OCC 97.000 24.0GC 26.000 2.3C0 2. 500 9.OCC 9. OOC 1 .200- 1.300 973.000 507.COO 116.000 105.000 0.310 0.230 33 34 10 10 294.000 290.000 7.800 9.100 100.000 110.000 280.000 260.000 420.000 410.000 650.000 700.COO 330.000 2 10.000 140.000 120.000 170.000 200.000 430.000 500.000 290 .000 260.000 40.000 390.000 150.000 340.000 10.100 3. 000 2.000 1.000 7.000 I.000 0.006 0.004 42.300 55.400 1.000 1. 000 0.100 0 .200 0.170 0.280 0.230 0.220 1 .400 2.500 50.000 50.000 8.000 9 .000 0.160 0. C 0.0 0.0 500.000 400.000 1.000 1 .000 100.000 0.0 8.000 0.0 0.200 0.100 104.000 43.000 151 .000 85.000 24.000 22.000 1 .600 2 .200 9.000 9. 000 1.200 1.400 441.000 580.000 125.OOC 90.000 0.220 0.230 35 10 243.000 10.500 140.000 320.000 480.000 760.000 290.000 160.000 240.000 570.OOC 400.000 500.000 220.000 3.000 1.000 1 .000 0.005 47.700 1.000 1.600 0.260 0. 120 2.800 180.000 38.000 0.0 0.0 800.000 2.000 0.0 8.00 0 0.400 99.000 228.000 45.000 2. 300 11.000 2. 800 527.000 273.OOC 0.220 .246 ITEM CLUSTER 36 37 ATTRIBUTE GROUP 11 11 NUTRIENTS: KCAL 384.000 433.000 PROT (GM) 8.OCC 9.000 TRY (MG) 100 .000 110.000 THR ( MG) 230.OCC 260.000 ISO (MG) 370.OCC 420.OCC LEU (MG) 620.OOC 690.000 LYS (MG) 180.OOC 210.000 MET (MG) 100.000 120.000 CYS (MG) 160.OOC 180.000 PHE (MG) 440.000 490.000 TYR ( MG) 270.OCO 310.000 VAL (MG) 3 4 0.CCC 390.000 HIS (MG) 160.OCC 180.000 FAT-T (GM ) 9.4CC 12.000 SFA (GM) 2.OOC 3.000 PUFA (GM) 7.00C 8. 000 CHOLE (GM ) 0 .0 0.0 CHO-T (GM) 73.300 71.500 SUCR (GM) 24.100 0. 200 CHO-F (GM) 1.1CC 0.400 THI A (MG) 0 .040 0. 010 RIBC (MG) 0.21C 0.040 NIACIN (MG ) 1. 500 1.000 VIT-B6 (MG) 70.000 70.000 FOLIC (UG) 24.000 24.000 VIT-B12(UG) 0 .0 0.0 VIT-C (MG) 0.0 0.0 PANTO (UG) 500.OCC 500.OCO BIOTIN (MG) 1.000 1 .000 VIT-A (IU) 0.0 0. 0 VIT^D (IU) 0 .0 0.0 VIT-E (MG) 0.100 0. 100 CA (MG) 40.000 21.000 P ( MG) 149.OOC 90.000 -MG (MG) 36.CCC 25.000 FE (MG) 1.50C 1.200 I (MG) 2.OOC 2. OCO ZN (MG) 0.800 0.700 NA (MG ) 670.OOC 1100.000 K (MG) 384 .000 120.000 CU (MG) 0.180 0.170 38 39 40 11 11 11 386.000 568.000 354.000 12.700 5.300 10.200 80.000 60.000 120.000 510.000 210.000 340.000 590.000 230.000 470.000 1670.000 270.000 840.000 370.000 280.000 340.000 240.000 70.000 120.000 130.000 50.000 180.000 1000.000 250.000 500.000 620.000 100.000 300.000 660.000 290.000 540.000 330.000 80.000 220.000 5.000 39.800 1.600 1.000 10.000 0.0 4.000 28.000 1 .000 0.0 0.0 0.0 76.700 50 .000 80.500 1.000 0.200 20.100 2.200 1 .600 1.600 0.420 0.210 0.640 0. 120 0.070 0.140 2.200 4.800 4.900 200.OOC 180.000 290.000 0.0 9.000 18.000 0.0 0. 0 0.0 0.0 16.000 0.0 400.000 500.000 500.000 1.000 7.000 1.000 0.0 0.0 0.0 0.0 O.C 0.0 4.400 4.300 0.500 11.000 40.000 41.000 281 .000 139.000 309.000 173.000 43.000 96.000 2 .700 1 .800 4.400 14.000 13.000 14.000 2.000 2.500 2.400 3.000 1000.000 1032.000 200.000 1130.000 120.000 0.310 0.220 0.450 247 ITEM CLUSTER 41 42 43 44 45 ATTRIBUTE GROUP 12 12 12 12 12 NUTRIENTS: KCAL 377.000 286.000 261.000 216.000 266.000 PROT (GM) 24.2CC 24.200 28 .600 25.300 25.800 TRY (MG) 300 .000 250.000 350.000 290.000 330.000 THR ( MG) 1100.OCC 960.000 1300.000 1100.000 1180.000 ISO (MG) 1300.OCC 1130.OOC 1500.000 1300.000 1330.000 LEU (MG) 2100.OOC 1770.000 2500.000 2100.000 1990.000 LYS (MG) 2200.OCC 1890.000 2600.000 2200.000 2060.000 MET (MG) 640.OCO 540.000 760.000 630.000 630.000 CYS (MG) 320.OCC 270.000 380.000 320.000 340.000 PHE (MG ) 990 .000 1000.000 1170.000 1040.000 1000.000 TYR ( MG) 9C0.00C 740.000 1060.000 900.000 890.000 VAL (MG) 1500.CCC 1190.000 1800.000 1400.000 1260.000 HI S (MG) 890.OOC 750.000 1050.000 8 70.000 720.000 FAT-T (GM) 30. 300 20.300 15.400 12.000 17.300 SFA (GM) 14.000 10.000 7.000 6.COO 10.000 PUFA (GM) 14.000 9.000 7.000 5.000 5 .000 CHOLE (GM) 0.070 0.070 0.07C 0.070 0.070 CHO-T (GM) 0.0 0. 0 0.0 0 .0 0.0 SUCR (GM) 0 .0 0.0 0.0 0. 0 0.0 CHO-F (GM) O.C 0.0 0.0 0 .0 0.0 THIA (MG) 0.040 0. 090 0.080 0.020 0.150 RI BC (MG) 0.190 0.210 0.220 0.240 0.270 NIACIN (MG) 3.80G 5.400 5.600 3.400 5.600 VIT-B6 (MG) 350.000 460.000 410.000 100.000 320.000 FOLIC (UG) 3.OOC 5.000 4.000 3 .000 3.000 VIT-B12(UG) 1.800 0.900 2.650 1. 840 3.100 VI T-C (MG) 0.0 0.0 0.0 0.0 0.0 PANTO (UG) 4CO.0CC 400.OCO 500.000 600.000 600.000 BIOTIN (MG) 3.OCC 3.000 3.000 3.OOC 6.000 VIT-A (I U) 60.OCC 40.OOC 30.000 20.000 0.0 VIT-D (IU) 0.0 0.0 0.0 O.C 0.0 VIT-E (MG) 0.200 0.400 0.200 0. 100 0.200 CA (MG ) 10.000 11.000 12.000 20.000 11.000 P (MG) 121.CCC 194.000 250.000 106.000 212.000 MG (MG) 23.OCC 21.OCO 28.000 27.000 22.000 FE (MG) 3.ICO 3.200 3.500 4.300 1.800 I (MG) 6.000 6. OCO 6.000 6.000 3.000 ZN (MG) 2.500 4.300 3.000 3.100 5.400 NA (MG) 60.OOC 47.000 60.000 1010.000 70.000 K (MG) 370.000 450.000 370.000 87.000 290.000 CU ( MG) 0.08C 0.080 0.080 0 .210 0.240 248 ITEM CLUSTER 46 47 ATTRIEUTE GROUP 12 12 NUTF-IENTS: KCAL 373.000 193.000 PR07 (GM) 22.600 18.300 TRY (MG) 3 GO.OCC 180.000 THR (MG) 1050 .OCO 760.000 ISO (MG) 1160.OCC 920.000 LEU (MG) 1650.000 1430.000 LYS (MG) 1850.OCO 1550.000 MET (MG) 570.000 450.000 CYS ( MG) 210.OCC 300.000 PHE (MG) 890.000 700.000 TYR (MG ) 810.000 720.000 V AL (MG) 1190.CCC 960.000 HI S (MG) 780.OCC 600.000 FAT-T (GM ) 30.600 12.300 S FA (GM ) 12.000 4.000 PUFA (GM) 15.000 6.000 CHOLE (GM) 0 .070 0.070 CHO-T (GM) 0.0 0.900 SUCR (GM) 0.0 0. C CHO-F (GM) 0 .0 0.0 THIA (MG) 0.500 0. 530 RIBO (MG) 0 .230 0. 190 NIACIN (MG ) 4.900 3. 800 VIT-B6 (MG) 320 .000 360.000 FOLIC (UG) 3.OCC 3.000 VIT-B12(UG) 0.500 0. 500 VIT-C (MG) 0.0 0.0 PANTO (UG) 500.OCC 400.OCO BIOTIN (MG) 5 .000 5.000 VIT-A (I U) 0.0 0. 0 VIT-D (IU) 0 .0 0.0 VIT-E ( MG) 0.20C 0. 200 CA (MG) 10.CCC 11.000 P (MG) 232.000 156.000 MG (MG ) 23.CCC 15.OCO FE (MG) 2 .900 2.700 I (MG ) 10.OOC 8.000 ZN (MG) 2.700 1.900 NA (MG) 65.OOC 1100.000 K (MG) 3^0.CCC 340.000 CU ( MG) 0.09C 0.440 48 49 50 12 13 13 611.000 198.000 161.000 30.400 21.700 15.300 320.000 250.000 230.000 1000.000 890.COO 790.000 1300.000 800.000 1000.000 2400.000 1100.000 1340.000 2000.000 1300.000 1630.000 470.000 400.000 480.000 360 .000 210.000 250.000 1400.000 600.000 490.000 780.000 540.000 650.000 1400.000 750.000 920.000 820.000 440.000 530.OOC 52.000 11.700 9.300 17.00 0 4.000 3.000 30.000 6.000 6.000 0.C80 0. 070 0.040 3 .200 0.0 3. 200 3.200 0. 0 0.0 0.0 0.0 0.0 0.510 0.040 0.020 0. 340 0. 120 0.07C 5.200 4.400 2 .400 100.000 300.000 250.000 2 .000 2 .000 4.000 0.700 0. 790 0.260 0.0 4.000 0.0 1300.000 900.000 900.000 8.000 11.000 6.000 0.0 230.000 70.000 0.0 0.0 0.0 0 .300 0.100 0. 200 14.000 21.000 6. 000 224.000 247.000 113.000 25.000 18.000 11.000 3.300 1.500 0.900 27.000 5.000 6.000 5.100 4.900 2.400 1021 .000 400.000 154.000 236.000 138.000 140.000 0.520 0 .230 0. 180 249 ITEM CLUSTER 51 52 53 54 55 ATTRIBUTE GROUP 13 13 14 14 14 NUTRIENTS: KCAL 250.OCC 263.000 98.000 176.000 165.000 PPOT (GM) 30 .600 27.000 15.800 16.600 19.600 TRY (MG ) 370.OCC 330.OCO 160.000 170.000 200.000 THR (MG) 1290.000 1140.000 680.000 720.000 840.000 ISO (MG) 1600.000 1420.000 800.000 840.000 1000.000 LEU (MG) 2200 .000 2050.000 1200.000 1250.000 1490.000 LYS ( MG) 27C0.CCC 2450.000 840.000 1460 .000 1730.000 MET (MG) 8 CO.000 750.000 460.000 480.COO 570.000 CYS (MG) 410.000 370.000 210.000 220.000 260.000 PHE (MG • 1200.OCC 1080.OCO 580.000 620.000 730.000 TYR (MG) 1080.OCC 1080.000 430.000 450.OCO 530.COO VAL (MG) 1500.OCC 1320.000 480.000 880.000 1040.000 HIS (MG) 880.000 730.000 30.000 760.000 900.000 FAT-T (GM) 11.900 16.400 2.500 8.900 6.400 SFA (GM) 3.000 4.000 1.000 2.000 2.000 PUFA (GM) 8.OCO 10.000 I .000 6 .000 4.000 CHOLE (GM) 0.C7C 0. 075 0.080 0. 070 0. 060 CHO-T (GM) 2 .800 0.0 1.900 6.500 5. 80C SUCR (GM • 0.0 0. 0 0.0 0.0 0.0 CHO-F (GM ) 0 .0 0.0 0.0 0.0 0.0 THIA (MG) 0.060 0. 110 0.010 0.040 0.040 RIBO (MG ) 0.360 0. 200 0.110 0.070 0.07 0 NIACIN (MG) 9.200 11.400 1 .100 I .600 3.200 VIT-B6 (MG) 4C0.0CC 400.000 80.000 50.000 140.000 FOLIC (UG ) 6.OCC 10.000 3 .000 16.000 16.000 VIT-B12(UG> 0.42C 0.420 19.100 1. 000 1.000 VIT-C (MG) 0.0 2.000 11.000 2.000 2.000 PANTO (UG) 9C0.C0C 900. 000 300.000 300.000 300.000 BIOTIN (MG ) 11.000 11.000 20.000 3.000 3.000 VIT-A ( IU) 170.OCC 170.000 110.000 0.0 320.000 VIT-D (IU) 0.0 0.0 3.000 0. 0 0.0 VIT-E (MG) 0.200 0.300 0.300 0.600 0.600 CA (MG ) 12.000 11.000 55.000 11.000 40.000 P (MG) 243.OCC 300.000 184.000 167.000 247.000 MG (MG ) 19.OCC 25.OCO 113.000 18.000 36.000 FE (MG) 1.800 2.100 4.100 0.400 1.200 I (MG ) 7.000 6. 000 90.000 34.000 62.000 ZN (MG) 4.600 2. 800 1.60C 0.300 0.300 NA (MG) 88.OOC 93.000 1010.000 180.000 177.000 K (MG) 428.OCC 443.CCO 140.000 390.000 348.000 CU (MG) 0.33C 0. 180 0.0 0.140 0. 150 250 ITEM CLUSTER 56 57 58 5S 60 ATTRIBUTE GROUP 14 14 14 14 14 NUTRIENTS: KCAL 171.OCC 76.CCO 239.000 182.000 210.000 PROT (GM) 25 .200 8.500 8.600 27.000 19.600 TRY (MG) 25 0. CCO 90.000 90.000 270.000 200.000 THR (MG) 1080 .000 370.000 370.000 1160.000 840.000 ISO (MG) 12 9 0.OOC 430.000 440.000 1350.000 980.000 LEU (MG) 1920.000 640.000 650.000 2030.000 1470.000 LYS (MG) 2220.OOC 280.000 280.000 2350 .000 1710.000 MET (MGI 730.OCC 250.000 250.000 780.000 570.000 CYS (MG) 340.OOC 110.000 120.000 380.000 280.000 PHE (MG ) 940.OCC 320.000 320.CCO 1000.000 730.000 TYR (MG) 690.000 230.000 230.000 730.000 530.000 VAL (MG) 1330.OCC 450.000 460.000 1430.000 1040.000 HIS (MG) 1160.000 960.000 960.000 700.000 510.000 FAT-T (GMl 7. OOC 2. 200 13 .900 7 .400 14.000 SFA (GM) 3 .CCC 1. 000 4.000 2.000 4.000 PUFA ( GM) 3.000 1.000 8.000 3 .000 4.000 CHOLE (GM) 0.06C 0.230 0.230 0. 060 0.060 CHG-T (GM) 0.0 4.900 18.600 0.0 0.0 SUCR (GM) 0.0 0.0 0.0 0.0 0.0 CHO-F (GM) 0 .0 0. 100 O.C 0. 0 0.0 THIA ( MG) 0.05C 0. 020 0.170 0.160 0.030 RIBO (MG ) 0.07C 0. 200 0.290 0. 060 0.140 NIACIN (MG) 8.3CC 0.800 3 .200 9.800 7. 300 VIT-B6 (MG) 340.OCC 40.OCO 40.000 300.000 300.000 FOLIC (UG) 16.000 3.000 3.000 7.000 7.OOC VIT-B12(UG) l.COC 18.OCO 18.000 1 .000 6.900 VIT-C (MG ) 4.000 26.000 39.OCO 5.000 9.000 PANTO (UG) 3 CO.OCC 200.000 200.000 500.000 600.000 BIOTIN (MG) 8 .000 9.000 9.000 12.000 12.000 VIT-A ( I U) 680.000 260.000 440.000 160.000 230.000 VIT-D ( IU ) 0.0 10.CCO 5.000 400.000 370.000 VIT-E (MG) 0 .60C 0.300 0.600 1.400 0.500 CA (MG) 16. OOC 28.000 152.000 80.000 154.000 P (MG) 248.OOG 124.000 241.000 414.000 289.000 MG (MG) 23.OOC 17.000 17.000 41.000 27.000 FE (MG) 0.800 5.600 8.100 1. 200 0.900 I ( MG) 46.OCC 48.000 69.000 37.000 51.000 ZN (MG) 1.000 52. CCC 52.000 1. 700 0.700 NA (MG) 134.OCC 62.000 206.000 116.000 407.OOC K (MG) 525.OCC 70.OCO 203.000 443.000 366.000 CU (MG) 0 .190 3.430 4.270 0. 800 0.290 .251 ITEM CLUSTER 61 62 63 64 65 ATTRIBUTE GROUP 14 14 14 14 14 NUTRIENTS: KC AL 176.CCC 203.000 94.000 223 .000 197.000 PROT (GM) 21.600 24.000 21.000 12.300 28.800 TRY ( MG) 220.OCC 210.000 240.000 150.000 290.000 THR (MG) 930.000 920.000 1040.000 560.000 1240.000 I SO (MG) 1040.OOC 1070.000 1230.000 690.000 1470.000 LEU (MG ) 1620.OCC 1590.OCO 1840.000 1050.000 2160.000 LYS (MG) 1890.OCC 1850.000 2130.000 1030.000 2530. COO MET (MG) 630. OCC 610.OCO 700.000 360.000 840.000 CYS (MG) 310.OOC 280.000 320.000 180.000 390.000 PHE (MG ) 8 CO.OOC 880.000 850.000 460.000 1000.000 TYR (MG) 580 .000 570.000 650.000 360.000 780.000 VAL ( MG) 1150.OCC 1120.000 1280.000 450 .000 1530.000 HIS (MG) 5*0.OCC 990.OCO 530.000 270.000 1550.000 FAT-T (GM) 9.30C 11.100 0.800 11.OCC 8. 200 SFA (GM) 3. OCC 5. OCO 0.0 7. 000 3.000 PUFA (GM) 3.OCC 5.000 0.0 . 0.0 4.000 CHOLE (GM ) 0.06C 0.070 0. 140 0.140 0.060 CHO-T (GM ) 0.0 0.0 0.500 18.600 0.0 SUCR (GM) 0.0 0.0 0.0 0 .0 0.0 CHO-F (GM) 0 .0 0.0 0.100 0. 100 0.0 THIA ( MG) 0.21C 0.030 0.020 0.030 0.050 RIBO (MG ) 0.08C 0. 200 0.030 0. 030 0. 120 NIACIN (MG) 12.70C 5.400 3.300 2.000 11.900 VIT-B6 (MG) 7 CO. COG 180.OCO 50.000 60.000 430.000 FOLIC (UG) 7 .000 32.000 2.000 2.000 1.000 VIT-B12(UG ) 7.OCC 10.000 0.690 0.720 2.200 VIT-C (MG) 5 .OOC 0. 0 11.000 7. 000 10.000 PANTO (UG) 7C0.C0C 900.000 200.000 300 .000 300.000 BIOTIN (MG) 12.OCO 24.OCC 10.000 10.000 3.000 VIT-A (IU) 190.OOC 220.000 40.000 30.000 80.000 VIT-D (IU) 4CO.0CC 5C0.000 150.OCO 150.000 250.000 VIT-E (MG) 1 .400 0.600 0.500 0.60C 0. 500 CA (MG) 14.OCC 437.000 78.000 38.000 8.000 P (MG) 245.000 499.000 208.000 111.000 234.000 MG (MG) 33.CCC 39.000 42.000 61.000 27.000 FE (MG) 1.4CC 2. 9C0 1.700 1. COO 1.900 I (MG) 37.OOC 37.000 65.000 66.000 16.000 ZN (MG) 1.3CC 2.9C0 1.500 1 .000 0.400 N A (MG) 134.000 823.000 126.000 213.000 662.000 K (MG) 512.OCC 590.000 203.000 197.000 249.000 CU (MG ) 1 .300 0.040 0.57C 0.37C 0. 120 252 ITEM CLUSTER 66 67 ATTRIBUTE GROUP 14 15 NUTRIENTS: KCAL 133.OCC 229.000 PROT (GM) 27.400 26.400 TRY ( MG) 250.OCC 390.000 THR (MG) 1C60.0GC 1240.CCC I SO (MG) 1260.OOC 1360.000 LEU (MG ) 1850.OCC 24CO.OC0 LYS (MG) 2170.OCC 1950.000 MET (MG) 720.OOC 610.000 CYS (MG) 330 .000 320.000 PHE ( MG) 860.OCC 1300.000 TYR (MG) 670.000 980.000 VAL (MG) 1310.OOC 1690.000 HIS (MG) 1330. CCC 1240.000 FAT-T (GM) 3.OCC 10.600 SFA (GM) l.OCC 3. OCO PUFA (GM) 1.000 6.000 CHOLE (GM) 0.060 0. 300 CHO-T (GM) 0.0 5.300 SUCR (GM) 0.0 0.0 CHO-F (GM ) O.C 0. 0 TH IA (MG) 0.02C 0.260 RIBO (MG) 0.05C 4. 190 NIACIN (MG) 6.600 16.500 VIT-B6 (MG) 900.000 670.000 FOLIC (UG) 3.000 294.000 VIT-B12(UG> 3.OOC 80.000 VIT-C (MG) 7.OCC 27.000 PANTO (UG) 500.OOC 7100.000 BIOTIN (MG ) 3. OCC 96.OCO VIT-A (IU) 50.OOC 53400.000 VIT-0 (IU) 2 50.000 50.000 VIT-E (MG) 0 .200 0.600 CA (MG) 4. OOC 11.000 P (MG) 177.OCC 476.000 MG (MG) 29.OOC 22.000 FE (MG) 1,300 8. 800 I (MG) 23.OOC 19.000 ZN (MG) 0.50G 7.000 NA (MG) 37.OOC 184.000 K (MG) 181.OOC 380.000 CU (MG ) 0.50C 3.700 68 6S 70 16 17 17 245.000 304.000 476.000 27.500 12.400 18.100 370.000 120.000 160.000 1270.000 510.000 720.000 1430.000 620.OCC 880.000 2450.000 970.000 1290.000 2275.000 1100.000 1460.OOC 685.000 300.000 380.000 350.000 200.000 230.000 1235.000 460.000 600.000 1020.000 480.000 590.000 1745.000 650 .000 920.000 1145.000 400.000 490.000 13.000 27.20C 44.200 5.000 10.000 16 .000 7.000 15.000 23.000 0.190 0.070 0.070 2.700 1.600 0.0 0.0 I .600 0.0 0.0 0. 0 0.0 0.170 0. 150 0.790 2.210 0.200 0.340 11. 100 2.500 3.700 540.000 110.000 190.000 149.000 4.000 4.000 41 .330 I .300 1.400 14.000 0.0 0.0 3800 .000 400.000 600.000 50.000 5.000 5. 000 26715.000 0.0 0.0 25.000 0.0 0.0 0.400 0. 100 0.200 12.000 5 .000 7.000 363.000 102.000 162.000 25.000 9. COO 16.000 6.200 1. 500 2.400 13.000 8.OOC 8.000 5.000 1 .500 0.600 122.000 1060.000 958.000 375.000 212 .000 269.000 1.890 0. 080 0.150 253 ITEM CLUSTER 71 72 73 74 75 ATTRIBUTE GROUP 17 18 18 18 18 NUTRIENTS : KCAL 307.OCC 90.000 118.000 108.000 118.000 PROT (GM) 16.2CC 5.700 7.800 8. 100 9.800 TRY ( MG ) 240.OCC 50.000 70.000 80.000 100.000 THR (MG) 760.CCC 250.OCO 340.000 320.000 370.000 ISO (MG) 840.OOC 320.000 450.000 390.000 490.OOC LEU (MG) 1470.OCC 490.OCO 670.000 600.000 680.000 LYS (MG) 1180.000 420.000 580.000 530.000 620.000 MET (MG) 3 80.OCC 60.000 80.000 120.000 140.000 CYS (MG) 200 .000 60.000 80.000 90.000 100.000 PHE (MG) 820.OOC 310.000 400.000 450 .000 530.000 TYR (MG) 6CO.OCC 220.000 300.OOC 200.000 380.000 VAL (MG) 1010 .OCC 340.000 480 .000 450.000 540.000 HIS (MG) 760.OCC 160.000 220.000 270.000 330.000 FAT-T (GM ) 25.60C 0.400 0.600 0.8C0 5. 100 SFA (GM) 10.000 0. 0 0.0 0.0 1.000 PUFA (GM) 14.000 0.0 0.0 0.0 4.000 CHOLE (GM) 0.346 0.0 0.0 0.0 0.0 CHO-T (GM ) 1.800 16.400 21.200 18.100 10.100 SUCR (GM) 0.0 0.700 0.700 1 .300 3.400 CHO-F (GM ) 0.0 0.900 1.5CC 1. 800 1.400 THI A (MG) 0.2CC 0.050 0. 140 0. 200 0.31C RIBO (MG) 1.3CC 0.040 0.070 0.110 0.130 NIACIN (MG) 5.700 0. 600 0.700 1.400 1.200 VIT-B6 ( MG) 930.OCC 330.000 140 .000 50 .000 40.000 FOLIC (UG) 6.OCC 6.000 8.000 26.000 38.000 VIT-B12(UG) 2.36C 0.0 0 .0 O.C 0.0 VIT-C (MG) O.C 0. C 0.0 17.000 17.000 PANTO (UG) 5900.OOC 100.000 200.000 300.000 700.000 BIOTIN (MG ) 111.OOC 4.000 6. OCO 10.000 30.000 VIT-A (IU) 6350 .000 0. 0 0.0 350.000 660.000 VIT-D (ID 15. OCC 0.0 0 .0 0.0 0.0 VIT-E (MG) 0.700 0. 100 0.200 0.200 0.700 CA (MG) 9.000 29.000 50.000 24.000 60.000 P (MG ) 238.OCC 109.000 148.000 146.000 191.000 MG (MG) 23.OCC 27.000 37.000 19.OOC 194.000 FE (MG) 5.4CC 1. 800 2.700 2.100 2 .500 I (MG) 23.000 2.000 3. OCC 7.000 4.000 ZN (MG) 7.5CC 1.100 1 .500 0.800 1.100 NA (MG) 291 .000 236.000 7.000 1. 000 2.000 K (MG) 232.CCC 264.000 416 .000 379.000 487.000 CU (MG) 3.05C 0. 100 0.240 0.280 0. 810 254 ITEM CLUSTER 76 77 78 79 80 ATTRIBUTE GROUP 19 19 19 19 19 NUTRIENTS : KCAL 598.OOC 561.000 346.000 581.000 568.000 PROT (GM ) 18.6CC 17.2CC 3.500 27.800 26.300 TRY (MG) 180.OOC 430.000 32.000 360.000 340.000 THR (MG) 6 10 .OCC 690.OCO 126.000 870.000 820.000 ISO (MG) 870.000 1140.000 175.000 1300.000 1250.000 LEU (MG) 1450.OCC 1410.000 260.000 2000.000 1850.000 LYS (MG) 580.000 740.000 148.000 1100.000 1090.000 MET ( MG) 260.OCC 330.000 69.000 280.000 270.000 CYS (MG) 380.CCC 480.OCO 60.000 480.000 450.000 PHE (MG) 1100.OOC 950.000 170.000 1600.000 1500.000 TYR (MG) 620.OCC 660.000 106.000 1200.000 1090.000 VAL (MG) 1120.OOC 1480.000 205.000 1600.000 1520.000 HIS (MG) 5 2 0. CCC 390.OCO 670.000 780.000 740.000 FAT-T {GM ) 54.200 45.700 35.300 49.400 48.400 SFA (GM) 4. OCC 8.000 30.000 9.000 10.000 PUF* (GM) 47.OOC 35.000 2.000 39.000 34.000 CHOLE (GM) 0.0 0.0 0.0 0 .0 0.0 CHO-T (GM * 19.5CC 29.300 9.400 17. 200 17.600 SUCR (GM ) 2.300 3.000 4.700 4. 500 4.500 CHO-F (GM ) 2.600 1.4C0 4.000 1 .900 1 .900 TH IA (MG) 0.240 0.430 0.050 0. 130 0.990 RIBO (MG) 0.92C 0.250 0.020 0. 130 0.130 NIACIN (MG) 3.500 1.800 0. 500 15.700 15.800 VIT-B6 ( MG) ICO.CCC 400.000 40.000 330.000 400.000 FOLIO (UG) 45.OCC 2 5. CCC 27.000 13.000 25.000 VIT-B12(UG ) 0.0 0.0 0.0 0.0 0.0 VIT-C (MG) 0.0 0. 0 3.000 0.0 0.0 PANTO (UG) 500.000 1300.000 200.000 1700.000 2800.000 BIOTIN (MG) 8.000 30.000 6.000 39.000 34.000 VIT-A ( IU) 0 .0 100.000 0.0 0. C 0.0 VIT-D (IU) 0.0 0. 0 0.0 0.0 0.0 VIT-E (MG) 15.CCC 5. ICC 1.000 6. 700 6.500 CA (MG) 234.000 38.000 13.000 63.000 59.000 P (MG) 5C4.C0C 373.OCO 95.000 407.000 409.000 MG (MG) 269.000 267.000 44.000 174.000 168.000 FE (MG) 4.7CC 3. 800 1.700 2.000 2.000 I (MG) 2.000 3.000 2.000 12.000 11.000 ZN (MG) 1.50C 1.000 3.000 2 .200 2.100 NA (MG) 4 .CCC 15.000 23.000 607.OOC 5.000 K (MG) 773.OOC 464.000 256.000 670.000 674.000 CU (MG ) 0.68C 0. 760 0.390 0. 570 0.690 255 ITEM CLUSTER 81 82 83 84 85 ATTRIEUTE GROUP 19 20 20 2C 20 NUTRIENTS: KCAL 651.OCC 93.000 274.000 94.000 141.000 PROT (GM ) 14.8CC 2. 600 4.300 2. 100 2.100 TRY (MG) 170.OCC 30.000 40.000 20.000 40.OOC THR (MG) 580.000 ICO.CCO 170.000 90.000 100.000 ISO (MG) 760.000 120.000 180.OCO 90.000 100.000 LEU (MG) 1220.OCC 130.000 210.000 110.000 120.000 LYS (MG) 490.000 140.000 220.000 110.000 100.000 MET (MG) 3C0.0CC 30.000 50.000 30.000 40.000 CYS (MG) 320.OCC 30.CCC 40.000 20.000 30.000 PHE (MG) 760.OOC 120.000 190.000 90.000 120.000 TYR (MG) 580.COC 50.OCO 80.000 40.000 100.000 VAL (MG) 950 .000 140.000 200.000 110.000 160.000 HIS (MG) 4C0.00C 40.000 60.000 30.000 40.000 FAT-T (GM) 64.000 0. 100 13.200 4.300 0.500 SFA (GM) 4 . 0 C C 0.0 3.000 2.000 0.0 PUFA (GM) 50.CCC 0.0 10.000 1. 000 0.0 CHOLE ( GM) 0.0 0.0 0 .020 0.015 0.0 CHO-T (GM ) 15.8CC 21.100 36.CCO 12.300 32.500 SUCR (GN ) 3.OCC 0.100 0.200 0. 100 7.20C CHO-F (GM) 2. ICC 0. 6C0 1.000 0.400 0.900 THIA (MG) 0 .330 0.100 0. 130 0.C80 0.09C RIBO (MG) 0.13C 0.040 0.080 0.050 0.070 NIACIN (MG) 0 .900 1. 700 3. 100 1.000 0.700 VIT-B6 (MG) 730.OCC 200.000 180.000 90.000 170.000 FOLIC (UG ) 58.OCC 12.OCO 9.000 12.000 19.000 VIT-E12(UG) 0 .0 0.0 0 .0 O.C 0.0 VIT-C (MG) 2.OCC 20.000 21.000 9.000 22 .000 PANTO (UG) 900.000 400.000 500.000 200.000 700.000 BIOTIN (MG) 3 7.000 2. 000 1 .000 2.000 2.000 VIT-A (IU) 30 .000 0.0 0.0 170.000 8100.000 VIT-D (IU) 0.0 0.0 0.0 0.0 0.0 VIT-E (MG) 22.CCC 0. C 0.300 0.200 2.000 CA (MG ) 99.OOC 9.000 15.000 24.000 40.000 P (MG ) 380.OCC 65.COO 111.000 48.000 58.000 MG (MG) 1.34.000 22.000 17.000 14.000 12.000 FE (MG) 3. ICC 0.700 1.300 0.400 0.900 I (MG) 3.OOC 4.000 11.000 3.000 3.000 ZN (MG) 2.8CC 0.200 0.200 0.100 0.700 NA (MG) 2. OOC 4. 000 6.000 331.000 12.000 K ( MG) 450.OCC 503.000 853.000 250.COO 300.000 CU (MG) 0.90C 0. 150 0.270 0. 100 0.170 256 ITEM CLUSTER 86 ATTRIBUTE GROUP 20 NUTRIENTS : KCAL 114.OCC PROT (GM) l.CCC TRY (MG) 20.000 THR (MG) 50.OCC I SO (MG) 50 .000 LEU (MG) 60.0GC LYS (MG) 50.OOC MET (MG) 20.OOC CYS (MG) 20.OCC PHE (MG) 60 .OOC TYR (MG ) 50. OCC VAL (MG ) 80.000 HIS (MG) 20.OCC FAT-T (GM) 0 .200 SFA (GM) 0.0 PUF/S (GM) 0.0 CHOLE (GM) 0 .0 CHO-T (GM) 27.5CC SUCR (GM ) 14.9CC CHO-F (GM) 0.6 CC THIA (MG) 0.030 RIBO ( MG) 0.030 NIACIN (MG) 0.6 CC VIT-B6 ( MG) 70.OOC FOLIC (UG) 19.OCC VIT-B12(UG) 0.0 VIT-C (MG) 8.CCC PANTO (UG ) 400.OOC BIOTIN (MG) 2.000 VIT-A (IU) 5000.000 VIT-D (IU) 0.0 VIT-E (MG ) 0.2CC C/i (MG) 13.OCC P (MG ) 29.OCC MG (MG) 18.000 FE (MG) 0.7CC I (MG) 3.000 ZN (MG) 0.5CC NA (MG ) 48. CCC K (MG) 120.OCO CU (MG) 0.06C 87 88 89 90 20 21 21 21 93.000 167.000 25.000 24.000 2. 600 2. 100 1.600 1.400 30.000 14.000 20.000 20.000 100.000 1.000 60.000 50.000 120.000 1.000 70.000 70.000 130.000 I .000 90.000 80.000 140.000 74.000 80.000 80.000 30.000 12.000 20 .000 20.000 30.OCC 1.000 20.000 10.000 120.000 1.000 60.000 40.000 50.CCO I.000 30.000 30.000 140.000 1.000 80.000 70.000 40.000 1.000 30.000 30.000 0.100 16.400 0. 200 0.200 0.0 3.000 0.0 0.0 0.0 9. OOC 0. 0 0.0 0.0 0.0 O.C 0.0 21.100 6.300 5.400 5.200 0. 100 1.600 0.400 0.400 0. 600 1.600 1.000 1 .000 0. 100 0.110 0. 07C 0.030 0.040 0.200 0.090 0.050 1. 700 1.600 0. 500 0.300 200.000 420.000 60.000 40.000 12.COO 30.000 5.000 12.000 0.0 0.0 0.0 0.0 20.OCO 14.000 12 .000 4.000 400.000 1100.000 200.OOC 100.000 2.000 6.000 I .000 1 .000 0. 0 290.000 540.000 470.000 0.0 0.0 0.0 0.0 0. 0 1.500 0. 800 0.0 9.000 10.000 5O.C0C 45.OOC 65.CCO 42.000 37.000 25.000 22.000 37.000 22.000 14.000 0. 700 0.600 0.600 1.500 4.000 2.000 3.000 1.000 0.200 2.400 0.300 0. 300 4. OCO 4.000 4.000 236.000 503.000 604.000 151.000 95.000 0. 150 0.390 0. 090 0.090 257 ITEM CLUSTER 91 92 93 94 95 ATTRIBUTE GROUP 21 21 21 21 21 NUTRIENTS: KCAL 26.OOC 24.000 20.000 33.000 13.OOC PROT (GM! 3. ICC 1.3 CO 1.100 3.600 0.900 TRY (MG) 30 .000 10.000 10.000 50.000 10.000 THR (MG) 120.OCC 40.000 30.000 160.000 40.000 ISO (MG) 120 .000 50.000 40.000 150.000 40.000 LEU (MG) 150.OCC 50.000 40.000 290.000 70.000 LYS (MG) 140.CCC 60.000 50.000 140.000 50.000 MET (MG) 50.OCC 10.000 10.000 40.000 0.0 CYS (MG) 50.OCC 30.OCO 20.000 60.000 10.000 PHE (MG) 100.OOC 70.000 60.000 120.000 40.OCC TYR (MG) 110.OCC 30.OCO 20.000 140.000 30.000 VAL (MG) 160.000 40.000 30.000 210.000 50.000 HIS (MG) 180.OOC 20.000 20.000 80.000 20.000 FAT-T (GM) 0.30C 0. 200 0.200 0.700 0.100 SFA (GM) 0.0 0.0 0.0 0.0 0.0 PUFA (GM ) 0.0 0. 0 O.G 0. 0 0.0 CHCLE (GM) 0.0 0.0 0.0 0.0 0.0 CHO-T (GM ) 4.5CC 5.400 4.300 5. 100 2.900 SUCR (GM) 0.20C 0.300 0.200 0.200 0.200 CHO-F (GM) 1 .50C 0. 800 0.800 1 .000 0.500 THI A (MG) 0.090 0.050 0.040 0.110 0.060 RIBO (MG) 0.20C 0.050 0.040 0.200 0.060 NIACIN (MG) 0.8CC 0.300 0.300 1.200 0.300 VIT-B6 (MG) 170.OOC 160.000 130.000 200.000 60.000 FOLIC (UG) 22.CCC 55.OCO 11.000 24.000 200.000 VIT-B12(UG) 0 .0 0.0 0.0 0.0 0.0 VIT-C (MG) 90.OOC 47.000 33.000 76.000 6.000 PANTO (UG) 500 .000 200.000 200.000 500.000 200.000 BIOTIN (MG) l.OCC 2.000 1 .000 1.000 3.000 VIT-A (IU) 2500.000 130.000 130.000 78C0.000 330.000 VIT-D ( IU) 0.0 0.0 0.0 0.0 0.0 VIT-E (MG) 1.9CC 7. 8CC 7.600 5. 900 0.300 CA (MG) 88.OCC 49.000 44.000 188.OOC 20.000 P (MG) 62.OOC 29.OCO 20.000 52.000 22 .000 MG (MG) 21.000 15.000 12.000 42.OCO 11.000 FE (MG ) 0.8CC 0.400 0.300 0.800 0.500 I (MG ) 4.000 3. 000 2.000 3. 000 10.000 ZN (MG) 0.2CC 0.300 0.200 0.700 0.100 NA (MG ) 10.CCC 20.000 14.000 25.000 9.000 K (MG) 267.OOC 233.000 163 .000 262.000 175.000 CU (MG ) 0. ICC 0. 120 0.C4C 0.310 0.09 0 258 ITEM CLUSTER ATTRIBUTE GROUP NUTRIENTS : KCAL PROT (GM ) TRY (MG) THR (MG) ISO (MG) LEU ( MG) LYS (MG) MET (MG) CYS (MG) PHE (MG) TYR (MG ) VAL (MG) HIS (MG) FAT-T (GM) SFA (GM) PUFA (GM) CHCLE (GM) CHO-T (GM ) SUCR (GM) CHO-F (GM) THIA (MG) RIBO (MG) NIACIN (MG) VIT-E6 (MG) FOLIC (UG) VIT-B12(UG) VIT-C (MG) PANTO (UG) BIQTIN (MG) VIT-A (IU) VIT-D ( IU) VIT-E (MG ) CA (MG ) P (MG) KG (MG ) FE (MG) I (MG) ZN (MG) NA (MG ) K (MG) CU (MG) 96 97 21 21 23.OCC 88.000 2.2CC 4. 7C0 40.000 40.000 60.OCC 180.000 70.000 220.000 60.OOC 300.000 110.OCC 220.000 20.OOC 40.000 30.OOC 50.000 70.OCC 180.000 120.000 110.OCC 100.OOC 190.000 40. OCC 80.000 0.400 0.400 0.0 0.0 0.0 0. C 0.0 0.0 4.OOC 16.800 0 .300 6. 4C0 0.9CC 2. 300 0.080 0. 090 0. 14C 0. 060 0.6CC 0. 800 130 .OCC 50.000 8.CCC 15.OCO 0.0 0.0 48.OCC 8. 000 200 .000 200.000 1.000 2.000 8CO.CCC 690.000 0.0 0.0 1 .7 C C 0. 0 138.OOC 26.000 32.OOC 76.OCO 17.OOC 13.000 I. 800 1.900 4.000 2. 000 O.2C0 1.400 18.CCC 236.CCO 220.OOC 96.000 0.09C 0. 170 98 99 21 21 68 .000 18.000 5.100 1.000 40.000 10.000 80.000 40.000 240.000 40.000 320 .000 40.000 240.000 40.000 40.000 10.000 60.000 20.000 200.000 50.OCC 120.000 40.000 210.COO 30.000 80.000 10.000 0.300 0.20C 0.0 0.0 0.0 0. 0 0.0 0.0 11.800 3.800 4.500 0. IOC 1.900 1.400 0.27C 0. 060 0.090 0.070 1.700 0. 500 130 .000 210.000 15.000 2 .000 0.0 0.0 13.000 96.000 300.000 200.000 2 .000 I .000 600.000 420.000 0.0 0.0 0.300 0. 500 19.000 9.000 86.000 16.000 21.CCC 12.000 1.900 0.500 3.000 9. 000 0.900 0.100 115.000 9.000 135.000 149.000 0.210 0.070 100 21 24.OCO 2.700 40.000 120.000 130.000 210.000 170.000 50.000 50.000 130.000 90.000 150.000 60.000 0.600 0.0 0.0 0.0 3.600 0.300 0.900 0.020 0. 120 0.300 70.000 29.000 0.0 14.000 100.000 2.000 8000.000 0.0 0.100 118.OOC 26.000 43.000 2.600 14.000 0.500 236.000 250.000 0.100 259 ITEM CLUSTER 101 102 103 104 105 ATTRIBUTE GROUP 21 21 22 22 22 NUTRIENTS: KCAL 23.000 26.000 37.000 31.000 42.000 PROT (GM) 3.OCC 3. 2CO 1.000 0.900 1.100 TRY (MG) 50 .OOC 50.000 10.000 10.000 10.000 THR (MG) 130.OCC 140.000 20.000 30.000 40.000 ISO (MG) 140 .000 150.000 40.000 30.COO 40.000 LEU (MG) 230.000 250.000 30.000 50.000 60.000 LYS (MG) 190.CCC 200.CCC 50.000 40.000 50.000 MET (MG) 50.OCC 50.000 10.000 10.OCC 10.000 CYS (MG ) 60.OCC 60.000 10.000 20.000 30.000 PHE (MG) 140.OOC 150.000 20.000 30.OCC 40.000 TYR (MG) 90.000 100.000 30.000 20.000 20.000 VAL (MG ) 170.OOC 180.000 30.000 40.000 50.000 HIS ( MG) 60.OCC 70.000 20.000 10.000 20.000 FAT-T (GM) 0.300 0. 300 0.100 0.200 0.200 SFA (GM) 0.0 0.0 0.0 0 .0 0.0 PUFA (GM) 0.0 0. 0 0.0 0. 0 0.0 CHOLE (GM) o.c 0.0 . 0.0 O.C 0.0 CHO-T (GM ) 3.700 4.300 8.800 7.100 9.700 SUCR (GM ) 0.300 0.300 1 .300 1.200 1.700 CHO-F (GM) 0.800 0.600 0.800 I.000 1 .000 THIA (MG) 0 .07C 0.100 0.C10 0. 050 0.060 RIBO (MG) 0.150 0. 200 0.030 0.050 0.050 NIACIN (MG) 0 . 4 C C 0. 6CC 0. 100 0. 500 0.600 VIT-86 (MG) 190 .OCC 280.000 50.000 30.000 150.000 FOLIC (UG ) 29.OOC 75.OCC 20.000 3 .000 15.000 VIT-B12(UG) 0 .0 0.0 0.0 0.0 0.0 VIT-C (MG) 19.OOC 51.000 3.000 6.000 8.000 PANTO (UG) 100.000 300.000 100.000 300.000 300.000 BIOTIN (MG) 2. OCC 7.000 1 .000 2.000 3.000 VIT-A (IU) 7900.OCC 8100.OCO 20.000 10500.000 1000.000 VIT-C (IU) 0.0 0.0 0 .0 O.C 0.0 VIT-E (MG) 1.1CC 2.9C0 0.0 0.500 0. 500 CA (MG ) 113.OOC 93.000 19.000 33.COO 37.OOC P (MG) 44.OCC 51.000 18.000 31.000 36.000 MG (MG) 42.OOC 57.000 15.000 6.000 18.000 FE ( MG) 2. ICC 3. 100 0.700 0.600 0.700 I (MG) 3.CCC 9.000 5.000 2.000 2.000 ZN (MG) 0.50C 0.700 0.400 0.300 0.500 NA (MG) 52.OCC 71.OCO 236.000 33.000 47.000 K (MG) 333 .OCC 470.000 167.000 222.000 341.000 CU (MG) 0.8 00 0. 140 0.210 0.090 0.090 260 ITEM CLUSTER ATTRIBUTE GROUP NUTRIENTS : KCAL PROT (GM ) TRY (MG) THR (MG) ISO (MG) LEU (MG) LYS (MG ) MET (MG) CYS (MG ) PHE (MG ) TYR ( MG) VAL (MG) HIS (MG) FAT- T (GM) SFA (GM) PUFA (GM) CHCLE (GM ) CHO- T (GM) SUCP (GM) CHO- F (GM) THI /S (MG) RI BO (MG) NIACIN (MG) VI T-B6 (MG) FCLIC (UG ) VIT- 812(UG) VIT- C (MG) PANTO (UG) BIOTIN ( MG ) VIT- A (IU) VIT- D (IU) VIT- E (MG) CA (MG) P (MG) MG (MG) FE ( MG) I (MG ) ZN (MG) NA (MG) K (MG) CU (MG) 106 22 83.0GC 3.2CC 20.000 130.OCC 120.OCC 250.OCC 120.OCC 60.OOC 50.OCC 180 .000 110.OCC 200.000 80.OCC l.OOC 0 .0 0.0 0.0 18.8CC 0.300 0.7CC 0.11C - 0.100 1.300 290 .OCC 2. OCC 0.0 7.OOC 400.000 2.OCC 4C0.0CC 0.0 1.2CC 3 .000 89.OOC 31.000 0.600 4. OCC 0.3CC 0.0 165.OOC 0.090 107 22 82.000 2. 100 10.000 90.000 80. OCO 230.000 80.OCO 40.000 40.000 120.000 70.000 130.000 50.000 0. 600 0.0 0. 0 0.0 20.000 0.300 0.500 0. C30 0. 050 1. OCO 200.000 2.000 0.0 5.000 300.000 2.000 330.OCC 0.0 0. 100 3.000 56.000 20.000 0.600 2. OCC 0.300 236.000 97.000 0. 060 108 22 84.000 2.600 20.000 110.000 100.000 290 .000 ICO.COO 50.000 40.000 150.000 90.000 170.000 70.000 0.800 0.0 0.0 0.0 19.800 0.300 0.800 0.030 0.050 G.900 200.000 2.000 0.0 4.000 200.000 2 .000 350.OCO 0.0 C.100 5.000 49.000 21.000 0.500 2.000 0 .300 236.000 97.000 0.060 109 22 14.000 0.900 10.000 20.000 30.000 40.CCO 30.000 10.000 10.000 20.000 20.000 30.000 10.000 0.100 O.C 0.0 0.0 3.100 0. 700 0.600 0. 050 0.080 0. 800 60.000 2.000 0.0 10 .000 200.000 2 .000 390.000 0.0 2.400 25.OOC 25.000 15.000 0 .400 4. 000 0.400 1 .000 141.000 0.080 110 22 21.000 I .000 10.000 30.000 30.000 40.000 40.000 10.000 10.000 20.000 10.000 30.000 20.000 0.200 0.0 0.0 0.0 4.300 0.300 0.400 0. 050 0.C30 0.700 90.000 26.000 0.0 17.000 200.000 2.000 900.000 0.0 0.0 6.OOC 19.000 11.000 0.500 0.0 0. 100 130.000 217.OOC 0.130 261 ITEM CLUSTER ATTRIBUTE GROUP NUTRIENTS: KCAL PROT (GM) TRY (MG) THR (MG) ISO (MG) LEU (MG) LYS (MG) MET (MG) CYS (MG ) PHE (MG ) TYR (MG) VAL (MG ) HIS (MG) FAT-T (GM) SFA (GM) PUFA (GM) CHOLE (GM) CHO-T (GM) SUCR (GM) CHO-F (GM ) THI A (MG ) RIBC (MG) NIACIN (MG ) VIT-B6 (MG) FOLIC (UG) VIT-E12(UG) VIT-C (MG) PANTO (UG) BIOTIN (MG) VIT-A ( IU) VIT-D ( IU) VIT-E (MG) CA (MG) P (MG) MG (MG) FE (MG) I (MG) ZN (MG) NA (MG ) K (MG) CU (MG) 111 22 22.OOC l.iOO 10 .000 40.CCC 30.CCG 50.OCO 50.OCC 10.OOC 10.CCC 30 .000 20.OCC 30 .000 20.OOC 0 .200 0.0 O.C 0.0 4.70C 0.3CC 0.5CC 0.060 0.040 0.700 ICO.OCC 18.OOC 0.0 23.OCC 300.OOC 4. OOC 900.000 O.C 0.3CC 13.OCC 27.OOC 13.000 0.5CC 2.000 0.20C 3. OCC 244.OCC 0.11C 112 22 23.000 0. 800 10.000 20.000 10.CCC 40.000 40.000 10.000 10.000 10.000 20.OCO 30.000 10.000 0. 200 0.0 0.0 0.0 4. 9C0 0.700 0. 9CC 0.040 0.050 0. 300 70.000 1. OCC 0.0 22.OCO 100.000 I.000 0.0 0.0 0. 0 35.000 24.OCC 14.000 0. 400 3. 000 0. 100 34.CCC 188.000 0. C40 113 23 14.000 0.600 0.0 20.000 20.000 30.000 30.000 10.000 10.000 10.000 20.000 20.000 0.0 C. 100 0.0 0.0 0.0 3.200 0.100 0.30C 0.C30 0.040 0.200 40.000 14.000 0 .0 11.000 300.000 3.000 0.0 0.0 0. 100 17.OOC 18.00 0 10.000 0.300 2.000 0.100 6.000 160 .000 0.050 114 23 17.000 1.900 10.COO 40.000 420.000 220.000 60.000 130.000 20.000 20.000 30.000 300.000 20.000 0.100 0.0 0. 0 0 .0 2.400 O.C 0.600 0.020 0.250 2. 000 60.000 8. 000 O.C 2.000 1000.000 7.000 0. 0 40 .000 0.0 6.000 68.000 8. OCC 0.500 O.C 0.400 400.000 197.OOC 0.260 115 23 29.OOC 1.200 20.000 20.000 20.000 30.000 60.000 10.000 20.000 40.000 40.000 30.COO 10.000 0. 100 0.0 0.0 0.0 6. 500 2.200 0.600 0.030 0.030 0.200 100.000 25.000 0.0 7.000 100.000 2.000 40.000 0.0 0.200 24.000 29.000 8.000 0.400 3.000 0.600 7.000 110.000 0.080 .262 ITEM CLUSTER lie 117 118 119 120 ATTRIBUTE GFOUP 24 25 25 25 26 NUTRIENTS: KCAL 64.OCC 116.000 146.000 11.OCC 384.CCC FPGT (GM) 3.2CC 1.4CC 0.700 0.700 0.0 TRY (MG ) 20.000 10.000 10.CCC 10.000 0. 0 THR (MG) 80.OOC 40.000 20.000 20.000 0.0 ISO (MG) 140 .000 40.000 20.000 20.000 0.0 LEU ( MG) ISO.CCC 60.000 30.000 30.000 0.0 LYS (MG) 150.OCC 60.CCC 30.000 30.000 0.0 MET (MG) 30.OOC 10.000 10.000 10.000 0.0 CYS (MG ) 50.OOC 20.OCO 10.000 10.000 0.0 PHE (MG) 100 .000 30.000 20.000 20.OOC 0.0 TYR (MG) 8 0.OCC 60.000 30.000 30.000 0.0 V AL (MG) 150.OOC 50.000 20.000 20.000 0.0 HIS (MG) 60.OOC 0. 0 0.0 0.0 0.0 FAT-T (GM) 0.30C 12.7CO 0.400 0. 200 100.000 SF A (GM) 0.0 2.000 0 .0 O.C 23.000 PUFA (GM) 0.0 10.CCC 0.0 0.0 72.000 CHCLE (GM) 0.0 0.0 0.0 O.C 0.0 CHO-T (GM ) 13.4CC 1.300 36.500 2.200 0.0 SUCF (GM ) 1.600 0. 0 33.4CC 0.0 0.0 CHO-F (GM) 1.2 CO 1.300 0.500 0.500 0.0 THIA (MG) C.12C 0. 0 0.0 0. 0 0.0 RIBO (MG) 0.07C 0.0 0.02C 0.020 0.0 NIACIN (MG) 1 .ICC 0. 0 0.0 0. 0 0.0 VIT-B6 (MG) 100.OCC 20.000 10.000 10.000 0.0 FOLIC (UG) 15.000 13.OCO 3.000 3.000 0.0 VIT-E12(UG) 0 .0 0.0 0.0 O.C 0. 0 VIT-C (MG) 8.000 2.000 6.000 16.000 0.0 PANTO (UG) 300 .000 0.0 200.000 200.000 0.0 BIOTIN (MG) 2. OCC 1.000 1 .000 1.000 0.0 VIT-A (IU) 4950.OCC 3 00. OCO 90.000 100.000 0.0 VIT-D (IU) 0.0 0.0 0.0 0.0 O.C VIT-E (MG) 0.5CC 0. 0 O.C 0.0 2 .300 CA (MG) 25.OOC 61.000 12.000 26.000 0.0 P (MG) 63. OCC 17.000 16.000 21.000 0.0 MG (MG) 25.000 12.000 1.000 1.000 1.000 FE (MG) 1.3CC 1.600 1 .200 I .000 0.0 I (MG) l.OCC 17.OCC 17.000 17.000 24.000 2N (MG) 0 .6CC 0.300 0.500 O.5C0 0.800 NA (MG ) 53.OCC 2400.OCC 823.000 1420.000 0.0 K (MG) 191.000 55.000 200.000 200.000 0.0 CU (MG) 0.12C 0. 370 0.210 0.020 0.030 263 ITEM CLUSTER 121 122 123 124 125 ATTRIBUTE GROUP 26 26 26 26 27 NUTRIENTS: KCAL 902.000 884.000 884.000 884.000 716.000 PROT (GM) 0.0 0.0 0.0 0.0 0.600 TRY (MG) 0.0 0.0 0.0 0. c 0.0 THR ( MG) O.C 0.0 0.0 0.0 0.0 ISO (MG) O.C 0. 0 0.0 o.c 0.0 LEU (MG > 0.0 0.0 0.0 o.c 0.0 LYS (MG) O.C 0. 0 0.0 0.0 0.0 MET (MG) 0 .0 0.0 0.0 o.c 0.0 CYS (MG) 0.0 0.0 O.C 0.0 0.0 PHE (MG) 0.0 0.0 0.0 0. c 0.0 TYR (MG) 0.0 0. 0 0.0 0 .0 0.0 VAL (MG) 0.0 0.0 0.0 0. 0 0.0 HI S (MG) 0 .0 0.0 0 .0 o.c 0.0 FAT-T (GM) ICO.OCC 100.OCC 1 CO.000 100.000 81.000 SFA (GM) 38.OOC 25.000 11.000 15.000 46.000 PUFA (GM ) 56.OCC 71.OCO 83.000 72 .000 29.000 CHGLE (GM ) 0.090 0.0 o.c 0.0 0.270 CHO-T (G M) 0.0 0.0 0.0 0 .0 0.400 SUCR (GM) 0 .0 0.0 0.0 o.c 0. 0 CHO-F (GM) 0.0 0.0 0.0 0 .0 0.0 THIA (MG) 0 .0 0. C 0.0 0. 0 0.0 RIBO (MG) 0.0 0.0 0.0 o.c o.c NIACIN (MG) 0.0 0. 0 0.0 0.0 0 .0 VIT-B6 (MG ) 20 .OOC 0.0 0.0 0.0 0.0 FOLIC (UG) 0.0 0.0 0.0 0.0 0.0 VIT-B12(UG) 0 .0 0.0 0.0 0. 0 0. 100 VIT-C (MG) 0.0 0.0 0 .0 0 .0 0.0 PANTO (UG) 0.0 0. 0 0.0 0.0 0.0 BIOTIN (MG) 0.0 0.0 0.0 0.0 10.000 VIT-A (IU) 0.0 0. 0 0.0 0.0 3300.000 VIT-D (IU) 0 .0 0. 0 0.0 0.0 40.000 VIT-E (MG) 1.200 43.600 14.400 12.100 1 .900 CA (MG) 0.0 0. 0 0.0 o.c 20.000 P ( MG) 0.0 0.0 0.0 0 .0 16.000 MG (MG) l.OCC 1.000 1.000 1. 000 2.000 FE (MG) 0.0 0.0 0.0 o.c 0.0 I (MG) 3. OCC 4. COO 7.000 4.000 9.000 ZN (MG) 0.500 0.500 0.500 0.500 0.30C NA (MG) O.C 0. 0 0.0 0.0 987.000 K (MG) 0.0 0.0 o.c 0.0 23.000 CU ( MG) 0.020 0.070 0.070 0.070 0.030 264 ITEM CLUSTER 126 ATTRIBUTE GROUP 27 NUTRIENTS: KCAL 720.000 PROT (GM) 0.6CC TRY (MG) O.C THR (MG) 0.0 ISO (MG) 0.0 LEU (MG) 0.0 LYS (MG ) O.C MET (MG) 0.0 CYS (MG) 0.0 PHE (MG) 0 .0 TYR ( MG) O.C VAL (MG) 0.0 HI S (MG) 0.0 FAT-T (GM ) 8 1 .000 SFA (GM) 19.000 PUFA (GM) 60.OCC CHOLE (GM) 0 .0 CHO-T (GM) 0.4CC SUCR (GM) O.C CHO-F (GM ) 0 .0 THIA (MG) 0.0 RIBC (MG) 0.0 NIACIN (MG) 0.0 VIT-B6 (MG) 0 .0 FOLIC (UG) 0.0 VIT-B12(UG) O.C VIT-C (MG) 0.0 PANTO (UG) 0.0 BIOTIN (MG) 0.0 VIT-A (I U) 33 CO.CCC VIT-D ( IU) 0 .0 VIT-E (MG) 12.5CC CA (MG) 20.000 P (MG) 16.OCC MG (MG) 2. OCC FE (MG) 0.0 I (MG) 7. OCC ZN (MG) 0 .300 NA (MG) 987.OCC K (MG) 23 .000 CU (MG) 0.03C 127 128 129 130 28 28 28 28 173.000 228.OOC 162.000 106.000 7.900 1 .700 3.900 2.000 110.OCO 20.000 90.000 20.000 290.000 50.000 230.000 70.000 510.CCC 80.000 3 50.OCO 60.000 760.000 130.000 580.OCO 80.000 570.000 40.000 260.000 80.000 200.000 20.000 110.000 10.000 50.000 30.000 120.000 20.000 410.000 90.000 190.000 50.000 390.000 60 .000 270.000 30.000 560.OOC 70.000 350.000 60.000 260.000 30.000 150.000 30.000 13.OCO . 19.500 12.500 0.400 7.000 9.000 7.OCC 0. 0 5.000 9.000 4.000 0.0 0.070 0.01C 0. C40 0.0 6.400 11 .100 8.800 25.400 0. 0 O.G 0. 0 16.600 0.0 0.0 O.C 0.500 0. 030 0.060 0.040 0.090 0.210 0.040 0. 170 0.070 0. 0 0.0 0.200 1.600 40.000 50.000 50.000 110.000 9.000 1 .000 I .000 27.000 0.4CC 0. 160 0.160 0.0 1.000 0.0 0.0 15.000 200.000 200.000 600.000 200.000 4.000 0.0 4.000 4. 000 550.000 0.0 460.000 1400.000 1.000 0.0 O.C 0.0 0.500 0.200 0.100 0.200 234.000 0.0 115.OOC 22.000 172.000 11 .000 93.000 50.000 17. OCO 2.000 14.000 21.000 0.300 0.600 0.200 0.80C 8. OCO 1.000 7.000 6.000 0.900 0.500 0.400 1.100 518.000 1000.000 379.000 1042 .000 106.000 106.OOC 139.000 363.000 0. 070 0.010 0.040 0.510 265 ITEM CLUSTER 131 132 133 134 135 ATTRIBUTE GROUP 28 29 29 29 29 NUT FIENTS: KCAL 21.000 718.000 552.000 435.000 368.000 PROT (GM) I.CCC 1. 100 0.200 1 .000 21.500 TRY (MG) 10.OCC 0.0 0.0 0. 0 290.000 THR (MG) 30 .OCC 0.0 0 .0 O.C 810.COO ISO (MG) 30.CCC 0. 0 0.0 0.0 1480.000 LEU (MG) 40.OOC 0.0 0.0 0.0 2140.000 LYS (MG) 40.OCC 0. 0 0.0 0.0 1550.000 MET (MG) 10 .000 0. 0 0.0 0.0 580.000 CYS ( MG) 10.OOC 0.0 0.0 0.0 120.000 PHE (MG ) 20 .OOC 0.0 0.0 O.C 1200.000 TYR (MG) 10.OCC 0.0 0.0 0 .0 1030.000 VAL (MG) 30.OCC 0. 0 0.0 0. 0 1580.000 HIS (MG) 20 .OCC 0.0 0.0 o.c 700.000 FAT-T (GM ) 0.2CC 79.9C0 60.000 42.300 30.500 SFA (GM) 0 .0 14.000 10.000 8.C0C 17.000 PUFA (GM) 0.0 57.000 44.000 30.000 11.000 CHOLE (GM) 0 .0 0.050 O.C 0.050 0.150 CHO-T (GM) 4.300 2.200 6.900 14.400 2.000 SUCR (GM ) 0.3CC 0.0 0.0 10.000 0. 0 CHO-F (GM) 0 .4CC 0.0 0.0 0.100 0.0 THIA (MG) 0.G5C 0. 020 0.0 0.010 0.030 RI BC (MG) 0.030 0.040 0.0 0.030 0.61C NIACIN (MG) 0.7 00 0.0 0.0 0.0 1 .200 VIT-B6 (MG) 90.000 0. 0 0.0 0.0 170.000 FOLIC (UG) 26.OCC 0.0 0.0 0 .0 11.000 VIT-B12(UG) 0.0 0.0 0.0 0. 0 1.400 VIT-C ( MG) 17.OCC 3.000 0.0 3.000 0.0 PANTO (UG) 2 CO.OCC 100.000 0.0 100.000 1800.000 BICTIN (MG) 2.000 3.000 0.0 ' 3.000 3.000 VIT-A (I U) 900.OOC 280.000 0.0 220.000 1240.000 VIT-D (IU) 0 .0 8.000 0.0 8.000 30.000 VIT-E (MG) 0.0 11.900 9.100 5.300 0.800 CA (MG ) 6. OCC 18.000 10.CCO 14.000 315.000 P (MG ) 19.OOC 28.000 4.000 26 .000 339.000 MG (MG) 11.OCC 2. OOC 7.000 2.000 20.000 FE (MG) 0.5CC 0.500 0.200 0.20C 0.500 I (MG) 0.0 27.COG 2.000 27.000 11 .000 ZN (MG ) 0.10G 0. 500 0.40C 0.500 2.200 NA (MG) 130.OCC 597.000 2092.000 586.000 666.000 K (MG) 217.000 34.000 15.000 9.000 78.000 CU ( MG) 0.13C 0.240 0.040 0.240 0. 160 266 ITEM CLUSTER ATTRIBUTE GROUP NUTRIENTS: KCAL PRQT (GM) TRY (MG 1 THR (MG) ISO (MG) LEU (MG) LYS ( MG) MET (MG ) CYS (MG) PHE (MG ) TYR (MG) VAL (MG) HIS (MG) FAT-T (GM) SFA (GM) PUFA (GM) CHOLE (GM) CHC-T (GM) SUCR (GM ) CHO-F (GM ) THIA (MG) RIBG (MG) NIACIN (MG) VIT-B6 (MG) FOLIC (UG) VIT-B12(UG) VIT-C (MG) PANTO (UG) BIOTIN (MG) VIT-A (IU) VIT-D (IU) VIT-E (MG) CA (MG ) P (MG) MG (MG) FE (MG ) I (MG) ZN (MG) NA ( MG) K (MG) CU (MG) 136 13 7 3C 30 65.OCC 36.000 3.500 3.600 5 0.OCC 50.CCC 160.000 160.000 230.OCC 190.000 3 50 .000 360.000 280.OOC 280.000 EO.OCC 90.000 30.OOC 30.000 170. OCC 170.OCC 180.OCC 180.000 240.000 250.OCO 90.000 100.000 3.5CC 0.100 2.OOC 0. 0 I.OCC 0.0 0.C14 C. CC2 4.90C 5.100 0.0 0. 0 0 .0 0.0 0.03C 0.040 0 .170 0. 180 0.1CC 0.100 40. CCC 40.000 9. OOC 9.000 0.4CC 0.4CC l.OOC 1.000 3C0.CGC 400.OCO 4.000 2.000 140.OCC 0.0 41.000 41.000 0.1CC 0.0 118.OOC 121.OCO 93 .OCC 95.000 13.000 15.000 0.0 0.0 7. OCC 7.000 0 .400 0.400 50.CCC 52.000 144.OCC 145.OCO 0.15C 0.020 128 139 30 31 59.000 138.000 4.200 3.300 60.000 45.000 190.000 150.000 270.000 210.000 420.000 325.000 330.000 255 .000 100.000 75.000 40.000 30.000 2C0.000 160.000 200.000 165.000 290.000 225.000 110.000 85.000 2.000 12.100 1.000 7. 000 1 .000 1 .000 0.002 0. C08 6.000 4.600 0.0 0.0 0.0 O.C 0.040 0.030 0.210 0.160 0.100 0.100 40.000 35.000 9.000 15.000 0.400 0.330 1.000 1.000 400.000 300.000 3.000 4.000 80.000 495.000 41.000 28.000 0.100 0 .400 143.00 0 110.000 112.000 87.000 17.000 12.000 0.10C 0.0 8.000 7.000 0.400 0.400 61 .000 47.000 175.000 133.000 0.020 0.160 140 31 211.000 3.000 40.000 140.000 190.000 300.000 230.000 70.000 30.000 150.000 150.000 210.000 80.000 20.600 11.000 8.000 0.070 4.300 0.0 0.0 0.03 0 0. 150 0.100 30.000 20.000 0.250 l.OCC 300.000 4.OCC 840.000 15.000 0.700 102.000 80.000 10.000 0.0 6.000 0.300 43.000 122.000 0.170 267 ITEM CLUSTER 141 142 143 144 145 ATTRIBUTE GROUP 31 31 32 32 32 NUTRIENTS: KCAL 252.OCC 137.OCO 47.00 0 41.000 23.000 PR 07 (GM) 2.20C 7.000 0.100 0.50C 0.400 TRY (MG) 20.OCC 100.000 0.0 1 .000 2 .000 THR (MG ) 100.000 320.000 3.000 1.000 1.000 ISO (MG) 140.OOC 450.000 5 .000 1 .000 I.000 LEU (MG) 22 0. CCC 690.000 5.OCC 1. OCO 1. 000 LYS (MG) 170.000 550.000 4.000 6.000 8.000 MET (MG) 50.OCC 170.OCO 2.000 0. 0 1.000 CYS (MG) 20.OOC 60.000 1.000 l.OCC 1.000 PHE (MG ) 110.000 340.COO 3. OOC 11.000 11 .000 TYR (MG ) 110.OOC 360.000 2. COO 6.000 6.000 VAL (MG) 150.OCC 480.000 3.000 I.000 1.000 HIS (MG ) 60.000 190.000 2.000 1. 000 1.000 FAT-T (GM) 37.6CC 7.900 0.0 0 .100 0. 100 SFA (GM) 21.OCC 7. OCC 0.0 0.0 0. 0 PUFA (GM) 13.OCC 3.000 0.0 0.0 0.0 CHOLE (GM ) 0.12C 0. 110 O.C 0.0 0.0 CHO-T (GM) 3 .100 9.700 11.900 9. 800 7.600 SUCR (GM) 2. OOC 0.0 5.500 2.700 0.100 CHO-F (GM) 0 .0 0. 0 0. 100 0. 0 0.0 THIA (MG) 0.02C 0. 040 0.010 0 .030 0.030 RIBC (MG ) 0.11C C.34C 0.020 0. 020 0.010 NIACIN (MG) 0.0 0.200 0.100 0.20C 0. 100 VIT-B6 (MG) 20.CCC 50.OCO 30.000 10.000 50.000 FOLIC (UG) 15.000 1.000 0.0 1.000 2.000 VIT-B12CUG) O.IOC 0. 160 0.0 0.0 0.0 VIT-C (MG) 0.0 1.000 l.OCC 34.000 42.000 PANTO (UG) 2C0.00C 600.000 100.000 100.000 100.000 BIOTIN (MG) 3 .OCC 8.000 O.C 1.000 0.0 VIT-A ( IU) 1540.OCC 320.000 40.000 10.000 20.000 VIT-D (IU ) 11.OCC 79.OCC 0.0 0. 0 0.0 VI T-E (MG) 4.9CC 0.200 0.0 0.0 0. 0 CA (MG ) 75. CCC 252.000 6.000 8.000 7.000 P (MG) 59 .000 205.000 9.000 14.OOC 10.000 MG (MG) 7.OCO 33.000 4.000 7.000 9.000 FE (MG) O.C 0. 100 0.600 0.400 0.200 I ( MG) 4.OOC 16.000 2 .000 1 .000 5.000 ZN (MG) 0.2CC C. 700 0. 100 0.0 0.200 NA (MG) 32.000 118.000 1.000 1.000 1.000 K (MG) £9.OCC 303.000 101.000 162.000 141.000 CU (MG ) 0 .120 0.090 0.020 0.010 0.080 268 ITEM CLUSTER ATTRIBUTE GROUP NUTRIENTS: KCAL PROT (GM) TRY ( MG) THR (MG ) I SO (MG) LEU (MG ) LYS (MG) MET (MG) CYS (MG) PHE (MG) TYR (MG) VAL ( MG) HIS (MG) FAT- T (GM) SFA (GM) PUF/ (GM) CHOLE (GM ) CHO- T (GM ) SUCR (GM) CHO- F (GM) THIA ( MG) RIBG (MG) NIACIN (MG) VIT- B6 (MG) FOLIC (UG ) VIT- B12(UG) VIT- C (MG) PANTO (UG) BIOTIN (MG ) VI T- (IU) VIT- D (I U) VIT- E (MG ) CA (MG) P (MG) MG (MG) FE (MG ) I (MG) ZN (MG) NA (MG) K (MG) CU (MG ) 146 147 32 32 44. OCC 48.000 O.lOO 0. 800 0.0 3.000 0.0 1.000 0.0 1.000 0.0 1. OCC 0 .0 21.000 0.0 2.000 0.0 1.000 0.0 9. 000 0.0 15.000 0.0 1.000 0.0 1. 000 0 .0 0. 200 0.0 0. 0 0 .0 0.0 0.0 0.0 11 .400 11.200 4.9CC 3.200 0 .0 0. 100 0.0 0.070 0.01C 0. 02 0 0.100 0.300 10.OCC 40.OCC 2.OOC 4. 000 0.0 0.0 7 .000 40.000 O.C 200.000 0.0 1. 000 0 .0 200.000 0.0 0. 0 0.0 0.0 1.000 10.000 1.000 18.000 l.OCC 11.000 O.C C.4CC 7.OOC 1.000 0.10C 0.2C0 0 .0 1. 000 16. OCC 199.000 0 .010 0.050 148 149 32 32 45.000 49.000 0.700 1.000 3 .000 3 .000 1.000 1.000 I.000 1.000 1.000 1. 000 21.000 24.000 2.000 3.000 1.000 1.000 9.000 12.000 15.000 21.000 1 .000 1 .000 1.000 1. 000 0.100 0.200 0.0 0. 0 0.0 O.C O.C 0.0 1C.7CC 12.200 3.200 4.200 O.C 0. 500 0 .090 0.100 0.010 0.040 0.300 0.400 30.000 60.000 4.000 45.OOC 0.0 0.0 45.000 50.000 200.000 300.000 0.0 2. 000 200.000 200.000 0.0 0.0 O.C 0.200 9.000 41.000 16.000 20.000 12 .000 11.000 0.100 0.400 1.000 O.C 0. 100 0.100 1.000 1.000 186.000 200.000 0.050 0.090 150 33 21.000 1.000 10.000 30.000 30.000 40.000 40.000 10.000 10.000 20.000 10.000 30.000 20.000 0.200 0.0 O.C 0.0 4.300 0.300 0.400 0.050 0. 030 0.70C 90.000 26.000 0.0 17.000 200.000 2.000 900.000 0.0 0.0 6.000 19.000 11.000 0.500 0.0 0.100 130.000 217.000 0.130 269 ITEM CLUSTER ATTRIBUTE GROUP NUTRIENTS: KCAL PROT (GM) TRY (MG) THR (MG) I SO (MG) LEU (MG ) LYS (MG) MET (MG) CYS (MG) PHE (MG) TYR (MG) VAL (MG) HIS (MG) FAT-T (GM) SFA (GM) PUFA (GM) CHOLE (GM) CHO-T (GM) SLCR (GM CHO-F (GM ) THIA (MG ) RIBO (MG) NIACIN (MG) VIT-B6 (MG) FOLIC (UG) VIT-B12(UG ) VIT-C (MG) PANTO (UG) BIOTIN (MG ) VIT-A (IU) VIT-D (IU) VIT-E (MG) CA ( MG ) P (MG) MG (MG) FE (MG) I (MG) ZN (MG) NA (MG) K (MG) CU (MG) 151 152 34 34 39.OCC 31.000 0.0 0.0 0.0 0.0 O.C O.C O.C 0.0 0.0 0. 0 0.0 0.0 0.0 0.0 0.0 0.0 o.c 0.0 0.0 0. 0 0.0 0.0 0.0 0. 0 0 .0 0.0 0.0 0.0 0 .0 0.0 0.0 0.0 10.OCC 8. 000 10.OCC 8.000 0.0 0. 0 0.0 0.0 0.0 0. 0 0 .0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 o.c 0. 0 0 .0 0.0 CO 0. 0 0.0 0.0 0.0 0. 0 0.0 0.0 8.000 8.000 15.CCC 15.OCC 1.000 1.000 0.400 0.4CC 1 .000 1.000 0.1CC 0.100 6 .000 6.000 o.c 0.0 0.04C 0.03C 153 154 35 35 I .000 2.000 0.0 O.C 0.0 0 .0 0.0 0. 0 0.0 0.0 o.c 0.0 0.0 0.0 0.0 0.0 O.C 0. 0 0.0 0 .0 0.0 o.c 0.0 o.c 0.0 0.0 0.0 o.c 0.0 0.0 0.0 o.c 0.0 0.0 O.C 0.400 0.0 o.c 0.0 0.0 0.0 o.c 0.0 0.010 0.300 0.0 10 .000 0 .0 0.0 0. 0 0.0 0.0 0.0 0. 0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0. 0 2 .000 0.0 4.000 0.0 5.000 2.000 0.100 o.c 4.000 16.000 0.0 0.0 1.000 o.c 36 .000 25.000 0.02C 0. 010 155 36 42.000 0.300 0.0 0.0 0. C 0.0 O.C 0.0 0.0 0.0 0.0 O.C 0.0 0.0 0.0 0.0 0.0 3. 800 0.0 0.0 0.0 0.030 0.600 60.000 0.0 0.0 0.0 100.000 0.0 0.0 0.0 0. 0 5.000 30.000 10.000 0.0 1.000 0.100 7.000 25.000 0.070 270 ITEM CLUSTER ATTRIBUTE GROUP NUTRIENTS: KCAL PROT (GM) TRY (MG) THR (MG) ISO (MG) LEU (MG ) LYS (MG) MET (MG) CYS (MG ) PHE (MG) TYR (MG) V AL (MG) HJS (MG) FAT-T (GM) SFA (GM) PUFA (GM) CHOLE (GM) CHO-T (GM ) SUCR (GM) CHO-F (GM ) THIA (MG) RIBC (MG) NIACIN (MG) VIT-B6 (MG) FOLIC (UG) VIT-E12(UG) VIT-C (MG) PANTO (UG) BIOTIN (MG) VIT-A (IU) VIT-D ( IU) VIT-E (MG) CA (MG) P (MG) MG (MG) FE (MG) I (MG) ZN (MG) NA (MG) K (MG) CU (MG) 156 157 36 36 249.000 137.000 0.0 0. ICC 0.0 0.0 0.0 0. 0 0 .0 0.0 0.0 0. 0 0.0 0.0 0.0 0.0 O.C 0. 0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 .0 0.0 0.0 0.0 0 .0 0.0 0.0 0.0 0.0 7. 7CC 0.0 7.700 0.0 0. 0 0 .0 0. 010 0.0 0.020 0.0 0.200 O.C 40.000 0.0 0. 0 0 .0 0.0 0.0 0. 0 0.0 0.0 O.C 0. 0 0 .0 0.0 0.0 0.0 0.0 0. 0 8.000 8.000 10.OOC 10.OCC 0 .0 5.000 0.400 0.400 1.000 1.000 0.1CC 0.100 1.000 4.000 2.OCC 75.000 0.C8C 0. 08C 158 159 26 37 85 .000 26.000 0. 100 1. 400 0.0 10.OCC 0.0 70.000 0.0 80.000 0.0 80.000 0.0 180.000 0.0 40 .000 0.0 60.000 0.0 60.000 0.0 30.000 0.0 60.OOC 0.0 20.000 0.0 0. SCO 0.0 0 .0 0.0 0. 0 0.0 0 .003 4.200 3.300 4.200 O.C 0.0 0.100 0.0 0.010 0.010 0.010 0. 100 0.300 40.000 30.000 0.0 0. 0 0 .0 O.C 0.0 0.0 0.0 100.000 0.0 1 .000 0.0 20.000 0.0 0.0 0.0 0. c 9.000 4.000 10.000 15.000 8.000 4.000 0.400 0.200 1.000 1.000 0.100 0.300 5.000 408.000 92.000 23 .000 0.110 0. 130 160 37 73.000 3.000 40.000 130.000 290.000 230.000 260.000 90.000 120.000 120.000 50.000 170.OCC 40.000 4.200 I.000 2.000 0.009 5.900 0.0 0.100 0.20C 0.110 0.300 30.000 4.000 0. 200 1.000 200.000 1.000 250.000 I.000 0.0 70.000 62.000 9.000 0.200 4.000 0.400 430.000 106.000 0.050 271 ITEM CLUSTER ATTRIBUTE GROUP NUTRIENTS: KCAL PROT (GM) TRY (MG) THR (MG) ISO (MG) LEU ( MG) LYS (MG) MET (MG) CYS (MG ) PHE (MG) TYR (MG) VAL (MG) HIS (MG) FAT-T (GM) SFA (GM) PUFA (GM) CHOLE (GM) CHO-T (GM ) SUCR (GM) CHO-F (GM) THIA (MG) RIBO (MG) NIACIN (MG) VIT-B6 (MG) FOLIC (UG) VIT-B12(UG) VIT-C (MG) PANTO (UG) BIOTIN (MG) VIT-A (IU) VIT-D (IU) VIT-E (MG) CA (MG) P (MG) MG (MG) FE (MG) I (MG) ZN (MG) NA .(MG) K (MG) CU (MG) 161 162 37 37 59.OCC 36.000 3.5CC 0.8C0 30.OCC 10.000 130.OCC 30.OOC 190.000 30.000 250.OCO 30.000 210.000 80.000 30.OCC 30.000 3 0. CCC 10.CCC 8C0 .OCO 20.000 80. OCC 10.OCC 190.000 30.000 80.000 20.000 1.300 1.000 0.0 - 0. 0 O.C 0. 0 0.002 0.002 8.40C 6.4C0 2 .6CC 0.400 0.2CC 0.200 0 .100 0. 020 0.060 0.020 0.6 CC 0.500 50.OCC 20.000 l.OCC 4. OCC 0.16C 0.0 0.0 5. CCO 100 .000 100.000 l.OCC 3.000 180 .000 410.000 l.OOC I.000 0 .0 0. C 12 .OCC 6.000 61.000 14.OCC 6.000 7.000 0.6CC 0. 300 1.000 3.000 0.4CC 0. 300 384.OCC 396.OCO 110.OOC 94.000 0.09C 0. 160 163 164 37 37 32 .000 3 .000 2. 100 0.400 20.000 5.000 90.000 18.000 70.000 21.000 170.000 33.000 230.000 35.OOC 40.000 10.000 30.000 5.000 90.000 7.000 30.000 14.000 120.000 22.000 30.000 14.000 C.900 0.100 0.0 0.0 0.0 0. 0 0.003 0.0 3.900 0.100 0. 100 0.0 0.200 0.0 0.020 0. 0 0.020 0.010 0.400 0. 500 30.000 0.0 4.000 0.0 0.0 0.0 2.000 0.0 100.000 0.0 1.000 0.0 100.000 0. c 0.0 0 .0 0. 100 0. 0 5.000 o.c 20.000 6.000 11.CCC 1.000 0.300 0. 100 2.000 1.000 0 .300 0.0 427.000 48.000 66 .000 2.OOC 0.C90 0.0 165 38 259.000 5. 700 90.000 210.000 330.000 550.000 170.000 90.000 100.000 310.COC 220.000 310.000 200.000 0.200 0.0 0.0 0.0 59.400 35.OOC 0.0 0.0 0.110 0.100 10.000 2. 000 0.040 0.0 200.OCC 4.000 0.0 0.0 0.0 95.000 119.000 15.000 0.300 3.000 0. 200 146.000 60.000 0.040 272-ITEM CLUSTER 166 167 168 169 170 ATTRIBUTE GROUP 38 38 38 38 38 NUTRIENTS : KCAL 322.OOC 380.000 379.000 411.000 337.OOC PROT (GM) 6.30C 4. 3 CO 4.800 6.400 4.100 TRY (MG) 130.000 80.000 90.000 130.000 80.000 THR (MG) 360.OOC 210.000 230.000 360.000 210.000 ISO (MG) 510.000 340.000 340.000 520.OCO 300.000 LEU (MG) 760.OCC 530.000 530.000 770.000 470.000 LYS (MG) 1150.OCC 270.OCC 270.000 1160.000 240.000 MET (MG) 200.OCC 120.000 120.000 210.OOC 110.000 CYS (MG ) 180.OCC 120.OCC 120.000 180.000 100.000 PHE (MG) 340.000 240.000 260.000 350.OCC 230.000 TYR (MG) 350.OCC 250.000 250.000 350.000 220.000 VAL (MG) 540.000 350.000 350.OCC 550.COO 310.000 HIS ( MG) 2 CO.OCC 330.000 330 .000 200.000 290.000 FAT-T (GM) 9.6CC 17.600 15.300 18.7CC 11.300 SFA (GM) 3.OCC 9.000 6.000 5 .000 5.000 PUFA (GM) 6.OOC 8. OCC 9.000 12.000 6.000 CHCLE (GM) 0.080 0.070 0. 100 0. 160 0. C9C CHO-T (GM) 52.400 55.60C 59.700 54.700 57.600 SUCR (GM) 19.100 43.000 26.700 28.100 36.200 CHO-F (GM) 0. ICC 0. 300 0.600 0. 100 0.200 THIA (MG) 0.180 0.020 0.130 0. 040 0.020 RIBO (MG) 0.16C 0.080 0.140 0.110 0.080 NIACIN (MG) 1.4C0 C.2CC 0.800 0.200 0.200 VIT-B6 (MG) 40.OOC 50.000 80.000 40.000 40.000 FOLIC (UG) 8 .OOC 22.OCC 3.000 8.000 8.000 VIT-E12(UG) 0.0 0.0 0. 130 0.0 0.0 VIT-C (MG) 0.0 0. 0 0.0 0.0 0.0 PANTO (UG) 200 .000 200.000 400.000 300.000 300.000 BIOTIN (MG) 5. OOC 6.000 8.000 3.000 5.000 VIT-A (IU) 160.CCC 430.OCC 120.000 290.000 140.000 VIT-0 (IU ) 6 .OCC 9.000 13.000 20.CCC 2. 000 VIT-E (MG) 0.20C C. ICC 0.700 1 .100 0.500 CA (MG) 61 .OOC 54.000 72.000 40.000 91.OOC P (MG) 174.OCC 52.000 113.000 104.000 182.000 MG (MG) 15.000 24.000 16.000 13.000 20.000 FE ( MG) 1.6 00 0.800 2 .600 0.800 0.600 I (MG ) 7. CCC 7.000 7.000 7.000 7.000 ZN (MG ) 0.6CC 0. 500 0.600 0.600 0. 500 NA (MG ) 421 .OCC 420.OCO 158.OCC 178.000 227.000 K (MG) 109.OCC 119.000 496.000 78.000 109.000 CU (MG ) 0.08C 0.310 0.100 0.060 0.100 273 ITEM CLUSTER ATTRIBUTE GROUP NUTRIENTS: KCAL PROT (GM) TRY (MG) THR ( MG) ISO (MG) LEU (MG) LYS (MG) MET (MG) CYS (MG ) PHE (MG) TYR ( MG) VAL (MG) HI S (MG) FAT-T (GM) SFA (GM) PUFA (GM) C HC LE (GM) CHO-T (GM) SUCR (GM) CHO-F (GM) THIA (MG) RIBC (MG) NIACIN (MG ) VI T-B6 (MG) FOLIC (UG) VIT-B12(UG) VIT-C (MG) PANTO (UG) BIOTIN (MG) VIT-A (IU) VIT-D ( IU) VIT-E (MG) CA (MG) P ( MG) MG (MG) FE (MG) I (MG) ZN (MG) NA (MG) K (MG ) CU (MG) 171 39 256.OOC 2 .20C 30.000 60.000 ICO.OOC 170 .OCC 50.CCC 30 .OOC 40.OOC 120 .000 80.000 ICO.000 40.OOC 11.100 3.OCC 8. OOC 0.030 38.ICC 10.900 0.4CC 0.02C 0.020 0.40C 40 .OOC 4.OCC 0.0 1.000 ICO.OCC 1 .OOC 20.OCC 0.0 0.900 8.000 22.000 4.CCC 0.30C 4.OCC 0 .40C 301.000 80.000 0.C8C 172 39 211.000 4.000 60.000 170.000 250. CCC 340.000 260.OOG 110.000 70.000 220.000 160.000 260.000 90.000 11.2 CO 3.000 8. 000 0. 100 24.500 15.100 0. 500 0. C3C 0.100 0. 500 40.000 4.000 0. 0 0. 0 5 CO.CCO 5.000 24 70.OCC 10.000 0. 900 51.000 69.000 6. 000 0. 500 3. OCC 0.400 214.OCO 160.000 0.050 173 40 480.COO 5.100 60.OOC 150.000 230.000 390.000 120.000 70.000 100.000 280.000 170.000 220.000 100.000 20. 200 4.000 15.000 0.100 71.000 37.100 0 .100 0.030 0.050 0.400 50.000 11.000 0.0 0.0 400.OCO 5.000 80.000 1.000 0.500 37.OOC 163.000 15.000 0.7C0 10.000 1.700 365.OOC 67.OOC 0.150 174 40 .358.000 3.900 50.OCO 110.000 190.000 300.CCC 90.000 50.OOC 80.000 210.000 130 .000 170.000 80 .000 5. 60C l.COO 3.000 0.060 75.400 25.700 1 .700 0. 040 0.07C 0.300 90.000 10.000 0. C 0 .0 300.000 5.CCC 110.000 1. COO 0.400 78.000 60 .000 23.000 1 .000 10.000 0.900 252.000 198.000 0.190 175 41 391.OOC 4.600 60.000 160.000 240.000 380.000 170.000 80.OCC 140.000 250.000 200.000 240.000 90.000 18.600 4.000 13.000 0.045 51.400 16.100 0. 100 0. 160 0. 16C 1.200 40.OCC 9.000 0.0 0.0 400.000 3.000 80.000 3.000 0.400 40.000 190.000 16.000 1.400 7.000 0.700 501.000 90.OOC 0.110 274 ITEM CLUSTER 176 ATTRIBUTE GROUP 41 NUTRIENTS: KCAL 414.000 PROT (GM) 6.3CC TRY (MG) 80.000 THR (MG) 220.OCC ISO (MG) 320.OCC LEU (MG) 510.OCC LYS (MG ) 220.OCC MET (MG) 110.000 CYS (MG) 70.OCO PHE (MG) 340.000 TYR (MG) 160.OCC V AL (MG > 260.OOC HI S (MG) 90.OCC FAT-T (GM ) 26.7CC SFA (GM) 6.000 PUFA (GM) 19. CCC C HOLE (GM ) 0 .040 CHO-T (GM) 37.7CG SUCR (GM) 15.ICC CHO-F (GM) 0.20C THIA (MG) 0 .16 C RI BO (MG) 0.17C NIACIN (MG) 1.3C0 VIT-B6 (MG) 40.000 FOLIC (UG) 8. OOC VIT-B12(UG) 0 .0 VIT-C (MG) 0.0 PANTO (UG) 5 CO.OCC BIOTIN (MG) 3.000 VIT-A (10) 60.000 VIT-D (IU) 1.000 VIT-E (MG) 0.7CC CA (MG) 38.OOC P (MG) 76.OCC MG (MG) 16.OCC FE (MG) 1.5CC I (MG) 7.OCC ZN (MG) 0.70C NA (MG) 234.OCC K (MG) 80 .000 CU (MG) 0.11C 177 178 41 42 322.000 207.OCC 6.300 4.000 130.000 60.OOC 360.000 180 .000 510.OCC 260.COC 760.000 400.000 1150.OCO 310.000 200.000 100.000 180.000 40.000 340.000 190.000 350.000 200.000 540.CCO 280.000 200.000 110 .000 9. 6CC 12.500 3.000 7.000 6. 000 4.000 0.080 C.C4C 52.400 20.600 19.100 15.6CC 0. 100 0.0 0. 180 0.C40 0. 160 0. 190 1.4C0 0. 100 40.000 30.000 8.000 1.000 0. 0 0.250 0.0 1 .000 200.OCC 500.COO 5.000 4.000 160.OCC 520.000 6.000 5.000 0.200 0.100 61.000 123.OOC 174.000 99.000 15.OOC 18.000 1.600 0.100 7.OOC 12.000 0.600 0.500 431.000 40.000 109.000 112.000 0.080 0.020 179 180 42 42 134.000 31.OOC 0.900 0.0 10.000 0.0 40.000 0.0 60.000 0. 0 90.COO 0.0 70.000 0.0 20.000 O.C 10.000 0.0 40.OCO 0.0 50 .000 0.0 60.000 0.0 20.OOC 0. 0 1 .200 0.0 0.0 0.0 0.0 0.0 O.C 0.0 30.800 8.000 29.300 8.000 0 .0 0.0 0. 010 0.0 0.030 O.C 0.0 0.0 30.000 0. 0 3.000 0.0 0. 0 0.0 2 .000 0.0 300.000 0.0 1.000 0.0 60.000 0.0 0.0 0.0 0.100 0.0 16.000 8.000 13.000 15.000 9. 000 1.000 0.0 0.400 2 .000 1 .000 0.200 0. 100 10.000 6.000 22.000 0.0 0.020 0.030 275 ITEM CLUSTER ATTRIBUTE GROUP NUTRIENTS: KCAL PROT (GM) TRY (MG) THR (MG) ISO (MG) LEU (MG) LYS (MG ) MET (MG ) CYS (MG) PHE (MG ) TYR (MG) VAL (MG ) HIS (MG) FAT-T (GM) SFA (GM) PUFA (GM) CHOLE (GM ) CHO-T (GM ) SUCR (GM) CHO-F (GM) THIA (MG ) RIBC (MG) NIACIN (MG) VIT-B6 (MG) FOLIC (UG) VIT-B12(UG> VI T-C (MG) PANTO (UG) BIGTIN (MG) VIT-A (IU) VIT-C (IU) VIT-E (MG) CA (MG) P (MG) MG (MG) FE (MG) I (MG) ZN (MG) NA (MG) K (MG) CU (MG) 181 182 43 43 124.000 59.000 3.4CC 4. 200 60.CCC 60. CCO 190.OCC 190.000 270.OCC 270. CCC 400 .OOC 420.000 270.OCC 330.000 110.000 100.000 240.OCC 40.000 220.OCC 200.OCC 200.OCC 200.000 280.OCC 290.OCO 110.OOC 110.000 3. OCC 2. 000 2.000 1.000 l.OCC I.000 0.C5C 0.002 22.80C 6.000 15.OCC 0. 0 0.1CC 0.0 0.020 0. 040 0.15C 0.210 0. ICC 0.100 50 .000 40.000 7.C0C 9.000 0.25C C. 4CC 0.0 1.000 5CC.0CC 4C0.OCC 5.OOC 3.000 130.OCC 80.000 2.000 41.000 0.7CC 0. 100 102. OOC 143.OCC 95. OCC 112.000 23. OCC 17.CCC 0.30C 0.100 7.OCC 8. OCO 2 .600 0.400 129.OOC 61.000 136.OOC 175.000 0.12C 0.020 183 184 44 44 58.OCC 91.000 0.200 0.200 0.0 O.C 6.000 6.OCO 11.000 11 .000 10.000 10.000 8.000 8 .000 3.000 3.COO 1 .000 1 .000 6.000 6. COO 3.000 3.OCC 7.000 7.000 3.000 3.CCC 0.60C 0.100 O.G O.C 0.0 0 .0 O.C O.C 14.500 23 .800 3.300 12.600 1.000 0.5CC 0.030 0.020 0.020 0.01C 0. 100 0.0 30.000 30.000 2.000 2 .000 0.0 0.0 4.000 1.000 100.000 100.000 1.000 l.OOC 90.000 40.000 0.0 0.0 0.600 0 .100 7.C0C 4. 000 10 .000 5.OCC 5.000 4.000 0.300 0.500 3.000 13.000 0.0 0.900 I .000 2.000 110.OOC 65.OOC 0.080 0 .010 185 44 85.000 I. 100 18.000 27.000 56.000 59.000 55.000 11.000 16.000 34.000 33.000 65.000 1.000 0.200 0.0 0.0 0. 0 22.200 8.700 0.50C 0.050 0.060 0.700 320.000 27.000 0.0 10.OOC 300.000 4.000 190.000 0.0 0.400 8.000 26.000 31.000 0.70C 8.000 0.200 1.000 370.000 0.130 276 ITEM CLUSTER 186 187 188 189 190 ATTRIBUTE GROUP 44 44 44 44 44 NUTRIENTS: KCAL 30.OCC 41.CCC 49.000 78.000 38.000 FROT (GM) 0.70C 0.500 1.000 0.400 0.6CC TRY (MG) 1 .OCC 1. OCC 3.000 1.000 1.000 THR (MG) 1 .OCC 1.000 1.000 1.000 l.OOC ISO (MG ) l.OCC 1.000 1.000 1 .000 1 .000 LEU (MG ) 1.000 1.000 l.OCO 1 .000 1.000 LYS (MG) 15.OCC 6.000 24.000 I .000 1.000 MET (MG) 2 .000 0.0 3.000 1. 000 1.000 CYS (MG) l.OCC 1.000 1 .000 1 .000 I.000 PHE (MG ) 21 .OCC 11.CCO 12.000 8.000 12.000 TYR (MG) 12.000 6.000 21.000 10.000 15.OOC VAL (MG) 1 .OCC l.OCO 1.000 I .000 1.000 HIS (MG) 1 .000 1.000 1.000 1.000 l.OOC FAT-T (GM) 0.100 0. 100 0.200 0.100 0.100 SFA (GM) 0.0 0.0 0.0 0. c 0.0 PUFA (GM) 0.0 0. 0 0.0 0 .0 0.0 CHCLE (GM ) 0.0 0. 0 O.C 0. 0 0.0 CHO-T (GM) 7.5CC 10.600 12.200 20.100 9.700 SUCR (GM ) 4.4CC 2.9G0 4.200 16.300 5. 900 CHO-F (GM) 0.3C0 0.200 0.500 0.400 0.600 THI A (MG) 0.04C 0. 040 0.100 0.010 0.020 RI BO (MG) 0 .030 0. 020 0.040 0.020 0.050 NIACIN (MG) 0.6CC 0.200 0.400 0.600 1.000 VIT-B6 (MG) SO.OCC 30.000 60.000 20.000 20.000 FOLIC (UG) 8.00C 2.000 45.000 11.000 11.000 VIT-B12CUG> 0.0 0.0 0.0 0. 0 0.0 VIT-C (MG) 33 .OCC 38.000 50.000 3.OOC 7.000 PANTO (UG) 3 CO.CCC 300.000 300.000 100.000 200.000 EIOTIN (MG) 3 .000 3.000 2.000 0.0 2.000 VIT-A ( IU) 34C0.0CC 80.000 200.000 430 .000 1330.000 VIT-D (IU) 0 .0 0. 0 0.0 0. 0 0.0 VIT-E ( MG) 0.1CC 0.300 0.200 0.0 0.500 CA (MG 1 14. OCC 16.OCO 41.000 4. 000 9. 000 P (MG) 16.OCC 16.000 20.000 12.000 19.000 MG (MG) 14.OCC S. 000 11.000 6.000 11.000 FE (MG) 0 .400 0.400 0.400 0.30C 0.500 I (MG ) 2.CCC 1.000 0.0 16.000 6.000 ZN (MG) 0 .100 0. 100 0.100 0.0 0.0 NA (MG) 12.CCC 1.000 1.000 2 .000 1.000 K (MG ) 251.CCC 135.OCC 2C0.000 130.000 202.000 CU (MG) 0.04C 0.040 0.090 0.07C 0.050 277 ITEM CLUSTER 191 192 193 194 195 ATTRIBUTE GROUP 44 44 45 46 47 NUTR IENTS: KC AL 74.OCC 52.000 289.000 59.000 304.000 PROT (GM) 0 .3CC 0.400 2.500 1.500 0.300 TRY (MG) 5.OCC 5. CCC 61.000 0.0 0.0 THR (MG) 1.000 1.000 61.000 30.000 0.0 I SO (MG) 1.000 1.000 74.000 20.000 0.0 LEU (MG) 1.000 1.000 77.000 50.000 0.0 LYS (MG) 9.OCC 9.000 65 .000 80.000 0.0 MET (MG) l.OCC 1. CCO 27.000 10.000 0.0 CYS (MG) 1 .OCC 1.000 1.000 O.C 0.0 PRE (MG ) 8.OOC 8. OOC 75.000 40.000 0.0 TYR (MG) 8.OOC 8.000 19.000 10.000 0.0 VAL (MG ) i.OOC I.000 94.000 40.000 0.0 HIS (MG) 1.000 1.000 49.000 10.000 0.0 FAT-T (GM) 0.1CC 0.200 0.200 0.0 0.0 SFA (GM) 0.0 0.0 0.0 0. 0 0.0 PUFA (GM) 0.0 0.0 0.0 o.c 0.0 CHOLE (GM • 0.0 0. 0 O.C 0.0 0.0 CHO-T (GM) 19 .400 13.700 77.400 14.100 82.300 SUCR (GM) 13•10 C 7.400 14.200 14.100 1.900 CHO-F (GM ) 0 .300 0.400 0.90C 0.0 0.0 THIA ( MG) 0.080 0.090 0.110 0.0 0.0 RI BO (MG) 0.C2C 0. 030 0.C8C 0.0 0.040 NIACIN (MG) 0.20C 0.200 0.500 0.0 0.300 VIT-B6 (MG) 70.OCC 90.OCO 240.000 0. 0 20.000 FOLIC (UG) 2 .OCC 1.000 9.000 0.0 3.OCO VIT-B12(UG) 0.0 0.0 0.0 0.0 0.0 VIT-C (MG ) 7.000 17.000 1.000 o.c 1.000 PANTO (UG) 2 CO.OOC 200.000 100.000 0 .0 200.000 BIOTIN (MG) 1.000 2.000 5.000 0.0 0.0 VIT-A ( IU) 50.OCO 70.000 20.000 0.0 0.0 VIT-D (IU) 0.0 C. 0 0.0 0.0 0.0 VIT-E (MG) 0.0 0.600 0.300 0.0 0.0 CA (MG ) 11 .OCC 17.000 62.000 0.0 5.000 P (MG) 5.OOC 8.000 101.000 0.0 6.000 MG (MG) 8.000 12.000 31.000 I .000 4.000 FE (MG) 0 .300 0. 500 3. 500 0. 0 0.500 I ( MG) 2.OCC 16.000 3 .000 1 .000 2.000 ZN (MG) 0.2CC C.20C 0.2C0 0. 500 0.900 NA (MG) 1 .OCC 1.000 27.000 51.CCC 5.000 K (MG) 96.OCC 146.000 763.000 0.0 51 .000 CU (MG) 0.150 0.070 0.230 0.0 1.670 278 ITEM CLUSTER 196 ATTRIBUTE GROUP 47 NUTRIENTS: KCAL 2 32.CCC PROT (GM) 0 .0 TRY (MG) 0.0 THR (MG) O.G ISO ( MG) O.C LEU (MG) 0.0 LYS (MG) 0.0 MET (MG) 0.0 CYS (MG ) 0 .0 PHE (MG) 0.0 TYR (MG) 0 .0 VAL ( MG) O.C HIS (MG) 0.0 FAT-T (GM) 0 .0 SFA (GM • O.C PUFA (GM) 0.0 CHOLE (GM) 0.0 CHO-T (GM) 60.000 SUCR (GM) 53.6CC CHO-F (GM ) 2 .CCC THIA (MG) 0.09C RIBO (MG) 0.120 NIACIN (MG) 1.200 VIT-B6 (MG ) 2C0.00C FOLIC (UG) 10 .000 VIT-B12(UG) 0.0 VIT-C (MG) O.G PANTO (UG) 4G0.0CC BIOTIN (MG ) 9. OCC VIT-A (IU) O.C VIT-D (IU) 0.0 VIT-E (MG ) 0 .20C CA ( MG) 290.OCC P (MG) 69. CCC MG (MG) 81.OCC FE (MG ) 6.OCC I (MG) 4.OCC ZN (MG) 4.600 NA (MG ) 37.OOC K (MG) 1063.OCC CU (MG) 1.170 197 156 199 200 47 48 48 49 385.000 273.000 290.000 399.000 0.0 0.100 O.C 4.000 0.0 0.0 0 .0 60.000 0. 0 0.0 0. 0 180.000 0.0 0 .0 0.0 260.000 C. 0 O.C 0.0 400.000 0.0 0.0 0.0 310.000 0. 0 0.0 0.0 100.000 0.0 O.C O.C 40.000 0.0 0.0 0.0 140.000 0.0 0.0 0. 0 200.000 0.0 0.0 0.0 280.000 0. 0 0.0 0. 0 110.000 0.0 0 .100 O.C 10.200 0. c O.C 0.0 5.000 0.0 0.0 0.0 5.OOC 0. 0 0.0 0.0 0.0 99.500 70.600 75.000 76.600 99.500 53 .000 4 .500 64.400 0. 0 0.0 0.0 0.200 0.0 0.010 O.C 0.030 0. 0 0.030 0.0 0. 170 0.0 0. 200 0.0 0.20C 0.0 30.000 0.0 20.000 0.0 1.000 0. 0 4.000 0. 0 0.0 0.0 0.0 0. c 4.000 0. 0 0.0 0.0 100.000 0.0 0.0 0. 0 1.000 0.0 5.000 0.0 10.000 0.0 10.000 0. 0 0.0 0.0 40.000 0.0 0.0 0.0 0.300 0.0 21 .000 46.000 148.000 0.0 7.000 16.OCO 122.000 0.0 4.000 2.000 0.0 8. 000 1.5CC 4. 100 1.400 0.0 1.000 4.OOC 7.CCO 0. 0 0.500 1.300 1 .100 1.000 17.OOC 68.OOC 226.000 3.000 75.000 4.000 192.000 0.030 0. 11C 0. C90 0.040 279 ITEM CLUSTER ATTRIBUTE GROUP NUTRIENTS: KCAL PROT (GM) TRY ( MG) THR (MG) ISO (MG) LEU (MG ) LYS (MG) MET (MG) CYS (MG) PHE (MG) TYR (MG) VAL (MG) HIS (MG) FAT-T (GM ) SFA (GM) PUFA (GM) CHOLE (GM) CHO-T (GM ) SUCR (GM) CHO-F (GM ) THI A (MG) R I BO (MG ) NIACIN (MG) VIT-B6 (MG) FOLIC (UG) VTT-B12(UG) VIT-C (MG) PANTO (UG) BIOTIN (MG ) VIT-A (IU) VIT-D (ID VIT-E (MG) OA (MG) P (MG) MG (MG) FE (MG) I (MG ) ZN (MG) N A (MG) K ( MG) CU (MG ) 201 202 49 49 520.OCC 385.000 7.70C 0.0 40.OCC 0.0 90.OCC 0. 0 120.OCC 0.0 240.CCC 0. C 130.OCC 0.0 20.OCC 0.0 80.OOC 0.0 160.OCC 0. 0 SO.CCC 0. 0 140.OCC 0.0 40.OCC 0. 0 32.3CC 0.0 19.OCC 0. 0 12.OOC 0.0 0.015 0.0 56 .900 99.500 43.OCC 59.500 0.4CC O.C 0.06C 0.0 0.34C 0. 0 0.300 0.0 20.OCC 0.0 8 .000 0.0 0.0 0. 0 O.C 0. 0 100.OOC 0.0 22.CCC 0. 0 270.OCC 0.0 88.OCC 0.0 1.100 0.0 228.CCC 0.0 231.CCC 0.0 82.OCC 0.0 1. ICC C. ICC 14.OCC 8.000 2.6CC 0. 0 94.000 1.000 384. OOC 3.000 1 .000 0.030 203 204 49 50 273.000 122.000 0. 100 6. 100 0.0 50 .000 0.0 260.000 0.0 350.000 0.0 520.000 0.0 450.CCC 0.0 60.000 O.C 20.000 0.0 310.000 0.0 180.000 0 .0 370.000 0.0 200.000 0.100 2.60C 0.0 1 .000 0.0 l.OCC 0.0 0.001 70.600 19.000 53 .000 3 .600 G.O 1.400 0 .010 0.08C 0.030 0.030 0.200 0.600 30.000 380.000 1.000 10.000 0 .0 0.0 4.000 2.000 100.000 100.OCO 1.000 6. 000 10.000 130.000 0.0 0.0 0.0 0. 100 21 .000 54.000 7.000 92.000 4.000 28 .000 1.500 1.800 1.000 4.000 0.500 I .400 17.000 463.000 75.000 210.000 0.110 0.210 205 51 192.000 7.300 110.000 330.000 480.000 790.000 530.OOC 190.000 160.000 300.000 330.000 480.000 270.000 9. 900 • 3.000 7.COO 0.02 0 18.000 0.300 0.100 0.C30 0. 060 1. 200 110.000 5.000 0.490 0. 0 400.000 2.000 410.COC 0.0 0. 400 10.000 48.000 11.000 1. 000 3.000 1.000 366.000 93.000 0. 060 280 ITEM CLUSTER 206 ATTRIEUTE GROUP 52 NUTRIENTS: KCAL 89.000 PROT (GM) 6.40C TRY (MG) 80.000 THR ( MG) 260.OCC ISO (MG ) 470.OCC LEU (MG) 590 .OCC LYS (MG) 470.OCC MET (MG) 150.OCC CYS (MG) 110.OCC PHE (MG ) 350 .000 TYR ( MG ) 250.OCC VAL (MG) 360.OCC HIS (MG ) 690.OCC FAT-T (GM) 4.3CC SFA (GM ) 2 .OCC PUFA (GM) 2.000 CFXLE (GM ) 0.020 CHO-T (GM) 6.2CC SUCR (GM) 0 .300 CHO-F (GM) 0.4CC THIA (MG) 0.06C RIBC (MG) 0.07C NIACIN (MG ) 1.9 CO VIT-E6 (MG) 110.000 FOLIC (UG) 3. CCC VIT-E12(UG) 0 .65C VI T-C ( MG) 7.OOC PANTO (UG) 5 CO.CCC BIOTIN (MG) l.OCC VIT-A (IU) 980.OOC VIT-D (IU) 0.0 VIT-E (MG) 0.3CC CA (MG ) 12.000 P ( MG) 75.OOC MG (MG) 20.CCC FE (MG) 1.20C I (MG) 3.OCC ZN (MG) 1 .OCC NA (MG) 37.OCC K (MG) 250.000 CU (MG) 0.02C 207 208 209 210 53 54 55 56 38.000 102.CCC 133.000 173.000 2.600 12.400 7.500 12.800 30.000 120.OOC 70.000 160.000 90.000 440.000 300.000 520.000 110.OOC 540.COO 410.000 660.000 170.000 800.000 630.OCC 940.000 190.OCC 900.COO 620.000 960.000 60.000 270.000 140.000 300.000 30.000 130.000 90.000 180.000 490.000 260.000 360.000 100.000 70.000 360.000 270.000 400.000 100.000 500.000 430.000 500.000 60.000 300.000 240.000 300.000 0. ICO 4.000 6.100 8.500 0.0 1.000 3.OOC 3. 000 0. 0 2.000 3.000 4.000 0.006 0.C2C 0.015 0.050 7.100 4.000 12.200 11.300 1.000 0.50C 8. 900 1.200 0.300 0.300 0.600 0.400 0. C2C 0.030 0. 030 0.070 0.040 0.090 0.070 0. 18C 0.4C0 1.700 1.300 5.200 180.000 180.000 100.000 300.OOC 5.000 5.000 9.000 6.000 0.660 0.660 0.230 0. 220 5. 000 4.000 2 .000 4.000 500.000 5C0.000 100.000 500.000 2.000 5.000 2.000 8.000 60.CCO 110.000 60.000 590.000 0.0 0.0 0.0 0.0 0. 0 1.200 0.200 0.200 18.000 23.000 32.000 41.000 34.000 117.000 126 .000 145.000 18.000 18.000 26.000 19.000 0. 500 I.000 1 .700 1.200 3. OCO . 4.000 5.OOC 5.000 0.500 1.900 1.8C0 2.500 290.000 287.OOC 531.000 344.000 167.000 189.000 233.000 112.000 0. 110 0.190 0 .330 0.220 281 ITEM CLUSTER 211 212 213 214 215 ATTRIEUTE GROUP 57 58 59 6C 61 NUTRIENTS: KCAL 131.OCC 112.000 181.OCO 215.000 245.000 PROT (GM) 8.OCC 8. 400 8.800 8.400 9.500 TRY (MG) 80 .OCC 100.000 100.000 170.000 160.000 THR (MG) 250.000 5C0.OCO 390.000 520.000 400.000 ISO (MG ) 460.000 630.000 460.COC 750.000 690.000 LEU (MG > 580.OOC 920.000 720.000 1110.000 1090.000 LYS (MG) 460 .000 1090.000 780.000 730.COO 590.000 MET ( MG) 140.CCC 330.000 220.000 280.000 240.000 CYS (MG) 110.OCC 17C. OCC 110.000 170.000 160.000 PHE (MG) 330.OCC 340.000 360.000 440.000 490.000 TYR (MG) 250. OCC 130.OCO 3C0.000 560.000 520.000 VAL (MG) 370. OOC 590.000 490.000 860.000 700.COO HIS (MG ) 680.OCC 320.000 310.000 370.000 330.000 FAT-T (GM) 6 .700 3.000 11.300 11.100 7. 100 SFA (GM) 3. OCC 1. 000 5 .000 5 .000 2.000 PUFA (GM) 3.CCC 1.000 5.000 6. OCO 3.000 CHOLE (GM) 0.02C 0.020 0.020 0.040 0.040 CHO-T (GM ) 9.8CC 12.7 00 1C.7C0 20.100 35.400 SUCR (GM) 1 .40C 1.200 0.100 0. 100 3.900 CHO-F (GM) 0.3CC 0. 300 0.50C 0.100 0.300 THIA (MG) 0.100 0.0 70 0.C10 0. ICO 0. 060 RIBO (MG) 0. 140 0.090 0.090 0.200 0.170 NIACIN (MG) 1 .700 2.300 2. 100 0. 900 1.000 VIT-B6 (MG) 230.OCC 250.000 80.000 40.000 50.000 FOL IC (UG ) 7.OCC 10.OCO 9.000 6. 000 3.000 VIT-E12(UG) 0.36C 0.210 0.920 0.35C 0. 200 VIT-C (MG) 4. OCC 4. CCC 10.000 0.0 6. 000 PANTO (UG) 300.000 600.000 500.000 200.000 300.000 BIOTIN (MG) 2.OOC 4.000 2.000 2 .000 1.000 VIT-A (IU) 430 .OOC 130.000 10.000 430.000 440.000 VIT-D ( IU) 0.0 0.0 0.0 20.000 8 .000 VIT-E (MG) 0.2CC C. 2CO 0. 100 0.500 0.300 CA (MG) 19.000 26.000 13.000 181.CCC 156.000 P (MG ) 117.CCC 87.000 67.000 161.000 156.000 MG (MG) 19.000 21.000 19.000 26.000 27.OOC FE (MG) 1 . 3 C C 1. 100 2.000 0.900 0.900 I (MG) 4.000 3.000 4.COO 5. OCO 9.000 ZN (MG) 1.900 1.200 1.200 0.300 I. 100 NA (MG) 393.OCC 400.OCC 540.000 543.000 647.000 K (MG) 115.OCC 176.000 200.000 120.OCC 114.000 CU (MG) 0.17C 0. 14 C 0.140 0.040 0.340 282 ITEM CLUSTER 216 217 218 219 220 ATTRIEUTE GROUP 62 63 64 65 66 NUTRIENTS: KCAL 76.OCC 134.000 140.000 120.000 O.C PROT (GM) 2.2CC 7.500 4.50C 20.000 0.0 TRY (MG J 30.OOC 80.000 70.OCC 230.000 0.0 THR (MG) SO. OCC 300.000 120.000 880.000 0.0 ISO (MG) 110 .000 350.000 150.000 1040.000 0.0 LEU (MG) 150.CCC 530.000 230.000 1630.000 0.0 LYS (MG) 70.OCC 520.CCC 240.000 1740.COO 0.0 MET (MG) 30.OCC 160.000 70.000 490.000 0.0 CYS (MG ) 40. OOC 100.OCO 40.000 250.000 0.0 PHE (MG) 110 .OOC 310.000 130.000 340.CCO 0.0 TYR (MG) 70.OCC 220.000 100.000 680.000 0.0 VAL (MG) 130 .000 350.000 150.OOC 1110.000 0.0 HI S (MG) 50.OCC 220.000 100.000 690 .000 0.0 FAT-T (GM) 0.6CC 4.700 7. 100 3. 000 0.0 SFA (GM) O.C 2.000 3.000 I.000 0.0 PUFA (GM ) O.C 3. CCC 3.COO 1.000 0.0 CHCLE (GM) 0.006 0.020 0.C09 O.C02 O.C CHO-T (GM) 15.4CC 15.6CC 14.200 5.000 0.0 SUCR (GM) 5.50C 4.200 0.0 0.0 0.0 CHO-F (GM) 0.2CC 0. 300 0. 0 1 .100 0.0 TH IA (MG) 0.140 0. 100 O.C 0.010 0.0 RIBO (MG) 0.11C 0.120 0.0 0.230 0.0 NIACIN (MG > 1.8CC 1.600 0.0 11.400 0.0 VIT-B6 (MG) 50 .OCC 150.000 200.000 O.C 0.0 FOLIC (UG) l.OCC 6. CCC 1.000 0.0 0.0 VITrei2(UG) 0.250 0.220 0.0 0.0 0.0 VIT-C (MG) 4.OCC 9.000 0.0 0.0 0.0 PANTO (UG) 300 .OOC 200.000 400.OOC 0.0 0.0 BIOTIN (MG) 0.0 1.000 10.000 0 .0 0.0 VIT-A ( IU) 370.OCC 640. OCC 0.0 0. 0 0.0 VIT-D (IU) l.OOC 0.0 O.C 0.0 0.0 VIT-E (MG) 0.4CC 0.3CC 0. 100 0.100 0.0 OA (MG) 16.OCC 50.000 20.000 O.C 253.OOC P (MG > 35.OCC 95.000 39.000 297.000 200.000 MG (MG) 11.000 17.000 9.000 57.OOC 190.000 FE (MG) 1. IOC I. 500 1 .200 4.600 0. 100 I (MG ) 5.CCC 3.000 3.000 43.000 10000.000 ZN (MG) 0.10C 1.400 0 .900 0.800 0. 500 NA (MG ) 382.OCC 4C7.0CC 665.000 2400.000 38758.000 K (MG) 121.000 268.000 0.0 100.000 4.OOC CU (MG) 0.12C 0. 17 0 0.050 0.090 0.380 ITEM CLUSTER 221 ATTRIEUTE GROUP 67 NUTRIENTS : KCAL PROT (GM) TRY (MG) THR ( MG) ISO (MG) LEU (MG) LYS (MG) MET (MG) CYS (MG) PFE (MG ) TYR (MG) VAL (MG ) HIS (MG) FAT- T (GM) S FA (GM) PUFA (GM) CHCLE (GM ) CHO- T (GM) SUCR (GM) CHO- F (GM) THIA (MG) RIBC (MG) NIACIN (MG) VIT- B6 (MG) FOLIC (UG ) VIT- E12(UG ) V IT-C (MG ) PANTO (UG) BIOTIN (MG ) VIT- A (I U) VIT- D (IU) VIT- E (MG) CA (MG ) P (MG) MG (MG) FE (MG ) I (MG) ZN (MG) NA (MG) K (MG) CU (MG ) 392 .OOC 9.4CC 40 .000 110.OCC 150.CCC 3C0 .OOC 160.OCC 20 .000 ICO.000 200.000 110.OOC 190. CCC 50 .CCC 10.6CC 6 .000 4.OCC 0.0 73 .9CC 12. CCC 0 .800 0 . C 8 C 0.410 0.5CC 20.OCC 80 .OCC 0.0 1.000 ICO.OOC 32 .CCC 10.OCC 1C60.C0C 0 .4CC 275.OCC 290 .OOC 371.OCC 1.4CC 8 .CCC 2.60C 382 .OCC 605.OCC 3.69C 284 APPENDIX D ABRIDGED FOOD-COMPOSITION FILE Listing of 22 nutrients selected from the food-composition file (Appendix C) for each of the 127 item clusters designated in Appendix B. ENERGY (kcal) PROTEIN (gm) TOTAL FAT (gm) SATURATED FAT (gm) POLYUNSATURATED FAT (gm) TOTAL CARBOHYDRATE (gm) SUCROSE (gm) FIBER (gm) THIAMIN (mg) RIBOFLAVIN (mg) NIACIN (mg) PYRODOXINE (ug) FOLACIN (ug) ASCORBATE (mg) RETINOL (iu) CHOLECALCIFEROL (iu) TOCOPHEROL (mg) CALCIUM (mg) PHOSPHORUS (mg) MAGNESIUM (mg) IRON (mg) POTASSIUM (mg) 286 APPENDIX E NUTRIENT-LIMITS FILE Tables of daily minimum (Table E-l) and daily maximum (Table E-2) nutrient limits disaggregated for age, sex, activity pattern, and pregnancy status. 287 Table E-l. Minimum nutrient limits Code # Sex .ab Proteinaef Histidinefh Isoleucinefi Age Activity Energy' Pattern kcal/kg /day gm/kg/day mg/kg/day 111 Male 19-35 A 42.6 112 B 40.6 113 C 35.7 114 D 32.6 121 36-50 A 41.4 122 B 39.3 123 C 34.4 124 D 31.3 131 51-65 A 40.2 132 B 38.0 133 C 33.3 134 D 30.1 141 66 + A 38.9 142 B 36.9 143 C 32.0 144 D 28.6 211 Female 19-35 A 37.3 212 B 35.2 213 C 33.8 214 D 30.7 221 A 36.3 222 B 34.1 223 C 32.7 224 D 29.6 231 51-65 A 35.0 232 B 32.9 233 C 31.4 234 D 28.4 241 A 33.9 242 66 + B 31.8 243 C 30.4 244 D 27.3 205 Preg. 1st + 100 206 Preg. 2nd + 3rd + 100 207 Lact. + 500. .57 .10 mg/kg/day .10 .52 kcal/day * 20 gm/day kcal/day + 20 gm/day kcal/day + 24 gm/day Values for minimum limit (and for recommended energy intake) are ob tained from: Bureau of Nutritional Sciences, Department of National Health and Welfare. Dietary Standard for Canada. Information Canada, Ottawa, 1975. To calculate lower and upper limit on energy intake per day for an individual, the recommended intake for a given age, sex, and activity level is multiplied by kilograms ideal body weight, then, the multi plied figure is summed with the additional caloric requirements for pregnancy or lactation, and lastly, this figure is reduced or raised five percent. For longer period of time, the daily limits are multi plied by the number of days considered. Body weight is measured as kilograms ideal body weight. 288 Table E-l. Minimum nutrient 1imits (cont'd) Code # 111 112 113 114 121 122 123 124 131 132 133 134 141 142 143 144 211 212 213 214 221 222 223 224 231 232 233 234 241 242 243 244 205 206 207 • fi Leucine mg/kg/day 14 Lysine^1' mg/kg/day 12 Methionine + Cystine mg/kg/day 13 fi Phenylalanine + Tyrosine mg/kg/day 14 fi Threonine mg/kg/day fi Where increased minimum allowance associated with pregnancy and lacta tion is not stated, the increase, if any, is calculated on the basis of increased body weight or caloric requirement. Maximum intake limits for pregnant and lactating females are the same as for non pregnant females. Minimum protein limit for the abridged nutrient consumption file (Appendix F) is 0.80 gm/kg/day for males, and 0.73 gm/kg/day for females. This figure is based on the average amino acid composition of the Canadian diet as stated in the Dietary Standard For Canada, Bureau of Nutritional Sciences, Department of National Health and Welfare, Information Canada, Ottawa, 1975. 289 Table E-1 Code # Minimum nutrient limits (cont'd) 111 112 113 114 121 122 123 124 131 132 133 134 141 142 143 144 211 212 213 214 221 222 223 224 231 232 233 234 241 242 243 244 205 206 207 Tryptophan 1 Valine 1 mg/kg/day mg/kg/day Fata:i Saturated^ P/Sk Polyunsat.aj %kcal/day 3.5 10 Fat %kcal/day 0 Ratio Fat %kcal/day For determining limits on intake per day, multiply the table value by the individual's ideal body weight in kilograms. Additional minimum protein requirement for pregnancy and lactation based on the composition of the average daily protein intake of Canadians as per the Dietary Standard for Canada, Bureau of Nutritional Sciences, Department of National Health and Welfare, Information Canada, Ottawa, 1975. Value for minimum limit on histidine obtained from: cation, Dr. P.J. Stapleton, 1978. personal communi-Values for minimum limit obtained from: FA0/WH0, Ad Hoc Expert Committee. Energy and Protein Requirements. WHO Tech. Rep. Ser. 522, Geneva, 1973. No. • 29Q Table E-l. Minimum nutrient limits (cont'd) Code # Cholesterol9 Carbo-J1 SucroseJm Fiber"0 Thiamin30 Niacin30' mg/day hydrate % kcal/day gm/100 mg/1000 NE/1000 % kcal/day kcal/day kcal/day kcal/day 111 112 113 114 121 122 123 124 131 132 133 134 141 142 143 144 211 212 213 214 221 222 223 224 231 232 233 234 241 242 243 244 205 206 207 55 0.4 0.5 6.6 +0.2 mg/day +0.4 mg/day +2 NE/day +7 NE/day To convert %kcal/day to gm or mg/day, use the average caloric intake for an individual of given age, sex, weight, activity level, and pregnancy status; and the conversion value of 9kcal/gm for fat, or 4kcal/gm for carbohydrate. Value for limit obtained from: American Heart Association. Diet and Coronary Heart Disease. 1973. Value for limit obtained from: Select Committee on Nutrition and Human Needs, U.S. Senate. Dietary Goals for the United States. Government Printing Office, Washington, D.C., 1977. 29J Table E-l. Minimum nutrient limits (cont'd) Code # 111 112 113 114 121 122 123 124 131 132 133 134 141 142 143 144 211 212 213 214 221 222 223 224 231 232 233 234 241 242 243 244 205 206 207 m Riboflavin90 Vit-B6aq mg/1000 kcal/day mg/day 2.0 Folacinc ug/day 200 Vit-B12c ug/day 3.0 Vit-Cao mg/day 30 Panto thenate mg/day 5.0 Biotinc ug/day 40 1.5 + 0.5 mg +0.3 mg/day + 0.5 mg +0.6 mg/day + 0.6 mg + 50 ug + 50 ug + 50 ug + 1.0 ug + 1.0 ug +0.5 ug + + 20 mg + 20 mg 20 mg Value for minimum limit on sucrose obtained from: Food and Nutrition Board, NRC/NAS, Recommended Dietary Allowances, 1974. Value for limit obtained from: Cheney, M.C. Food Enrichment: Nutritional Standard for Synthetic and Modified Foods. Paper pre sented at Current Topics in Food and Nutrition - Workshop. University of British Columbia, Vancouver, B.C., July 27, 1976. Average caloric intake per day for an individual of given sex, age, size, activity level, and pregnancy status is used for conversion of table value to gm/day, mg/day, or NE/day. Potential niacin equivalents derived from conversion of excess trypto phan have not been considered in evaluation of niacin intake. 292 Table E-l. Minimum nutrient limits (cont'd) Code # Vit-A' Vit-D iu/day iu/day 111 112 113 114 121 122 123 124 131 132 133 134 141 142 143 144 211 212 213 214 211 222 223 224 231 232 233 234 241 242 243 244 205 206 207 a s 5000 100 Vit-Ed Calcium0 mg/day mg/day 9 800 Phosphorus mg/day 800 Ca/P" Ratio .8 Magnesium mg/dg/day 4.5 4000 700 700 +1000 +2000 iu iu +100 +100 +100 iu iu iu +1.0 +1.0 +2.0 mg mg mg +500 +500 +500 mg mg mg +500 mg +500 mg +500 mg +25 mg/day +25 mg/day +75 mg/day H The increased requirement for pyridoxine and vitamin-C with the con sumption of birth control pills has not been considered in determining the minimum limit. A tenfold increase in requirement has been suggested in the Dietary Standard for Canada. Bureau of Nutritional Sciences, Department of National Health and Welfare, Information Canada, Ottawa, 1975. r Value for limit obtained from: Food and Nutrition Board, National Research Council. Recommended Dietary Allowances, 8th edition. National Academy of Sciences, Washington, D.C, 1974. Minimum vitamin-D allowance should be increased 100 iu/day for those individuals confined indoors or otherwise deprived of sunlight for long period of time, except pregnant and lactating females whose rec ommendation has already been increased to 200 iu. 293 Table E-l Code # 111 112 113 114 121 122 123 124 131 132 133 134 141 142 143 144 211 212 213 214 221 222 223 224 231 232 233 234 241 242 243 244 205 206 207 Minimum nutrient limits (cont'd) afu j i • ao Iodine ug/1000 kcal/day 50 Iron" Zinc" Sodium"'" Potassium mg/day mg/day mg/kg/day mg/kg/day afu 10 10 20 Copper mg/day 2.0 14 +15 mg/day +15 mg/day +25 ug/day +1 +1 +1 mg mg mg +3 mg +3 mg +7 mg +164 mg/day +164 mg/day +864 mg/day +280 mg/day +280 mg/day +1480 mg/day v Values for limits on Ca/P ratio arbitrarily assigned for test purposes. Minimum maintenance level has been used for establishing the minimum sodium and potassium limit as per the Dietary Standard for Canada. Bureau of Nutritional Sciences, Department of National Health and Welfare, Information Canada, Ottawa, 1975. Value for the maximum limit arbitrarily assigned at twice the minimum limit in the absence of empirical values. This was done for test purposes and does not imply that these values represent empirical values. 294 Table E-2. Maximum nutrient limits Code # Sex Age Activity Pattern ab Energy kcal/kgVday 111 Male 19-35 A 42.6 112 B 40.6 113 C 35.7 114 D 32.6 121 36-50 A 41.4 122 B 39.3 123 C 34.4 124 D' 31.3 131 51-65 A 40.2 132 B 38.0 133 C 33.3 134 D 30.1 141 66 + A 38.9 142 B 36.9 143 C 32.0 144 D 28.6 211 Female 19-35 A 37.3 212 B 35.2 213 C 33.8 214 D 30.7 221 36-50 A 36.3 222 B 34.1 223 C 32.7 224 D 29.6 231 51-65 A 35.0 232 B 32.9 233 C 31.4 234 D 28.4 241 66 + A 33.9 242 B 31.8 243 C 30.4 244 (j A 27.3 205 Preg. 1st 206 Preg. 2nd+3rd 207 Lact. •f7 Protein u gm/kg/day 1.14 Histidine mg/kg/day 1057 fw 1.04 957 The maximum limit for amino acids is the amount of any one amino acid that could be attained in excess of the minimum requirement for all other amino acids without exceeding the maximum protein intake limit. x The maximum limit on polyunsaturate fat equals the upper limit on total fat intake. ^ The maximum limit on cholesterol intake is arbitrarily set at twice the American Health Association recommended upper level of cholesterol intake. z The maximum limit on total carbohydrate intake allows 10% of energy expenditure for minimum protein requirement and minimum fat requirement. * Value for maximum limit of vitamin-C intake arbitrarily assigned for test purposes. 295 Table E-2. Maximum nutrient limits (Cont'd) f w . • f w • • f w 1• Leucine Lysine mg/kg/day mg/kg/day Code # Isoleucine mg/kg/day 111 112 113 114 121 122 123 124 131 132 133 134 141 142 143 144 211 212 213 214 221 222 223 224 231 232 233 234 241 242 243 244 205 206 207 1057 1061 1059 Methionine + Cystine mg/kg/day 1060 fw Phenylalanine + Tyrosine mg/kg/day 1061 fw 957 961 959 960 961 + Values for maximum limits of vitamin-D and vitamin-A intake arbitrarily assigned for test purposes. Determination of the maximum limit of vitamin-A does not consider non-toxicity of the provitamin forms. Value for maximum limit of vitamin-E intake arbitrarily assigned for test purposes. The maximum limit on sodium intake is extrapolated from the suggested maximum sodium intake per day of 3 grams, obtained from the Select Committee on Nutrition and Human Needs, U.S. Senate. Dietary Goals for the United States. Government Printing Office, Washington, D.C., 1977. The goal of6-3 gm/day is transformed to mg/kg/day on the basis of a reference individual of 60 kg. 296 Table E-2. Maximum nutrient limits (cont'd) Code # 111 112 113 114 121 122 123 124 131 132 133 134 141 142 143 144 211 212 213 214 221 222 223-224 231 232 233 234 241 242 243 244 205 206 207 Threonine mg/kg/day 1054 fw Tryptophan mg/kg/day 1050 fw Valinefw mg/kg/day 1057 Fatjk SaturatedjkP/Sv %kcal/day 30 Fat %kcal/day 10 Ratio 954 950 957 297 Table E-2. Code # 111 112 113 114 121 122 123 124 131 132 133 134 141 142 143 144 211 212 123 214 221 222 223 224 231 232 233 234 241 242 243 244 205 206 207 Maximum nutrierit 1imits•(cont'd) y Carbo-JZ hydrate Polyunsat Fat %kcal/day Cholesterol* mg/day 30 600 %kcal/day 90 Sucrose0' Fiberuy Thiamin %kcal/day gm/100 mg/1000 kcal/day kcal/day ov 15 0.8 1 .-0 298 Table E-2. Maximum nutrient 1imi ts (cont'd) Code # 111 112 113 114 121 122 123 124 131 132 133 134 141 142 143 144 211 212 213 214 221 222 223 224 231 232 233 234 241 242 243 244 205 206 207 Niacin07 NE/1000 kcal/day 13.2 Riboflavin mg/1000 kcal/day 1.2 ov Vit-B6' mg/day 4.0 Folacin ug/day 400 Vit-B12 ug/day 6.0 v Vit-C mg/day 500 Panto-v thenate mg/day 10.0 299 Table E-2. Maximum nutrient 1imits:(cont'd) Code # 111 112 113 114 212 122 123 124 131 132 133 134 141 142 143 144 211 212 213 214 221 222 223 224 231 232 233 234 241 242 243 244 205 206 207 Biotinv Vit-A+ ug/day iu/day 80 20000 Vit-D iu/day 600 Vit-E mg/day 1600 ++ Calcium mg/day 1600 Phosphorus mg/day 1600 Ca/P* Ratio 1.2 1400 1400 300 Table #-2. Maximum nutirent limits (cont'd) Code # 111 112 113 114 121 122 123 124 131 132 133 134 141 142 143 144 211 212 213 214 221 222 223 224 231 232 233 234 241 242 243 244 205 206 207 Magnesium mg/kg/day 9.0 Iodine^ ug/1000 kcal/day 100 T V Iron mg/day 20 Zincy mg/day 20 Sodiurn mg/kg/day 50 Potassium mg/kg/day 40 Copper mg/day 4.0 28 3Q1 APPENDIX F ABRIDGED NUTRIENT-LIMITS FILE Listing of 48 maximum and minimum nutrient limits selected from the nutrient-limits file (Appendix E) to coordinate with the abridged food-composition file (Appendix D) and the abridged food-item file (Appendix B). MINIMUM ENERGY PROTEIN TOTAL FAT SATURATED FAT POLYUNSATURATED FAT P/S RATIO TOTAL CARBOHYDRATE SUCROSE FIBER THIAMIN RIBOFLAVIN NIACIN PYRIDOXINE FOLATE ASCORBATE RETINOL CHOLECALCIFEROL TOCOPHEROL CALCIUM PHOSPHORUS CA/P RATIO MAGNESIUM IRON POTASSIUM * see Appendix E for the values 1imits. 3Q2 MAXIMUM* ENERGY PROTEIN TOTAL FAT SATURATED FAT POLYUNSATURATED FAT P/S RATIO TOTAL CARBOHYDRATE SUCROSE FIBER THIAMIN RIBOFLAVIN NIACIN PYRIDOXINE FOLATE ASCORBATE RETINOL CHOLECALCIFEROL TOCOPHEROL CALCIUM PHOSPHORUS CA/P RATIO MAGNESIUM IRON POTASSIUM associated with each of these nutrient 303 APPRENDIX 6 ATTRIBUTE-GROUP MATRIX Table of attribute group assignments for 221 item clusters. Item a Attrib. k - lbc k = 2 k = 3 k=4 k = 5 k = 6 k = 7 Cluster Group 0-| =67 ^2 = ^ ^3 = ^ ^4 = ^ ^5 = ^ ^6 = ^ Jy = 4 Code"# Code # 001-005 001 cheese dairy dairy-eggs non-sweet solid 006-009 002 cultured milk  010-011 003 e^s 012-013 004 ready-to-eat cereal breakfast grain 014-016 005 cooked  017 006 pancake-waffle  018-019 007 pastas dinner 020-023 008 other cereal 024-030 009 bread bread-rolls other meal 031-035 010 rolls  036-040 011 crackers  041-048 012 mammal carcass meat meat protein foods 049-052 013 poultry 053-066 014 fish 067 015 1 iver organ meat 068 016 other organ  069-071 017 variety meat  072-075 018 peas-beans plant protein 076-081 019 nuts 082-087 020 potatoe vegetables vegetable 088-102 021 greens  103-112 022 yellow-red 113-115 023 other color 116 024 mixed  117-119 025 veg. products  120-124 026 cooking fats fats and oils 125-126 027 table fats  127-131 028 sauces  132-135 029 salad dressing _ 136-138 030 fresh fluid milk dairy beverages fluid 139-142 031 cream  143-149 032 fruit fruit-vegie 150 033 vegetable  151-152 034 misc. -sugar  153-154 035 misc.-tea 155-158 036 misc.-alcohol Item3 Attrib. Cluster Group Code # Code # 159-164 037 165-170 038 171-172 039 173-174 040 175-177 041 178-180 042 181-182 043 183-192 044 193 045 194 046 195-197 . 047 198-199 048 200-203 . 049 204 050 205 051 206 052 207 053 208 054 209 055 210 056 211 057 212 058 213 059 214 060 215 061 216 062 217 063 218 064 219 065 220 066 221 067 J1 = 67l k = 2 55 k - 3 L = 52 k = 4 J4 = 38 k = 5 J5 = 35 6 27 k = 7 Jy = 4 cakes cakes-pies gram soups desserts sweet solid pies cookies cookies-other other pastry frozen non-frozen dairy not dried fruit dried simulated sweeteners sweets spreads candies miscellaneous The abridged attribute-group matrix which corresponds to the abridged food-item file (Appendix B) is derived from Appendix G. k equals the hierarchy level number. j equals the number of attribute classes in the hierarchy level. 

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.831.1-0094593/manifest

Comment