Uses and abuses of food composition data

Proceedings of a Symposium held at the University of New South Wales, 25 and 26 September, 1989, organised by the Dietitians' Association of Australia (NSW), the Council of Australian Food Technology Associations and the Department of Food Science and Technology, UNSW.

H. Greenfield, editor

Introduction S4 Composition of foods, Australia R. English S5 Food composition data and clinical dietetics D. Foote S8 Food composition data in clinical research J. Allen S10 The uses of food composition data in nutritional K.I. Baghurst and epidemiology P.A. Baghurst Sll Food composition data and the National Heart Foundation's Food Approval Program B. Haddy S14 Sampling foods for nutrient composition studies J. Cunningham S16 Data production and data quality assurance in a nutrient P. Scheelings and analysis program D. Buick S18 Compilation and scrutiny of food composition data K. Cashel S21 Food composition data in the promotion of lean meat in Australia L. Fantini S25 Food composition data in the promotion of canned foods A.K. Wailes S29 The development of a special purpose food composition data base for industry J.A. Barnes S32 Food composition data use for food legislation purposes R. Peters S36 The New Zealand food composition data bases B.A. Burlingame S39 Recommendations S44

This publication is reproduced with permission of the copyright holder, H. Greenfield.

Copies of these proceedings are available from A/Prof. H. Greenfield, Department of Food Science and Technology, University of New South Wales, PO Box 1, Kensington NSW 2033 at a cost of $A12 (includes postage). Make cheques and money orders payable to Dept. of Food Science and Technology, UNSW.

Introduction

he idea for this Symposium arose view of the importance of regional Tf rom the realisation in 1987 that cooperation, there is also a paper on although publication of the new New Zealand's food composition Australian food composition tables, program. A number of recommendations Composition of foods, Australia (Cashel were drawn up by four workshops and & others 1989) was imminent, there had agreed to in plenary session. These been very little public discussion by recommendations conclude the printed intending users about the new analytical proceedings. data and the associated new tables and In the USA, a National Nutrient computerised data base. Those who Databank Conference is held annually in had been involved in the analytical work different states (eg Morgan 1980) and and subsequent publications at the many other countries are now also University of New South Wales were beginning to hold national meetings (eg conscious of the level of FNRI 1985). In Australia the first such misunderstanding among some readers meeting was held at the University of who had written with requests for New South Wales in 1980 (Greenfield & clarification of points related to the Wills 1981) and it is hoped that other meaning, derivation or use of the data. national meetings will be held from time These requests had already exposed to time to increase knowledge about the the lack of training among some users in national food composition data base and the scientific meaning and appropriate to learn how to exploit it fully and use of nutrient composition data. appropriately. It is anticipated that the It was therefore decided to hold a computerised data base will become an meeting which would enable scientists everyday tool for all working in the field who had been involved in the Australian of nutrition in Australia in the coming analytical program to give outline years. This makes it even more presentations on how the work had important that users become more been carried out, and to enable some informed and knowledgeable about the major users in various fields to discuss data and their appropriate use, as well their requirements and current uses of as computer-literate. The power of the data. The Symposium was held on 25 computer data base and the situation and 26 September, 1989, at the created by ongoing releases of data and University of New South Wales and was the associated costs could lead to a attended by about 200 people from proliferation of different versions in use, industry, health and consumer and easy access by untrained users, organisations, hospitals, academia, and both situations with the potential for research institutions. These Symposium abuse or inappropriate use. It is also proceedings contain the papers clear that nutrient composition data will presented on sampling, on data grow in importance for nutrition production and analytical quality education and consumer education. assurance, and on data scrutiny and Industry has long recognised the compilation. In addition there are papers importance of creating its own nutrient outlining the use of food composition data bases and the mutual benefits of data and data bases or tables in the process to industry, research and dietetics, clinical research, nutritional the consumer alike (Gurr 1989). Other epidemiology, health promotion, food professional groups must not lag behind legislation and the food industry, in knowledge about the nutrient particularly in product advertising. In composition of foods and the applications of the data, and should not hesitate to make their voices heard on needs for data and data formats. The Symposium organisers wish to thank those who participated in the

Symposium, including those who Nutrient Analysis Software for displayed posters or who demonstrated software packages, those who assisted Dietitians and Food Scientists with publicity (eg Food Australia, the The DCI Consultant's “Smart Diet!” Nutrition Society of Australia, the Home (c) is a powerful nutritional analysis Economics Association of Australia, and tool while at the same time simple to regional groups of the Dietitians' operate. The graphical output provides Association of Australia), the a clear profile of a nutrient analysis, organisations who sponsored the and at a glance the strenghts and participation of speakers weaknesses of a food product, (Commonwealth Department of menu,diet or recipe can be focussed Community Services and Health, on for appropriate modification. Canned Food Information Service Inc, Key features of “Smart Diet !” (c) National Heart Foundation) and the include: organisations who assisted with the Full support of NUTTAB89 (c) costs of publication of these — Proceedings: CSR Refined Sugars and database and updates. the Australian Meat and Live-stock — An easy to use menu system Corporation. Finally we thank the many suitable for operators with students and staff of the University of minimal computer experience. New South Wales, who assisted with — Additional nutrient databases organisation on the two days of the which are user-definable. Symposium. — Project, menu and recipe References databases which are user- Food and Nutrition Research — Exclusive “Dynamic Graphing” Institute/National Research Council of and printout capability. the Philippines. 1985. Proceedings of A Hot-Line telephone help the first national workshop on food — Continuingf development and composition data generation, — support. compilation and use. Manila: FNRI. To arrange for a demonstration or Greenfield, H & Wills, RBH (eds). 1981. obtain further information please Tables of food composition: an contact: Australian perspective. Food Technol. Wailes and Phone: (03)4199077 Aust. 33: 101–31. Associates Gurr, MI. 1989. Has the food industry Unit 4, 19–35 Fax: (03)417 1227 benefited from the science of nutrition? Gertrude Street Proc. Nutr. Soc. 48: 159–63. Fitzroy, Vic. 3065 Morgan, KJ (ed). 1980. Proceedings of the fifth national nutrient data bank conference. East Lansing MI: Michigan State University.

Composition of foods, Australia

R. English

his paper continues the story about Professor RH Thorp (Chairman), Te stablishing a national data base on Australian Federation of Consumer the nutrient content of the Australian Organisations food supply. At an earlier meeting held Miss June Bullock, Royal Prince at the University of New South Wales Alfred Hospital (Greenfield & Wills 1981) I presented a Dr Basil Hetzel, CSIRO Division of paper Australian food tables — some Human Nutrition glimpses into the past (English 1981). At Mr Lesley Ion, Commonwealth that time the food analytical program for Department of Health the current revision had just Dr WG Murrell, CSIRO Division of commenced and the paper detailed the Food Research history of the development of food Mr John Neuhaus, Government composition data bases in this country. Analyst (NSW) The present paper tells the story behind Mr RC Norris, Australian Government the present revision program at least up Analytical Laboratory to the publication of the 1989 edition of Mrs Barbara Smith, SA College of Composition of foods, Australia (Cashel Advanced Education & others 1989). Professor Stewart Truswell, 1978 Working Party on Food University of Sydney Composition Data Dr Ron Wills, University of NSW Miss Margaret Corden (Secretary), The history of the revision program Commonwealth Department of commenced in 1978 when the Working Health Party on Food Composition Data of the The members made a number of National Health and Medical Research recommendations including: the range Council (NHMRC) was established and of foods should be extended from the held its first meeting on 22 August 1978. 1970 edition of the food tables and Its terms of reference were: To draw up cover ethnic, processed and fast or a plan for the central collection of data take-away foods; a loose-leaf manual on food composition in Australia; to form of publication was desirable in identify the nutrients and other food which data on nutrients/foods would be components (eg additives) for which grouped by page/section to facilitate data are to be collected and the form in updating; and, procedures to be taken to which they should be recorded and identify appropriate laboratories to distributed to the public; and, to identify participate in the analytical program. analytical and research facilities which could participate in a collaborative 1979 Working Party on Food survey of food composition by direct Composition Data analysis. The membership of that In 1979, the Nutrition Committee of Working Party comprised: NHMRC reviewed the report of the Working Party and agreed to Mrs Ruth English is Chief Nutritionist, recommend the establishment of a Commonwealth Department of Community steering committee to develop Services and Health, PO Box 9848, guidelines for the collection of food Canberra, ACT 2601. composition data; urged early appointments to the secretariat and and the development of the NUTTAB technical positions recommended by the computer data base. Working Party; and, requested that the Analytical program Council give consideration to the allocation of funds for laboratory This program can be divided into two analysis. phases, the first being funded by the The first milepost in the revision NHMRC and the second by the program was achieved in 1979. This Department. Table 1 lists the food was the establishment of a small analytical programs commissioned by Working Party to establish guidelines for the Council between 1980 and 1984, the collection of dietary data. The first together with the laboratories and chairman of the Working Party was investigators involved, and includes the Professor Stewart Truswell 1979–1982. additional grants made. Dr Alan Johnson was Chairman from Funding for the analytical program 1982 until his retirement in 1987. The was provided by the Council, on a yearly current membership of the Working basis only. This meant a major Party is: application had to be made to the Dr Ross Richards (Chairman), October Session of the Council each Materials Research Laboratory — year for funding for the following year — Tasmania, Defence Science and a mountain of paper work! Technology Organisation At the end of 1984, the NHMRC Mr Wally Hauser, Australian decided the food analytical program was Government Analytical Laboratory more appropriately funded by the Mr John Harris, Council of the Department than from the Medical Australian Food Technology Research Endowment Fund and funding Associations was not approved for a program in Mrs Ruth English, Commonwealth 1985. However, after very considerable Department of Community Services paperwork, the Department stepped into and Health the breach and since 1986 has provided Ms Karen Cashel (Secretary), a major funding source through a Commonwealth Department of program grant for food analyses to be Community Services and Health undertaken by the Australian The present amended terms of Government Analytical Laboratory. To reference for the Working Party are to date the Department has provided recommend to the Nutrition Committee $400,000 for this program with an on the program for the revision of allocation of $220,000 in 1989–90, a Tables of composition of Australian total of $620,000. Work undertaken foods, and to advise on the laboratory within the Department's program grant aspects of the program. includes analyses of fish and fish products, cereals and cereal products, Programs since 1979 milk and milk products, processed fruits The work program for the current and vegetables, legumes, and eggs. revision of the Australian food tables is The data for these foods will be included composed of four components: the in future editions of Composition of analytical program; the processing and foods, Australia. validation of food composition data and Planning and supervising the food preparation for publication; the analytical program are very demanding establishment of the Australian Nutrient of professional time as they involve Data Bank to store and process data; developing a national sampling plan for foods to be analysed, supervising the food analytical program, reviewing food designed and tested in late 1985, analytical methodology, completing proved to be user-unfriendly and the administrative paper work for funding for system was then redesigned with each year's program, requisitioning the modified objectives. The present system actual analytical work, and following up has been up and running for two years the work program and payment of now but with changes to mainframe accounts to acquit annual funds before computer systems in the Department, the end of the financial year. the ANDB now requires major Nutrients analysed in the program modifications to adapt to the new include: water, protein and amino systems and a case can always be acids*, starch, sugars, total fats, fatty made for extensions to the system. acids, cholesterol, sodium, potassium, The ANDB provides for the storage iron, calcium, zinc, magnesium, and retrieval of all information relating to phosphorus*, manganese*, copper*, food analytical data (Carmody 1987). retinol, carotenes, thiamin, riboflavin, Data are stored on the specifics of the niacin, vitamin C, vitamin B6*, vitamin food, its sampling, handling and B12*, biotin* and dietary fibre* (*denotes preparation, the laboratory, the analysed on a selected range of foods). analytical methods used and the use of The list of nutrients has been further the analysis in generating representative expanded to include a more complete data for the nutrient composition of a range of minerals and vitamins in food. The system provides several current and future analytical programs. advantages. The revised tables Validation and processing of data for Composition of foods, Australia are publication prepared direct from the computer. This allows for a variety of presentation Major tasks associated with this process formats and an efficient means of commence when the raw data from the changing or expanding these formats. laboratory are received in the Updating and expanding the range of Department and they are described in both food items and nutrients are limited detail by Cashel (1990). For this task, a only by the capacity of the computer, major administrative problem has been and specialised needs for information obtaining adequate staff resources to can be met. expedite the work. NUTTAB Australian Nutrient Data Bank Another project associated with the The establishment by the Department of revision program of the food tables and the Australian Nutrient Data Bank the establishment of the ANDB has (ANDB) has certainly expedited the been the release in late 1987 of a storage, access to, validation and computer data base NUTTAB providing processing of nutrient data both for nutrient composition data on a range of NUTTAB (Commonwealth Department items, covering all major food groups. of Community Services & Health 1987) This data base was released to facilitate and the publication of Composition of the use of Australian food composition foods, Australia. The development of the values and to encourage the use of a ANDB has covered a five year period common tool in dietary analysis in with significant input from two Australia. nutritionists, Karen Cashel and Janine The 1987 edition of NUTTAB was Lewis, as well as computer scientists based on the food composition tables from the Information Services area of developed to analyse the diets in the the Department. The first prototype, 1983 national dietary survey of adults (English & others 1987) and the 1985 (Commonwealth Department of national dietary survey of schoolchildren Community Services & Health 1989). (English & others 1989). This edition With updating, there is an increase in comprised 50 percent Australian data the proportion of date based on the with the balance being modified data current food analytical program. The from the British food tables (Paul & latest edition now contains some 67 Southgate 1978). percent of Australian data. There have been two updates of NUTTAB, in 1988 and 1989

Table 1. Commissioned food analytical program Year Title and investigators Funds 1980 Fruit juices and other beverages (SA Department of Services and Supply) $22,901 1981 Take-away foods (Professor Wills and Dr Greenfield, University of New South Wales) $32,692 Fruit juices and other beverages (Professor Truswell, University of Sydney) $26,710 Dietary fibre (Professor Walqvist and Dr Jones, Deakin University) $21,492 1982 Meat (Professor Truswell, University of Sydney) $35,349 Fruit and vegetables (Professor Wills and Dr Greenfield, University of New South Wales) $31,985 Dietary fibre (Professor Wahlqvist and Dr Jones, Deakin University) $22,000 Standard weight for measure (Mrs Smith, SA College of Advanced Education) $ 3,750 1983 Meat, fruit and vegetables (Professor Wills and Dr Greenfield, University of New South Wales) $83,000 Dietary fibre (Professor Wahlqvist and Dr Jones, Deakin University) $20,350 Standard weight for measure (Mrs Smith, SA College of Advanced Education) $ 4,960 Preparation of data for publication (Mrs English, Department of Health) $13,500 1984 Meat, fruit and vegetables (Professor Wills and Dr Greenfield, University of New South Wales) $96,000 Dietary fibre (Professor Wahlqvist and Dr Jones, Deakin University) $20,000 Standard weight for measure (Mrs Smith, SA College of Advanced Education) $ 6,160 Amino acids (Mr Stanhope, Victorian State Chemistry Lab) $ 7,500 Microbiological assays (Dr Davis, Royal Perth Hospital) $ 3,600 Preparation of data for publications (Mrs English, Department of Health) $10,000

Composition of foods, Australia program. It should be noted however, that to this time it is estimated that the The most visible evidence of the cost of analyses and the establishment progress made by the national program of the ANDB alone is approximately is the new publication Composition of $1,550,000. foods, Australia (Cashel & others 1989). This publication provides nutrient data Concluding remarks on the food programs commissioned Up to 1981 I had been an onlooker as and reported to mid-1987, namely, fruit, various editions and reprints of the vegetables, meats, poultry, offal, take- Tables of composition of Australian away and savoury snack products. foods (Thomas & Corden 1970) had The loose-leaf format provides a one been prepared in the Nutrition Section page per food overview of the prior to 1979 and had appreciated the proximate, vitamin and mineral content demanding and unenviable nature of the of the food both per 100 g edible portion task. For the 1989 edition of and per defined food measure. It also Composition of foods, Australia and provides a number of specialised NUTTAB in which I have been directly appendices eg on fatty acids, sugars, involved I would like to pay tribute to the amino acids, organic acids. officers who have worked in the It is proposed that the initial edition of Nutrition Section on the day to day detail Composition of foods, Australia will be of this enormous task, Karen Cashel supplemented by approximately annual and Janine Lewis with able clerical issues of further data. These will be support from Kerry Pearce. mainly additions of new foods. Currently Acknowledgments to the many data from the fish and cereal programs people associated with the program are being prepared for release in 1990. appear in the publication Composition of However, revisions of updates of food foods, Australia, notably the members of items included in the 1989 issue may be the Nutrition Committee and Working included as new formats or special Party on Food Composition Data, Dr appendices as requested by users. In Heather Greenfield, Professor Ron fact, consideration is already being Wills, Mrs Sue Cassidy and Mr Geoff given to the minimum information Hutchison. Lastly, I would like to needed to update the data on meats in acknowledge the contribution of the the near future. National Health and Medical Research With each new edition of Council and the Department and Composition of foods, Australia it is express my appreciation to these proposed to update NUTTAB so that organisations for having faith in and for within the next two or three editions, this supporting the program to revise the data base will comprise all Australian national reference resource on the data. Other proposed tasks include the nutrient composition of the Australian long overdue revision of the booklet of food supply. simplified food composition tables and References development of a reference publication on serving sizes of Australian foods. Carmody, J. 1987. Development of the Whether these proposals come to Australian Nutrient Data Bank — fruition will be very dependent on the computer aspects. English, RM & monetary and staff resources made Lester, I (eds). Proceedings of the First available to the Section to continue this OCEANIAFOODS Conference. Thomas, S & Corden, M. 1970. Tables Canberra: AGPS. 51–61. of composition of Australian foods. Canberra: AGPS. Cashel, K, English, R & Lewis, J. 1989. Composition of foods, Australia. Canberra: AGPS. SERVE Cashel, K. 1990. Compilation and Nutritional Analysis Software scrutiny of food composition data. Food SERVE Australia 42: 8, 521–24/42: 521–24. Provides detailed nutritional analyses of foods, meals, diets and Commonwealth Department of recipes, recommended daily Community Services & Health. 1987. intakes, patient records, foods NUTTAB87. Nutrient data table for use search by nutrient value and much in Australia. Disk and tape format. more … SERVE provides powerful Canberra: Commonwealth Department features and flexibility to meet your of Community Services & Health. specific requirements yet is easy to Commonwealth Department of use. Community Services & Health. 1989. SERVE NUTTAB89. Nutrient data table for use Is widely used throughout Australia in Australia. Disk format. Canberra: in hospitals, industry, universities, AGPS. health care centres and private English, R. 1981. Australian food tables practice. – some glimpses into the past. Food Telephone help and support are Technol. Aust. 30: 103–63. provided to all our users. Software and nutrient updates are regularly English, R, Cashel, K, Bennett, S, available. Berzins, J, Waters, A–M & Magnus, P. SERVE 1987. National dietary survey of adults: 1983. No. 2. Nutrient intakes. Canberra: Is the best dietary software at the AGPS. most reasonable price, just $499, including the latest Australian food English, R, Cashel, K, Lewis, J, Bennett, tables (NUTTAB). S, Berzins, J & Penm, R. 1989. National A sample diskette is available for dietary survey of schoolchildren (aged $20 (refundable on purchase). 10–15 years): 1985. No. 2. Nutrient Enclose a cheque and specify 5.25" intakes. Canberra: AGPS. or 3.5" diskette. Greenfield, H & Wills, RBH (eds). 1981. For more information send name Tables of food composition: an and address to: Australian perspective. Food Technol. M & H Williams Pty Ltd Aust. 33: 101–31. 33 Green Valley Avenue, St. Ives, Paul, AA & Southgate, DAT. 1978. The NSW 2075 composition of foods. London: HMSO. or phone (02) 449 2593

Food composition data and clinical dietetics

D. Foote

n important tool in the repertoire of foods consumed or may not describe A the clinical dietitian is familiarity with foods adequately. A weighed food diary the ingredients and nutrient composition should be done over some weekdays of a wide variety of commonly eaten and some weekend days to cover the foods. Formal use of food composition possible variations between these data is made in several ways in the periods. Nutrient intakes are calculated clinical dietetic setting. from the records of foods consumed Assessment of nutrient intakes over the period of the diary then averaged to give daily intakes. Nutrient intakes of an individual are Results from the calculations of most commonly assessed from a dietary nutrient intakes from either a dietary history or weighed food diary. Using the history or a weighed food diary would be dietary history method the dietitian compared to the recommended dietary questions the subject on his or her usual intakes (RDIs) for Australians in the eating patterns including types and appropriate sex and age group for the amounts of foods and liquids consumed, subject (National Health & Medical and the frequency of consumption of the Research Council 1986). foods eaten. A dietary history will Advice may be given to the subject include an indication of the variation in regarding desirable changes to the diet, intake from day to day and of the range based on results of the assessment. If of foods typically eaten at each meal or there appears to be a trend to low or snack time. The dietitian will make excessively high intake of any nutrients estimates of the amounts of food such advice may be given, however the consumed. The food composition tables dietitian needs to be very cautious in would then be used to calculate the making judgements based on such nutrients obtained from a typical day's calculations. intake. These results would only be a guide to the subject's usual intake, being Planning special diets based on so many estimates. Most special diets involve controlling the Using the weighed food diary method amounts of certain foods eaten the subject (or caregiver) weighs or according to the relevant nutrients for measures, using metric cups and the medical condition being treated. spoons, all food and liquids actually Reduced or increased intakes of consumed over a period of three or four nutrients may be needed. Food days. This gives much more accurate composition data allow dietitians to amounts than a dietary history, formulate exchange lists giving however, there may be some problems. allowable serving sizes of Subjects may alter their usual eating interchangeable foods, or allowed and patterns to make the recording easier, disallowed food lists for use in special or may omit to weigh or record some diets. These lists may be based on different nutrients: carbohydrate for Deborah Foote is Deputy Chief Dietitian, diabetic diets, protein for renal diets, Royal Alexandra Hospital for Children, sodium for renal or cardiac diets, Camperdown, NSW 2050. potassium or phosphate for renal diets, particular amino acids for metabolic The Composition of foods, Australia disorders, eg phenylketonuria (PKU) (Cashel & others 1989) has been and so on for a large number of released with data for only some of the conditions. food groups from new Australian Commercial food lists analyses. The new tables still include at the back many foods remaining from the In conjunction with diet sheets which list old tables (Thomas & Corden 1977) for basic foods that are allowed or which the nutrient composition data are disallowed on special diets, dietitians from foreign sources. The dietitian must will frequently supply commercial food remember that in different countries the lists. Both nutrient and ingredient same food name can be given to very composition data of manufactured foods different foods which would have widely are used to collate lists of foods by variable nutrient composition, or the brand name which can be included in same foods have different names. It is various special diets. Data are used important to keep in mind the variability either to calculate appropriate serving of many local foods when analysing an sizes of commercial foods to be used individual's dietary intake. Nowadays with exchange lists eg diabetic and PKU many dietitians use computerised diets, or to determine allowable foods on nutrient analysis of packages frequently diets eliminating particular ingredients, based on NUTTAB (Commonwealth eg gluten- and wheat-free, milk- or egg- Department of Community Services & free, sucrose-free diets. Health 1988). The information included Recipe analysis in this computer data base is not as extensive as in the printed tables and Another use of food composition data is therefore has more limitations in its use. in analysing recipes. These calculations From a paediatric dietitian's point of can be used either for food items which view the main difficulty is the lack of are to be included on a hospital menu or compositional data on many of the foods in cookbooks to be supplied to patients commonly consumed by children eg on various special diets. Nutrients in baby foods, both commercial and home each of the ingredients in a recipe are prepared, and infant formulas are still summed and the average content per not included. Some food groups still serving is calculated. have only a very limited range of Assessment of nutritional adequacy individual foods in the tables, eg of hospital menus breakfast cereals, biscuits, In many hospitals, the cycle menu is confectionery, ice creams, compared planned by the food service manager with the enormous range available who would frequently ask the dietitian to commercially. review the menu for its nutritional Abuses of food composition data adequacy. The dietitian would look for Use of food composition tables by compliance with the Australians dietary inexperienced unsuitably trained people guidelines (Commonwealth Department eg health workers with little or no of Health 1981) and may also calculate training in nutrition and food skills, leads possible nutrient intakes from various to taking imprecise diet histories or food combinations of food choices from the records and then attempting to calculate menu. It would be assumed that full detailed nutrient intakes from this meals would be chosen and consumed. sketchy information. Such people will Problems using the Composition of frequently be unable to make correct or foods, Australia best choice of the most similar food when a consumed food is not listed in patients' nutrient intakes adequately or the food tables. Their knowledge of food to give individualised dietary instructions and nutrients may not be sufficient to which patients can follow on a long term know how alike various foods are in basis. nutrient composition or which are the References most important nutrients to access in different situations. Cashel, K, English, R & Lewis, J. 1989. Any practitioner (adequately trained Composition of foods, Australia. or not) can be guilty of making Canberra: AGPS. unjustifiably strong judgements based Commonwealth Department of on calculations from a diet history or Community Services & Health. 1988. measured food intake over a small NUTTAB88. Nutrient data table for use number of days. Both methods of in Australia. Disk or tape format. determining food intake are notoriously Canberra: Commonwealth Department open to inaccuracies. Accurate of Community Services & Health. assessment of food intake over long periods is necessary before a really Commonwealth Department of Health. definite statement can be made that a 1981. Dietary guidelines for Australians. diet is inadequate. Nutrient deficiencies J. Food Nutr. 38: 111–9. can really only be diagnosed in the National Health & Medical Research presence of corroborating biochemical Council. 1986. Recommended dietary and clinical evidence. intakes for use in Australia. Canberra: Knowledge of the nutrient AGPS. composition of foods forms the backbone of clinical therapeutic Thomas, S & Corden, M. 1977. Metric dietetics. Without access to such data tables of composition of Australian dietitians would be unable to assess foods. Canberra: AGPS.

Food composition data in clinical research

J.R. Allen

n clinical research in nutrition, food necessary to know the intake of Ic omposition data are used to evaluate particular nutrients accurately. the adequacy of the nutrient intake of Food tables individuals or groups. Adequacy means the probability of individuals or groups Errors in the use of food tables can be having an adequate nutrient intake minimised by following a few simple (Beaton 1986). It is difficult to assess rules, eg read all the descriptive notes nutritional status from food intake data accompanying a set of food tables; note alone. Nutrient intake data are usually particular items which vary between correlated with biochemical and food tables, such as the factors which anthropometric measurements. The final have been used to calculate the step is to test for a relationship between metabolisable energy in the diet from the observed nutrient intake and the protein, fat and carbohydrate, the health or disease state of the study multiplication factor(s) for nitrogen to sample. Accuracy in the measurement calculate the protein content of different of food and hence nutrient intake is foods and so on; observe the method for essential to establish true relationships calculation of niacin and retinol (Pennington 1988). equivalents for these may also vary The accuracy of the nutrient intake between food data bases. The measurement is dependent on the convention for missing values also quality of food composition data (Guthrie needs to be known as a calculation of 1989). The quality of the food tables is nutrient intake which leaves missing determined by the number of food values as zero will be an under-estimate samples analysed, the analytical (Anderson 1986). The extent of this methods used and the quality control in underestimation has been calculated the laboratory. However, the variability using the British food tables (Paul & in the measurement of food intakes is Southgate 1978) for a group of pregnant greater than the variability in the food and lactating women (Black & others composition data (Paul & Southgate 1985). It is better to estimate a value 1988). from a related food than leave a zero. Sometimes there is a need for direct Dietary records food analysis, that is, to create the laboratory's own set of food tables. This There should be a check that the dietary situation occurs when there is no records to be analysed contain enough available nutrient information about a specific information to enable the correct particular food or when the nutrients food item to be identified in the tables. being investigated cannot be calculated When this specific information is missing accurately from food tables (vitamin C, it is necessary to make rules about the sodium, folates). In metabolic balance choice of a food. These rules need to be studies, direct food analysis is documented and used consistently in all data coding. For foods or nutrients not Jane Allen is a research dietitian, James listed, information may be obtained from Fairfax Institute of Paediatric Clinical other food tables, manufacturers, food Nutrition, Royal Alexandra Hospital for analysis, or substitution with a similar Children, Camperdown NSW 2050 food or a combination of foods to achieve a particular nutrient balance. comparison of the different food tables The methods of food analysis for the in use. A large unresolved problem in missing food need to be similar to those translating food composition data into used for the other foods in the diet actual nutrient intakes is the lack of otherwise an inaccurate calculation of allowance for nutrient bioavailability or nutrient intake will result. interaction between different foods. Recipes require accurate information More information is needed about the on the weights of all the uncooked proportion of the nutrients actually ingredients and the cooked weight of the absorbed from food as eaten. total recipe. The reduction in weight with References cooking is mostly water loss. If the total cooked weight of the recipe is missing, Anderson, SA (ed). 1986. Guidelines for the proportion of the whole recipe eaten use of dietary intake data. Bethesda can be used in the calculation. When MD: Federation of American Societies foods gain fat it is preferable to know the for Experimental Biology. fat content before and after cooking Beaton, GH. 1986. Toward (Paul & Southgate 1988). harmonization of dietary, biochemical, Manufacturers' information and clinical assessments. The meanings of nutritional status and requirements. Information from manufacturers needs Nutr. Rev. 44: 349–58. to be continually updated as food ingredients are frequently changed, Black, AF., Paul, AA & Hall, C. 1985. often for economic reasons. It is Footnotes to food tables. 2. The important to record the date of the underestimations of intakes of lesser B manufacturers' nutrient analysis. Usually vitamins by pregnant and lactating only a partial nutrient analysis is women as calculated using the fourth supplied so the missing values need to edition of McCance and Widdowson's be obtained from other sources, as The composition of foods. Hum. Nutr. outlined above. If a proportionate list of Appl. Nutr. 39A: 19–22. ingredients (including water) is given, Guthrie, HA. 1989. Interpretation of data the nutrient content can be calculated on dietary intake. Nutr. Rev. 47: 33–8. from a recipe. Paul, AA & Southgate, DAT. 1978. The Poor knowledge of nutrition composition of foods. London: HMSO. Another source of error which is difficult to avoid is the use of food tables by Paul, AA & Southgate, DAT. 1988. people without the appropriate Conversion into nutrients. Cameron, ME knowledge of food and nutrition to make & Van Staveren, WA (eds). Manual on correct decisions in nutrient analysis of methodology for food consumption food intake data. This problem seems to studies. Oxford: Oxford University have increased with the arrival of Press. 121–44. computerised food tables. Pennington, JAT. 1988. Associations For the future, food tables may need between diet and health: The use of to include non-nutritive components food consumption measurements, such as pesticide levels (Anderson nutrient databases, and dietary 1986). Other needs are ways to reduce guidelines. J. Am. Diet. Assoc. 88: time and cost in data processing, 1221–4. improvement of analytical procedures for certain nutrients (eg folates), and a The uses of food compositional data in nutritional epidemiology

K.I. Baghurst and P.A. Baghurst

rom the late 1940s until the early calculated values. Again, however, the F1 970s, a series of reviews and matching for iron was poor with only studies were published assessing 50% remaining in the same extreme values obtained by use of food tables tertiles and 20% being grossly compared to chemical analyses misclassified. Marr, in her review in (Bransby & others 1948, Groover & 1971, concluded that Bransby's data others 1967, Buzina & others 1966, and the other data available to that time, Pekkarinen 1967, Whiting & Leverton showed that, whilst absolute agreement 1960, Eagles & others 1966, Marr for every individual is not achieved with 1971). The views expressed in these the use of food tables, there is a high papers about the validity of food tables degree of correlation for most nutrients, for assessment of nutrient intake varied such that it is possible to compare widely but where tables were available groups of individuals in terms of their for locally produced food products, broad distributional properties and to correlations between calculated and rank them, two common needs in analysed data were generally epidemiologic or population-based considered to be good. For example, the research. early studies of Bransby and co-workers In recent years, nutritional in the 1940s, which compared the methodology issues have continued to results obtained for 33 adults using receive wide coverage both in the chemical analyses of three-day weighed epidemiological and nutritional literature. records with food table analyses, There have been extensive discussions produced correlation coefficients of 0.93 on techniques and the validity of various for energy, 0.91 for protein, 0.81 for fat, methods to assess usual or current 0.89 carbohydrate, 0.80 for calcium but intake and how best to express the only 0.49 for iron. There was some bias resulting data. However, despite the in the data, in that the calculated values large number of recent publications and for energy, carbohydrate and fat tended the many national and international to be higher and for protein, calcium and workshops and symposia on these iron, lower, than the analysed figures. topics in recent years (Baghurst & However, the differences were small, Baghurst 1981, Block 1982, Stuff & again with the exception of iron (2053 vs others 1983, Byers & others 1985), 2088 calories; 76 g vs 68 g protein; 81 g there has been relatively little discussion vs 84 g for fat; 255 g vs 265 g for of food composition data bases, their carbohydrate; 1.1 g vs 1.0 g for calcium use and abuse. and 19 g vs 11 g for iron). An analysis of The nature of epidemiological work the relative ranking of subjects using tertiles also showed that 70–80% of The lack of discussion about the validity subjects remained in the same extreme and use of food tables in the tertile of intake comparing analysed and epidemiological field, probably relates, in part, to the fact that no realistic Dr Katrine Baghurst and Dr Peter Baghurst alternative approach to the analysis of are Principal Research Scientists at CSIRO Division of Human Nutrition, PO Box 10041, population-based dietary intake data Gouger Street, Adelaide SA 5000. exists. Epidemiological work, by its nature, involves studies of large notes detailing the source of data, numbers of people. These studies can selection of items for analysis, the range be cross-sectional or prospective in of values found, the analysis techniques nature, in which case a measure of used and any assumptions made when current usual dietary intake is often calculating values for items or nutrients required, or they can be retrospective, in not directly measured. They also which case past usual intake is the item carefully annotate missing data for a of interest. The studies may involve a given food. This information is rarely case-control or cohort design but in both available in computer packages. Indeed cases, the main purpose is usually to in many computer packages missing compare relatively large groups of values are entered as zero rather than individuals. Therefore the need for using best estimate figures. In a major absolute accuracy is not as great as recent review of the use of dietary intake when trying to assess, for example, the data, Anderson (1986) estimated that carbohydrate intake of a single diabetic use of zero figures for missing data was patient in a clinical setting. A systematic the major potential error source in food error in the food data base, such as may tables. Unless the users of tables have occur when an analytical technique for a an original copy of the annotated data, given nutrient gives consistently high or and unless it is also possible for them to low values across all foods, will not amend errors in any computerised data markedly affect the final outcome of a base they may use, biases will inevitably case-control or cohort study where be introduced into their dietary intake comparison between groups is the analyses. Unfortunately, many users of issue, but may cause problems to the computerised food tables will not avail user wishing to assess absolute intakes themselves of the printed or original in an experimental or clinical setting. On computerised tables because of the the other hand, random errors or biases additional costs. in the food composition data base can The increasing use of food data cause problems even in comparative bases by non-specialists also reinforces studies. For example, if one particular the need to annotate tables and food or category of foods is eaten more programs carefully to prevent common often in one group than in the other, coding errors such as the use of codes errors in analyses for those foods will for dried items (coffee grounds, cocoa bias the comparison, making differences powder, gravy powder or custard between the groups bigger or smaller powder) when the appropriate code is than they are truly. for the rehydrated form. There is also a Computerisation of data bases need to avoid ambiguous labelling such as ‘beef, fat’, a code descriptor used in A major area of concern in relation to one widely used table to mean the fat the use of food tables in epidemiological from beef. In our experience, the code work is that they are often used by for ‘beef, fat’ has often been used people not trained in any way in their mistakenly when attempting to code use and, indeed, often with no nutritional fatty beef. This problem may be training at all. The increasing availability compounded in computerised programs of predesigned computer packages for if abbreviated food descriptors or analysing diets will only compound this abbreviated food tables are used. It problem by further distancing the user should be possible to overcome these from the producers of the tables. problems by careful naming of items or Food compositional data bases, in by the use of warning messages where their printed form, often have extensive known potential problems exist. In short, carotenoids (including not only those computer programs which include the which are precursors of vitamin A such use of food tables need to be more as β- and α-carotene and β- user-proof. cryptoxanthin, but also non-provitamin A Specific needs of epidemiologists carotenoids such as lutein and lycopene). In the cardiovascular field, The needs that nutritional the pressing need is for further data epidemiologists have in relation to food relating to fatty acid profiles of fats and compositional data bases are, in most oils from various sources, including local respects, similar to those of other users. varieties of fish, and to the various The data base would contain accurate dietary fibre components of foods (when and comprehensive nutrient and non- measurement issues relating to the nutrient analyses (nitrates, nitrites, latter have been resolved). oxalates, phytic acid, additives, Many nutritionists involved in contaminants etc) for all major food population-based research would prefer items eaten by Australians, including a to have accurate data for a wide range wide range of foods commonly of nutrients in key or basic foodstuffs consumed by ethnic communities and than to have limited nutrient information by the more traditional aboriginal on a wide range of foods, some of which groups. are very rarely eaten and only contribute The range of nutrients and non-nutritive marginally to overall nutrient intake in components of the food supply that are the population group of interest. In this of interest to epidemiologists, and regard, it is particularly important to especially to cancer and cardiovascular have these accurate analytical data, epidemiologists, is far in excess of those together with details about their currently available or indeed reliability and the range of values for immediately planned for the Australian specific nutrients in given foods, where data base. For example, despite the fact those foods are major contributors to that Australia has recently developed population intakes of the nutrient. For recommended levels of intake for example, it is more important for the nutrients such as vitamin E and assessment of population figures of β- selenium, we have little or no data on carotene intake to have exhaustive data the levels of these nutrients in Australian for β-carotene in carrots which are the foods. Indeed, there are few overseas major and dominant population source data on these nutrients or other non- of this nutrient. Small inaccuracies in nutritive components such as nitrites, tables due to analyses of samples nitrates and nitrosamines, which are of limited in terms of their source, cultivar, great interest in cancer epidemiology. stage of maturity etc will have a marked The relevance of the overseas data that effect on the accuracy of population do exist is difficult to assess as items intake figures for carotenes whereas such as selenium and nitrates and similar or even larger inaccuracies nitrites are highly dependent on the soils related to other minor sources such as in which foods are grown and on related nectarines would have a negligible farming practices. Others such as effect. The same concern relates to the vitamin E are also dependent on the representativeness of analytical data for commercial usage of various fats and dietary fats and oils in relation to fatty oils. Other nutrients which would be of acid profile and vitamin E for which they particular interest to cancer are dominant sources; to liver and epidemiologists for similar reasons, kidney for items such as retinol and include manganese and various vitamin B12; and to eggs for cholesterol. face in the coming years will be how to It is also essential to have accurate and cope with a data base in continuous detailed data for basic foodstuffs such evolution. This is a particular problem in as breads which can come in a wide epidemiological work where changes in variety of types and which contribute intake over time are being monitored or substantially to a wide range of where one wishes to compare the nutrients. results from various studies. It will Food tables are often used by clearly be necessary for all research epidemiologists and other research workers to describe in detail in their workers to aid in the design of publications, which of the various questionnaires relating to food intake. In versions of the Australian or other tables many instances, the researcher is they have used in their analyses, while seeking to determine which food to those undertaking longitudinal studies include in a questionnaire, or which to should use a consistent data base with take into consideration in computing the possibility of having to recompute dietary intakes, and would be aided in earlier data sets. this by the inclusion, in both printed and The rolling nature of data release in computerised data bases, of nutrient Australia may also cause problems with content per usual serve size as well as the use of computer packages for per 100 g. In addition, many dietary analysis. Due to their high cost, inexperienced users need an indication purchasers will be reluctant, or unable, of the relative importance of to repurchase such packages each time consumption of various foods when NUTTAB, the computerised form of the assessing nutrient intakes at the Australian food tables (Commonwealth population level. This would prevent the Department of Community Services & common occurrence of a key food or Health 1989), is updated especially if foods being omitted from questionnaires the need for reprogramming has on the basis of relatively low levels of increased the costs. Clearly a system nutrient per 100 g or serve, when in fact that can be updated readily and cost- the frequency of consumption of that effectively will be needed in the coming food is high and the usual serve size is years to avoid non-standardised usage. large. In short, it must be remembered Conclusion that the nutrient contribution that a food makes to a given diet, is a function not In conclusion, a reliable and just of its chemical composition, but also comprehensive food composition data serve size and frequency of base for Australian foods is essential for consumption. nutritional epidemiology research and An additional aid to the user would be for other population-based surveys. the inclusion of condensed lists of Although it is possible to assess nutrient nutrients as well as the more common status, in some instances, using page-by-page, screen-by-screen or line- biomedical indices or chemical analysis by-line food lists. For example, a few of duplicate diets, the size of the sample pages or screens listing fat levels in all populations and the wide range of foods, either per 100 g or per serve, nutrients usually of interest, prohibits would be a time-saving device for those this approach in all but a small number wishing to concentrate on fat intakes. of applications. The usefulness of the new Australian food tables (Cashel & Continuous updating of tables others 1989) could be enhanced One problem which epidemiologists, considerably if funds were available to and other users in Australia, will have to speed up the release of data for other major food groups and to include Cashel, K, English, R & Lewis, J. 1989. analyses, at least for key foods or food Composition of foods, Australia. groups, for components such as vitamin Canberra: AGPS. E, manganese, selenium and nitroso- Commonwealth Department of compounds for which there are no local Community Services & Health. 1989. data and few relevant overseas data NUTTAB89. Nutrient data table for use available. in Australia. Disk format. Canberra: AGPS. References Eagles, JA, Whiting, MG & Olsen, RE. 1966. Dietary appraisal — problems in Anderson, SA (ed). 1986. Guidelines for assessing dietary data. Am. J. Chin. use of dietary intake data. Bethesda Nutr. 19: 1–9. MD: Federation of American Societies for Experimental Biology. Groover, ME, Boone, L, Houk, PC & Wolf, S. 1967. Problems in the Baghurst, KI & Baghurst PA. 1981. The quantitation of dietary surveys. J. Am. measurement of usual dietary intake in Med. Assoc. 201: 8–10. individuals and groups. Trans. Menzies Found. 3: 139–60. Marr, J. 1971. Individual dietary surveys: purposes and methods. World Rev. Block, G. 1982. A review of validations Nutr. Diet. 13: 105–64. of dietary assessment methods. J. Am. Diet. Assoc. 87: 43–7. Pekkarinen, M. 1967. Weighing methods in dietary surveys. Voeding 25: Bransby, ER, Daubney, CG & King, J. 26–31. 1948. Comparison of nutrient values of individual diets found by calculation from Stuff, JE, Ganza, C, Smith, ED, Nichols, food tables and by chemical analysis. BJ & Montandon, CM. 1983. A Br. J. Nutr. 2: 232–6. comparison of dietary methods in nutritional studies. Am. J. Clin. Nutr. 37: Buzina, R, Keys, A, Bordarec, A & 300–5. Fidanza, F. 1966. Dietary surveys in rural Yugoslavia. III. Comparison of Whiting, MG & Leverton, RM. 1960. three methods. Voeding 27: 99–105. Reliability of dietary appraisal: comparisons between laboratory Byers, T, Marshall, J. Fiedler, R, analysis and calculation from food Zielezny, M & Graham, S. 1985. tables. Am. J. Publ. Health 50: 815–23. Assessing nutrient intake with an abbreviated dietary questionnaire. Am. J. Epid. 122: 51–4.

Food composition data and the National Heart Foundation's Food Approval Program

B. Haddy

he National Heart Foundation (NHF) from various quarters. The initiative is Ti s a nongovernment, non-profit the Food Approval Program which offers community health organisation formed food manufacturers and producers the in 1959. For 30 years the Foundation opportunity to use a special National has been encouraging people to stop Heart Foundation Approved logo on smoking, exercise regularly, maintain their products. normal weight and eat sensibly. Some When the Foundation launched this of these messages seem to have Program earlier this year, it generated worked as there has been a steady fall lively debate within the food industry in premature heart disease deaths in and amongst health professionals. This Australia since 1967 resulting in a review of the Program discusses the saving of some 103,000 lives in the 20- various benefits which it can offer the year period to 1987. However, much Australian community and offers some remains to be done as cardiovascular responses to several common criticisms disease still accounts for just under 50% of the Program. of all deaths and Australia still ranks First of all the Program has been high in international comparisons for based, to some extent, on a Heart Food coronary heart disease (NHF 1987a). Festival which was conducted by the Until recently probably the most Foundation in April 1987 and which controversial aspect of the Heart offered food manufacturers the Foundation's activities was the anti- opportunity to use a special logo on food smoking area. Many individuals and products (NHF 1987b). Although only a organisations became more hostile short term campaign, a subsequent towards the Foundation as its anti- evaluation showed that over 40% of the smoking messages, when combined population were aware of the project. with the efforts of other similar Positive feedback was also received organisations, became more and more from a broad cross-section of the food successful. This threatened the industry (NHF 1987c). It was realised profitability of the tobacco industry and that an on-going food approval program tobacco retailers, restaurateurs claimed had a real chance for success in terms that they would lose revenue if a of both consumer and industry smoking ban was implemented and acceptance. Therefore the Foundation's sporting groups believed their traditional Diet and Heart Disease Committee funding sources were under threat. commenced work on a set of The NHF Food Approval Program acceptability criteria for approved foods, a laboratory was appointed to test all Given this background it was expected foods prior to approval and a licence fee that when the Foundation undertook an structure was established to provide on- initiative to change behaviour in the going funds for the promotion of the dietary area it would meet opposition Program and related education activities. A simple, distinct logo Brian Haddy is Food Program Manager, featuring a stylised tick was also National Heart Foundation, PO Box 2, devised. Woden ACT 2606 to the food industry has already had a considerable effect in many companies. Acceptability criteria For the first time, real guidelines are The acceptability criteria for various now available for companies wishing to foods rely heavily on nutrient develop and promote low saturated fat composition data. The earlier and low salt products and the approval experience with the Heart Food Festival logo offers a tangible marketing had originally forced the Foundation to incentive. look carefully at individual food Consumers have also responded composition data. For the Food very positively to the Program as it Approval Program the NHF Diet and provides a quick, easy guide to the type Heart Disease Committee carefully of foods the NHF recommends as part studied each type of food in turn to of healthy eating. By influencing some of arrive at levels it considered appropriate the food choices people make, the for total fat, saturated fat and sodium, in Foundation hopes to improve the particular, as well as for dietary fibre, national diet and lower the incidence of cholesterol and added sugar levels for diet-related health problems. some food categories. In many cases Food choices are also being affected the acceptability criteria developed have through the Foundation's input into the been aligned with definitions in the advertising of approved products. All National Health and Medical Research companies participating in the Program Council's Food Standards Code (1987) must sign a legal agreement with the for low or reduced levels of fat and salt. NHF which stipulates prior approval for In other cases, usual serving sizes, the all advertising copy featuring the broader nutritional value of the food approval logo. Through this mechanism, concerned and the availability of lower the NHF tries to help ensure food fat or lower salt substitutes in the advertising is factually correct and marketplace all contributed to the responsible. establishment of the criteria (NHF A final plus for the Program is that it 1989). is generating a significant food analysis data bank which NHF hopes to publish Food analysis at some future time. Assuming the co- The Foundation appointed the South operation of food companies concerned, Australian Regional Laboratory of the the NHF expects to be able to publish Australian Government Analytical brand-specific data which would Laboratories (AGAL) some nine months undoubtedly be of great help to many ago to undertake all its analysis work to dietitians. test the foods prior to approval. Six Criticisms of the Program different batch samples must be submitted for any food for which a One of the common criticisms of the manufacturer is seeking approval and Program by some sections of the food AGAL uses widely-accepted analytical industry is that the NHF is artificially techniques for determining moisture, dividing foods into “good” and “bad”. In total fat, fatty acid profiles, sodium, total response, the Foundation has never sugar, sugar profiles, cholesterol and claimed that consumers should only eat dietary fibre. NHF Approved foods and has widely acknowledged that many foods not in Program benefits the Program can be part of a healthy The effect of promoting the Food diet. The Program is intended to simply Approval Program acceptability criteria be a general guide to low saturated fat, low salt foods and our accompanying maximise its potential for improving the education resources highlight this point. nation's health status. Some nutritionists have criticised the References Program because of its perceived focus on individual foods rather than the total National Health & Medical Research diet. The Foundation is very mindful of Council. 1987. Food Standards Code. this concern and is therefore continuing Canberra: AGPS. to put the Program into the context of National Heart Foundation. 1987a. overall healthy eating. Program posters, Heart Facts Report 1987. Canberra: literature and publicity reinforce this NHF. point further. National Heart Foundation. 1987b. Conclusion Heart Food Festival 1987. Report No. 1. The NHF believes the Food Approval Canberra: NHF. Program will become one of the most National Heart Foundation. 1987c. Heart successful nutrition education initiatives Food Festival 1987. Report No. 2. ever undertaken in Australia. However, Canberra: NHF. this success will not occur overnight and requires input form both the food National Heart Foundation. 1989. industry and health professionals. This Guidelines for acceptability. Canberra: will allow finetuning of the Program to NHF.

Sampling of foods for nutrient composition studies

J.H. Cunningham

ampling of foods for analysis, and habits. In general, food samples S their subsequent preparation prior to obtained from a single source or analysis, are critical steps in the produced under special conditions are production of good quality and useful not suitable for generation of data for food composition data. If sampling and food composition tables as they are sample preparation are done incorrectly rarely representative of the foods then all subsequent analyses are a available to the majority of the public waste of time and money, as a mistake (Greenfield & Southgate 1985). Before in sampling can only be corrected by any samples of food are purchased for a repurchasing and repreparing the food composition program, a sampling sample. This review of the process of plan should be developed and assessed sampling and sample preparation for critically to ensure that the foods nutrient composition studies is written purchased will be representative of from a practical viewpoint pointing out those available. The analyst and important considerations and likely program supervisor should be closely pitfalls. Sampling considerations are involved in the development of the relevant not only to those who will sampling plan. In Australian studies the actually be involved in the generation of usual practice for obtaining data or the preparation of tables; users representative samples has been to of food composition tables should also sample foods available in the major be aware of the factors that influence urban areas due to constraints of cost sampling and hence the analytical data and distance. In contrast, however, the ultimately reported. bush foods program at the University of In the present context, sampling is Sydney analysed samples primarily the selection and collection of items of collected in the Northern Territory by food defined in number, size and nature Aborigines, Aboriginal health workers to represent the food under and dietitians (Brand & others 1983). consideration; and the consolidation and Type and source of food to be reduction of the collected items to form analysed the portion for analysis (Horwitz & others 1978). Within the general category of foods to The aim of sampling for food be analysed a survey should be carried composition purposes is to obtain an out of the range of products available analytical sample that is representative and their sources. The same general of the foods available to, or consumed type of food may be available from a by, the population concerned range of outlets (restaurants, (Greenfield & Southgate 1985). This aim supermarkets, delicatessens, take-away implies that the participants in the food shops, fast food chains, health food program have a detailed knowledge of shops), in a range of forms (fresh, the population in question and its food frozen, canned, dried) and from a number of manufacturing sources Dr Judy Cunningham is an officer of the (factory-processed, home-made etc). Commonwealth Department of Industry, For example, in a 1981 study of Greek Technology & Commerce, 51 Allara Street, Canberra ACT 2601. foods samples were purchased from delicatessens, restaurants and takeaway food bars (Greenfield & others purchasing, to reduce bias in samples 1983). and the subsequent analytical results. If a food is factory-produced it is likely Type of representative sample to be relatively consistent in nutrient composition regardless of where it is Ideally if one were to purchase ten units purchased. Fresh produce, restaurant of a food sample (eg Valencia oranges) foods or home-made foods are much then each unit would be analysed more likely to show wide variations in separately and the final result calculated composition according to their source as the mean of these units. Such and the chosen sampling plan must samples are referred to as single reflect this variation. The more variable replicate samples. As this is obviously in nutrient composition a food is likely to an expensive approach it is infrequently be, the greater the number of samples used and an approach described as that should be purchased to achieve a single composite sampling is more representative result. For example, in common. With single composite the University of NSW nutrient analysis samples, the ten units of Valencia programs the approach used has been oranges would be mixed together to to purchase four units of standardised produce a composite sample for food, such as a manufactured breakfast analysis. A third class of sample is the cereal, and ten units of non- multiple composite sample where foods standardised foods, such as meat pies from different regions, seasons, cultivars produced in different small cake shops or brands are combined together to form (Wills & others 1988). one composite sample. Continuing the One important variable that may analogy of oranges, Washington Navel affect non-factory produced foods is the and Leng Navel oranges would be socio-economic status of the area in combined with Valencia oranges to form which a sample was purchased. In a a composite sample called ‘oranges’. large program of analysis of the The multiple composite approach is composition of Australian fresh meats, a most suitable for foods that are not sampling plan was devised, based on staples of the diet, such as specialty fruit published demographic and social or frozen pizza (Greenfield & Southgate, studies, dividing Sydney into ten ranked 1985). If the food in question is a socio-economic areas and then significant component of the diet selecting one suburb out of each of separate samples are usually drawn for these ten areas from which to purchase different cultivars, different growing meats (Greenfield & others 1987). Some areas, different seasons and different foods of the type being examined may processing and preparation methods. only be available in certain suburbs or Trial runs towns; for example some fresh delicatessen-style products may be A trial run of the entire procedure should readily available in a yuppie suburb but always be done to identify and resolve may be unheard of in a middle class problems prior to the actual purchase of mortgage belt suburb. Similarly foods samples for analysis. Typical problems consumed by particular ethnic groups and questions include: what type of may only be readily available in certain transport is needed? (is a small car cities or certain welldefined suburbs. satisfactory or will a delivery-size van be Suppliers of foods for analysis should required?); what type of containers are not be made aware of the purpose for needed for sample storage during transportation? (perishable products need to be placed directly into insulated containers); how long will it take to visit All relevant details about the sample every outlet required? (in large cities it should be entered immediately into a can take one or more working days just bound register. Loose pages are not to collect the samples); if samples are acceptable as they can become lost. being flown in from overseas, customs Registration details can be transferred and quarantine clearances will be at a later time to a computer system but necessary; is there enough storage it is often impractical to site a computer space in the refrigerator and freezer at close to a messy food preparation area. the laboratory to store all the samples Details that should be recorded about before and after preparation?; how long the sample include: common names(s), does it take to prepare one laboratory sometimes the same food will have sample?; what equipment is needed to different common names in different prepare and homogenise the sample?; areas or among different ethnic groups is there sufficient time remaining on the and as many common names as day of preparation to perform the possible should be reported; systematic immediate analyses of moisture and name for plant foods or seafoods (in vitamin C? (reagents and equipment particular); serving or purchase size in should be prepared in advance for these grams or millilitres; edible portion weight analyses); is there a washing-up area together with a description of the readily available?; if samples are to be material constituting both the edible and cooked are there hot plates, ovens and inedible portions; description of the state utensils available? of the food as purchased (whether raw, Sample preparation cooked, frozen etc.); description of the components of the foods (eg the Sample preparation often generates proportion of pastry in a pie) with many problems. When dealing with photographs if possible; date marking or perishable foods it is preferable to product code if available; place and date purchase only a few samples at a time of purchase; name and address of so that the preparation and immediate manufacturer, if available; label claims analyses can be completed rapidly and and ingredient lists, if provided; details without mistakes. A range of standard of the cooking method, if the sample is kitchen equipment is needed to prepare subsequently cooked (Greenfield & samples prior to analysis. This Southgate 1985). equipment includes cutting boards, knives, spoons, spatulas, strainers, Cooking of samples bowls, blenders or homogenisers, Care must be taken in the selection of plastic storage bags, aluminium foil and cooking method when samples are to be indelible marking pens. Plastic utensils cooked prior to analysis. Popular are desirable where possible to reduce cookbooks should be consulted to the risk of metal contamination. Ideally, obtain recipes. In the case of foods from a food composition laboratory will have another culture, advice may have to be a dedicated set of utensils as well as a sought from members of this culture. dedicated sample preparation area. The For packaged foods it is acceptable to sample preparation area should be free follow the manufacturer's instructions. If from environmental contamination such oil, fat or water are added during the as dust and insects, particularly when preparation this should be recorded and, trace metals are to be analysed. in the case of fats, the type of fat noted. Registering the sample Other details such as the amount of sample cooked, time and temperature of cooking and the type of cooking method and utensils must be recorded. Samples sample exposure to light, air and heat should be weighed before and after as these will lead to destruction of cooking (Greenfield & Southgate 1985). vitamins. At least 500 g to 1000 g of Homogenising the sample sample should be obtained if possible. This should provide sufficient material If reproducible, reliable results are to be for the inevitable repeat analyses. obtained the laboratory sample must be reduced to a homogeneous mass. A Storing the sample poorly homogenised sample will Samples should be stored in such a invariably give nonreproducible results. manner as to minimise deterioration or A study by Wills & others (1980) found change in nutrient composition. Usually that domestic food processors were as samples are frozen although some effective as special laboratory researchers have used freeze-drying homogenisers in obtaining (Brand & others 1983). Frozen samples homogeneous samples for a range of must be stored in containers that are foods. Many foods are extremely difficult impervious to water. Strong polythene to homogenise and food processors bags are satisfactory and cost-effective; may not be entirely satisfactory. In this the sample should be placed within case additional equipment and trials are three bags to minimise the risk of required. For hard, dry foods such as leakage, each bag being carefully grains and legumes it is generally sealed (eg with a heat sealer) and necessary to use a laboratory mill or a clearly labelled with the product identity. coffee grinder. However the extensive Plastic bags may not be suitable for high grinding required can result in metal water content foods such as fruits, as contamination of the food leading to the sample may swell when frozen and high iron levels. A solution to the burst the bag. Rigid plastic dishes with problem of metal contamination may be strong seals may be used as an to use samples coarsely ground, by a alternative. Headspace within solid mortar and pestle for mineral analysis, containers should be minimised to and a finely ground sample for the reduce the likelihood of sample remaining analyses. Although sieving oxidation. Glass containers should be will remove large pieces of food some avoided where possible due to the risk component of the food, such as the of breakage. Code names on sample bran, may be selectively removed also. labels should be avoided unless an Extensive blending or grinding may efficient and permanent record for produce sufficient heat to damage some deciphering the code is available. The vitamins. Sticky or sugary foods may homogenised samples should be require coarse grating and mixing rather divided into at least two separate sub- than homogenisation. This is, however, samples which should, preferably, be likely to result in a greater degree of stored in different freezers to avoid the variation between analytical duplicates loss of all sample material in the event in subsequent analyses. Excessive of a thaw-out. During frozen storage blending of high oil content foods such some of the water or fat in samples may as nuts should be avoided as the oil separate out; in this case it is necessary may separate. Similarly some fruits such to reblend the sample prior to taking an as bananas or papaya should be aliquot for analysis. mashed rather than blended at high Conclusion speed. During the process of sample During the process of sampling and preparation it is important to minimise sample preparation compromises will have to be reached. Indeed the entire Greenfield, H, Kuo, YL, Hutchison, GI & process of sampling can be regarded as Wills, RBH. 1987. Composition of a compromise because it is clearly Australian foods. 33. Lamb. Food impossible to sample every piece of Technol. Aust. 39: 202–7. food throughout the country. Provided Greenfield, H & Southgate, DAT. 1985. the sampling process is approached A pragmatic approach to the production with knowledge, a clear mind and of good quality food composition data. careful planning this most vital of steps ASEAN Food J. 1: 47–54. in a food composition program will, however, be successfully negotiated. Horwitz, W, Cohen, S, Hankin, L, Krett, J, Perrin, CH & Thornburg, W. 1978.

Analytical food chemistry. Inhorn, SL (ed). Quality assurance practices for health laboratories. Washington DC: References American Public Health Association. Brand, JC, Rae, C, McDonnell, J, Lee, 545–646. A, Cherikoff, V & Truswell, AS. 1983. Wills, RBH, Balmer, N & Greenfield, H. The nutritional composition of Australian 1980. Composition of Australian foods. Aboriginal bushfoods. I. Food Technol. 2. Methods of analysis. Food Technol. Aust. 35: 293–8. Aust. 32: 198–204. Greenfield, H, Lerogiannis, V, Makinson, Wills, RBH & Greenfield, H. 1988. J & Wills, RBH. 1983. Composition of Laboratory instruction manual for food Australian foods. 19. Greek foods. Food composition studies. Kensington NSW: Technol. Aust. 35: 84–6. University of NSW, Dept. of Food Science & Technology.

Data production and quality assurance in a nutrient analysis program

P. Scheelings and D. Buick

nutritional analysis program extends The provision of reliable nutrient data A beyond the collection of a trolley of requires the analytical laboratory to foods at the corner supermarket and understand the objectives and analysing the products for a variety of limitations of the food composition nutritional components. Such a program program, the end use and application of requires an understanding of the size the data, and the knowledge and and complexity of the project and a qualifications of the commissioning careful assessment of the resources, client. Likewise the client should have skills, courage and discipline required to some appreciation of the complexities of progress the task to completion the analytical task, be able to assess the (Scheelings 1987). It is as much about proficiency of the contract laboratory, program planning as it is about analysis have some knowledge of the natural and more about data evaluation than variability of nutrients in foods, reporting. understand the statistical basis for This review covers the process of data sampling procedures and preferably generation and identifies the critical have some knowledge of the principles elements of quality assurance for a of analytical quality control and nutrient analysis program. It is based on assurance in chemical measurements. four years involvement in food Sampling compositional analysis for the Australian food tables program, as well as for Much has been written on the nutritional labelling of packaged foods importance of representative sampling and nutrient stability studies in and the value of considered sampling processed foods, leading to plans (Taylor 1987, Garfield 1989). considerable experience in the Representative sampling for nutritional associated activities of sampling, analysis for food composition tables transport and storage, analytical poses particular difficulties. strategies, methodology and data A fundamental difference between reliability. regulatory or quality control analysis and food composition analysis is the size of Data production the sample lot or population. In Data production can be defined simply regulatory analysis, the size of the as the generation of analytical data by sample population is well defined. The machine, analyst or organisation. The sampling plan merely needs to provide usefulness of the data, however, is a for a representative sample of the function of its reliability and requires population or consignment, taking into correct program design, planning and consideration the size or volume of the execution, and thorough data lot and the physical state and evaluation. homogeneity of the product. In on-going Dr Pieter Scheelings and Don Buick are, programs such as the National Heart respectively, the Regional Director and Foundation Food Approval Program Senior Chemist, Australian Government (Haddy 1990), the sample population is Analytical Laboratories, 338–340 Tapleys less defined in terms of size, and Hill Road, Seaton SA 5023. analysis is performed on a representative test sample which the specified sampling plans; the describes as best as possible the samples are delivered to the laboratory product at the time of sale. In the Food and checked for integrity, Approval Program representative documentation and transport duration; sampling involves the submission to the sample information and analytical laboratory of at least six individual requirements are detailed in a sample samples taken from different production register and a laboratory worksheet is batches. The laboratory sample is produced and sent with the samples to prepared by compositing and the analyst or supervisor. homogenising the samples submitted Analyses and test samples taken from the homogenate using appropriate sampling The priority of the task is assessed procedures. The analytical data so taking into consideration client produced are normally valid for an requirements, sample stability and agreed period of time (eg 12 months) laboratory workload. In nutritional after which period the product is analysis, care must be taken to ensure resampled and submitted for re- that sufficient time is available to analysis. complete analyses for vitamins with Due to the high cost of analyses and limited stability. Highest priority must be the extensive range of nutrients given to moisture and the more labile normally involved, data produced for vitamins (eg vitamin C, carotenes) which food composition tables, however, are often analysed in individual demand greater durability. In these samples. Then client samples are analytical programs the size of the composited, where necessary, and sample population is even less definitive homogenised to produce a laboratory and the samples taken for analysis can sample. Test portions are taken from the only represent a snapshot of a changing laboratory sample for analysis. An equal population. It is important that this number of laboratory samples are limitation is both recognised and frozen and used as duplicates or retest accommodated in any sampling plan. samples in case of dispute or unusual results. The appropriate methods are Critical steps in data production selected which provide reliable data for The normal sequence of activities in the the particular food matrix. Universal production of data by the laboratory is methods which are applicable to a wide described below. range of food commodities are Commissioning preferred. In addition, alternative methods such as bio-assays may be The client submits a request for an employed as check methods for some analytical service, the laboratory advises vitamins. When analyses are completed, on its capacity to undertake the work and acceptable recoveries and and provides a quotation and estimated duplicates are obtained (ie evidence that turn-around time; then a contract is the analytical system is in statistical signed if parties are agreeable and the control), the data are reported on the laboratory is advised on sampling worksheet and submitted to the protocol, documentation requirements supervisor. and transport. Data checking and reporting Sampling The supervisor confirms the validity of Samples are taken by the client or the the results by checking calculations, contract laboratory in accordance with examining chromatograms and other raw data. When all the results are occupational safety and health, a available, the data are again scrutinised professionally stimulating work and checked against expected or environment, and, sound technical and reported values. Outlying data may be administrative management. confirmed by re-analysis. Final data are Quality control entered on the computer, and a report of analysis prepared for the client. A report Quality control can be defined as the should provide sampling details where overall system of activities whose appropriate, analytical results, precision purpose is to provide a quality service or data and methodology and be signed by product which meets the need of the the supervising analyst. If required, user. Quality control in an analytical laboratory staff should be available to laboratory covers personnel, equipment interpret the data or discuss aspects of and methodology and promotes the methodology employed with the procedures such as replicate analyses client. and use of spiked controls and split samples. Quality assurance in nutritional analysis programs Quality assurance In the context of analytical services, Quality assurance describes a system of data production is concerned with the activities whose purpose is to provide output of quality data, ie data which are assurance that the quality control tasks reliable, timely and cost-competitive. are being performed effectively. This is Most clients have a clear understanding normally achieved through regular of time and an even clearer one of cost quality audits, use of blind samples, but generally lack an appreciation of the participation in proficiency studies and statistical concepts of data quality. external laboratory assessments. The Analysts and service laboratories are effectiveness of the quality assurance able to provide a quantitative measure process determines the reliability of the of data quality by reporting the data produced by the laboratory. The uncertainty limits of the measurements term reliability means the data produced involved and describing the quality by a laboratory are accurate (close to assurance parameters employed. This the correct value) and precise (able to is formalised through the existence and be reproduced). use of a quality assurance plan which is Selection of analytical methods an integral part of any Good Laboratory Practice (GLP) protocol. The important In selecting analytical methods the elements of a GLP protocol include the following parameters need to be given employment of competent and qualified due consideration (Horwitz & others personnel, a modern wellequipped 1978): specificity, sensitivity, laboratory environment, the use of detectability, accuracy and precision. appropriate and validated methods, the The specificity of a method defines the use of experienced analysts, proven level of matrix interference with the proficiency in the analytical methods, analyte, whilst method detectability routine use of appropriate analytical determines the limit at which an analyte standards and reference materials, can be detected with confidence and comprehensive sample and data sensitivity is the change in response per recording and reporting procedures, unit of analyte concentration. Other protocol for sample chain of custody, criteria which influence the usefulness of attention to customer service, attention a method include ease of operation, to staff training and development and large dynamic range, speed and cost, portability and ruggedness.

FOOD ANALYTICAL SERVICES • Nutritional Profiles • Export/Import Certification • NHF Heart Food Approval Program

AGAL's National network of Laboratories has:

• NATA registration • In house AQA and R & D facilities • A reputation for quality service

• Proximates • Dietary Fibre • Vitamins • Pesticides • Sugars • Cholesterol • Minerals • Aflatoxins • Starch • Fatty Acids • Preservatives • Microbiology For further information and quotations on our extensive services in Food Analysis, please contact Pieter Scheelings or Don Buick on (08) 356 3399 or facsimile (08) 353 4015 Address: 338–340 Tapleys Hill Road SEATON S.A. 5023

In addition to using official or Statistical control validated methods with known accuracy and precision, the analyst must employ Overall the analyst must ensure that the a number of quality control actions to methods are always in statistical control. optimise data quality. These include the Until a measurement operation has use of accurate and clean glassware, attained a state of statistical control, it calibrated equipment, pure or standard cannot be regarded in any logical sense chemicals and reagents, reagent blanks, of measuring anything at all (Taylor recoveries, duplicate (or replicate) 1987). analyses and control samples and In effect this means that in the routine appropriate reference samples. application of the method, the accuracy and the precision of the measurements Certified biological reference as described for the method should be materials. maintained at all times. The accuracy of Certified biological reference materials the method is established by (RMs) are specially prepared well- determining the difference between the characterised homogenous stable food true or known value of the analyte and (or other biological) materials containing the measured value. The precision of a analytes at certified levels measured by method is measured by the standard highly accurate methods (Wolf 1985). deviation of a set of results from the RMs are used to validate measurement mean derived from a number of systems and analytical results. Most repeated applications of the method. RMs in common use in food A convenient laboratory tool for composition work are for inorganic assessing the performance of the nutrients and the lack of RMs for organic measuring process (method and nutrients has been recognised as a analyst) is a control chart on which severe problem for nutrient analysis. values of control or reference samples analysed over a period of time are recorded (Taylor 1987). By maintaining reference food materials and a shortage the control chart in real time, the analyst of suitable collaborative laboratories in is able to detect changes in the Australia. Few (if any) methods performance of the method at an early employed in nutritional analysis are stage. Significant changes in the absolute but generally rely on a precision or bias of the process are comparative measurement between a obvious from eyeballing the chart. standard (or reference) substance and Warning limits and control limits based the analyte. The unavailability of similar on previously determined precision data reference substrates increases are included on the chart to facilitate uncertainties about the accuracy of the corrective action. Control charts are not measurement. Likewise, whilst within- limited to measurements on control and laboratory or operator precision can be reference samples, but can also be readily determined, the shortage of used to monitor the range between collaborating laboratories reduces duplicate measurements. This is opportunities for comparison of particularly useful when reference between-laboratory precision. samples or suitable controls are Conclusion unavailable. A range control chart however is limited in that it can only Because of the diversity of food indicate changes in analyst precision as substrates, the extensive range of any change in method bias will affect analytes involved and the differing each duplicate result. The control chart stabilities of analytes, in particular illustrates whether the measurements vitamins, the application of standard process is in statistical control. As a analytical quality assurance procedures consequence the analyst can assume in nutritional analysis cannot be fully confidence in data derived from single optimised. There is therefore a greater determinations. Having established that reliance on the experience and intuition the analytical process is in control, the of the analysts, a professional laboratory analyst derives more useful data from a environment and organisation, effective set number of single determinations program management and control, due than half the set number of duplicate attention to staff training and continuity tests, the analytical effort remaining the of expertise and an appreciation of and same. However, this situation is mainly commitment to the principles of GLP. In applicable when large numbers of addition there is a greater requirement similar samples are being routinely for analysts to check and evaluate tested. analytical results against published values and to support data with details Application of quality assurance of sampling, methodology, method concepts to nutritional analysis precision and quality assurance. Food Whilst commitment and adherence to composition analysis for national food GLP procedures provide the first and composition tables is and will continue most important level of quality to be a professionally demanding task. assurance, the complexity of nutritional References analysis due to the diverse range of foods and the wide range of analytes Garfield, FM. 1989. Sampling in the adds another dimension of difficulty. The analytical scheme. J. Assoc. Off. Anal. limiting factors in assessing a Chem. 27: 405–11. laboratory's performance in Australian Haddy, B. 1990. Food composition data nutritional analysis programs include the and the National Heart Foundation's general unavailability of certified Food Approval Program. Food Aust. Lester, I (eds). Proceedings of the first 42:8, s 14–15. OCEANIAFOODS conference. Canberra: AGPS. 39–44. Horwitz, W, Cohen, S, Hankin, L, Krett, J, Perrin, CH & Thornburg, W. 1978. Taylor, JK. 1987. Quality assurance of Analytical food chemistry. Inhorn, SL chemical measurements. Chelsea MI: (ed). Quality assurance practices for Lewis Publishers Inc. health laboratories. Washington, DC: Wolf, WR (ed). 1985. Biological American Public Health Association. reference materials: availability, uses 545–646. and need for validation of nutrient Scheelings, P. 1987. Food composition measurement. New York: John Wiley analysis—getting started. English, R & and Sons.

Compilation and scrutiny of food composition data

K. Cashel

he compilation of food composition At the time the Australian food Td ata into tables or a data base is a composition program began there were process that begins with defining the few published data on Australian foods aims, objectives and uses of the food available in the literature. The Australian composition data base. The result is a program was therefore built around a specialised, documented data base commissioned analytical program aimed tailored to meet these defined at producing, within practical and specifications. Compilation is a complex financial constraints, the most task involving the collection of data; the representative data possible. This commissioning of food analytical required particular attention to sampling programs; the documentation and procedures. storage of data for each food from each Sampling procedures for source; the search for information representative data necessary to compile these individual data appropriately and meaningfully into Since the sampling procedure aimed to representative food items; the assembly ensure that the food sample analysed of these into a data base to produce was reasonably representative of the derived values, documentation and food item available to the consumer for description of items for the user; and the consumption, consideration was given preparation of derived data bases for to geographical density of the specific user groups. At every step in population, where food is produced and this process the data are scrutinised for purchased and the factors that might acceptability against data quality criteria influence the variability of the nutrients specified for the data base by the in the food, at the time of purchase, and compiler (Southgate & Greenfield 1988). during and after sampling. Many factors This data scrutiny is itself a process influence nutrient content, but not all of requiring a knowledge of the factors that these factors apply to all types of food. influence foods and their composition A flexible system is therefore necessary including laboratory procedures and and for each new food type a specific analytical methods, the use of statistical sampling protocol is developed. and research techniques as well as data In Australia while regional variation in comparison procedures. foods can be expected, sampling across This review of the factors considered the capital cities of a number of States during data scrutiny and compilation is does not necessarily provide the most based on the experience of producing appropriate or representative food Composition of foods, Australia (COFA) sample. Many primary products and (Cashel & others 1989) and NUTTAB89 manufactured foods are produced in a (Department of Community Services & small number of centres and then Health 1989). The programs for these nationally distributed, and so do not publications have been outlined reflect the local growing conditions previously (Cashel 1989, English 1990). around each city, eg mangoes and other tropical fruit. Further, nutrient changes may be introduced into fresh produce Karen Cashel is a Senior Nutritionist, during packing, storage and transport to Department of Community Services & the central analysing laboratory resulting Health, PO Box 9848, Canberra ACT 2601 in data which do not reflect the nutrient The submitted laboratory reports are content of foods as purchased by scrutinised in the Nutrition Section of the consumers. These and other Department and queries are referred considerations including budget both to program advisers and back to limitations resulted in much of the the laboratories. This may continue for sampling being undertaken in the region months, even years, and may require of the analysing laboratory. re-analysis and sometimes re-sampling The aim was to obtain a sufficiently to finalise the analytical data. wide range of samples from differing Factors in data scrutiny batches to obtain a representative nutrient profile for the product. For The new data were drawn from the food processed and manufactured foods, analytical program, from published where possible, varieties or brands articles on Australian foods, from the sampled are based on national market food industry and from unpublished shares while for some foods, sampling research. A number of factors, from different production runs was more discussed below, determine not only the important. Samples were purchased suitability and usability of the data but from consumer preferred retail outlets also correct interpretation of the data. In by an unidentified purchaser. For some addition overseas food composition of the fruit and vegetable items these tables and papers were consulted to outlets included major markets to obtain provide a global context for data exotic produce, to assist with identifying comparison and validation. cultivars and to allow sampling from The food description is vital different growing regions. Foods were information for any user. It must clearly analysed within their product shelf-life, identify the food item analysed and as defined by use-by-dates. For most should therefore include an objective foods a composite sample was prepared identification, ie scientific name, for analysis from multiple individual components of mixed foods; physical samples. state, ie raw, cooked or processed; and cooking or processing description. All Laboratory programs too often the food descriptor consists of Data scrutiny begins back at the ‘mango’, ‘hamburger’, or ‘mince meat’ laboratory. The demands made from a and the reader is left to guess the form project of this size and complexity are of the food analysed (raw, canned, immense and each laboratory involved dried), the animal (or other) food base of has needed 18–24 months to set up the ‘mince meat’; whether the mango methods, solve problems and develop was peeled or unpeeled; which expertise in all aspects of the program. components made up the ‘hamburger’ Far more is expected of the laboratory and, or course, the exact meaning of than just analysing a food sample. The that useful descriptor ‘cooked’. Misuse laboratory must collect samples to a of such terms can be highly misleading. written protocol and must provide a For example the data for both battered range of information about the item and fried oysters and shrimps in analysed including purchase date, Thomas and Corden (1977) are identical purchase site, sample descriptions and to those for breaded and fried oysters copies of labels, sample sizes, handling, and shrimps in the 1963 edition of the gross composition, edible portion, US tables (USDA 1963). However, not analytical sample preparation, weight only are the composition of fried changes, analytical methods used as battered foods and fried breaded foods well as the nutrient levels determined. different due to differing proportions of coating to food and to differing fat relative to the commercially available absorption levels (Makinson & others fruit. 1987) but the food called shrimp in the Further, without cultivar identification, US is commonly called prawn in local data on the potato, for example, Australia. The use of the term ‘shrimp’ in would have been difficult to evaluate in Australia suggests a very different sized a global context. unit of food and a different flesh to The edible portion information is vital coating ratio. This kind of factor has to to knowing to what the analytical data be considered when evaluating data relate. The compiler needs to know against data from other countries. which components of the food were If the food is processed, included in the analytical sample; which consideration must be given to whether components were rejected; and what its composition is affected by food the proporations of these were. This legislation which may vary both by State information assists both in further and with time. This information is identifying the food as purchased and pertinent not just to the relevance of the the food as analysed. All too often this data but also to checking the accuracy information is not given or only the of the analytical values eg margarines edible portion factor is given, usually as and the meaning of terms such as a percentage. The edible portion can at ‘table’, ‘polyunsaturated’. least provide some basis for assessing The sampling description required what proportion and, to a lesser extent, includes whether the food was as what part of the food is considered as generally available to consumers; edible. Different cultures may eat whether the food was selected to different parts of a food and this may specified criteria and, if so, what these change with time. For example, the fibre were; when and where the food was debate has affected what is commonly purchased or obtained; how many considered as ‘edible’ in our community. samples of what size were obtained, Also, the food as purchased may and, if relevant, brand names. change with time and, without an This highly pertinent information objective description the edible portion, assists to identify season, region of expressed only as a percentage, will be origin, market availability, consumer quite inappropriate and even misleading availability as well as nutrient to the user. For example, fresh carrots implications. For example, the vitamin C used to be marketed in bunches with content of a fresh-picked apple analysed their tops resulting in a lower edible immediately is inappropriate to portion value than that of current represent the vitamin C content of market-trimmed carrots. Similarly, for applies stored for more than three days meat chops changes in retail trimming regardless of whether the apple is retail practices affect the proportions of edible or home produced. The vitamin C flesh to bone in the food as available. content of a fresh-picked apple falls to a Analytical sample information is stable level after picking (Wills & El- required to identify whether a single or Ghetany 1986). Data for fresh-picked composite sample were formed apples would therefore be misleading if (Greenfield & Southgate 1985) and, if used to represent apples as available to composite, the number of units used the vast majority of the population. and the proportions in which they were Similarly data from a study of vitamin C combined. The preparation of the in oranges grown on experimental root analytical sample indicates the stocks may be quite inappropriate approach taken to obtain data when the dietary fibre content is representative of the food as purchased. determined by the current AOAC Laboratory handling and preparation recommended method will give an information required by the compiler apparently higher value for total dietary includes how the food sample was fibre than exists in the fresh food item handled and stored both in the (Bingham 1987, Englyst & others 1987). laboratory and, if transported significant The analytical method used must be distances, during transport; how the valid, appropriate to the food and analytical sample was prepared and identify all required components of the stored; and when and in what order the nutrient. Analytical technology and analyses were carried out. Sample associated methods are continually handling and storage should be evolving as knowledge improves about consistent with, at a minimum, usual the active forms of a nutrient and the home practices, ie consistent with specificity with which they can be manufacturer's instructions and quantified. For some nutrients, eg appropriate to the food. If the food was dietary fibre, vitamin A, this naturally transported, eg from interstate, details of leads to a healthy debate about the how the food was prepared and components which should be included in packaged and information about the the generic name of the nutrient and time lag between purchase and their relative physiological activities. laboratory delivery must be available. Significant inconsistencies are seen Obviously some nutrients are affected in reported levels of nutrients due to the by these handling practices. use of differing analytical methods which Moisture and vitamin C analyses, in vary in the active principles they particular, should be performed as soon measure. For example, the method after receipt of the food as possible. used to analyse vitamin C and accepted Further, for a food that must undergo for the official US tables measures only significant transport and time delays, ascorbic acid (USDA 1976–1989) while weighing of the food at the time of that in the Australian (Cashel & others purchase and packaging, and again at 1989) and British (Paul & Southgate the laboratory provides an indication of 1978) tables measures both ascorbic the degree of physical change in the and dehydroascorbic acid. Further, the food. As many foods have a high US food composition tables do not moisture content small changes can include analytical data for carbohydrate have a large effect on the final but data calculated ‘by difference’. Most expression of the nutrient concentration food industry laboratories also prefer the of a food as originally obtained. For ‘by difference’ method. Such ‘measures’ other nutrients, inappropriate handling include dietary fibre as part of the and storage may produce artifacts in the carbohydrate measure and affect the data resulting in both losses and factors used to calculate energy. apparent gains in nutrients. Factors Carbohydrate and its components may affecting loss of nutrients such as the be reported as monosaccharide effect of ultraviolet light on riboflavin equivalents (eg Paul & Southgate 1978) levels in milk are well known, but certain or as grams of component (eg Cashel & preparation and analytical method others 1989). Nutrients determined and combinations may also result in methods of expression can vary even apparent gains. The freeze-drying of a within a food composition table, high starch food will result in the especially when released in a series formation of resistant starch which, over a number of years (eg dietary fibre data in both the US tables (USDA 1976– industry data are often a combination of 89) and UK tables (Paul & Southgate limited analyses and values calculated 1978, Holland & others 1988). For some from product formulations. The nutrients eg folate, the debate over the analytical data may be local or from the revelance of the data by existing parent company overseas. The rest of methods makes suspect all the available the data may be derived from any data. For others, eg vitamin A, the new number of sources, the most popular analytical technologies challenge being the previous official Australian previous assumptions about the isomers tables (Thomas & Corden 1977), the US being measured and their biological tables and the UK tables. Many of the relevance. These factors must all be food tables of the world, including considered when scrutinising data Thomas and Corden (1977), use or are against comparable foods from based on the US tables with or without elsewhere in the world. specific referencing and without Quality control and mode of adaptation except for changing the food expression must also be considered name to the local preferred name. before analytical values, from any Nutrition texts can use one or more laboratory, should be accepted as sources, usually a major national table, correct. Data such as range, standard sometimes with a small proportion of deviation and coefficient of variation local data and even with selective and must be used with caution as these may arbitrary changes to one or more relate to biological variability or seasonal nutrients (eg Briggs & Wahlqvist 1984). variability of the nutrient or to intra- and If sources are not specified the user inter-analyst, laboratory or method has no information at all on the reliability variability. of the data, their Australian relevance, There are internal consistency the analytical methods used (and thus checks for reliability such as the sum of the components measured) or the fatty acids and proximates to assess currency of the data and cannot go back consistency of relative levels of nutrients to the primary source or the analyst for in data for the same food. Moisture has more information. Lack of adequate already been identified as a critical referencing is a breach of scientific nutrient in terms of expressed levels of conventions and source nutrients on a wet weight basis. acknowledgement and even of Apparent differences in data may be copyright. Food composition data are a due to differing levels of hydration. national and international resource and Moisture too may be the only indicator their integrity needs to be recognised, of a dried or partially dried food. In valued and protected. published data it is also important to Compiling food composition data check the factors used for determining calculated values, eg nitrogen factors, The process of compiling food fat factors, energy factors. composition data into the data base The data source is generally obvious aims at ensuring internal consistency when dealing with analytical research and compatibility of data and requires papers and laboratory reports. For data the steps described below. from food composition tables, nutrition Repeating, extending or rejecting texts, consumer reports and food data analyses refers to data needed to industry information the source may not provide the minimum range of nutrients be clear and yet is critical to interpreting necessary for inclusion in the tables the data and their usefulness. Food and/or for checking on data that the compiler questions as a result of the Englyst to the AOAC method for use in scrutiny process. COFA was a pragmatic decision based For example, the Australian meat on factors such as regular availability of data as published by the researchers analytical resources and the cost of (Greenfield 1987) are not necessarily analysis and is regularly reviewed in the identical to those in COFA. Data from light of the current scientific debate. one group could not be used because of Levels of significant figures have uncertainty about the representivity of recognised and recommended the meat samples analysed, while fatty conventions for nutrient data acid data from another group were no presentation (Greenfield & Southgate longer adequate as the capillary gas 1990). Analysts may, however, report chromatography technique became data to a higher or lower number of preferred. Therefore, a separate significant figures thus necessitating in program commissioned later determined some cases the rounding of data for the fatty acid content of beef by capillary tables. Most computer generated tables, gas chromatography and provided eg the US tables, have a fixed level of further gross composition and other reporting with most nutrients at a nutrient data. These data were then standard two or three decimal places. combined with the data from the initial The Australian Nutrient Data Bank program for inclusion in the new tables. system specifies a level of reporting for Data may be rejected if the method each nutrient at the time of printing but used was incompatible with the does not yet have the facility to round specified or validated methods. For selectively to the nearest five or ten example, the data identified as ‘residue’ units, for example. ‘fibre’ and ‘dietary fibre’ in the Greenfield Calculation factors for generating data and Wills published papers (Greenfield on some nutrient components eg & others 1981, Wills & El-Ghetany 1986) energy, or for adjusting to quantity per were not included in COFA as the unit of food using fat and nitrogen method was not validated. The modified factors are generally based on Englyst method (Englyst & others 1982, internationally recognised conventions Jones & others 1985) was initially used (Paul & Southgate 1978). The literature in the commissioned Australian program is monitored as these factors can be and the AOAC method (Prosky & others affected by developments in analytical 1985) is now used. methods. These two methods are not Changes to the food may have compatible and do not measure the occurred since the food item analysed same components of dietary fibre. The was purchased. For the commissioned Englyst method is highly specific and Australian programs the delay between measures chemical defined components the decision to analyse a food and of dietary fibre and excludes resistant publishing the data should be less than starch and substances measuring as three years but has been up to eight lignin from its reported total dietary fibre years for COFA. The interval comprises values. The AOAC method provides a the time needed to research available measure of total dietary fibre (TDF) that products and to draw up a sampling includes resistant starch and lignin. The protocol (usually, 6–8 weeks); the food modification to provide some measure analysis program (usually 12 months, of soluble and insoluble fractions is also especially if foods are seasonal); used in the Australian program (Prosky laboratory checking and reporting & others 1988). The shift from the (usually 6–12 months); data scrutiny, evaluation and preparation (6–12 identification of the food to the months per program); and the consumer is generic, eg apricot, baked publication process (now about 3 bean, milk coffee biscuits or cheddar months). The article in a scientific cheese. For others, such as apples, journal may be a little younger (date of both approaches may be taken, ie data food sampling should be sought). Even provided on individual cultivars in this relatively short time there may (Jonathon, Granny Smith etc) plus have been changes in a food item due generic ‘apple’ from weighted means of to changes to the fortification formula, the individual cultivars. ingredients, food legislation, additives The potential use of the data is (eg β-carotene being used as a food considered when making such colour), wholesale or retail practices (eg decisions. For example, a major use of butchering) or agricultural practices (eg food composition data is in dietary preferred carcase composition or major studies where subjects report the foods plant cultivar). Such changes must be consumed. The likely ability of the monitored, and assessed and the data consumer to identify a food (eg by handled accordingly in order to cultivar) at the time of purchase or represent the food accurately in the consumption is considered in presenting published tables. There may be a need data. Also, if a statistical analysis of for reanalysis of some nutrients, eg β- different data on a range of shaped, carotene, or to modify the way the food sized and branded extruded snack is described. For example, an extruded products shows no difference in their snack product may stay the same in analytical values then the inclusion of a composition but change in size or shape generic named item in the food table of the pieces or in name and package provides useful and reliable data to size of the product. users that allows for the cosmetic The best representation of a food in differences. the tables requires considerable Terminology development to describe thought, discussion and research. The the foods sampled and the parts actually arithmetic mean of all analytical values analysed is a significant aspect of for a nutrient in one food does not preparing data for users, as is grouping necessarily produce an appropriate the foods for location and index representation of the food as available purposes in tables. The aims are to to the consumer. Some of the data may develop criteria to allocate foods be appropriate to a specific growing objectively to a grouping; to use terms region, season, or factory output, or to unambiguously and consistently to selected batches of a processed describe the food, the edible and product. Other data may be from a inedible parts and the basis of the composite of samples more broadly sampling; to describe the analytical drawn from the food supply. Allowance sample; and to describe the analytes for these differences may result in and the analytical methods used. weighting the data to allow for such Working descriptors and definitions differences or the use of footnotes to have been developed for these specific highlight specific seasonal or regional purposes in the new Australian food differences. Similarly if there are data on tables and are described and defined in individual brands or cultivars of a food the text. these may be weighted on the basis of Gross composition and food market share or production ratios. This measures are critical to providing is usually done if the major basis for nutrient information in practical terms for users. Gross composition data, that is, and unchanging. Modern users are ingredients and proportions of generally far removed from the ingredients in mixed foods or production and compilation of the basic components and proportions of edible composition data, and, in addition, the and inedible parts of fresh or single ease of ‘plugging’ into a computer and ingredient foods, are regularly generating instant analysis of food and monitored. Changes to these may affect diets further seduces their questioning edible portions (eg carrot tops, facilities. Users must be prepared to butchering techniques), nutrient become more knowledgeable by composition (butchering techniques and reading the explanatory text practices), weight changes associated acompanying food composition tables or with cooking and meaningful description a data base. In addition they must be of the food as available. prepared to become better informed Where possible sampling protocols about the background to the production and the details recorded for food and compilation of food composition samples anticipate that such changes data as this will enhance the quality and occur. For example, the muscle tissue appropriateness of the uses to which and separable fat components were such data are put in the scientific analysed separately for meats and practice of nutrition and dietetics. details of the proportions of muscle, References separable fat and bone recorded (Greenfield 1987). This allows for Bingham S. 1987. Definitions and representation of meat cuts at different intakes of dietary fibre. Am. J. Clin. Nutr. levels of fat trimming and for revision of 45: 1226–31. nutrient data of meat cuts as available Briggs, D & Wahlqvist, M. 1984. Food retail based on gross composition data, facts. Ringwood: Penguin Books without further expensive analytical Australia Ltd. work. Full descriptions and associated Cashel, K. 1989. Australian nutrient data weights of measures of foods, tables. Food Aust. 41: 1034–5. particularly standard measures of foods, Cashel, K, English, R. & Lewis, J. 1989. are part of the food analytical program. Composition of foods, Australia. Such measures are routinely collected Canberra: AGPS. as part of the laboratory reports and a funded three year weight for defined Commonwealth Department of measure program was commissioned to Community Services & Health. 1989. obtain such information (English 1990). NUTTAB89. Nutrient data table for use Other investigations of these measures in Australia. Disk format. Canberra: from a variety of sources have also AGPS. been undertaken. English, R. 1990. Composition of foods, Correct understanding and use of Australia. Food Aust. 42: 8, 55–57. food composition data Englyst, H, Trowell, H, Southgate, DAT It should be clear from this account that & Cummings, J. 1987. Dietary fibre and compiling national food tables involves resistent starch. Am. J. Clin. Nutr. 46: more than cataloguing numbers 873–4. collected through the commissioned Englyst, H, Wiggins, HS & Cummings, analytical programs or other sources JH. 1982. Determination of the non- and wrapping a cover around them, and starch polysaccharides in plant foods by that the data provided are not stagnant gas-liquid chromatography of Prosky, L, Asp, N-G, Furda, 1, De Vries, constituent sugars as alditol acetates. JW, Schweizer, TF & Harland, BF. Analyst 107: 307–18. 1985. Determination of total dietary fibre in foods and food products: collaborative Greenfield, H (ed). 1987. The nutrient study. J. Assoc. Off. Anal. Chem. 68: composition of Australian meats and 677–9. poultry. Food Technol. Aust. 39: 181– 240. Prosky, L, Asp, N-G, Schweizer, TF, De Vries, JW & Furda, I. 1988. Greenfield, H, Lee, YH & Wills, RBH. Determination of insoluble, soluble and 1981. Composition of foods. 11. Muesli. total dietary fibre in foods and food Food Technol. Aust. 33: 564–8. products: Interlaboratory study. J. Greenfield, H & Southgate, DAT. 1985. Assoc. Off. Anal. Chem. 71: 1017–23. A pragmatic approach to the production Southgate, DAT & Greenfield, H. 1988. of good quality food composition data. Guidelines for the production, ASEAN Food J. 1: 47–54. management and use of food Greenfield, H & Southgate, DAT. 1990. composition data: an Infoods project. Guidelines for the production, Food Sci. Nutr. 42F: 15–23. management and use of food Thomas, S & Corden, M. 1977. Metric composition data. An INFOODS manual tables of composition of Australia foods. (in press). Canberra: AGPS. Holland, B, Unwin, ID and Buss, DH. US Department of Agriculture. 1963. 1988. Cereal and cereal products. Composition of foods. Raw, processed, London: The Royal Society of prepared. Handbook 8. Washington DC: Chemistry. USDA. Jones, GP, Briggs, DR, Wahlqvist, ML & US Department of Agriculture. 1976– Flentje, LM. 1985. Dietary fibre content 1989. Composition of foods. Raw, of Australian foods. 1. Potatoes. Food processed, prepared. Handbook 8 Technol. Aust. 37: 81–3. series. Washington DC: USDA. Makinson, JH, Greenfield, H, Wong, Ml Wills, RBH & El-Ghetany, Y. 1986. & Wills, RBH. 1987. Fat uptake during Composition of Australian foods. 30. deep-fat frying of coated and uncoated Apples and pears. Food Technol. Aust. foods. J. Food Comp. Anal. 1: 93–101. 38: 77–82. Paul, AA & Southgate, DAT. 1978. The composition of foods. London: HMSO.

Food composition data in the promotion of lean meat in Australia

L. Fantini

eef and lamb have long played a overcome, as they added up to a B part in both Australia's rural industry fundamental lack of consumer and in the eating habits of Australians. confidence in the product: The introduction of cattle and sheep by Perception of inconsistency of both the first fleet in 1788, the opening of price and quality. Common concerns pasture land in western NSW and the and complaints were quality variation, success of grazing livestock led to beef from week-to-week or retailer-to-retailer, and mutton becoming staple items in the and inexplicable sharp rises and falls in diet of early Australians (Symons 1984). price; In those early years, the relative Competitive protein sources. Chicken, abundance of meat and its low price fish, vegetarian meals and the Pritikin could well have created an attitude diet had all infiltrated the Australian diet; towards red meat as a favoured food Trading hours. The inability to item, an attitude that was to last many purchase red meat during all normal years (Clements 1986). Given this retail trading hours in most States was attitude, it is not surprising that beef and cited as a major disadvantage, lamb continue to play a major, albeit particularly by working women, whereas less dominant role in the current competitive protein foods such as Australian diet. Until comparatively chicken and fish could always be readily recently, it was considered unnecessary purchased; to spend a great deal of time, effort and Lack of convenience. Consumers money promoting meat to Australians, regarded red meat as messy and time- and until the early 1980s the major role consuming to prepare. It did not meet of the Australian Meat and Live-stock the needs of working women who Corporation (AMLC), formerly the preferred quick, convenient foods which Australian Meat Board, was to promote reduced time and effort in the kitchen; the export of Australian meats and Poor merchandising. By comparison livestock. with most other food categories there Changing food preferences was nothing new happening in the area of retail presentation; In 1984–5, comprehensive market Nutrition. Concern was expressed research (McKinna & others 1984, The about nutritional aspects, ie excessive Campaign Palace 1985) painted a very fat, cholesterol, kilojoules, and the dismal picture of Australian consumer vague and undefined feeling that “too attitudes towards beef and lamb. No much red meat isn't good for you”; single factor could be isolated as solely Ethnic foods. Fashion in food was responsible for the move away from red being created by a surge of new meat, but rather, a number of inter- restaurants featuring different, exciting related factors emerged identifying and varied cuisines, very few of which marketing problems which needed to be included red meat; Luisa Fantini is Manager — Nutrition and Lifestyle changes. Red meat was not Education, Australian Meat & Live-stock seen as contemporary or fitting into the Corporation, GPO Box 4129, Sydney NSW modern lifestyle. The traditional family 2001. Sunday dinner of roast and three vegetables was no longer appropriate came from a lack of knowledge of as meals were often served on the run; Australian meat composition and its Lighter meals. Red meat was seen, production, such as: that meat particularly by women, as too heavy and composition (in terms of fat and indigestible in contrast to the many light cholesterol) could vary from country to alternatives available; country; that Australia's meat production techniques (essentially pasturerearing) Lack of variety. The product was seen generally favoured the production of as lacking in variety, with very little lean, non-marbled meats; that grain- novelty; in short, dull and boring. feeding or lot-feeding is not as A number of marketing strategies extensively used in Australia as it is in were undertaken to counteract these the USA, and as a result marbling is problems, particularly in the area of practically non-existent in Australian advertising and merchandising, but beef as compared to US beef; and, that effective response to the nutrition improvements had been made by the controversy was hampered by a lack of industry, such as the introduction in the up-to-date data on the fat and 1960s and 70s of European breeds cholesterol content of meat. In addition, which favoured carcases of a lower total the only data on the fat contribution fat cover. made by meat to the diet was apparent The employment of dietitians by the consumption data (ABS 1986). This AMLC in 1984–5 was seen by the indicated that meat as available industry as a vital step towards the provided around 37% of the fat in the dissemination of accurate nutrition diet. In reality, meat as consumed information on beef and lamb. contributed much less. This is because Therefore, the attainment of up-to-date apparent consumption statistics data on the composition of beef and generally overestimate fat contribution lamb became the first priority. by not taking into account the loss of fat from the carcase during the preparation Meat composition data of retail cuts (Fantini & MacDonald The existing official food composition 1987). Media reports on health and tables for Australia (Thomas & Corden popular nutrition paperbacks (mainly 1977) did not reflect the meats based on US meat data) created a Australians were consuming in the perception of beef and lamb as less 1980s but were based on British, than optimal protein sources for today's American and/or 40-year-old Australian consumer. This perception was equally data. The out-of-date and non- common amongst health professionals, applicable data contributed to the strong as evidenced by the fact that only perception that red meat consumption recently was “chicken (no skin)” meant a high fat intake and therefore a incorporated as standard advice into high health risk, whereas white meat (eg nutrition education materials produced fish or chicken) was nutritionally more by an independent health authority desirable (Chris Adams Research (Australian Nutrition Foundation 1989). 1986). Until now, the advice has been “lean Despite the concern regarding the meat, poultry, fish”, incorrectly indicating nutrient composition of red meat and its that poultry was already lean and thus effects on the nutrient intake of acceptable. Australians, no new data became The industry was aware that the available until the mid 1980s. The first unwarranted criticism of the fat and available data came from the work of cholesterol content of beef and lamb Sinclair and O'Dea (1987). Their final report to the Australian Meat Research as well as fresh and frozen chickens, Committee in 1986 (subsequently whole and as chicken pieces. These published in 1987) showed that the results showed that some portions of average lean intramuscular fat content chicken were higher in fat relative to of beef cuts was 3.1%, with a range beef or lamb, even after the removal of from 1% to 11% over 460 samples. chicken skin. It was clear that only Lean lamb intramuscular fat content was skinless chicken breast at 3% fat (raw) around 4.6% on average with a range of was lower in fat than the four beef cuts 2.6% to 12.3% over 36 samples. These examined. Therefore, lean beef and figures were lower than those published lamb, trimmed of visible fat, could not be in the existing official food tables condemned as high fat foods. (Thomas & Corden 1977) for lean These initial results from separate boneless beef (6.6%) and lean boneless research groups considerably aided the lamb (6.3%). The work of Sinclair & AMLC in promoting the nutritional O'Dea (1987) was also the first to supply benefits of beef and lamb in a more information on the fatty acid profile of specific manner. However, the AMLC lean meats. There had been a strong refrained from any major advertising perception that not only was red meat activities as release of official research high in fat but it was also high in on the nutrient composition of meats saturated fat with little or no was imminent. These data were to polyunsaturated or monosaturated fatty further consolidate the findings of acids present. Their results showed that Sinclair and O'Dea (1987) and Thornton lean beef contained major amounts of and co-workers (1987). oleic acid, a monounsaturated fatty acid The publication of those official data (27–44% of total fatty acids) and from the work of Greenfield, Truswell significant amounts of the and co-workers on the composition of polyunsaturated fatty acids, linoleic, Australian meats and poultry (Greenfield linolenic and arachidonic. 1987) allowed the AMLC to implement Sinclair and O'Dea (1987) also its marketing activities fully promoting revealed little or no difference in the the nutritional benefits of beef and lamb. intramuscular fat content of grain-vs These data and the newly revised and grassfed beef, even when comparing released official food composition tables carcases of two different fat cover (Cashel & others 1989) form the basis depths. This work indicated that unless for all current activities of the AMLC in significant marbling was present, the the promotion of the nutritional benefits grain-fed beef generally available on the of beef and lamb. Australian domestic market was similar Recently, further analytical work by in fat content (around 4%) to pasture fed the Australian Government Analytical meats, once selvedge fat was removed. Laboratories (AGAL) in Adelaide was This contrasts sharply with US grain-fed carried out in conjunction with AMLC beef which has a minimum fat content of participation in the National Heart 9%. Foundation's Food Approval Program. This work was supplemented by Results indicated that all cuts submitted research conducted at CSIRO by satisfied the NHF's guideline of less Thornton and co-workers (1987) who than 10% fat. Cholesterol and fat data examined the fat content of popular cuts were similar to those of Greenfield, of meat, raw and cooked, including T- Truswell and co-workers (Greenfield bone, rump, blade and topside, lamb 1987). These data add further weight to loin chops, rolled lamb shoulder roasts, the argument that lean meat, trimmed of low-fat food with a deserved place in the visible fat, is a low fat food. Australian diet. Three TV commercials Advertising using meat composition were created in the 12 months following data the release of the publication: Weigh Out, Pasta and Potatoes and Jane The AMLC's concept of “lean meat” is a Flemming. central marketing tool in the promotion The first, Weigh Out, comparing a of beef and lamb in the 1980s and 90s. lean beef meal with a roast chicken “Lean meat” is readily identified by meal stirred much passion amongst consumers as “new meat” (AMLC health professionals and stimulated 1988a) and thus meets their needs for consumers to seek much needed convenience, variety, and lighter, more information. As previously discussed, nutritious meals, in contrast to “old the perception of beef (and meat” which is fatty and lacking in consequently lamb) being unhealthy variety. was rife in the early 1980s. Market Over the past four years there has research (Chris Adams Research 1986) been a meat retail revolution in had indicated that consumer perception Australia. Today, meat in butcher shops of chicken as a healthy food was all and supermarkets is different, ie it is embracing. It was seen as nutritionally leaner, often pre-prepared, and visually superior to beef, and by consequence, stimulating. This is partly the result of a lamb. While Australians were acquiring single-minded marketing and advertising the habit of fat-trimming beef (having strategy called Short Cuts focused on associated beef with a high fat, high quick, easy-to-prepare, nutritious whole cholesterol diet) substantially fewer meat meals complete with vegetables people understood that chicken with and cereals. The advertising and skin was also a culprit in a high fat diet. merchandising campaign established In fact the 1983 dietary survey of adults lean meat as the hero for the busy (Cashel & others 1986) indicated that working woman looking for a solution to removing beef fat was more common the evening meal dilemma. In 1988–89 amongst respondents (70%) than the Short Cuts proposition was removing chicken skin (30%). Further, expanded to include more traditional having positioned lean meat as its meals but with a modern touch. These central marketing tool, the AMLC could are the Today's Steak and Lamb Roast not logically use anything other than campaigns. Again, the feature is lean lean meat depicted as fully trimmed of meat in complete meals, eg lean steaks fat, in its new phase of nutrition and lean roasts. In its mainstream advertising, using the new meat advertising the AMLC was already composition data. concerned with promoting lean meat Advertising, if it is to be successful and had spent a great deal of time, must evoke emotion. It is not truly effort and money establishing in designed to be educational. While consumers' minds that beef and lamb, Weigh Out was indeed based on facts, trimmed of fat, were nutritious and the real success of the commercial lay convenient, fulfilling a contemporary in the emotional response from need for modern women. consumers and health professionals The release of the publication The alike, it challenged people's established nutrient composition of Australian meats beliefs. It was designed to make people and poultry (Greenfield 1987) heralded search out more information. Nutrition the start of a new and aggressive education is complex, but advertising campaign to re-position lean beef as a must encapsulate a single idea in a 30- lean meat as being trimmed of all visible second message. That single idea was fat. that chicken, in its commonly consumed In addition, these arbitrary trim levels form (with skin), contained more fat and are dependent on the amount of fat cholesterol than consumers may have originally present on the cut when thought and is no magic food. The purchased. Thus, traditional lamb cuts, commercial also established the idea usually purchased with all fat attached that lean beef is not fatty and full of since some states regulate to prevent cholesterol but instead compared well to the removal of the lamb pink branding at the consumer's benchmark of a low-fat point of sale, are considerably higher in protein food, chicken. As a result of the fat content at the 75% trim and 50% trim debate which raged after the levels than beef, yet lean only data for commercial's airing, consumers now fat content of beef and lamb are not understand that eating chicken with skin dissimilar. is nutritionally no better than eating red The effect of plate waste (more fat meat with the fat. trimming on the plate) must also be The Pasta and Potatoes and Jane considered for lamb in particular. This Flemming commercials, whilst less aberration should be taken into account controversial, also relied on new meat when these tables and the NUTTAB composition data for the claims for fat, data base are used by health iron and vitamins. professionals. The current food tables (Cashel & Nutrition education using meat others 1989) and NUTTAB composition data (Commonwealth Department of Health & Community Services 1989) depict four Many other important and ongoing levels of trim for beef, lamb and pork: educational activities of the AMLC also lean only, 75% trimmed, 50% trimmed rely on the new compositional data, and and as purchased (ie with all fat as aim to cement these nutritional facts in purchased at retail). The AMLC on the the minds of consumers and health other hand, in its advertising and professionals alike. general promotional activities, uses lean In 1985, the AMLC entered its new only data primarily. Health professionals phase of nutrition education. Materials should be aware that the levels of trim were of a general nature, positioning indicated in the tables have been lean beef and lamb as important arbitrarily selected by nutritionists within components of a balanced diet and the Department as representing current incorporating the general dietary advice levels of fat trimming by Australians. of less fat, more fibre. However, the Health professionals should also be release of the revised meat composition aware that the compositional data for data enabled such material to contain these levels of trim are derived by more specific information on beef and calculation and not by analysis. Several lamb, and their contribution to the studies (Cashel & others 1986, Worsley Australian diet. & Crawford 1985) have now indicated The AMLC's Meat, the facts brochure that fat-trimming of meat is standard (AMLC 1988b) is a good example. A practice, although none of these studies publication of this type could not have have quantified the amount of fat been produced three or four years ago, trimmed. The AMLC believes that as official meat data were not then consumers understand the concept of available. Recently, the AMLC has undertaken the Meat, the facts program for general practitioners to give them accurate information on the fat and References cholesterol content of beef and lamb Australian Bureau of Statistics. 1986. while reminding them of the iron, zinc, Apparent consumption of foodstuffs and protein and B vitamins provided by nutrients. Australia, 1984–85. Catalogue these foods. This activity has included No. 4306.0. Canberra: ABS. general mailing of materials to GPs, seminars in major capital cities on the Australian Meat & Live-stock nutritional facts about meat, and the Corporation. 1988a. Meat marketing production of medical press trends. Sydney: AMLC. advertisements. Revised meat Australian Meat & Live-stock composition data as well as the release Corporation. 1988b. Meat, the facts. of the reports from the dietary survey Sydney: AMLC. (Cashel & others 1986, English & others 1987) have been crucial to this activity. Australian Meat & Live-stock Other AMLC activities relying on Corporation. 1988c. Healthy food for these composition data include the healthy hearts. Sydney: AMLC. calculation of fat and energy in recipes Australian Nutrition Foundation. 1989. specifically designed for the health Nutrition Time. Sydney: ANF. conscious, as in the AMLC's Healthy food for healthy hearts booklet (AMLC Cashel, K, English, R, Bennett, S, 1988c) and in the slimmers section of Berzins, J, Brown, G & Magnus, P. the new Short Cuts 2 cookbook 1986. National dietary survey of adults: (Murdoch Magazines/AMLC 1989). 1983. No. 1. Foods consumed. Further, AMLC's participation in the Canberra: AGPS. Australian Nutrition Foundation's Cashel, K, English, R & Lewis, J. 1989. Nutrition Time 89 program featured Composition of foods, Australia. editorial copy based on lean lamb Canberra: AGPS. composition data. Similarly, participation in the National Heart Foundation's Food Chris Adams Research Pty Ltd. 1986. Approval Program (Haddy 1990) has Evaluation of the short cuts campaign only been possible because of the and its effect on consumer attitudes knowledge of the fat content of lean, towards red meat. Sydney. Unpublished trimmed meat. report. Conclusion Clements, FW. 1986. A history of human nutrition in Australia. Melbourne: It is clear that the AMLC's promotional Longman Cheshire Pty Ltd. activities for beef and lamb rely heavily on new meat composition data. The Commonwealth Department of AMLC has been single-minded in its Community Services & Health. 1989. depiction, both visually and in print, of NUTTAB89. Nutrient data table for use lean meat as a central marketing in Australia. Disk format. Canberra: concept. The release of these new food AGPS. tables should further assist both the English, R, Cashel, K, Bennett, S, food industry and health professionals to Berzins, J, Water, A-M & Magnus, P. get information about food to 1987. National dietary survey of adults: consumers. However, care needs to be 1983. No. 2. Nutrient intakes. Canberra: taken by health professionals in the AGPS. interpretation and use of meat data as presented in the new food tables. Fantini, L & Macdonald, N. 1987. Trends in meat consumption in Australia 1938– 39 to 1983–84. Food Technol. Aust. 39, Food Approval Program. Food Aust. 187–190. 42:8, 514–15. Greenfield, H (ed). 1987. The nutrient McKinna, D. et al. Pty Ltd. 1984. composition of Australian meats and Consumer attitudes, opinions, poultry. Food Technol. Aust. 39: 181– perceptions and usage of red meat — 240. an exploratory study with particular emphasis on beef. Melbourne. Haddy, B. 1990. Food composition data Unpublished report. and the National Heart Foundation's McKinna, D. et al. Pty Ltd. and The Campaign Palace. 1985. How to Beveridge & Cummings Pty Ltd. 1984. A sell more meat by using Short Cuts. quantitative assessment of consumer Sydney. Unpublished report. attitudes, opinions and behaviour with regard to red meat compared with other Thomas, S & Corden, M. 1977. Metric protein rich foods. Melbourne. tables of composition of Australian Unpublished Report. foods. Canberra: AGPS. Murdoch Magazines Pty Ltd/Australian Thornton, RF, Tume, RK, Larsen, TW, Meat & Live-stock Corporation. 1989. Johnson, GW and Sexton, WR. 1987. Short Cuts 2. Sydney: AMLC. Fat content of popular cuts of meat: Cooked and raw. CSIRO Food Res. Q. Sinclair, AJ & O'Dea, K. 1987. The lipid 47: 30–7. levels and fatty acid compositions of the lean portions of Australian beef and Worsley, A & Crawford, D. 1985. lamb. Food Technol. Aust. 39: 228–31. Awareness and compliance with Australian Dietary Guidelines: a Symons, M. 1984. One continuous descriptive study of Melbourne picnic. A history of eating in Australia. residents. Nutr. Res 5: 1291–308. Ringwood: Penguin Books Australia Ltd.

Food composition data in the promotion of canned foods

A.K. Wailes

he Canned Food Information Service More recently the industry conducted TI nc (CFIS Inc) is an organisation another survey amongst 601 grocery promoting Australian food packaged in buyers. There were many contradictions steel cans. Through general publicity in the attitudes of consumers to canned CFIS Inc aims: to convince consumers foods particularly what constitutes and potential consumers of Australian typical buying patterns. The researchers canned food of the merits of this food concluded that canned food is a product delivery system and the foods category with wide utility, yet one with contained, and to dispel misconceptions which few people wish to be associated. and so generate increased purchases; The underlying basis for this negativity to ensure that Australian canned foods is concern with nutritional aspects of the remain a regular ingredient in meals product. They found that: the more rather than an occasional or standby health/nutrition conscious a person, the food; and, to influence those canning more negative he or she was likely to be companies responsible for the choice of to canned food; generally younger future packaging direction in order to people were more health and nutrition create a climate of greater confidence in conscious; and, younger people were the future of the steel can. correspondingly more negative to The CFIS Inc nutrition education canned food, particularly 16–24 year program is aimed at increasing the olds. awareness of the influencers of public Similar attitudes exist amongst health opinion particularly health professionals professionals, eg the Target on healthy and media writers. Canned food has an eating which has been used by a image problem with both the public and number of State Health Commissions in health professionals. A market research public education programs places survey conducted by the fruit canning canned food generally among the fair industry in 1987 was undertaken to choices rather than excellent choices. evaluate the attitude of consumers to Examination of the nutrient canned fruit. Two of the key finding composition of canned and, where were: nutritional balance is important, possible, equivalent fresh food would some canned food is permissible but answer the question of whether these canned fruits are not very important in attitudes were justified. Unfortunately at many households because they do not the time these market research studies connote fresh fruit taste or nutrition to were undertaken there were very few the majority; and, fresh fruit is an hard analytical data available on the important part of almost every composition of Australian canned foods household's diet from the nutrition point in comparison with other available of view, but it is difficult to select food foods. pieces and as a result fruit is frequently Nutritional value of canned foods — disappointing. Study I Alison Wailes is a registered dietitian and In 1988 CFIS Inc commissioned a study nutrition consultant to the Canned Food Information Service Inc, 130 Park Street, to examine the nutrient content of six South Melbourne VIC 3205. products, including both fruit and vegetables. The analyses were carried sugar and salt to taste as appropriate. out by the Australian Government The canned products were cooked by Analytical Laboratory, South Australia, a heating half the composite sample to NATA registered laboratory. 90°C, holding for one minute then cooling. This process was chosen since canned food is already cooked during Methods processing. In general, standard methods of the This study aimed to compare the American Association of Official nutrient content of food as consumed Analytical Chemists (AOAC 1984) were after undergoing different preparation used, and full details including and processing methods. modifications to methods appear in the Products tested were: pineapple report (CFIS 1988). Riboflavin and (canned in juice and syrup, and fresh); thiamin were determined by HPLC using corn niblets (canned, fresh and frozen); fluorescence detection (Wimalasiri & tomatoes (canned and fresh); peaches Wills 1985). Carotenes were determined (canned in natural juice and fresh); by reverse phase HPLC after hydrolysis orange juice (canned and fresh); pears and solvent extraction. Minerals were (canned in natural juice and fresh). determined by atomic absorption Because results applicable to the real spectrophotometry of an acid solution world were wanted, the products tested following a dry or wet ashing. were those available on the market at the time of the study. There was no Results attempt to match cultivars if this required Table 1 derived from the study report acquiring samples not available on the (CFIS 1988) illustrates the comparison retail market at the time. Samples of between fresh, fresh cooked, canned, each product were purchased from six and canned heated peaches. different retail outlets in Adelaide The moisture levels decreased on suburbs. The six outlets included three cooking, the solids increasing supermarkets, these locations being proportionally accounting for the selected to provide a balance between increased energy (kJ) level on cooking lower and higher levels on the socio- and heating. Fat, protein and ash did not economic scale. Fresh produce differ significantly across all peach generally was of the cultivars most products tested. Vitamin C, thiamin, commonly on sale during January and riboflavin and α-carotene were all at February. All fruit and vegetables were similar levels in fresh and canned in good condition and of a ripeness peaches and both cooked and heated, ready for use. The canned products respectively. β-carotene in both canned included the major brands available on and canned heated peaches was the Adelaide retail market and were all comparable with fresh peaches. Australian grown and produced. Generic In general, there were very few labelled products were not used. significant differences in nutrient content Composite samples were prepared between fresh, fresh cooked and from equal weights of each can sub- canned and canned heated products sample after homogenisation. The studied. Often the differences that did samples of canned products were exist were within the experimental error prepared using the whole can contents. of the methods used. The exceptions The fresh food was cooked by the were vitamin C which is heat sensitive, methods considered most commonly sodium which was added at different employed using measured quantities of levels and nutrients affected by cultivar in tomato sauce with cheese were also differences. included. Because results applicable to the real world were wanted, the products tested Communication to the public were those available on the market at the time of the study. Samples of each It is notoriously difficult to communicate product were purchased from different complex scientific information to the retail outlets in Adelaide suburbs. The public. So that the information is not canned products purchased included distorted, it is important that the the major brands on the Adelaide retail message contain a single idea and is market and were all Australian expressed simply. The findings from this produced. The canned products were all first study were summarised in the prepared according to the directions on statement that: There were no practical the label. Soups were diluted with half differences between fresh and canned milk and half water and heated. The foods in most of the nutrients studied. recipes for home made soup and entree Even vitamin C was not lost on canning were obtained from standard Australian to the extent expected by many people. home economics texts. Press releases were written around Moisture and sodium determinations this single message and interest from were done on regular and salt-reduced the media was encouraging. During varieties before making composite interviews with the media it was possible samples. The composite samples each to move on from that simple statement contained equal weights of each sub- to explain more aspects of food samples after homogenisation. Total processing and its impact on the can contents were used. As in the first nutritional value of foods. study, AOAC methods of analysis were Nutritional value of canned food — generally used (CFIS 1988). Study 2 Results In 1989 CFIS Inc commissioned a The tabulated results which are second study to examine a range of available (CFIS 1989), cover prepared canned foods. Many of these proximates, carbohydrates, minerals, products have an identity in their own vitamins and fatty acid profiles. The right and are not readily compared with results also expressed the nutrients in a home cooked equivalents. defined serving size of the product. The Methods serving size has been determined from The Australian canned products tested the labels or from knowledge of current in this program were: soups: tomato usage. Because each product forms (regular and salt reduced), creamy part of a single meal, the amount of chicken (regular and salt reduced), each nutrient in the serving size has cream of pumpkin; entrees: baked been compared with the recommended beans (regular and salt reduced), dietary intake for women 19–54 years, spaghetti in tomato sauce with cheese, men 19–64 years and infants of 6 braised steak and onions; baby food: months of age. mixed vegetables (canned), berry The results showed that each product flavour yoghurt (canned), fruit of the has specific strengths and to assess the forest (refrigerated yoghurt). Homemade value of these strengths, and express recipes of pumpkin soup and spaghetti them in a more positive way, a range of criteria were chosen. The Foods Standard Code of the The Australian dietary guidelines National Health & Medical Research (Commonwealth Department of Health Council has set limits on the nutritional 1981) include the recommendation that claims which can be made about the total fat in the diet should be particular foods, and a reference reduced to less than 35% energy. quantity of the food must contain a Where NI < 1 for fat this indicates that defined portion of the daily allowance of the food will assist in meeting that goal. that nutrient. The daily allowances have Using these criteria this second study been derived from the RDIs (National has been able to demonstrate that the Health and Medical Research Council canned foods studied make a significant 1986). In summary the code provisions contribution to the RDI of nutrients for are: a food may claim the presence of adults including young people. Where vitamins or minerals if the reference canned products have been compared quantity contains one-sixth (16.6%) to with a home cooked equivalent, the one-half (50%) of the daily allowance; a differences in composition reflect food may claim to be a good source of differences in the recipes rather than vitamins or minerals if the reference differences due to processing. quantity contains not less than one-half Table 1. Proximate and vitamin (50%) of the daily allowance; and, a composition of peaches (per 100g) food may claim to be of value in the Fresh Canned Canned prevention or cure of disease due to the Nutrient Fresh lack of a vitamin or mineral if the cooked in juice heated reference quantity contains not less than Moisture (g) 87.0 85.5 85.7 84.4 100% of the daily allowance. Fat (g) 0.1 0.2 0.1 0.4 An alternative criterion is the use of Protein (g) 0.7 0.6 0.6 0.7 the Nutritional Index (NI). This is an Ash (g) 0.4 0.4 0.4 0.4 expression of the relative contribution to Energy (kJ) 170 188 177 231 the RDI of nutrients in a food compared Vitamin C (mg) 5.0 5.0 4.0 5.0 with the contribution of energy value of the food to the energy RDI. People α-carotene (µg) ND* ND TR‡ 6 understand the concept of “value for β-carotene(µg) 97 32 89 111 money” and this is similar being nutrient * ND = Not detected ‡ TR = Trace “value for kilojoules”. Promotional message concerning canned foods The example for baked beans (Figure 1) can be explained in terms of the range of criteria described above. Thus a The ratio of the percent nutrient typical statement for canned baked contribution to RDI in the serving size to beans is: Canned baked beans fit well the percent of energy to RDI in the into the Australian dietary guidelines, serving is the NI. Where NI > 1 this they are low in fat, sugar and indicates a higher nutrient contribution cholesterol-free. A 125 g serving could than energy contribution of the food in claim the presence of fibre and iron and the diet. In Figure 1 the height of the by Nutritional Index criteria is a useful bars for nutrients which are higher than source of protein, complex the energy bar indicates NI > 1 and that carbohydrate, potassium, calcium, the food is a useful source of these magnesium, zinc, and thiamin. Reduced nutrients. salt varieties are also available. Similar statements can be made Association of Official Analytical about the strengths of all the products Chemists. 1984. Official methods of tested, ie: soups, entrees and baby analysis. 14th ed. Washington DC: foods. Although according to Food Association of Official Analytical Standards Code criteria, claims about Chemist. the nutrients in these products cannot Canned Food Information Service Inc. be legally made, these criteria are useful 1988. Residual nutritional value (RNVI). in illustrating the strength of canned Research and summary report. * foods. Viewing the nutrients according to the Nutritional Index offers further Canned Food Information Service Inc. evidence of the benefits which canned 1989. Nutritional value research foods can play in a healthy diet. Thus (NVR2). Research and summary looking at the nutrient profile of a food is report.* more useful than focusing on particular nutrients in isolation. Commonwealth Department of Health. 1981. Dietary guidelines for Australians. Communication of results to the J. Food Nutr. 38: 111–9. public National Health & Medical Research The scientific information involved in Council. 1986. Recommended dietary these results is more complex than that intakes for use in Australia. Canberra: involved in the first study. It involves the AGPS. need to communicate that different products have different strengths National Health & Medical Research whether they are processed or home Council. 1987. Food Standards Code. cooked. In order to communicate this Canberra: AGPS. idea, visual bar charts showing a profile Wimalasiri, P. & Wills, R.B.H. 1985. of the food are more meaningful than Simultaneous analysis of thiamin and plain figures. It is also possible not only riboflavin in foods by high-performance to compare individual foods but to liquid chromatography. J. Chromatogr. superimpose the food profile so that a 38: 412–6. combination could illustrate all of the nutrients required. This idea still remains * Copies a2vailable on request from to be tested in the public arena. Canned Food Information Service Inc. 130 Park Street, South Melbourne VIC References 3205.

Development of a special purpose food composition data base for industry

J.A. Barnes

he 1980s have seen an ever growing To service such enquiries more Te mphasis on food and nutrition. effectively, one major manufacturing National nutrition policies have been company initiated a project to make formulated, health education programs nutrition and ingredient information on with major food components have its food products readily available to the proliferated, the media are full of public. This involved the production of a nutrition information and misinformation, nutritional information booklet generated and nutrition is increasingly becoming from product formulas and calculated by an area of concern to the Australian means of a modified nutritional consumer. As a result the demand for computation program. The problems access to accurate detailed nutritional encountered in this exercise, which information about the Australian food required close collaboration with the supply has escalated. suppliers of the program, form the basis The demand for the information has of this paper. been formalised as food legislation Project options requires full ingredient listing on packages and nutritional labelling to There were two options available to justify any nutritional claim made on generate the information, ie analysis or packaging or advertising. Requests to calculation. Considering the magnitude extend the labelling requirements range of the project (250 foods), analysis was from the anxious consumer, concerned quickly ruled out as raising too many about the quality of manufactured foods, complex issues such as sampling and to health authorities who believe that full analytical methodology, collection of access to nutrient information will samples from manufacturing outlets enable consumers to make more scattered round the country and informed decisions and thus select a interpretation of the laboratory results. more appropriate balanced diet. Further, the cost of performing analyses The penalties for not providing the would have been prohibitive. On current labelling information required, or for commercial prices, simple proximate providing incorrect information, can be analysis, plus analysis of sodium and so costly that reputable manufacturers potassium would cost about $200 per go to great lengths to ensure that the food if carbohydrate were calculated by required information is provided and is difference. A more comprehensive accurate. Indeed, manufacturers often analysis, including full fatty acid, vitamin go beyond the formal requirements, and mineral determinations would answering enquiries from the general increase the costs to between $700– public, consumer groups, health $1,000 per food. The number of professionals and government bodies samples required to obtain a meaningful regarding specific aspects of their figure to account for seasonal variations products. in all the ingredients and the differences which inevitably arise in the Jane Barnes is a consultant with manufacturing process meant that a Foodsense, 103 Cremorne Road, Cremorne project based on analysis would also not NSW 2090. meet other needs, discussed later. It was therefore rejected as a viable option information generated from a range of for the project. Calculation of the data computer programs. Proximate nutrients from ingredient information was the were found to vary by between 11% to other option and this option was 25% of their respective means when selected as being both cost-effective results were compared between eight and flexible as described later. It was systems. Quality control protocols were decided to produce the information introduced to eliminate coding brochure by computerised recipe judgements relative to portion size and calculation similar to the approach of selection of food items. However, Marsh (1983). variation between computerised Calculation systems calculation systems remained high. Differences in the nutrient data between In a comparison of available calculation data bases were considered to be a methods (Marsh 1983), particular source of the variability in the final attention was paid to the variations calculations produced in the study. found in the results due to the different A model to review nutrient data base retention and yield factors built into the capabilities was subsequently various programs to account for nutrient developed (Hoover & Perloff 1983) and changes and losses during the cooking this comprised several computing tasks. processes. No one method of One task was to verify the accuracy of calculation was found to be superior. entry of data for additional food items Indeed the manipulations involved to into the data base; another task was to account for retention and yield factors check the accuracy of calculation of a were considered too complicated to recipe comprised of nine ingredients. A incorporate into a program which would test of the review model showed that the have multiple users. Marsh's recipe calculation task produced the recommendation (1983) was to use a greater variability, the final protein simple summation method, ie addition of composition of the recipe varying from all ingredient nutrients, with a basic 9.1 to 12.7 g per 100 g. The variability computation built in to account for water was mainly attributed to differences in losses during cooking. the nutrient data in the data bases and To investigate the reasons for to differences in the yield factors used. differing computation results, Frank and In a later specific study of data base co-workers in 1984 had analysed eight development, Shanklin and co-workers representative 24 hour dietary recalls, (1985) again identified data base entries using two similar software programs as the cause of analysis variation containing the same basic USDA between programs. More recently nutrient data base extended by the Powers and Hoover (1989) evaluated individual operators. This comparison various calculation software programs. found major differences between the These workers acknowledged the calculated nutrient intakes. The major shortcomings of the summation method, contributor to the diversity in results was but their comparison did not enable the differing sources of additional them to identify and recommend a nutrient data and the criteria applied to superior method of computation. their selection. Taking this kind of comparison one step further, Hoover All the available nutrient computation (1983), reported the results of a project computer programs in Australia are carried out for the US National Nutrient based on the summation principal, so Data Bank Conference series. This selection of the program depended on project set out to compare the nutrient its capacity to fulfil other user needs of mixture of both cooked and uncooked the client. foods and as it was primarily developed to be a tool for analysis of diets as consumed, many basic uncooked foods, for which recent Australian nutrient data Additional criteria for selection of have been published were not included. data base computation program Only about 35% of the ingredients used The other requirements considered in the manufactured products could be during the selection of a suitable data located in the data base. To develop the base and analysis program for the complete nutrient data base required for project were: ease of update when the project necessitated extensive changes occurred in ingredients or additional data input. formulations; ability to handle new The first step was to draw up a list of product work particularly in the all the food ingredients used in the development of food items to meet range of products to be analysed, specific nutrient guidelines, eg the thereby identifying the gaps in the data National Heart Foundation's Food base. These groups were: those foods Approval Program (Haddy 1990); for which Australian data were available formulation of specialised food products; in the literature (eg raw meats); those tool in quality control procedures; and, foods for which published Australian facility to enable the identification of the data were available but in an presence of specific ingredients, eg inappropriate form (eg dried fruits and presence of salicylates or MSG in the vegetables for which data were only food. available on the fresh products); those The first four requirements would be met ingredients where incomplete Australian using an appropriate calculation type data were available (eg flours); and, system. To fulfil the fifth criterion a new items, for which little information program was required which had the was available on nutrient content or flexibility to extend the fields of the data availability (eg thickeners and base by at least nine. Then every food stabilisers). item listed could be coded for the Local data available presence or absence of particular The major source of published ingredients or compounds, such as Australian data was a series of papers wheat, egg, milk protein, salicylate etc. published between 1981 and 1989 (eg When a product recipe was finally Greenfield & Wills 1985). calculated a positive listing of all these specific ingredients could be produced. Dehydrated items Many of the items in the product range Package and nutrient data base were dry products (eg soup mixes) The program selected (Diet 1, Xyris which incorporated many dried Software) was specifically modified both ingredients. Factory specifications all to accommodate the extra data base generally included a moisture fields needed and to incorporate an specification for these ingredients. It additional calculation option which was therefore decided to calculate the would extend the range of foods in the dehydrated form from the hydrated data. nutrient data base, which was This was viable as only macro nutrients NUTTAB87 (Common-wealth were being considered and few changes Department of Community Services & of significance occurred to these Health 1987). This data base contains a nutrients during the dehydration process. A conversion formula was built ingredients were governed by strict into the computation program enabling ingredient specifications from the conversion of the data for any food to a company, only in a few instances were different hydration level. A similar the specifications found to be adequate calculation was not possible in products to enable the appropriate data to be with varying levels of other nutrients, eg recalculated. For example for the flours. textured vegetable proteins, moisture The diverse range of baked products and sodium levels could be found, but to be calculated employed wheats with data were inadequate for other nutrient protein levels varying from 9% in the calculations. This limited the application soft protein types to 15% in the high of the second option. The fourth option protein durum wheat types necessary was chosen but as manufacturers' data for successful pasta manufacture. were generated both by analysis and by NUTTAB only included flour with 12% calculation, it was important to protein content. Although protein levels determine which had been used in each were available for the flours, it was not case so that the data base could be possible to do a simple conversion as marked accordingly. Inevitably with the water content. judgements had to be made as to the Incomplete data best value to ascribe to an ingredient or additive. Since these judgements In the case of the flours discussed inevitably contained an element of above, appropriate figures were personal bias, it is easy to understand incorporated from food tables from other how the studies cited earlier attributed countries which had appropriate protein most of the final discrepancies to levels. Here ingredient specifications differences in the data base. As were most important to facilitate ingredients for the food industry are selection of an appropriate figure from continually being modified or developed, the consulted data. The major sources new data will always have to be of data were Paul & Southgate (1978), generated for computation programs. especially the supplements (eg Tan & The data base content is likely, others 1985), the German food tables therefore, to remain the weakest link in (Souci & others 1981) and the USDA the system. food tables (eg USDA 1989). Specific difficulties New foods for which no data exist Specific difficulties were posed by new There appeared to be four options to fill materials; by flavours and colours; by the gaps: commission analyses of the confidentiality; and by ingredient-free ingredients, or at least only accept data lists. based on analysis; generate all data by calculation from ingredient specifications New materials drawn up by the company food With the increasingly complex ingredient technologists; utilise suppliers' data, supply, new ingredients and additives even though some would be from are being developed about which little is analysis and some from calculation; or a known, particularly regarding nutrient combination of the last two mentioned. availability. A major example of these With an estimated 300 different little studied additives is the modified ingredients to be considered, the first starches. option would have been prohibitive from In the food supply of the 80s, the point of view of cost and time. modified starches became increasingly Although a large percentage of these common. Modified starches are particularly helpful in adding texture and However manufacturers, even when not organoleptic qualities to foods which inhibited by confidentiality, were often have reduced fat content. With the drive unable to supply the sodium information towards lower fat foods and the required for the data base having never sophistication of chemical technology it estimated this themselves. The figures is to be anticipated their usage will eventually used were often a mixture of increase. Therefore the question of how calculated data and manufacturers' resistant the modified starch is to estimations, supplied to satisfy our digestion in the human gut becomes requests. However, in most products the relevant and even critical in determining flavours and colours were present in its contribution to the overall nutrient such small quantities that the variations composition of a product. Particular in data would produce minimal impact products, like the gravies in the survey, on the final results. already contain significant quantities of In products such as soups where modified starches. The food labelling flavours represent a more significant regulations permit 10% variation in percentage of total weight, initial declared energy level and 20% variation calculations using the generalised in other nutrients, which may not be sodium estimations proved to be very generous enough when considering a different from five year old laboratory product with such a major contribution analysis figures for some of the from a nutrient of unknown products. The revised estimates agreed bioavailability. well with sodium analyses of the new Flavours and colours products generated by the company laboratory. Undoubtedly flavours and colours are The current demand for products with the biggest and most diverse group of lower sodium levels will generate more additives used within the food requests to flavour manufacturers manufacturing industry, and this group regarding the levels of sodium in their represented the largest group of products and increase demand for their ingredients in the project. Over the last lower sodium products. As a result it few years companies have endeavoured should become increasingly easy to to find flavours which may be listed as collect sodium data on these products. If “natural” and colours which likewise do fuller nutrient analysis calculations are not have to carry the stigma of required then further details of the “artificial”. This change has resulted in nutrient composition of the carriers will an even more extensive range of also be necessary. flavours and colours being employed. The carrier for a flavour is important in Confidentiality determining the ways in which the Part of the difficulty in obtaining flavour can be successfully used. Huge accurate ingredient data is the variations were found in sodium content importance of this information to the depending on the carrier present. manufacturer. Competition within the Initially it was assumed that savoury ingredient supply industry is fierce. flavours would contain up to 40 g Flavour companies now have vast sodium per 100 g and sweeter flavours ranges of flavours, many of which are never more than 20 g sodium per 100 g. designer flavours developed by the However, further enquiries of suppliers companies in response to individual soon revealed that in some instances requests from manufacturers. Many new, more natural, sodium-free or food technologists have individual lower-sodium carriers were available. additive preferences when building up No system is perfect. Recognition of the new product flavours. So in a large major limitations empowers the user company it was not surprising to find with the caution not to misinterpret the there were up to seven beef flavours final data. The quest for more employed which to the uninitiated information about the composition of appeared to be performing the same foods in the diet is laudable and to be function. pursued, and a data base computation The flavour manufacturers have to system can augment current knowledge. guard the uniqueness of their products, The individual figures can only ever be a to remain competitive. Often the best approximation based on the difference between one company's information available. As these tools are particular flavour and their competitior's further developed it would be hoped that is the flavour carrier. If the nutritional they become increasingly relevant and information were freely available, a useful. discerning chemist from the opposition References may well be able to work out the Commonwealth Department of formulation. So to respect the flavour Community Services & Health. 1987. house confidentiality at times an NUTTAB87. Nutrient data table for use educated guess as to the sodium in Australia. Disk or tape format. content had to be made. Canberra: Commonwealth Department Ingredient-free listings of Community Services & Health. Public interest in food sensitivities Frank, CG, Farris, RP & Berenson, GS. generates many of the requests to 1984. Comparison of dietary intake by 2 manufacturers. Consumers are computerized analysis systems. J. Am. particularly anxious to know whether Diet. Assoc. 84: 818–20. some specific ingredient of concern is Greenfield, H & Wills, RBH. 1985. A present or not in a food. In addition bibliography of UNSW studies on the some consumers are interested to know composition of Australian foods and if products contained added sugar or related areas. Food Technol. Aust. 37: animal fats. It was decided to compile a 460–1. positive listing of nine ingredients, wheat, gluten, milk protein, lactose, egg, Haddy, B. 1990. Food composition data yeast, added sugar, animal fat and and the National Heart Foundation's MSG. A listing for the presence of Food Approval Program. Food Aust. salicylates, anti-oxidants and amines 42:8, s14–15. was considered but not published at this Holland, B, Unwin, ID, & Buss, DH. stage. As each food item was loaded 1987. Cereals and cereal products. into the data base, the presence of any London: Royal Society of Chemistry. of these ingredients was recorded, eg flours all had a positive indication for Hoover, LW. 1983. Computerized wheat and gluten. Decisions had to be nutrient data bases: 1. Comparison of made about, eg whether eggs should nutrient analysis systems. J.Am. Diet. gain two positive listings, for egg and Assoc. 82: 501–5. animal fat, and, since MSG is a naturally Hoover, LW & Perloff, BP. 1983. occurring substance in foods such as Computerized nutrient data bases: 2. tomatoes, whether all tomato products Development of model for appraisal of should be listed with MSG. nutrient data base system capabilities. Conclusion J.Am. Diet. Assoc. 82: 506–8. Marsh, A. 1983. Problems associated Shanklin, D, Endres, JM & Sawicki, M. with recipe analysis. Proceedings of the 1985. Computer analysis programs. J. Eighth National Nutrient Data Bank Am. Diet. Assoc. 85: 308–13. Conference. Washington, DC: National Souci, SW, Fachmann, W & Kraut, H. Technical Information Service, US 1981. Food composition and nutrition Department of Commerce. tables 1981/82. Stuttgart: Paul, AA & Southgate, DAT. 1978. The Wissenschaftliche Verlagsgesellschaft composition of foods. London: HMSO. mbH. Powers, PM & Hoover, LW. 1989. Tan, SP, Wenlock, RW & Buss, DH. Calculation of nutrient composition of 1985. Immigrant foods. London: HMSO. recipes with computers. J.Am. Diet. USDA. 1989. Composition of foods. Assoc. 89: 224–32. Raw, processed, prepared. Handbook 8–20. Cereal grains and pasta. Washington, DC: USDA.

Food composition data use for food legislation purposes

R. Peters

his review of the use of food Consumer expectations Tc omposition data for food legislation purposes covers the following aspects: The subject of nutritional labelling of consumer expectations of food labelling foods has been given extensive (particularly nutritional and coverage in the media and the general compositional labelling); the regulatory public is becoming increasingly scene; labelling issues confronting the interested in knowing more about the consumer, the regulator, and the food it purchases and consumes, thus manufacturer; the uses and abuses of needing clear, informative food labelling food composition data; and a brief which is readily understood. In addition, perspective on possible future the initiatives and activities of various developments. government and community-based organisations have been instrumental in Food labelling raising the profile of nutritional Food labelling is a term that covers a information and food composition data, broad spectrum of information, both eg the Commonwealth Department of mandatory and voluntary, that is to be Health's dietary goals and guidelines found on packaged food. The (1981) and the recommendations of the mandatory information includes product Better Health Commission (1987). name, manufacturer's name and However, as noted by Downer (1989), address, country of origin, lot/batch these objectives were based on limited identification, net weight, and, in certain data. The opportunity now exists to instances, nutritional information. incorporate recent data to update those Additional voluntary information includes objectives. recipes, serving suggestions (pictorial To provide consumers with the representations), promotional material, opportunity to develop healthier diets by product descriptions and, of course, the purchasing foods with known nutritional brand name. contents, clearly understood and usable Ingredient labelling is the declaration information is needed on food labels. of components used to prepare food, Furthermore, this information needs to and nutritional labelling is the be consistent with the public awareness declaration of certain key constituents of programs being conducted by the food, as determined by analysis, that government and community-based are significant to metabolism. Food organisations. When national data composition data are derived from the become available about Australian detailed chemical analysis of a food. consumers' current level of Nutritional data are a sub-set of food understanding of ingredient labelling or composition data and both have a direct nutritional labelling, then strategies can link with ingredient listing. be planned to determine the nutritional information required on food labels, how it can be best presented, and how the purpose of the information can be Ross Peters is a Director with Flavourfresh lastingly communicated to all consumers Foods, 7 Kerrie Rd, Dundas, NSW 2117. in Australia, ie the dietary objectives can be matched with the label information, Special Dietary Uses, Low Joule Foods, and with the level of consumer Carbohydrate Modified Foods, Canned information. Foods for Infants and Young Children, Regulatory perspective Cereal-Based Foods for Infants and Young Children, Infant Formula, and Ingredient labelling was introduced in Low Sodium and Low Salt Content 1978 as a mandatory addition to food Foods). In addition, cholesterol labelling, although many manufacturers statements are prescribed for had already been providing this polyunsaturated oil and margarine. information voluntarily prior to 1978. The Unfortunately, there is little 1978 initiative saw the introduction of consistency in the prescribed formats for class names to describe food additives: nutritional information for these foods, mineral salts, colours, emulsifiers, food particularly within the Special Purpose acids. In 1987 these requirements were Foods category. Carbohydrate Modified complemented by the inclusion of the Foods require a percentage composition numbering system for food additives. An of the key components and an energy alternative was provided whereby the statement per 100 g; Formula Dietary number could be substituted by the Foods require a statement of the energy name of the food additive. The 1987 content consumed per day as well as initiative also provided more detail about the proportions of protein, fat and these class names, eg mineral salt carbohydrate in the food; Infant Formula (508), colour (160(a)), food acid (330); requires a table of 28 components or alternatively, mineral salt (potassium expressed per 100 mL as reconstituted; chloride), colour (β-carotene), food acid Low Sodium Foods may also have their (citric acid). The international sodium and potassium contents standardisation of additive numbers and expressed in millimoles as well as descriptions, through the Codex milligrams per 100 g. This situation Alimentarius in which Australia could obviously lead to confusion in the participates through the activities of the consumer's mind. Australian Quarantine and Inspection Service in Canberra, should assist in Labelling issues developing consumer awareness. Who needs the information? Nutritional labelling Every consumer, every man, woman Regulations covering the descriptive and child in Australia, has a need for format were agreed upon in 1986 and information about the nutritional value of cover any claims made relating to a food consumed. This need has become food's nutritional content, eg fibre more important and more complex as enriched, protein enriched (NHMRC more is known about the composition, 1989). Where a claim is made a table the metabolic value and the potential must appear describing the energy, hazards of food, and, more importantly protein, fat, carbohydrate (total and these days, there is an infinitely greater sugars component), sodium and choice of foods to consume at our finger potassium contents on per serving basis tips. and on a 100 g (or 100 mL) basis. What does the nutrition label mean? Similarly regulations incorporating nutritional declarations have been An example of the current confusion developed for Australian Standards: may exist in the consumer's minds if Vitamins and Minerals, and Special they are intending to purchase food to Purpose Foods (covering Foods for satisfy publicised needs to achieve a healthier, more balanced diet is given by the energy intake, often dependent upon sample labels for protein enriched bread the total sugar intake, will be important shown in Table 1. Bread A has more in any dietary consideration. If total protein, but more sodium, less sugar sugars are to be the preferred form for and less fat than Bread B. Both breads labelling, then most fruits and many are described as “protein increased”, but vegetables could be designated as for a consumer with a strong interest in being high in sugar content. This would purchasing the appropriate bread for be factually correct, but would require developing a healthier diet, the choice is substantial educational input to the not particularly straight-forward based population to explain this apparent on the information presented on the anomaly. label. Does the sodium content have Labelling of salt or sodium content precedence over protein content, or a lower fat content? Or should bread be Salt has been used as a method of considered a component of a broader preservation since recorded time. menu for a daily diet, or a weekly diet, or However, the relevance of salt to a life-time diet? And therefore, does it hypertension and its subsidiary side- matter if there are slight differences in effects is becoming well documented compositional data of this component and well publicised. The public health food? The critical point is, how can the concern arises through the addition of consumer assimilate the nutritional salt to foods rather than from that information presented to them on this naturally present in foods. Whilst the one food? There are many more product total sodium content of foods should be comparisons that would be equally given on a label in quantitative terms (g relevant. salt or sodium per 100 g of food), some have argued that millimoles of sodium Labelling of sugar content (or sodium chloride) can be better Sugar labelling has been a subject of integrated into a dietary objective. contention for some time. Should sugars Dietary fibre be labelled as “added sugars” or as “total sugars”? A number of natural, Dietary fibre has been a widely used unprocessed foods contain simple term in the past decade. In recent years sugars, such as lactose in milk, and there have been substantial fructose and sucrose in fruit. However, developments in our understanding of these sugars are not perceived as the term, the composition of dietary constituting health problems for the fibre, and its physiological effects on the average consumer. Indeed, general body. Traditionally, the term fibre has advice is to increase the consumption of been loosely applied to those fresh fruit for its mineral, vitamin and components of the plant cell wall which fibre content. Sugars in whole fruits are escape digestion by the enzymes of the not believed to be damaging to teeth or human intestine. More recently, a other aspects of metabolism because stricter chemical definition has been the way in which nature ‘packages’ developed which refers to non-starch sugar tends to reduce sugar contact polysaccharides, thereby eliminating with teeth and may delay its absorption minerals, lipids and lignins from into the blood (Anon 1987). consideration as fibre (Anon 1987). All Furthermore, the sugar content of fruit non-starch polysaccharide components can be highly variable depending upon escape digestion in the human small the cultivar and stage of maturity. intestine, moving on to the colon. This Further, from a nutritional point of view, definition of fibre could satisfy those who are concerned with regulatory aspects Table 1. Nutrition labels for two types including scientific accuracy, and those of protein enriched bread (per 100g) concerned with nutritionally meaningful Nutrient Bread A* Bread B** definitions. However, it is becoming Energy (kJ) 985 980 evident that the full meaning of dietary Protein (g) 9.3 9.2 fibre is yet to be unravelled. Fat (g) 1.9 2.2 Manufacturers are bound to label food Carbohydrate —total (g) 44.5 46.6 according to regulations, but is the —sugars (g) 0.2 2.5 information on the label meaningful to Dietary fibre (g) 3.4 not stated the consumer in terms of needs, and Sodium (mg) 605 432 metabolic effect? Potassium (mg) 136 112 * serving size 29.6 g Labelling of fat content in food ** per slice 28.0 g The drive for healthier food has seen Through media advertising, consumers the description “natural” debased to a are becoming more aware of the need point where it is difficult to know whether to monitor intake of polyunsaturated it really has the meaning the public fats, mono-unsaturated fats and expect. The term “all natural” is often cholesterol. As knowledge increases so used to indicate that the ingredients do the concerns for various aspects of used to prepare the product are natural. fats and fat-soluble components of our However, in an instance where a diet. The point comes through again, do product is highlighted as “all natural” but we have the necessary information to its ingredient statement lists potassium provide consumers with a clear chloride and calcium chloride, is the understanding of their dietary needs? claim “all natural” justified? These Uses and abuses of food particular substances exist in nature, but composition data many may not be satisfied that they are There may be need to adopt a natural. The critical issue is the need to “helicopter” viewpoint in addressing the have clear understanding among issues surrounding food labelling, so as consumers of the contribution to the diet to avoid the divisions among the many of ingredients and additives approved groups, each with a well planned for use in food, and their relation to the agenda to achieve defined objectives, nutritional goals for Australians. though often with a narrow perspective. The future for nutritional advice on The public is being bombarded with bits food labels and pieces of information concerning The issues surrounding information their diet and the opportunity to abuse presented to consumers through labels, food compositional data then arises. advertising and publicity are complex. The abuse may take the form of The need to know more about the food selective promotion of key nutritional we eat is undisputed. The problem lies groups in isolation from the broader in how to convey the necessary perspective of a balanced diet. The nutritional information in a form that is reverse also applies in the case of readily understood and that can be negative claims. The manufacturer must assimilated into the wider perspective of respond to the changing food better health and well-being, bearing in information environment and provide mind that the ultimate choice is that of pertinent statements or data that will the consumer. The difficulty with a maintain or develop market share for numerical tabulation of nutritional data certain products. on labels is that the consumer not only has to understand the units in which the developed accordingly. Access to nutrient is expressed, but then has to considerable food data is now possible collate the information on several but there may be little benefit in nutrients depicted in different terms, eg presenting such data through food kilojoules, grams, milligrams, millimoles. labels unless consumers can clearly If this has been achieved, consumers understand them and appropriately can then choose the brand of food collate and assimilate the information which conforms most closely with their against dietary objectives. There needs wishes. But then the consumers to be a national approach to address all somehow have to integrate information the issues to reduce the potential for on all their food purchases if they wish misunderstanding and abuse of food to maintain their diet in accordance with data. Meanwhile, there is a real risk that established guidelines. The integration consumers will become totally apathetic of such information has been towards the nutritional information recognised for some time, both prescribed by regulation on food labels. internationally and in Australia, and has References led to the statement that there is no such thing as an unhealthy food and Anon. 1987. Nutritional labelling of that the balance of food choices is the foods: a rational approach to banding. key to an appropriate diet. London: The Nutrition Advisory This leads to the opportunities for the Committee, The Coronary Prevention future. Perhaps the pictorial Group. representation of nutritional information, Better Health Commission. 1987. as is being developed in the UK, has Towards better nutrition for Australians. merit. However, it may be more Canberra: AGPS. pertinent at this moment to determine what Australians currently understand Commonwealth Department of Health. with respect to food, nutritional and 1981. Dietary guidelines for Australians. compositional labelling and what J. Food Nutr. 38: 111–9. consumers want to and need to know. Downer, A.H. 1989. Dietary goals, Such a survey would need to be guidelines and guesswork. Food Aust. appropriately detailed to cover the 41: 916. spectrum of the Australian population, and structured in such a way that the National Health & Medical Research gaps in understanding can be identified, Council 1989. Food Standards Code. and future regulations, public awareness Canberra: AGPS. campaigns and advertising guidelines

The New Zealand food composition data bases

B.A. Burlingame

SIR Biotechnology is the site of foods. Each has been designed for DNew Zealand's national nutritional specific purposes with unique features data bases and one of its main centres and capabilities. for nutrient compositional analyses. The The main database of food and feed data base project has been developed and maintained in conjunction with a The Main database was developed systematic nutrient analysis project. using recommendations and design Both are ongoing, with constant criteria from the International Feed updating. This overview of New Information Centre (INFIC) (Lelystad, Zealand's direction in the area of The Netherlands). Data on both human nutritional information systems covers foods and animal feeds are entered into the information structure, groups this data base and the staff are able to involved and their roles, work completed search the data base at the request of a to date, problems encountered, and client. New Zealand is possibly the only future prospects. country associated with INFOODS which developed its food composition Information structure data base from a foundation of an Nutritional information systems require animal feed data base. This did have involvement at international, regional some advantages, including on-site (Oceania) and national levels. As does personnel familiar with computerised Australia, New Zealand participates in a nutritional data systems. United Nations University program The NZ food composition database called INFOODS (the International Network of Food Data Systems), The NZ food composition database headquartered at Harvard University, contains nutrient composition on 1122 USA. The purpose of this group is to human foods in 22 categories. Of these make recommendations and establish foods, 42 percent are based on New protocols for the compilation of Zealand source data and five percent nutritional information systems. An early are based on Australian data from activity of INFOODS was the literature sources (Wills & others, establishment of regional groups 1981a, 1981b, Wills & Greenfield 1982, throughout the world. New Zealand Greenfield & others 1981, 1982) and became part of OCEANIAFOODS, appropriately referenced in the user's along with Australia, member countries guide (Burlingame & Milligan 1989). The of the South Pacific Commission, and remaining food records are made up of Papua New Guinea. British data from three sources (Paul & There are three New Zealand Southgate 1978; Paul & others 1980; nutritional data bases: the Main food Wiles & others 1980). A copyright and feed database, the NZ food arrangement for the use of some of composition database, and the NZ these data has been entered into with therapeutic database of brand-name HMSO and royalty payments are made by DSIR based on numbers of computer Barbara Burlingame is the Food & Nutrition and printed copies produced for sale. Programme Leader, Department of Scientific and Industrial Research, Private Bag, The data are available for purchase Palmerston North, New Zealand. in the form of the printed NZ food composition tables (Milligan & others inclusion in the users' applications 1989), and computer products called software or data base management FOODfiles and FOODsearch. An system and is supplied as a set of ASCII example of a typical food record in the files on either standard half-inch NZ food composition tables is given in magnetic tape or 5.25" floppy disks. Figures la and 1b. In this example, FOODfiles requires computing expertise nutritional information for wholemeal for installation. flour, 100% extraction rate, is presented FOODsearch, a ready-to-use as amounts per 100 g edible portion. software package, is presently being The first part of the record includes the developed. This product contains all the proximates, elements, vitamins and data files plus application software with cholesterol. The next two parts show capabilities for selection, display and amino acid and fatty acid data. The final calculations using user-defined scaling part of a food record presents additional factors. Average daily nutrient intakes of information and various record identifier individuals and populations can be codes. In the example in Figures 1a and determined, and compared to RDIs of 1b, additional information includes several different nations. Totals will be different measures of dietary fibre and flagged to indicate where nutrient values its important constituents. are missing from one or more of the The first edition of the printed version food records used in the calculation. is presently available as an unabridged (Information on FOODfiles and reference book, complete with statistical FOODsearch may be obtained from the variation, number of samples associated author). with the mean values and country The NZ therapeutic database of source data. It has 1600 pages and is brand-name foods similar in format to the US Department of Agriculture Handbook No. 8 series The NZ therapeutic database of brand- (USDA 1976–89). An abridged version name foods is different from traditional will be ready by mid-1990 which will food composition data bases and tables contain nutrient mean values only and in that it provides information on brand- come in a form similar to the British name foods only, and provides much tables (Paul & Southgate 1978). more information than simply the FOODfiles is also presently available. nutrient composition of the foods. This product consists of data files for FLOUR, WHEAT, 100% E. R., Wholemeal Amount in 100g edible portion Units Mean Std error No. Src. PROXIMATES Water g 12.31 0.17 62 z Energy kcal 305 - - c Energy kJ 1274 - - c Protein (Nitrogen × 5.83) g 11.84 0.11 38 z Total fat g 1.84 - 1 z Available carbohydrate g 64.15 - 1 z Dietary fibre g 8.56 - 1 z Ash g 1.57 - 2 z NUTRIENT ELEMENTS Sodium mg 8.34 - 2 z Magnesium mg 142 - 2 z Phosphorus mg 398 - 2 z Sulphur mg 136 - 2 z Chloride mg 38.7 - - b Potassium mg 529 - 2 z Calcium mg 59.3 - 2 z Manganese ug 4340 294 6 z Iron mg 5.42 - 2 z Copper mg 0,532 0.056 6 z Zinc mg 3.15 0.51 6 z Selenium ug 1.24 0.21 7 z VITAMINS Retinol ug 0 - - b Carotene ug 0 - - z Total vitamin A equivalents ug - - - - Thiamin mg 0.467 0.049 4 z Riboflavin mg 0.118 0.012 4 z Niacin mg 7.26 0.05 3 z Potential niacin from tryptophan mg 2.05 - - z Vitamin B6 mg 0.510 - - b Pantothenate mg 0.816 - - b Biotin ug 7.14 - - b Folate, total ug 58.1 - - b Vitamin B12 ug 0 - - b Vitamin C mg 0 - - b Vitamin D ug 0 - - b Alpha-tocopherol mg 1.5 - 1 z Vitamin E mg - - - - OTHER LIPIDS Cholesterol mg - - - -

g/100g edible portion mg/g Nitrogen

AMINO ACIDS Mean Std error No. Src. Mean Std error No. Src. Isoleucine 0.391 - 1 z 192 - - z Leucine 0.759 - 1 z 374 - - z Lysine 0.322 - 1 z 159 - - z Methionine 0.180 - 1 z 89 - - z Cystine 0.275 - 1 z 135 - - z Phenylalanine 0.511 - 1 z 252 - - z Tyrosine 0.350 - 1 z 173 - - z Threonine 0.341 - 1 z 168 - - z Tryptophan 0.123 - 1 z 61 - - z Valine 0.502 - 1 z 247 - - z Arginine 0.568 - 1 z 280 - - z Histidine 0.256 - 1 z 126 - - z Alanine 0.436 - 1 z 215 - - z Aspartic acid 0.625 - 1 z 308 - - z Glutamic acid 3.504 - 1 z 1726 - - z Glycine 0.474 - 1 z 233 - - z Proline 1.174 - 1 z 578 - - z Serine 0.606 - 1 z 299 - - z Figure 1a. Sample record from the NZ food composition tables (Milligan & others 1989).

Dietitians recognised that they on ingredients, nutrient composition, required additional information about manufacturer and product details, and food which would not be included in the the presence and absence of clinically NZ food composition database and significant food constituents (Figure 2). tables. At a joint meeting of the New Nutrient composition data are subject to Zealand Dietetic Association and the quality assurance checks. Each Nutrition Society of New Zealand in individual entry requires disclosure of 1984, a working group convened to whether the data were obtained from discuss their needs and thus conceived direct analyses, calculated from food the idea of the therapeutic database. composition tables, or from unknown This brand-name foods database means, by ticking the boxes on the form contains industry-supplied information labelled A, C, and U, respectively.

FLOUR, WHEAT, 100% E. R., Wholemeal g/100g edible portion g/100g total fatty acids

FATTY ACIDS Mean Std error No. Src. Mean Std error No. Src. 4:0 ------6:0 ------8:0 ------10:0 ------12:0 ------14:0 0.003 - - z 0.23 - - z 15:0 ------16:0 0.245 - - z 18.48 - - z 17:0 ------18:0 0.019 - - z 1.43 - - z 20:0 ------22:0 ------24:0 ------Total saturated 0.267 - - z - - - - 10:1 ------14:1 ------15:1 ------16:1 ------17:1 ------18:1 0.204 - - z 15.39 - - z 20:1 ------22:1 ------Total mono-unsat 0.204 - - z - - - - 18:2 0.711 - - z 53.62 - - z 18:3 0.138 - - z 10.41 - - z 18:4 ------20:2 ------20:3 ------20:4 ------20:5 ------22:2 ------22:4 ------22:5 ------22:6 ------Total poly-unsat 0.849 - - z - - - -

Additional information (in 100g edible portion) Units Mean Std error No. Src. Dry matter g 87.69 0.17 62 z Total nitrogen g 2.03 0.02 38 z Total available sugars g 3.77 - 1 z Starch g 60.38 - 1 z Hemicellulose g 7.15 - 2 z Pectin g 0.35 - 1 z Cellulose g 1.63 0.21 3 z Lignin g 1.04 0.32 3 z Neutral detergent fibre (Van Soest 1967) g 8.56 - 1 z Crude fibre (AOAC, 1970) g 2.41 0.07 16 z Boron ug 492 - 2 z Chromium ug 8.02 2.08 4 z Cobalt ug 1.14 - 1 z Nickel ug <72.7 - - z Molybdenum ug 18.3 - 2 z Cadmium ug 5.86 0.62 4 z Delta-tocopherol mg 0.158 - 1 z Beta- and gamma-tocopherol mg 2.98 - 1 z Density kg/1 0.690 - - 1c

FOODDATA Biotechnology Division DSIR Palmerston North N.Z. E16 (8:9 Z:00391) W1.0 31/01/1989 E. 26 Figure 1b. Sample record from the NZ food composition tables (continued).

Groups involved development of these databases by making grant money available, including Three main groups have been involved the NZ Heart Foundation, Cancer in important aspects of this work: the Society of NZ, Lottery Board, DSIR (Biotechnology Division and Palmerston North Medical Research Chemistry Division), the NZ Health Foundation and the NZ Dietitians' Department, and the Food Composition Association. Steering Committee. Other organisations have contributed to the

Figure 2. The therapeutic database of brand-name foods data entry form. Two sections of the data entry form are shown: presence/absence of clinically significant food constituents and food additives, P = present, A = absent, U = unknown; and nutritional information, A = analysed, C = from food composition tables, U = unknown.

DSIR Biotechnology has been the resources such that the nutrient primary site for organisation, and has analyses conducted according to the played an important role in funding, guidance of the Steering Committee will sampling, analyses, and presentation of ultimately replace all British data with the data. Chemistry Division has also New Zealand source data; maintaining been involved in sampling and analyses, the validity and integrity of the data particularly vitamin analyses. The Health selected for inclusion into the data Department historically has funded bases; satisfying the wide mix of data much of the work and health inspectors base users with the end products; and have assisted with the sampling of foods sustaining funding from both the DSIR from different regions of the country. and the Health Department as the Finally, the Food Composition Steering usefulness of all the data bases Committee has had the role of advising depends entirely on continual revision on various aspects of the work including and updating. the setting of food category priorities, Future selection of food constituents to measure, and procedures for sampling. The maintenance and upgrading of The Committee, which over the years nutrient composition data bases and has included nutritionists, dietitians, tables will carry on with continued health department officers, food support from the NZ Health Department technologists, and academics in food and the Department of Scientific and technology and human nutrition, has the Industrial Research. Analytical work will obligation to review the information continue to focus on major categories of needs of nutrition professionals and foods. Currently, bread and flour consumers, and consider these in analyses are being conducted and relation to the food composition work. future categories will include beef, lamb and meat products, traditional Maori and The composition of New Zealand Pacific Island foods, and fast foods. foods Supplements and addenda to the Most of the New Zealand source data printed and electronic data bases will be come from the systematic nutrient issued periodically, the first of which will analyses of foods, begun at be on breads and flour. Biotechnology Division in 1980, and Specialised data bases and printed conducted in collaboration with products will also be developed. Chemistry Division. To date, Educational materials for secondary comprehensive nutrient data have been school students are being prepared obtained for New Zealand produced jointly by staff of DSIR and a team of fruits, vegetables, dairy products, fish home economics teachers convened by and cereal grains. In addition to the the Education Department (now the systematic nutrient determinations, Ministry of Education). Specialised other data were collected from research printed products include the continuation laboratories around the country and of the soft-cover book series called the incorporated into the data base after Composition of New Zealand foods meeting all the established criteria for (Visser & Burrows 1983). Books two high quality data. (export fruits and vegetables) and three Problems (milk and milk products) will be published in early 1990 and the first The challenge of these projects are draft of book four (bread and flour) is in many and include managing the limited preparation, to be published in late Zealand food composition tables. 1990. Palmerston North, NZ: FOODDATA In New Zealand, as in many other Biotechnology Division, DSIR. countries, nutritional information is Paul, AA & Southgate, DAT. 1978. The important for many sectors of the composition of foods. London: HMSO. population. The future will see this information required and put to use by Paul, AA, Southgate, DAT & Russell, J. agriculturalists and geneticists in the 1980. First supplement to McCance and development of new breeds and strains; Widdowson's The composition of foods. business personnel involved in the London: HMSO. marketing of food products; food US Department of Agriculture. 1976–89. technologists in the development of Composition of foods. Raw, processed, new, healthier manufactured foods; prepared. Handbook 8 series. dietitians, epidemiologists and other Washington, DC: USDA. health practitioners in advising on and evaluating food intakes of individuals Visser, FR & Burrows, JK. 1983. and populations; policy makers in Composition of New Zealand foods 1. advising national bodies on nutritional Characteristic fruit and vegetables. goals and guidelines and on product DSIR Bulletin 235. Wellington, NZ: labelling; and most importantly by DSIR. consumers making wise purchasing Wiles, SJ., Nettleton, PA, Black, AE & decisions. Paul, AA. 1980. The nutrient References composition of some cooked dishes Burlingame, BA & Milligan, GC. 1989. eaten in Britain: a supplementary food The New Zealand food composition composition table. J. Hum. Nutr. 34: tables user's guide. Palmerston North, 189–223. NZ: DSIR Biotechnology. Wills, RBH & Greenfield, H. 1982. Greenfield, H, Chuah, LK & Wills, RBH. Composition of Australian foods. 18. 1981. Composition of Australian foods. Foods from Kentucky Fried Chicken. 12. Hamburgers. Food Technol. Aust. Food Technol. Aust. 34: 566–9. 33: 619–20. Wills, RBH, Makinson, H. & Greenfield, Greenfield, H, Wimalasiri, P, Ma, SNN & H. 1981a. Composition of Australian Wills, RBH. 1982. Composition of foods. 9. Meat pies, sausage rolls and Australian foods. 16. Foods from Pizza pasties. Food Technol. Aust. 33: 450–4. Hut restaurants. Food Technol. Aust. Wills, RBH, Wimalasiri, P & Greenfield, 34: 364–7. H. 1981b. Composition of Australian Milligan, GC, Webster, DW & foods. 5. Fried take-away foods. Food Burlingame, BA. 1989. The New Technol. Aust. 33: 26–7.

Recommendations The Symposium resolved that the 5. The Symposium noted that following list of recommendations which specialised needs for data on non- were produced by workshops and nutrients in foods existed but believed agreed to in plenary session should be that top priority should still be given to transmitted to Federal and State analysing more foods for the existing Ministers of Health: nutrient profile as soon as possible. The priorities for foods on which nutrient 1. The Symposium commended the data are needed are: non-alcoholic Commonwealth Department of beverages; dairy products; Community Services and Health highly confectionery; take-away foods (non- for the production of the new official branded); ethnic group foods; baby Composition of foods, Australia (COFA) foods; infant formulas; sauces and (Cashel & others 1989) and the related pickles; and frozen packaged meals. computer data base NUTTAB 6. The Symposium agreed that the (Commonwealth Department of priorities for nutrient data should be as Community Services & Health 1989). follows: proximate data and a This publication and the ongoing carbohydrate profile should be associated program were recognised to generated on all foods and top priority be of indispensable value for the should be given to this; other nutrients promotion of better health in Australia should be prioritised on a public health and essential tools for the ongoing needs basis; where analytical data monitoring of the nutritional quality of cannot be obtained by reason of the national food supply. expense, lack of facilities etc, data 2. The Symposium recognised the should be estimated from overseas data need for a commitment of ongoing for similar foods; calculation factors are resources to secure the future of the needed for nutrient losses and gains program, and that these resources not during processing, storage and only be for continuing analytical work preparation; a method to indicate but also, and most importantly, for the bioavailability of nutrients in particular qualified staffing resources required for foods is needed for compiled data; the scrutiny and management of the research into methodology for nutrients analytical data in the production of the such as folate, vitamin B12' vitamin B6 printed tables and computerised data and biotin should be carried out. base. The need for printing and 7. The Symposium urged that a computer costs was also recognised by mechanism be found for laboratories the meeting. commissioned under the Australian 3. The Symposium noted that many of analytical program to publish their data the analytical data which had been independently on completion of analysis produced under the program were still of particular groups of foods, in awaiting release and urged that high recognition of the delays necessitated priority be given to releasing these data by the data management and printing as additional volumes of COFA as soon processes in production of the printed as possible. tables and NUTTAB. 4. The Symposium also noted the 8. The Symposium noted the use, complexity and size of the first volume potential use and potential for abuse of of COFA and urged that a set of nutrient data in the labelling and simplified tables for consumer use be promotion of foods, and recognised the produced as soon as possible. need for consumer education programs to enable the public to use the AUSTRALIAN JOURNAL OF information to the best advantage in NUTRITION AND DIETETICS terms of health and value for money. The Journal of the Dietitians 9. The Symposium noted the size and Association of Australia price of the first volume of COFA and the availability of NUTTAB. It stressed The Australian Journal of Nutrition the need for a review of the pricing and Dietetics is a scientific quarterly structure of such an essential public publication with original contributions information resource; the desirability of providing a forum on all aspects of a package comprising tables and food, nutrition and dietetics, diskette to be available for purchase; the including: need for the tables and NUTTAB to • human nutrition conform with each other; the need for • food science and technology the tables to contain the additional • diet in prevention and treatment of feature of printed lists of foods by disease nutrient so that the comparative value of • food service foods could easily be recognised; the • nutrition education need to extend the computer format; • community nutrition and the usefulness of an independent The Journal was first published more glossary of food names. than forty years ago by the 10. The Symposium noted that the Department of Health, and was original advisory committee structure for handed over to the Dietitians the program had been of necessity Association of Australia in 1987. It is modified over time, and highlighted the well accepted throughout Australia, need for groups of end users to be and is the only Australian scientific requested to make formal inputs into the journal specialising in nutrition and program in the future. dietetics. The Australian Journal of Nutrition References and Dietetics welcomes reports, Cashel, K, Lewis, J & English, R. 1989. original research, reviews, short Composition of foods, Australia. articles and letters expressing Canberra: AGPS. responsible viewpoints for Commonwealth Department of consideration for publication. The Community Services and Health. 1989. Journal also includes advertising. NUTTAB89. Nutrient data table for use Subscription for Volume 47 (4 in Australia. Disk format. Canberra: issues), 1990 AGPS. Within Australia $40.00 Overseas surface mail $50.00 Contributions and subscriptions may be sent to: Australian Journal of Nutrition

and Dietetics PO Box 11, O'Connor, ACT 2601,

Australia.