Foods and nutrients in the Australian Food Composition Database

There are so many foods available to eat in Australia that it isn't possible to generate nutrient data for every single one. The Australian Food Composition Database primarily includes foods which are most commonly eaten by Australians or used as ingredients in other foods. A food may also be included if it contains a lot of a particular nutrient or if high quality data was available.

Each food published in the Australian Food Composition Database is assigned a unique numeric identification code called the Public Food Key (the Key). The Key was first introduced in Release 1, and replaced the previously used Food ID. The Key remains the same for each equivalent food across our food composition publications. For example, full fat cow's milk uses the Key 'F005634' in Release 1, Release 2 and in all future releases. The same Key will also be used in any future AUSNUT or Nutrition Panel Calculator (NPC) publications.

The Key helps to compare nutrient data for the same food over publication releases. It also allows for more streamlined updates for users of our database.

Foods published in the Australian Food Composition Database are assigned a name which is used to describe the product and to help users identify the correct data. The Food Name aims to provide key details about the food, capturing the most commonly available form of the food and, where relevant, its preparation. Exceptions to the commonly available form of the food may also be included. For example, sugar-sweetened soft drinks are simply referred to as 'soft drinks' whereas the intense-sweetened versions are referred to as 'soft drink, intense sweetened'. In situations where a food's common form of availability is not obvious, nutritionally relevant information is included; for example, boiled white rice is referred to as either 'boiled, with added salt' if salt is included in preparation, or 'boiled, no added salt' if no salt has been included.

Where vitamins and/or minerals are added to a food for fortification purposes, this is generally identified in the Food Name. Where a food is always supplied in a fortified form, such as bread-making flour with the mandatory addition of thiamin, folic acid and iodine, this information is not included in the food name but is included in the Food Description.

Very few brands are used in the Food Name. This is because the formulation of specific products changes over time and nutrient levels at the time of analysis may not reflect those in a particular brand some years later. In the few cases where a specific brand is mentioned, this is generally intended to provide guidance for the user in situations where there are a number of products available with similar appearance but with differing nutrient composition. The values reported should be regarded as reflecting the average composition of that class of food. If you require information on the nutrients in a specific product as currently available, you should check the product's nutrition information panel or consult the manufacturer. Very few foods are analysed as a single brand only. The exceptions include Vegemite„ and Milo„ and some breakfast cereals.

Scientific names of foods of plant or animal origin have been provided, where known. The scientific name generally comprises two parts, the genus name and the species name, and is provided to assist in clarity and avoid confusion. Common names may vary across different countries and even within different parts of the same country. In contrast, the scientific name remains the same all over the world and in all languages, avoiding confusion and difficulties of translation.

The food description provides more detail about the product, including its appearance, texture, production and preparation. For processed foods, the major ingredients and food additives used are provided, where known.

Each food in the Australian Food Composition Database is assigned a 5-digit classification code.

• The first two digits refer to a major food group used to describe similar products e.g. 12 Cereals and Cereal Products which includes a range of products such as grains, breads, breakfast cereals etc.
• The first three digits together refer to a sub-major food group under the broader 2-digit category e.g. 125 Breakfast cereals, ready to eat.
• All five digits together represent a minor food group which is the most specific level of grouping e.g. 12501 Breakfast cereal, corn based.

The food groups and food sub groups used in the Australian Food Composition Database publications are the same as those used for AUSNUT 2011-13.

The specific gravity is provided in the Food File and the online searchable database. It refers to the density of a product. It is calculated from the ratio of the mass of 1 mL of a food or beverage to that of 1 mL of water.

i.e. specific gravity (g/mL) = weight (g) of 100 mL of food or beverage / 100

You can use this information to estimate nutrient levels in 100 mL of a beverage or other liquid food. Liquid foods published in the Australian Food Composition Database are also presented based on a per 100 mL basis. Note that all solid foods have been allocated a specific gravity of 0 by default.

The majority of nutrient data published in the Australian Food Composition Database is analysed data. A small proportion of data comes from overseas food composition tables, the food industry, recipe calculations, food label information and imputing from similar foods. We use derivation codes to describe the source of the data.

Even though each food and beverage is assigned an overall derivation code, individual nutrient values for some foods may have been derived using a different technique. For example, a food described as being Analysed may have a small number of nutrient values that were imputed from similar foods. Although the derivation for each individual nutrient value is not systematically presented, the Sampling Details field often contains this information using the same terms as below to describe the data origin.

Analysed (1005 foods)

The majority of nutrient values presented in the Australian Food Composition Database have been determined by laboratory analysis of foods purchased in Australia. Most of the older analytical data is derived from foods which have been purchased in one or more capital cities, generally Sydney, Melbourne or Adelaide. However, for more recent analytical programs, samples have been purchased nationally. For packaged foods, four to eight separate purchases are usually chosen to reflect the market composition at the time of analysis. For unpackaged foods, generally six to 12 purchases are made. In nearly all cases, the purchased items are mixed together to form a single analytical sample (or 'composite') that reflects the average composition of that type of product at the time the sample was prepared. This method does not provide information on the variation of nutrient levels between samples. In the case of samples purchased for the 2006 and 2008 Key Foods Programs and the 19th, 20th and 22nd Australian Total Diet Studies, values are means of sub-samples purchased nationally, with the sub-samples analysed separately.

For some foods, particularly major foods such as breads, a number of analytical programs may have been conducted over time. The results presented in the Australian Food Composition Database are usually the average results of these programs. However, where we know that changes have occurred to the formulation, processing or growing conditions of a food, or where an improved analytical method is now available, only newer data is reported.

A small number of nutrient values in records with a derivation code of Analysed may have been determined by other techniques, such as imputation or borrowing, where analytical data for these nutrients was not available.

For a number of foods in the Australian Food Composition Database, Australian food companies and organisations have provided us with nutrient data for their products. This data is largely generated by analysis. The data is incorporated into our food composition databases in accordance with our standard validation methodologies.

Recipe (508 foods*)

A number of foods reported in the Australian Food Composition Database are 'recipe' foods. For these foods, an average recipe for the food, as commonly prepared in Australia, is developed and the overall nutrient profile for that food is calculated from the nutrient data for the individual recipe ingredients. The recipe also takes into account, where necessary, loss or gain of moisture and nutrients during processing. Examples of recipe foods include most toasted breads, prepared cordials and some home-prepared traditional foods such as Anzac biscuits.

Information on the recipes used in the Australian Food Composition Database including weight changes and nutrient retention factors are available in the Recipe File and Retention Factor File. These factors are used to adjust the nutrient profile of a food based on assumptions about moisture gain or loss and nutrient loss during preparation.

* Note that a large number of red meat records (210) are derived using a recipe approach based on analytical data for the separated lean meat and fat portions and gross composition data measured by the laboratory for each meat cut. The lean meat and fat data is then combined based on the laboratory analysis of the proportion of each in different cuts when all fat present at purchase is retained (untrimmed), and when fat on the exterior is cut off (semi-trimmed). This helps to reflect the varying amounts of fat that may be removed between purchase and eating.

Borrowed (61 foods)

Small amounts of data have been borrowed from food composition tables published by the governments of the United States (USDA, 2008-2018), the United Kingdom (Food Standards Agency, 2002-2015), New Zealand (New Zealand FOODfiles 2018, Lesperance, 2009, Athar et al, 2003), Denmark (Mƒ¸ller et al, 2005), and Singapore (Ministry of Health, 2000). Data have also been included from the Tables of Composition of Australian Aboriginal Foods (Brand Miller et al, 1993). Individual nutrient values may have been obtained by borrowing to fill important data gaps in an analysed nutrient profile. In general, nutrient data is only borrowed from overseas food tables where the food is imported into Australia or where it was considered there was a need for the nutrient data but suitable Australian data could not be identified.

Imputed (24 foods)

Imputation is the process of assuming that a nutrient value in one food can represent that in another similar food. For example, in the case of salmon canned in water with no added salt, nutrient values other than sodium and chloride will be imputed from those for salmon canned in brine. Imputation also includes the process of assuming that some foods contain none of a particular nutrient, based on knowledge of the composition of that food. Imputation has only been used where we have confidence in the validity of the assumptions made. It is more commonly used for particular nutrients in a food rather than for a food as a whole. For example for soft drinks, vitamin E has been imputed as zero as soft drinks contain no fat (vitamin E is a fat soluble nutrient) and are not labelled as containing added vitamin E.

Label (13 foods)

A small number of records contain nutrient data derived from label information. Label information has been included where:

• no analytical or other appropriate data were available for that food
• the food was considered significant in the diet for all or some of the population and
• there have been known changes to production practices such as fortification, since the original data were generated.

Values presented are, wherever possible, averaged over a number of brands and taken from the nutrition information panels for these products. Because nutrition information panels are only required to report nutrient data for a small number of nutrients, some values in these records may have been derived by other techniques such as imputation.

Calculated (5 foods)

Some records, particularly for cooked wine and dry noodles, have been calculated using techniques similar to the recipe approach described above, but without generation of a formal recipe. This method is used in the limited cases where the standard recipe generation method does not account for factors such as the loss of alcohol or substantial moisture changes.

Further information about where the nutrient data comes from, such as the number of samples purchased for analysis, the date and place of purchase, whether the data was imputed or borrowed etc, can be found in the sampling details field. This field may also highlight any specific issues with particular nutrient values, such as analytical problems, areas of uncertainty, or cautions regarding a sum of proximates (macronutrients) outside the expected range.

The Australian Food Composition Database reports up to 256 nutrients and food components per food. The range of nutrients presented for each food varies depending on the analytical data available. However, there is a core set of 54 nutrients for which all foods have data. These nutrients align with those reported in AUSNUT 2011-13 (a set of files that enables food, dietary supplement and nutrient intake estimates to be made from the 2011‒13 Australian Health Survey), plus vitamin D.

It's important to note that if a food does not have a value for a particular nutrient that does not mean that the nutrient is not present in the food. It means that analytical data for that nutrient in the food is not currently available to us.

Information relating to each nutrient including equations published in the Australian Food Composition Database can be found with the online searchable database as part of the 'Browse nutrient list' function.

The nutrient data is available as both an online searchable database and as an Excel file.

In the online searchable database, nutrient values are reported per 100 g edible portion for solid foods and per 100 mL edible portion for beverages and other liquid foods such as salad dressings. For the first time, the nutrients for some foods are also reported by common serve size in addition to per 100 g/mL values to reflect the nutrient content of what a user may have eaten. For example, one serve of a soft drink may be a 375 mL can.  

In the Excel file, nutrient data are reported in two ways:

TAB 1   Per 100 g – all foods and all beverages are reported per 100 g edible portion

TAB 2   Per 100 mL – beverages and other liquid foods only, reported per 100 mL edible portion.

In general, the analytical techniques used are widely-accepted methods conducted by experienced laboratories with National Association of Testing Authorities (NATA) accreditation. However, because nutrient data reported in the food composition databases have been generated over approximately 40 years, the techniques used for nutrient analysis may have changed over this time. See details of nutrients, including analytical methods or equations for equated components by downloading the nutrients details file. Equated components are values for nutrients that we have calculated using specific equations from other analysed components for that food. For example energy content is equated from protein, fat, carbohydrates and other nutrients. Further information on the methods of analysis, including limits of detection and reporting may be available from us on request.

Sometimes when foods are analysed, the result for a particular nutrient may be less than (represented by the symbol “<") the limit of reporting (LOR).  These are sometimes referred to as “trace" values. Taking protein as an example, if a value is reported as <0.2 g/100 g, this means that the analytical method used to analyse protein in this food is able to quantify levels of protein at or above 0.2 g/100 g. In this case, the analytical method was able to detect a small amount of protein, but not enough to determine exactly how much.

For data analysis purposes we need to decide how we will use these <LOR values. In Release 1, <LOR values are reported as zero. In our data management system, we record that a “less than LOR" value was reported and note the LOR specified by the laboratory. This allows us to assign different values to these “less than LOR" values as the need arises. For example, there may be occasions where assigning all “less than LOR" values to half of the LOR or setting them to be equal to the LOR, is more appropriate for our work.  ​