Food consumption data used in dietary exposure assessments Dietary exposureorintake assessmentis the technique of combining food consumption data with food chemical concentration or nutrient data to estimate dietary exposure to food chemicals, or intake of nutrients:
| Dietary exposure = food chemical concentration x food consumption |
At FSANZ, dietary exposure estimates are generally conducted using our customised software DIAMOND ( hyperlink to DIAMOND fact sheet ).
This information paper examines the food consumption data used by FSANZ. Information on food chemical concentration data sources can be found in the FSANZ document Principles and Practices of Dietary Exposure Assessment for Food Regulatory Purposes.
Where do the consumption data come from?
National nutrition surveys (NNSs) provide detailed, national data on consumption of foods, beverages and, in some cases, dietary supplements. This quantitative data on consumption is essential to the type of dietary exposure assessments we generally do at FSANZ.
We currently use the following national nutrition surveys from both Australia and New Zealand to estimate food consumption:
- 1995 Australian National Nutrition Survey
o 13,858 respondents
o 2 years and above
o 24-hour recall – one recall per respondent [1]
o 2nd 24-hour recall -10% of respondents
o Food frequency questionnaire for 12 years and above
o Short questions on food consumption patterns
o 13 month duration
o Multi-stage area sampling based on census districts
- 1997 New Zealand National Nutrition Survey
- 4,636 respondents
- 15 years and above
o 24-hour recall – one recall per respondent
o 2nd 24-hour recall -15% of respondents
- Food frequency questionnaire
- Short questions on food consumption patterns
- 12 month duration
- Stratified sampling
The following children’s national nutrition surveys have become available in DIAMOND in 2009:
· 2002 New Zealand Children’s Nutrition Survey
o 3,275 respondents
o 5-14 years
o 24-hour recall – one recall per respondent
o 2nd 24-hour recall -15% of respondents
- Quantified dietary supplement consumption
- 10 month duration
- School based sampling with non-proportional samples
· 2007 Australian National Children’s Nutrition & Physical Activity Survey
- 4,487 respondents
- 2-16 years
- Two 24-hour recalls – all respondents
- Short questions on food consumption patterns
- Quantified dietary supplement consumption
- 7 month duration
- Stratified sampling with non-proportional samples
Why not use more current consumption data?
We need extensive quantitative data to conduct detailed dietary exposure assessments using DIAMOND. The 1995 Australian and 1997 New Zealand NNS are the most recent comprehensive sets of quantitative data on food consumption patterns currently available to FSANZ for the majority of the population, apart from the children’s NNS data that became available for exposure assessments in 2009. Conducting dietary modelling based on 1995 and 1997 NNS food consumption data provides the best estimate currently available of dietary exposure across our populations because they are national studies, across wide age ranges, of all foods and beverages consumed.
While the 1995, 1997 and 2002 NNSs may not include information regarding food products that are now available in the market, for staple foods such as breads, cereals, milk etc, the data derived from these older surveys are likely to be still representative today. For some product segments, NNS data from the 1990s may not be appropriate for use in a dietary exposure assessment (e.g. of bottled water). In these cases, FSANZ may modify its modelling technique (e.g. through the selection of a similar food as a surrogate) or seek additional consumption data.
Additional qualitative data on food consumption are sometimes available and these may help us interpret food consumption data and verify assessments carried out using the NNS on a case-by-case basis.
What other consumption data are used in dietary exposure and intake assessments?
Whether or not FSANZ uses additional, non-NNS consumption data in a dietary exposure or intake assessment depends on a number of factors, including the purpose of the assessment, the nature of the hazard and risk, the quantity and quality of data available, and the existence and type of the relevant reference health standard.
One of the data sets available to FSANZ is the Roy Morgan Single Source Survey, which contains data on consumption incidence and frequency of consumption of core foods for approximately 50,000 people per year in Australia and 25,000 people in New Zealand. Data are available for the period of January 2001- December 2008 based on initial face-to-face interviews followed by self-completed surveys. However, the data do not quantify consumption for individuals and do not cover all foods and beverages consumed. We also have access to the Roy Morgan Young Australians Survey (YAS) , which provides similar data for 6,000 Australian children aged 6-13 years. FSANZ uses these up to date consumer surveys to validate the currency of the data from the 1995 and 1997 NNSs.
In addition to national nutrition surveys and Single Source, we may use a variety of data sources such as:
- food frequency questionnaires
- consumer research
- published scientific research
- market share data and retail sales figures
- data supplied with applications to change the Food Standards Code or submissions associated with these applications.
Limitations ofnational nutrition survey data
Dietary modelling based on past NNS food consumption data provides the best estimate of actual consumption of a food and the resulting estimated dietary intake of a nutrient for the population. However, the NNS data do have their limitations. These limitations relate to the age of the data and the changes in eating patterns that may have occurred since the data were collected. Generally, consumption of staple foods such as fruit, vegetables, meat, dairy products and cereal products, which make up the majority of most people’s diet, appears to have been fairly stable between the 1983/85 and 1995 Australian NNSs, although there were a few notable changes over this time, including small increases in cereals and fish consumption among adults [2] [i] . However there is greater uncertainty when assessing consumption of foods introduced to the market since these NNSs were conducted, or for which there may have been changes in food consumption.
Over time, there may also be changes to the ways in which manufacturers and retailers make and present foods for sale. Since the data were collected for the 1995 Australian and 1997 and 2002 New Zealand NNSs, there have been significant changes to the Food Standards Code to allow more innovation in the food industry. As a consequence, another limitation of the dietary modelling is that some of the foods that are currently available in the food supply were either not available or were not as commonly available in 1995-2002. Since the data were collected for these surveys, there has been an increase in the range of products that are fortified with nutrients, for example. For some dietary exposure assessments, these changes in the food supply will be particularly relevant and we will address these as far as possible in our assessment report, and identify where significant data limitations exist.
Using a single 24 hour period to represent long term eating patterns brings challenges in interpreting the data. Daily food consumption amounts for occasionally consumed foods based on 24 hour food consumption data tend to be higher than daily food consumption amounts for those foods based on a longer period of time. It’s possible to overestimate the exposure of high consumers, and impose overly restrictive risk management measures, if only a single 24 hour record is available for survey respondents. FSANZ has prepared a paper exploring this issue further ( hyperlink to Protecting High Consumers ).
While the results of NNSs can be used to describe the usual intake of groups of people, they cannot be used to describe the usual intake of an individual [3] . In particular, they cannot be used to predict how consumers will change their eating patterns as a result of an external influence such as the availability of a new type of food.
There are some foods reported in the NNSs for which the number of people who ate the food is too few to allow assessment of population-wide consumption patterns. In these cases it may be necessary to select a similar, more widely consumed food as a proxy. For example, peaches could be selected as a proxy food for nectarines. There may also be insufficient consumers of a food within a certain age group.
The NNNs were not designed to capture information on food consumption during pregnancy, so we are not able to model dietary exposure in pregnant women. As an alternative, FSANZ will generally model exposure among women of child-bearing age (16-44 years) as a proxy for pregnancy.
No NNS data are available for New Zealand children aged 2-4 years. Dietary exposure assessments produced for Australian 2-4 year olds will be assumed to be applicable to New Zealand children, unless there is clear evidence to the contrary.
For more information:
See the fact sheet “ Dietary exposure and intake assessments at FSANZ” ( hyperlink ) , or the document Principles and Practices of Dietary Exposure Assessment for Food Regulatory Purposes ( hyperlink ).
Or contact the FSANZ Dietary Modelling Team at http://www.foodstandards.gov.au/contactus/contactform.cfm
[1] A respondent is anyone who participated in the NNS
[2] Cook, T., Rutishauser, I. and Seelig, M. (2001a) Comparable data on food and nutrient intake and physical measurements from the 1983, 1985 and 1995 national nutrition surveys. Australian Food and Nutrition Monitoring Unit, Commonwealth Department of Health and Aged Care, Commonwealth of Australia, Canberra.
Cook, T., Rutishauser, I. and Allsopp, R. (2001b)The Bridging Study: comparing results from the 1983, 1985 and 1995 Australian national nutrition surveys. Australian Food and Nutrition Monitoring Unit, Commonwealth Department of Health and Aged Care, Commonwealth of Australia, Canberra.
[3] Rutishauser, I. (2000)Getting it right – how to use data from the 1995 National Nutrition Survey.Commonwealth of Australia, Canberra.