USUAL NUTRIENT INTAKES
Due to resource constraints and to minimise respondent burden, national nutrition surveys usually collect a small number of days of food consumption data from survey respondents. This information is, in turn, used to estimate population nutrient intakes. Population nutrition guidelines, such as the Nutrient Reference Values (NRVs) for Australia and New Zealand, are based on long term or ‘usual’ nutrient intakes.1 NRVs include specific reference values for both population nutritional requirements and upper safe levels. The small number of days of nutrient intake data from nutrition surveys needs to be adjusted to estimate long term or ‘usual’ intakes. This is important to allow a direct comparison with NRVs so that the proportion of the population with nutrient intakes under nutritional requirements, or above upper safe levels, is not overestimated.
SINGLE DAY COMPARED TO A 'USUAL' DAY
Figure 1 shows, in general terms, a population nutrient intake from a single day compared to a usual intake, and a general comparison to reference cut points which could represent NRVs. If cut-points A and B are set on the basis of long term or usual intakes, comparisons with unadjusted single day intakes are misleading. In the figure below, if cutpoint A was representative of a requirement for nutrient adequacy, the area below the red line and to the left of cutpoint A would represent the prevalence of inadequate intakes (see 1.3 Reporting against Nutrient Reference Values). If cutpoint B was representative of an upper safe level, the area to the right of cutpoint B and below the red line would represent the proportion at risk of adverse effects from excess intake. Although the means of the two distributions are similar, comparison of the unadjusted single day distribution (green line) to the cutpoints in place of the usual intake distribution (red line) misleadingly shows a greater proportion of the population below cutpoint A, and above cutpoint B. Adjusted intakes are also used to more accurately determine percentiles of usual population nutrient intakes.
Figure 1. Population nutrient intakes based on a single day compared to usual and comparison with nutrient reference values2
A single 24-hour dietary recall was collected for 2011-12 NNPAS from all respondents. Respondents who reported they were pregnant or breastfeeding at the time of the survey (n=228) were excluded for the usual nutrient intakes release, leaving a day one sample of 11,925, with a second 24-hour recall from approximately 64% of respondents (n=7585).
Nutrient intakes derived from 24-hour recall data do not represent the usual intake of a person because there is variation in day to day intakes. The second 24-hour recall is used to estimate and remove within-person variation in order to derive a usual nutrient intake distribution for the population. Usual nutrient intakes represent intakes over a long period of time.
METHODS OF CALCULATING USUAL NUTRIENT INTAKES IN 1995
The method to estimate usual nutrient intakes for the 1995 National Nutrition Survey (NNS) was by using a shrinkage or adjustment factor approach. The proportion of respondents with a second day of food consumption data (approximately 10% of the sample) was used to produce the adjustment factors. These adjustment factors were applied to the Day 1 nutrient intakes to remove within-person variation to bring the distribution of intakes in towards the mean and produce population estimates of usual nutrient intake.3 However, more recent and sophisticated methods are now available.
CALCULATING USUAL NUTRIENT INTAKES FOR THE 2011-12 NNPAS
For 2011-12 NNPAS the ABS has used the NCI method for estimating usual nutrient intakes, developed by the National Cancer Institute (NCI) in the United States of America. Testing was undertaken by the ABS to ensure that the NCI method was compatible with 2011-12 NNPAS data, to determine how it should be used, to ensure that it was producing reliable results, and understand in what circumstances it should not be used.
More information on the way in which usual nutrient intakes have been produced using the NCI model in the Australian Health Survey: Usual Nutrient Intakes, 2011-12 is available in the following pages of this User’s Guide.
1.1 Overview of the NCI Method
1.2 Model implementation: data used and model specification
1.3 Reporting against Nutrient Reference Values
1.4 Data Quality
1.5 Summary tables
1 National Health and Medical Research Council and New Zealand Ministry of Health, 2006, Nutrient Reference Values for Australia and New Zealand, <http://www.nrv.gov.au/home/introduction>, last accessed 5/2/2015.
2 Food Standards Australia and New Zealand, 2009, Principles and practices of dietary exposure assessment for food regulatory purposes, Canberra: Food Standards Australia New Zealand, <http://www.foodstandards.gov.au/science/exposure/Pages/dietaryexposureandin4438.aspx>, last accessed 5/2/15.
3 McLennan W & Podger A (1998), National Nutrition Survey. Nutrient intakes and physical measurements. Australia. 1995, Commonwealth of Australia, Canberra.