4363.0.55.001 - Australian Health Survey: Users' Guide, 2011-13  
Latest ISSUE Released at 11:30 AM (CANBERRA TIME) 09/05/2014   
   Page tools: Print Print Page Print all pages in this productPrint All RSS Feed RSS Bookmark and Share Search this Product  
Contents >> Nutrition >> Interpretation of Results >> Comparisons with 1995 NNS

COMPARISONS WITH 1995 NNS

In 1995 the ABS and the then Commonwealth Department of Health and Family Services undertook a joint project to conduct the last National Nutrition Survey. As can be seen in Comparisons with Other Nutrition Surveys, and the table below, the characteristics of both surveys are similar. 2011-12 NNPAS was designed to facilitate comparisons to 1995 NNS where possible. Direct comparisons from results of both surveys need to be undertaken with care, and taking a number of factors into consideration.


SUMMARY OF SURVEYS - 1995 NNS AND 2011-12 NNPAS
1995 National Nutrition SurveyNational Nutrition and Physical Activity Survey

Period of collectionFebruary 1995 to March 1996May 2011 to June 2012

ScopeAdults and children aged 2 years and over. Adults and children aged 2 years and over in 2805 collection districts.
Includes some persons in what would now be known as Very Remote Areas.Excludes persons in Very Remote areas.
Three stage area sample of private dwellings, maximum of 2 in scope people per household in urban areas and 3 people in rural areas and QueenslandThree stage area sample of private dwellings, maximum of 2 people per household, one adult and one child if applicable.
Respondents to the 1995 National Health Survey were invited to participate in the National Nutrition Survey.
All persons aged 65 years and over were selected.

Method of collectionFace to face 24 hour recall Face to face 24 hour recall
Three pass method developed by United States Department of Agriculture Agricultural Research ServiceEnhanced Five phase method developed by United States Department of Agriculture Agricultural Research Service
Pen and paper interviewCAPI
Second 24 hour dietary recall from 10% of respondentsSecond 24 hour dietary recall from 63% of respondents (CATI) at least eight days after

Administered byNutritionists trained as interviewers for this surveyTrained and experienced interviewers from ABS interviewer panel

OutputMean and median nutrient intakes from day 1, and percentile distributions of nutrient intake adjusted for within-person variation based on a second day of intake are available (excludes nutrients from supplements)Mean nutrient intakes from day 1 (excludes nutrients from supplements), and percentile distributions of nutrient intake adjusted for within-person variation based on a second day of intake (using an updated method) will be published

Low Energy Reporters10.6% males
18.7% females
18.9% males
23.0% females

Food composition databaseDeveloped by Australia New Zealand Food Authority specifically for 1995 NNS
29 nutrients
Developed by Food Standards Australia New Zealand specifically for 2011-12 NNPAS
44 nutrients

Response rate61.4%77%

Proxy interviewsNo child assistance with interviewChildren 2 – 4 years interviewed by proxy
5 – 11 year olds asked to provide recall with assistance from adult household member

Day of primary interviewSunday 5.83%, Monday 18.11%, Tuesday 18.65%, Wednesday 20.13%, Thursday 17.20%, Friday 12.50%, Saturday 7.58% Sunday 3.51%, Monday 17.94%, Tuesday 18.39%, Wednesday 18.20%, Thursday 16.48%, Friday 14.03%, Saturday 11.45%

Seasons of interviewSummer 21%, Autumn 29%, Winter 26%, Spring 24%Summer 27%, Autumn 31%, Winter 23%, Spring 19%

Sample size13,85812,153




SCOPE

Whilst sample selection techniques are similar between 1995 NNS and 2011-12 NNPAS, additional persons selected from households in rural areas are likely to have some impact on comparisons. For further analysis on the impacts of excluding rural respondents from NNS data see The Bridging Study: comparing results from the 1983, 1985 and 1995 Australian national nutrition surveys1.

Potential volunteer bias

In 1995 NNS respondents were a volunteer sample recruited from the National Health Survey. 2011-12 NNPAS respondents were sampled as a part of the 2011-2013 AHS, and as such were not volunteers. It is unknown what impact this may have on results, but given the length and complexity involved in a 24-hour dietary recall interview, it is likely that a more motivated sample in 1995 may demonstrate different reporting behaviours than respondents in 2011-12 NNPAS. An analysis of respondents that agreed to have their weight measurement taken in 2011-12 NNPAS reported average energy intakes 13% higher than those who weren't measured. It is not clear whether this relationship is independent from weight or other body size characteristics as opposed to potentially indicating a more complete dietary intake may be provided by respondents who are more willing participants.


METHOD OF COLLECTION

Interviewers vs Nutritionists

Nutritionists trained to interview were used in NNS, and an experienced ABS interviewer panel were used in NNPAS. This change in methodology is likely to have some impact on results. All efforts were made to compensate for not using nutritionists to interview, including using the highly automated and scripted AMPM in a computer assisted personal interview setting, with in depth training provided to the ABS interviewer panel before commencing interviews.

Differences in behaviour during interview are also likely. It is not known to what extent nutritionist interviewers disaggregated foods or prompted for further detail in 1995, however some issues with portion recording have been outlined below.

Instrument

Changes to the AMPM instrument between 1995 and 2011 meant moving from a three phase paper based questionnaire to a five phase computer based interview. Look up lists containing Australian brands and measures were included in 2011-12 to facilitate reporting by respondents.

Food Model Booklet

A food model booklet was produced to aid in reporting of measures in 2011-12 NNPAS (see 24-hour Dietary Recall). This, in addition to the extensive look up lists of brand sizes available in the computerised AMPM, has resulted in a much wider range of measures reported. This has impacted results across all food groups. Using tea as an example, in 2011-12 NNPAS the mean tea serve was 17% less than in 1995 NNS, but this is likely to be largely due to a change in the way in which measures were reported for tea, rather than an actual drop in the mean serve size of tea consumed. In 1995, 72% of records of tea were coded to a measuring cup (250 ml). In 2011-12 NNPAS this was under 3%, with 94% coded to a mug and fill level from the food model booklet (ranging from 80 ml to 330 ml). Thus, the 33% drop in median tea consumption (for consumers) is likely to be largely due to the change in the way in which measures were reported, rather than a true reduction in the amount of tea consumed.

Other food groups known to have been impacted by this change in methodology include vegetables, coffee, cereals and mixed dishes.


CODING AND CLASSIFICATION

Updates to the food classification since 1995 have impacted some estimates. A concordance is available as an excel spreadsheet from the Downloads page of this publication.

Changes in how foods have been split into separate components may also impact results. In 1995 NNS, many commonly consumed foods were split into their component parts, e.g. some burgers were split into bun, patty and other fillings. For 2011-12 NNPAS, these foods have been coded and reported as consumed, i.e. as they were eaten by respondents. This needs to be taken into consideration when comparing patterns of food consumption between the surveys, as inaccurate conclusions may be drawn about changes to consumption patterns.

Whilst comparisons can be made between the surveys, some changes in the approach to certain food groups needs to be taken into account when drawing conclusions. 2011-12 NNPAS classified foods as consumed, i.e. as respondents reported and ate foods. Whilst some foods were disaggregated to allow for coding (see Food and Measure Coding for details) other foods were left as reported. 1995 NNS split many foods into the components or ingredient parts, so consumption from these foods will be spread across multiple major food groups. A notable example of this is hot dogs. In 1995 the frankfurt and bun were split and coded to separate codes, whereas for 2011-12 NNPAS, the hot dog was coded as consumed to one of the hot dog codes in the Sandwiches and filled rolls minor food group.


CHANGES TO CODING APPROACH - HOT DOGS IN 1995 AND 2011-12

1995 NNS2011-12 NNPAS
Meat, poultry and game products and dishesCereal based products and dishes
Sausages, frankfurts and saveloysMixed dishes where cereal is the major ingredient
Frankfurts, and saveloysSandwiches and filled rolls, saturated fat ≤5 g/100 g
Frankfurter, wiener, or hot dog, cooked
73 grams
Hot dog, bread roll, frankfurt filling
176 grams
Cereals and cereal products
Regular breads, and rolls
Breads, rolls, white
Bread roll, white
86 grams
Note: Values provided are example amounts only based on average consumption amounts



Other foods known to be impacted include mixed pasta dishes, hamburgers and filled bread rolls.


FOOD COMPOSITION DATABASE

Nutrient values were applied to data from 2011-12 NNPAS from a custom food composition database prepared by FSANZ (AUSNUT 2011-13). This process was conducted similarly in 1995, with the production of a custom food composition database for the NNS by the then Australia New Zealand Food Authority (ANZFA) (now FSANZ). ANZFA later published AUSNUT 1999 based on the NNS custom food composition database. Differences in the nutrient values between the two databases reflect not only changes in the composition of foods, but also changes in available data and improvements in analytical methods over the period. For detailed information on nutrient values used in 2011-12 NNPAS see AUSNUT 2011-13. The major changes that should be considered when making comparisons between the 1995 NNS and 2011-12 NNPAS estimates for Nutrients follows.

Folate

Folate values reported in the 1995 NNS did not include allowance for added folic acid. However there have been substantial revisions to the folates data since AUSNUT 1999 so differences in reported levels are likely to be the result of improved methods of analysis.

Dietary fibre

Analytical methods for analysing dietary fibre have changed since 1995, so care should be taken when comparing results between surveys.

Energy

Published results from 1995 were calculated using energy without dietary fibre. Any 1995 NNS results published with 2011-12 NNPAS have been recalculated to include dietary fibre.

Sodium

Although AUSNUT 1999 did not initially report sodium, salt was included as an ingredient in many of the AUSNUT 1999 recipes. Salt is not used as an ingredient in the majority of AUSNUT 2011-13 home prepared mixed dish recipes (i.e. casseroles, curries, pasta dishes and stir-fries). Salt has been used as an ingredient in things like home-made bread and pastries as it serves a technological function. Sodium is also included in some restaurant and commercially prepared foods.


ENVIRONMENTAL CHANGES

Fortification

Comparisons between certain nutrients should take into account the mandatory food fortification programs for bread making (folate, thiamin and iodine), and edible oil spreads (Vitamin D). Voluntary fortification changes also need to be considered when analysing changes in nutrient sources. For more information on food fortification see www.foodstandards.gov.au/consumer/nutrition/vitaminadded/Pages/default.aspx

Changes in consumption patterns/Food Group Classification

Changes in how foods are consumed also need to be considered when making comparisons between 1995 NNS and 2011-12 NNPAS. Direct comparison of certain food groups such as vegetables may not be possible, due to changes in how Australians consume those foods. Greater consumption of mixed dishes containing vegetables, but which may not be classified under the major food group Vegetable products and dishes (e.g. chicken and vegetable stir fry) need to be taken into account when making direct comparisons of intakes of vegetables. Further analysis needs to be conducted before direct comparisons of some food groups within the food group classification can be made.


ENDNOTES

1 National Food and Nutrition Monitoring and Surveillance Project 2001,The Bridging Study: comparing results from the 1983, 1985 and 1995 Australian national nutrition surveys, Commonwealth of Australia, Canberra, Last accessed 05/05/2014, <https://www.health.gov.au/internet/main/publishing.nsf/Content/45463F09DD0FF996CA257BF0001E8C75/$File/bridging.pdf>. Back

Back to top

Previous PageNext Page