CSIRO Healthy Diet Score and Egg Consumption in Australia

Australians’ usual egg consumption is a document prepared by researchers at the CSIRO. The CSIRO is the premier Australian government-funded research organisation.

The conclusion of this document states:

Eggs provide a low cost, convenient source of protein and other key nutrients. Our results [from the Healthy Diet Score survey] suggest their inclusion in the diet is associated with a higher diet quality, in particular higher consumption vegetables and lower consumption of discretionary foods.

I sent an email on 28th October 2017 to the lead author of this document, Dr Gilly Hendrie. I have concerns regarding the methodology and conclusions of the study and requested some clarification. Hendrie is also the lead author of a number of journal articles relating to the Healthy Diet Index. The details of the email shown below. A response will be published when it is received.

A shorter version of this critique can be found at Eggs are not OK


The CSIRO Healthy Diet Score survey is an extension of the Short Food Survey which was developed by CSIRO and the University of South Australia.[1] [2] The Healthy Diet Score collates self-reported dietary intakes.

The Short Food Survey is a series of 38 short questions. The scoring reflects the daily amount of fruit, vegetables, breads and cereals, meat and alternatives, dairy and dairy substitutes and discretionary foods to dietary recommendations. For the quality components, the frequency of wholegrains consumption, reduced fat dairy consumption, frequency of trimming meat, fat type of spreads used, and water consumption (as a proportion of total fluids) as well as the variety of foods.

CategoryDaily criteria for minimumCriteria for maximumScore
MenWomen
Fruit0 servings> 2> 210
Vegetables (starches, salad, cooked)0 servings6510
Grains (bread, cooked cereals, breakfast)0 servings665
Meat and alternatives (red meat, poultry, fish, legumes, tofu, nuts, seeds and other meat alternatives, eggs)0 servings310
Dairy (milk, cheese, yogurt)0 servings5
Beverages (water, fruit juice, soft drinks)0% Water100% water100% water10
Discretionary Choices (fried foods, savoury snacks, crackers, sweet biscuits, cakes, bars, chocolate, lollies, ice-cream, alcoholic drinks) ≥ 3 (M) or 2½ (F) servings< 3 servings< 2½ servings20
Food Quality
• wholegrain breads0AlwaysAlways5
• low-fat milk0AlwaysAlways5
Use of healthy fats
• trimmed meat0AlwaysAlways5
• margarine0AlwaysAlways5
Variety
• Fruit0≥ 4≥ 42
• Vegetables0≥ 5≥ 52
• Dairy0≥ 5≥ 52
• Protein foods0≥ 13≥ 132
• Grains0≥ 9≥ 92

Categories of the Healthy Diet Score


Sample CharacteristicsCount% Sample% Aust Pop
Male23,50627.149.4
Female63,10572.950.6
18-30 years22,09025.513.8
31-50 years32,60737.628.1
51-70 years28,44332.822.6
71+ years3,47149.7
Underweight2,1652.51.7
Normal weight40,93747.335.5
Overweight26,58730.735.3
Obese16,92219.527.5

Selected sample characteristics of CSIRO Health Diet Score Survey
May 2015 to June 2016

Below are the sorted scores for each of the nine result categories for both male and females.

FemaleScoreMaleScore
Fluids93Fluids88
Vegetables73Fruit67
Meat and Alternatives70Meat and Alternatives66
Fruit69Variety65
Variety66Bread and Cereals62
Bread and Cereals61Vegetables61
Healthy Fats55Healthy Fats52
Dairy and Alternatives49Dairy and Alternatives47
Discretionary Foods33Discretionary Foods26

Scores of CSIRO Healthy Diet Survey for the nine result categories

Discretionary FoodDailyWeekly
Total3.42.42.723.917.118.9
MaleFemaleTotalMaleFemaleTotal
Alcohol10.50.76.63.84.5
Cakes and biscuits0.40.30.32.52.32.3
Chocolate and confectionary0.40.50.52.93.53.3
Fried potato products0.20.10.11.10.70.8
Ice-cream0.20.10.11.10.80.9
Muesli and snack bars0.20.10.11.30.91
Processed meat and products0.30.20.22.31.21.5
Savoury pies and pastries0.100.10.70.30.4
Savoury snacks0.30.20.21.81.61.7
Sugars, sweetened beverages0.40.20.32.51.51.8
Takeaway foods0.10.10.110.70.7

Average servings per day and week


Below are some observations regarding the sample results.

  • 73% of the participants are female.
  • 47% of the participants are normal weight compared with 35% of the Australian population.
  • 50% of the participants are overweight or obese compared with 63% of the Australian population.
  • The dietary information is obtained by self-reporting which is unreliable.
  • The sample is not a representative sample of the Australian population.
  • Based on the weight status, the sample population is healthier than the Australian population.
  • It is likely that health conscious people are more likely to participate in such a survey.
  • The only health outcome measured is weight status.
  • The discretionary food items have no nutritional value and are not beneficial. However, 20 points are added to the Healthy Diet Score for males that have less than 3 servings a day and females that have less than 2½ servings a day.

The reliability and validity of the Healthy Diet Score is discussed in a paper published in Nutrition & Dietetics (2017)[3]. The score is used to “assess compliance with the 2013 Australian Dietary Guidelines”. In this context, reliability refers to the comparisons of two self-reported surveys performed one week apart. The validity is measured by “comparing the first administration [of the self-reported dietary survey] to the average of three 24-hour [self-reported] dietary recalls”. A better test of validity is to ensure that the Health Diet Score is a true reflection of health outcomes.

According to the paper, the “validity analysis showed moderate correlation between SFS#1 and 24-hour recalls (ICC = 0.43 (0.21:0.62, P < 0.001)), with 51% agreement in allocation to tertiles of diet quality between methods (k = 0.262, P = 0.004). The survey overestimated the diet index score by an average of 12.7 points out of 100 [-20.11:42.94] in comparison to recalls”.

The ICC refers to Intra-class Correlations. The ICC shows how well two sets of data match each other. The result of 0.43 is usually interpreted as a fair correlation—not a moderate correlation as stated. Less than 0.40 is usually considered to be a poor correlation. There is only a 51% agreement in tertiles. This means that if the two set of results are divided into three categories (low, medium and high), then there is only a 51% agreement into which category the result lies. The survey overestimated the diet score by an average of 12.7 points which is a significant difference in the participant’s Healthy Diet Score. The range of discrepancy is -20.11 to 42.94 which is very large difference given that the HDS is a value between 0 and 100.

It is important to note that ICC only measures the how well two sets of scores match. It does not measure the most important criteria—how well the scores measure actual health outcomes.

The real test of validity, which is not addressed, is how closely the actual food intakes relate to health outcomes, as opposed to a theoretical dietary index.


This survey was used to justify the “Eggs are OK every day” campaign.[4]

This is despite the fact that there is only a fair correlation between two different self-reported dietary surveys that were performed a week apart and despite the fact that the Health Diet Score only measures compliance to dietary guidelines and not actual health outcomes.
Daily Egg ConsumptionHDS
Low (0-.0.3)55.6
Low-Medium (0.3-0.4)58.2
Medium (0.4-0.6)59.7
Medium-High (0.6-1.0)60.5
High consumption (> 1.0)61.3
Non consumer57.7

Daily egg consumption compared with average HDS

The justification for “Eggs are OK everyday”, is that the as the egg consumption increases, so does the Healthy Diet Score. However, the Healthy Diet Score has not been shown to be associated with an increase in health outcomes. Consider how minor the differences are in the HDS, especially given that the average discrepancy was 12.7 points between two self-reported surveys one a week apart.

If the weight status is compared with the egg consumption and Healthy Diet Score, a different observation is shown.

Weight StatusEggs per day - maleEggs per day - femaleMean Diet Score
Underweight1.04 (2.24)0.62 (1.20)59.6 (14.4)
Normal weight0.83 (1.34)0.77 (1.28)60.5 (12.6)
Overweight0.89 (1.65)0.83 (1.45)58.1 (12.5)
Obese0.98 (1.84)0.88 (1.58)55.7 (13.2)

The mean egg consumption and mean HDS compared with weight status.
The standard deviation is shown in brackets.

The only measured health outcome shows an increase in body weight as the number of eggs consumed increases and a decrease in the mean of the Healthy Diet Score.

Dairy, meat, fish, poultry are given positive ratings in the Healthy Diet Score despite Seventh-day Adventists’ studies consistently showing an improvement in health status as the diet becomes more vegan.[5]

Category%BMIType 2 diabetes
Odds ratio*
Vegan
No red meat, fish, poultry, dairy, eggs
4.223.60.51
Lacto-ovo vegetarians
Vegan with eggs and milk
31.625.70.54
Pesco-vegetarians
Vegan with fish, milk and eggs
11.426.30.70
Semi-vegetarians
Red meat, poultry less than once a week plus fish, milk, and eggs
6.127.30.76
Non-vegetarians
Red meat, poultry more than once a week plus fish, milk, and eggs
46.928.81

(*) After adjustment for age, sex, ethnicity, education, income, physical activity, television watching, sleep habits, alcohol use and BMI

These figures are after adjustment for BMI, physical activity and television watching which not independent variables.

Note that the comparisons are with Californian Adventists who are much healthier that the average Californian who are also much healthier than the average American.

The life expectancies of vegan Californian Adventist men and women are higher than those of any population studied.[6]


The Physicians’ Health Study commenced in 1981. As well as examining the role of aspirin and a beta-carotene supplement, on heart disease and cancer, it also examined the role of egg consumption on health. The result from a 20-year follow-up showed a significant correlation between egg consumption and all-cause mortality. A key finding is that physicians consuming 7 or more eggs per week had a 31% increase in all-cause mortality compared with those consuming less than 1 egg per week. For diabetic physicians, the association was much higher with the increase in mortality slightly more than doubled.[7]

A British study reported a 2.7 times greater risk of cardiac deaths consuming 6 or more eggs per week.[8]

Based on the above observations, the recommendation for “one egg a day” does not appear to be justified.

Footnotes

  1. Hendrie, G., Baird, D., Golley, S., & Noakes, M. (2016, September). CSIRO Healthy Diet Score 2016.
  2. Hendrie, G. A., Rebuli, M. A., & Golley, R. K. (2017). Reliability and relative validity of a diet index score for adults derived from a self‐reported short food survey. Nutrition & Dietetics, 74(3), 291–297.
  3. Hendrie, G. A., Rebuli, M. A., & Golley, R. K. (2017). Reliability and relative validity of a diet index score for adults derived from a self‐reported short food survey. Nutrition & Dietetics, 74(3), 291–297.
  4. Hendrie, G., Baird, D., & Noakes, M. (2016, October). Australians’ usual egg consumption. CSIRO Publications.
  5. Le, L. & Sabate, J. (2014) Beyond Meatless, the Health Effects of Vegan Diets: Findings from the Adventist Cohorts. Nutrients. 6 (6), 2131–2147.
  6. Fraser, G. E., & Shavlik, D. J. (2001). Ten Years of Life – Is It a Matter of Choice? Archives of Internal Medicine, 161(13), 1645–1652. https://doi.org/10.1001/archinte.161.13.1645
  7. Djoussé, L. & Gaziano, J. M. (2008) Egg consumption in relation to cardiovascular disease and mortality: the Physicians’ Health Study. American Journal of Clinical Nutrition. 87 (4), 964–969.
  8. Mann, J. I. et al. (1997) Dietary determinants of ischaemic heart disease in health conscious individuals. Heart. 78 (5), 450–455.

Leave a Reply

Your email address will not be published. Required fields are marked *