Assessing dietary diversity in
South Africa: What does it tell us?
NP Steyn, D Labadarios, JH Nel
Why do we measure dietary diversity?
• Diet quality
• Micronutrient screen
• Measure of food security
Surveys in RSA
• In RSA there have been 3 studies which have measured
dietary diversity at national level
• The same method has been used: an un-quantified 24
hour recall to list foods eaten prior to the interview
• Division of foods into 9 food groups
• Calculation of a dietary diversity score (DDS) with
• 0 = no dietary diversity and 9 = maximum dietary
diversity
• Study 1: Children 1-9 years in 1999
(N=2200)
• Study 2: Adults in 2009 (N=3287)
• Study 3: SANHANES in 2012 adults and
children(N=13 357)
Study 1
•
•
•
•
National Food Consumption Survey (NFCS)
N=2200 children 1-9 years
Quantified 24 hour recall done
NAR (nutrient adequacy ratio) calculated for each
nutrient = Average consumed over 100 % RNI
• MAR (mean adequacy ratio) calculated as the Sum of
NARs divided by the number of nutrients
• FVS = Food variety score = mean number of different
food items consumed from all possible items eaten
• DDS calculated as the score out of 9 groups with each
group only counting once
Results
• Mean FVS = 5.5 (SD 2.5)
• Mean DDS = 3.6 (SD 1.4)
• Mean MAR = 50% (ideal = 100%)
Groups consumed
•
•
•
•
•
•
•
•
•
•
Cereals & tubers & roots= 99.6%
Dairy =55.8%
Meats =54.1%
Eggs=13.3%
Fats =38.9%
Legumes & nuts=19.7%
Vitamin A rich fruit &vegetables=23.8%
Other fruit=22%
Other vegetables=38.8%
Other: tea, sugar, sweets, jam=87.6%
Correlations
• High correlation between MAR and FVS: r=0.726,
p<0.0001
• High correlation between MAR and DDS: r=0.657,
p<0.0001
• MAR, DDS and FVS showed significant correlations with
HFA and WFA z scores
• A DDS of 4 was shown to be best indicator of MAR less
than 50% since it provided the best sensitivity and
specificity.
Z scores with DDS
NARs with DDS
NARs with DDS
Conclusion
•
• Study showed a strong relationship between DD and
indicators of child growth
• Either FVS or DDS can be used as a simple and quick
indicator of the micronutrient adequacy of the diet.
STUDY 2
A Simple Method of Measuring
Dietary Diversity at Population Level
in adults
NP Steyn, D Labadarios, JH Nel
Social science that makes a difference
Date:
Introduction
• Determining dietary diversity in RSA adults has not been
possible to date since there are no national dietary data on
adolescents or adults
• However a national study on food consumption in children
(NFCS) in RSA showed a very monotonous type of diet with
specific deficiencies including: Energy, iron, zinc, calcium,
vitamins A, C, E, B6, B2, niacin and folic acid. Largely due to the
majority of the population consuming large amounts of maize
meal, bread and sugar with low intakes of animal protein and
fruit and vegetables.
• It is known that a diverse diet is more likely to contain all the
essential nutrients than a monotonous one, hence measuring
dietary diversity is a simple way of identifying the likelihood of
having an adequate diet in terms of essential micronutrients.
Objective of the study
To measure dietary diversity in South
Africans aged 16 years and older from all
population groups
Methods
A cross-sectional study representative of adults from all specified
ages, provinces, geographic localities, and socio-economic strata in
South Africa was used (n= 3287).
Trained interviewers visited participants at their homes during the
survey. Dietary data was collected by means of a face validated 24
hour recall which was not quantified.
A dietary diversity score (DDS) was calculated by counting each of 9
food groups. A DDS <4 was regarded as reflecting poor dietary
diversity and poor food security.
Individual dietary variety
• Use of an adapted FAO (2011) method using 9 food groups.
Groups were based on outcomes of the NFCS
• Starchy staples (cereals, roots, tubers)
• Vitamin A rich fruit and vegetables
• Other fruit
• Other vegetables
• Legumes and nuts
• Fats and oils
• Meat/poultry/fish
• Milk and milk products
• Eggs
• FAO have an organ meats group and a dark green leafy group
and combine other fruits and vegetables
Cut-off value for dietary diversity used
• This was based on a validation study done with data from the NFCS
with assistance from FAO (Steyn NP et al, PHN 2005).
• A dietary diversity score ( DDS) of at least 4 (groups) was shown to be
the lowest minimum requirement and provided a specificity of 70% and a
sensitivity of 75% of at least at 50% MAR of the overall diet in children.
Furthermore, z scores for weight and height for age of children rose
above zero at a DDS of 4.
• No validation data available on adults
24 Hour recall
• Each participant was required to list all foods and drinks
consumed on the previous day. No quantities were
recorded. Each item consumed from a specific food
group was counted once only. A DDS< 4 would
represent poor diversity. Some studies include a
minimum of 15 g per item per day.
Results
% Population in each province having a low DDS
(<4 groups)
% Population having a low DDS by area
Social science that makes a difference
% Population having a low DDs by SES status
Social science that makes a difference
% population having a low DDS by ethnic group
Mean DDS of different SES categories
•
Living Standard Measure
Low
Medium
High
Sample sizec
585
1320
1219
Mean DDS
2.93 [C]
3.84 [B]
4.72 [A]
95% CI
2.81 – 3.05
3.76 – 3.93
4.64 – 4.80
Mean DDS by geographic area
Geographic area
RSA
Urban, formal
Urban, informal Tribal
Rural,
All
Sample size
2024
309
599
355
3287
Mean DDS
95%CI
4.42 [A] 4.7
3.46 [B] 3.8
3.17 [C] 3.3
3.64 [B]
4.34 – 4.50
3.30 – 3.61
3.05 – 3.29
3.46 – 3.81
3.6
4.02
4.2
3.96 – 4.07
Mean DDS by ethnic group
Ethnicity
RSA
Black / African
Mixed ancestry
Indian/Asian
White
All
Sample size
1941
604
389
353
3287
Mean DDS
3.63 [C] 4.0
4.43 [B] 4.5
4.44 [B] 4.1
4.96 [A] 5.6
4.02 4.2
95% CI
3.55 – 3.71
4.30 – 4.56
4.29 – 4.58
4.82 – 5.10
3.96 – 4.07
Most commonly consumed food groups
% Consumers
95%CI
Cereals
99.7
99.5-99.9
Vitamin A rich fruit & veg
17
15-18
Other fruit
25
24-27
Other vegetables
52
50-54
Legumes & nuts
18
16-19
Fat & oils
38
36-40
Meat/poultry
78
77-80
Milk & milk products
56
55-58
Eggs
18
16-19
Odd ratios of factors associated with having a DDS<4
Factor
Odds Ratio
95% CI
Casual work
2.769*
1.447-5298
Buys at spaza
1.979*
1.150-3.406
Lives in traditional house
2.394*
1.121-5.116
Water source is river
7.060*
3.096-16.101
No toilet
3.350*
1.061-10.562
No electricity
2.310*
1.198-4.453
Odd ratios of factors associated with having a DDS<4
Factor
Odds Ratio
95% CI
Employed full time
0.672*
0.430-1.050
Supermarket close by
0.584*
0.349-0.939
Flush toilet
0.467*
0.155-1.406
Electricity in house
0.389*
0.208-0.727
Motor vehicle
0.326*
0.270-0.394
Mobile phone
0.473*
0.389-0.575
Survey 3 SANHANES
• Full report not yet available but preliminary data confirms
data from 2 earlier studies
Conclusions
• Overall the major adults consumed a diet low in variety
• Tribal areas & informal urban areas were worst affected
• Eggs, legumes and vitamin A rich fruit & vegetables
were least consumed
• Will include organ meats and dark green leafy
vegetables in future as groups
Recommend
• That this method is used as a screening
tool at clinics and health centers to identify
families at risk of malnutrition and/or
having poor food security