Transcript 2009 - VAM Resource Center

Food consumption analysis
Food Security Indicators Training
Bangkok 12-17 January 2009
Objectives
To describe the analysis of food
consumption
• To describe the analysis on food sources
• To discuss experiences/problems related
with the analysis of food consumption
•
Steps
1.
2.
3.
4.
Explore the module of food consumption
Calculate the FCS
Graph the result
Create the Food consumption score
groups
5. Validate the FCS with other indicators
6. Analyze the sources of food
Definitions
Dietary
diversity
The number of individual foods or food
groups consumed over a reference
period
Food frequency Number of days (in the past week) that
a specific food item has been consumed
by a household
Household Food The consumption patterns (frequency *
diversity) of households over the last
Consumption
seven days
Food consumption module
FC module info
Information:
 Weekly frequency of foods and food groups
 Sources of foods
 Numbers of meals
Indicators:
→ FCS – dietary diversity
→ Food and Food group frequency (0-7)
→ Average number of meals (children/adults)
→ Sources of food
Food consumption score - FCS
The Food Consumption Score is a composite
score based on dietary diversity, food
frequency and relative nutrition importance
of different food groups.
The FCS can be considered as a proxy of
food access and food security.
Data collection
The data have to be collected according to
usual food items consumed that are
specific to the country’s context.
• Food items are grouped into food groups
that are standard.
• The difference between foods and
condiments must be captured during the
data collection.
•
Calculation steps
1.
2.
3.
Using standard 7-day food frequency
data, group all the food items into
specific food groups.
Sum all the consumption frequencies of
food items of the same group, and
recode the value of each group above 7
as 7.
Multiply the value obtained for each food
group by its weight and create new
weighted food group scores.
Calculation steps
4.
5.
Sum the weighed food group scores,
thus creating the food consumption
score (FCS).
Using the appropriate thresholds, recode
the variable food consumption score,
from a continuous variable to a
categorical variable.
FCS
FCS =
astaplexstaple+ apulsexpulse+ avegxveg+ afruitxfruit
+ aanimalxanimal+ asugarxsugar + adairyxdairy+ aoilxoil
Where,
FCS
Food consumption score
xi
Frequencies of food consumption = number of days for
which each food group was consumed during the past 7
days
(7 days was designated as the maximum value of the sum of the frequencies of the different
food items
ai
belonging to the same food group)
Weight of each food group
Food groups and weights
FOOD ITEMS
1
Maize , maize porridge, rice, sorghum, millet pasta,
bread and other cereals
2
Cassava, potatoes and sweet potatoes
3
Beans. Peas, groundnuts and cashew nuts
4
Vegetables and leaves
5
Fruits
6
Beef, goat, poultry, pork, eggs and fish
7
Milk yogurt and other diary
8
Sugar and sugar products
9
Oils, fats and butter
10
Condiments
Food groups
Weight
Cereals and
Tubers
2
Pulses
3
Vegetables
1
Fruit
1
Meat and fish
4
Milk
4
Sugar
0.5
Oil
0.5
Condiments
0
Weights
Food groups
Weight
Justification
Energy dense, protein content lower and poorer
quality (PER less) than legumes, micro-nutrients
(bound by phytates).
Energy dense, high amounts of protein but of
lower quality (PER less) than meats, micronutrients (inhibited by phytates), low fat.
Main staples
2
Pulses
3
Vegetables
1
Low energy, low protein, no fat, micro-nutrients
Fruit
1
Low energy, low protein, no fat, micro-nutrients
Highest quality protein, easily absorbable micronutrients (no phytates), energy dense, fat. Even
when consumed in small quantities,
improvements to the quality of diet are large.
Highest quality protein, micro-nutrients, vitamin
A, energy. However, milk could be consumed
only in very small amounts and should then be
treated as condiment and therefore reclassification in such cases is needed.
Meat and fish
4
Milk
4
Sugar
0.5
Empty calories. Usually consumed in small
quantities.
Oil
0.5
Energy dense but usually no other micronutrients. Usually consumed in small quantities
Graph
Laos FCS
Staple
Vegetables
Anim protein
Oil
Sugar
Fruit
Pulses
Milk
Cumulative Consumption
Frequency
49
42
35
28
21
14
7
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
FCS
This graph aids in the interpretation and description of both
dietary habits and in determining cut-offs for food consumption
groups (FCGs).
How to create the graph
1. Truncate the FCS variable
2. Run a frequency of the FCS
3. Run a compare mean of the FCS and all
the food groups included in the FCS
4. Export frequency and compare mean in
excel
5. Calculate an average of the surrounding
values for each food group (to smooth the
graph).
6. Use the ‘area’ graph in excel
Graph cont’
Staple
Fruit
consumed (*)
(Days/week)
Anim protein
Oil
Pulses
Sugar
Vegetables
Milk
7.00
6.00
5.00
4.00
3.00
2.00
1.00
0
10
20
30
40
50
60
70
80
90
100
Food Consumption Score
(*) w eighted moving average over 7 point range
This graph shows the consumption frequency of different food groups
by FCS independently and not stacked as the previous graph.
How to create the graph
1. Use the same steps from the graph
above;
2. Use the ‘line’ graph in excel.
FCS thresholds
Once the FCS is calculated, the thresholds
for the FCGs should be determined based
on the frequency of the scores and the
knowledge of the consumption behaviour in
that country/region.
The typical thresholds are:
Threshold
Profiles
0 – 21
Poor food
consumption
0-28
21.5 - 35
Borderline food
consumption
28.5 - 42
>35.5
Acceptable food
consumption
>42.5
Thresholds with oil
and sugar eaten
on a daily basis
(~7 days per
week)
Why 21 and 35?
A score of 21 was set as barely minimum, scoring below 21, a
household is expected NOT to eat at least staple and
vegetables on a daily base and therefore considered to have
poor food consumption. Between 21 and 35, households are
assessed having borderline food consumption.
The value 21 comes from an expected daily consumption of
staple and vegetables.
» (frequency * weight, 7 * 2 = 14)+(7 * 1 = 7).
The value 35 comes from an expected daily consumption of
staple and vegetables complemented by a frequent (4
day/week) consumption of oil and pulses.
» (staple*weight + vegetables*weight + oil*weight +
pulses*weight = 7*2+7*1+4*0.5+4*3=35).
……Even though these thresholds are
standardized there is always room for
adjustments based on evidence……
How to adapt the thresholds
1. Consider the basic/minimum food
consumption in the country.
Ex. Laos diet is mainly rice and vegetables, but in some
country you can have oil and/or sugar consumed daily
2. Based on the data information and the
knowledge of the country try to define
the thresholds for poor and borderline
consumption.
3. The thresholds should be changed based
on evidence and should be remain the
same if you want to compare FCS of
different surveys.
Example
Examples of different thresholds:
• Sudan
–
•
Two different thresholds were used north
and the south Sudan
Haiti
–
26 & 46 were used because the
consumption of oil and sugar among the
poorest consumption were about 5 days per
week.
!!!! We have to be careful that changes from
the standard are very well justified and
reported otherwise we can be viewed as
changing the threshold ‘ to get the numbers
we want’ !!!!
Validation of the FCS
•
Run verifications of the FCS and FCGs by
comparing them to other proxy
indicators of food consumption, food
access, and food security:
Cash expenditures,
 % expenditures on food,
food sources,
CSI,
wealth index,
number of meals eaten per day, etc.
Which is the analysis that we should use to
compare 2 continuous variables?
Correlations
Correlations with FCS comparing FCS to other food security
proxies
Burundi
kcal/capita/day
CSI score
% total cash
expenditures on food
asset index
total cash monthly
expenditures (LOG)
Pearson Correlation
0.31
Sig. (2-tailed)
<0.01
Pearson Correlation
-0.27
Sig. (2-tailed)
<0.01
Pearson Correlation
-0.11
Sig. (2-tailed)
<0.01
Pearson Correlation
0.24
Sig. (2-tailed)
<0.01
Pearson Correlation
0.28
Sig. (2-tailed)
<0.01
Malawi
CSI score
No. of assets
No. of means (adults)
Total per cap. Cash
exp. (LOG)
Pearson Correlation
-0.30
Sig. (2-tailed)
<0.01
Pearson Correlation
0.40
Sig. (2-tailed)
<0.01
Pearson Correlation
0.33
Sig. (2-tailed)
<0.01
Pearson Correlation
0.31
Sig. (2-tailed)
<0.01
Proxy for food security
If the FCS captures several elements of food
consumption, food access, and food security
(such as in the previous slide’s example)
FCS is an adequate proxy for CURRENT
food security
Sources of food
We have information about source of single
food but we need an indication of sources of
all the food items consumed in the
households.
This indicator can be used as proxy of food
access.
( ex. dependency on market, food assistance
or own production)
Sources of food
• Transform the single sources (x variables as the
food items) into n variables as the different
sources of food;
– Own production, purchase, food assistance, borrow,
exchange, gathering, social network, etc.
• Doing this we will have the percentage of food
consumed coming from different sources
– Ex % coming from purchase and % from food aid
etc.
• In this computation the sources of food should be
weighted on the frequency of the food items
consumed.
Steps
1. Copy the food frequency value into new variable
called as the different sources.
IF (source_rice
IF (source_rice
IF (source_rice
IF (source_rice
IF (source_rice
execute.
=1)
=2)
=3)
=4)
=5)
ownproduction_rice =consumption_rice.
purchase_rice = consumption_rice.
foodaid_rice = consumption_rice .
gathering_rice = consumption_rice.
borrowrice = consumption_rice .
Do this computation for all the food items and all the
sources.
Steps
2. Add all the variables of different foods with the
same sources together in order to create the
unique variable of the specific source
COMPUTE ownproduction = ownproduction_rice +
ownproduction_tubers + ownproduction_eggs +
ownproduction_vegetable + ownproduction_meat +
ownproduction_fruit + ……
3. COMPUTE the total sources of food
totsource = ownproduction + fishing + purchase +
traded + borrow + exc_labor + exc_item + gift
+ food_aid +other.
4. Calculate the % of each food source
COMPUTE
COMPUTE
COMPUTE
COMPUTE
COMPUTE
COMPUTE
COMPUTE
COMPUTE
pownprod = (ownproduction / totsource)*100.
pfishing = (fishing / totsource)*100.
ppurchase = (purchase / totsource)*100.
pborrow = (borrow / totsource)*100.
pexclabor = (exc_labor / totsource)*100.
pexcitem = (exc_item / totsource)*100.
pfoodaid = (food_aid / totsource)*100.
pother = (other / totsource)*100.
Example
S ourc es of food
100
1
90
1
7
3
12
2
5
7
1
12
22
80
70
60
50
95
72
90
89
61
71
53
73
94
65
93
40
30
20
10
0
25
19
24
7
3
4
Urban
Urban
P hnom
P enh
23
R ural
P lains
Urban
5
2
R ural
Tonle S ap
% own produc ion
% fis hing and hunting
% borrowed
% ex c hange of items for food
% food aid
Urban
R ural
P lateau
22
19
Urban
R ural
Urban
C oas tal
% purc has e
% traded
% ex c hange of labor for food
% gift
% ex c hange other
R ural
Total
Questions?
Some examples
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
acceptable
limite
pouvre
1
2
3
4
quintiles de indice de richesse
5
groupes de
consommation
alimetaire
acceptable
limite
pauvre
0
7
Maize
Other Cereals
Beans, Peas
Fruits
Fish
Milk/Yoghurt
Sugar, Honey, Jam
14
21
28
35
42
Rice
Casssava, Sweet Pots, Bananas
Vegetables
Meats
Eggs
Oils/Fat/Butter
49
Da
Su N hu
la in k
ym a
a n wa
Ta iya
m h
ee
m
Er
b
Di il
al
An a
Ba ba
gh r
da
Ba d
Ka bil
rb
Sa W ala
la a s
h sit
Al
D
in
Na
Q
ad jaf
M issi
ut a
Th han
i– a
M Qar
iss
Ba an
sr
ah
To
ta
l
% of households
35%
30%
25%
81
71
81 80 82 77
83 86
poor
84
78 80 81
77
borderline
77
Mean
83
91 89
81
69
20%
15%
10%
5%
0%
100
90
80
70
60
50
40
30
20
10
0
FCS
Poor and Borderline FCG
Spearman's rho
food consumption
score
Correlation Coefficient
1
Sig. (2-tailed)
.
N
Correlation Coefficient
CSI
Sig. (2-tailed)
N
Correlation Coefficient
wealth index
Sig. (2-tailed)
N
per capita total
expenditure
Correlation Coefficient
Sig. (2-tailed)
N
per capita non foof
expenditure
Correlation Coefficient
Sig. (2-tailed)
N
Correlation Coefficient
total_Income
food
consumption
score
Sig. (2-tailed)
N
24975
-.111(**)
0
8877
.378(**)
0
24972
.406(**)
0
24971
.343(**)
0
24971
.430(**)
0
24934
Sources of all foods
100%
90%
80%
70%
60%
50%
40%
30%
D
17
21
8
p_pds
29
15
24
28
21
32
34
26
24
17
p_purchase
p_ow nproduction
p_family
other
21
To
ta
l
16
28
Ba
b
Ka il
rb
al
a
W
Sa a ss
it
la
h
A
lD
in
N
aj
af
Q
ad
is
si
a
M
ut
ha
Th na
i–
Q
ar
M
is
sa
n
Ba
sr
ah
19
22
Er
bi
l
D
ia
la
An
ba
r
Ba
gh
da
d
30
ah
uk
N
in
Su
aw
la
a
ym
an
iy
a
Ta h
m
ee
m
20%
10%
0%
Sources of PDS food basket
100%
80%
60%
40%
64
20%
67
62
40
47
33
54
52
66
63
60
48
41
39
70
58
49
49
16
ppds_pds
ppds_purchase
ppds_ownproduction
ppds_family
OTHER
To
ta
l
ad
is
si
a
M
ut
ha
na
Th
i–
Q
ar
M
is
sa
n
Ba
sr
ah
aj
af
Q
N
Ba
bi
Ka l
rb
al
a
W
as
Sa
si
t
la
h
A
lD
in
An
ba
r
Ba
gh
da
d
ia
la
D
Er
bi
l
N
in
Su
av
la
a
ym
an
iy
ah
Ta
m
ee
m
D
ah
uk
0%
Food sources - rural model
Plains
C oastal
Tonle Sap
Total
Plateau
0%
20%
40%
60%
80%
type of source
% own producion
% purchased+traded
% fishing and hunting
% other
100%
Food sources - urban model
Phnom Penh
C oastal
Total
Plains
Tonle Sap
Plateau
0%
20%
40%
60%
80%
100%
type of source
% own producion
% purchased+traded
% fishing and hunting
% other