Details on the study variables

CHAPTERS: METHODOLOGY

5.3 Details on the study variables

All analyses were conducted on the national dataset (children aged 1-9 years old) and by province and urban/rural area (urban formal and urban informal;rural tribal and rural commercial farm). Table 5.2 below details further on the study objectives and variables.

Inthis study, a household was regarded as food insecure when (according to the criteria described in Table 5.2)it had:

.:. low income, OR

.:. low energy intake of the selected child in the household,OR .:. low vitamin A intake of the selected child in the household, OR .:. low dietary diversity of the selected child in the household, OR .:. stunting of the selected child in the household,OR

.:. underweight of the selected child in the household, OR .:. experienced hunger.

Table 5.2: Details on the study variables by specificstudy objective

SPECIFICOBJECTIVES TOADDRESS NFCS STUDY SPECIFIC STUDY VARIABLES CUT OFF POINT

RESEARCH AIMS DATA SOURCE

1. To determinethe number and percentage of low Socio-demographic

*

Low income income<R120001

income households[low income= income< questionnaire year"

R120001year]

2. To determine (from 24 Hour Recall data) the number 24 hour recall

*

Energy (kJ/day) consumed by <2/3_ofRDA~ for and percentage of childrennot receiving adequate questionnaire child child'srequirement nutrient intakes [<2/3ofRDA for energy and

*

Vitamin A (ug RE 1day) for energy and

vitamin A]. consumedby child vitamm. . A^IO

8From Socio-demographicquestionnaire:Questionnumber 24:Householdincomepermonth- Optionswere: "None;RI00-500;R500-1 000;RI 000-3000; R3000-5000;

Over R5000,Don'tknow". Lowincome= <R12000cut ojJpoint asusedinNFCSreport(basedon1996Census data- wherelow incomewasdefinedasannualincome

<R9600 (Hirschowitz 2000),1999low income estimated at= R9600+inflation) = ~R I 2000.

9Energyintakeiswidelyusedas an indicator off oodsecurity,with varyingcutoffpoints, e.g.<80%(Haddad,Kennedy and Sullivan1994)versus <70%of recommended energy intake (Chung, Haddad, Ramakrishna and Riely 1997).RoseandOliviera (l 997) useaconservativecutojJpointof<50%oftheRDAfor energy and other nutrients in determininglow intakesdue tounderreportingin dietarysurveysand variabilityinvolvedin onedayrecalldata. A cutoffpointof<213of theRDA was selectedforthis study. Thecut ojJpoint of lessthan 213 of theRDAf orthespe cified nutrientintakewas alsousedin theNational FoodConsump tionSurvey report (Labadarios,ed. 2000).

10Energy intake is widely usedasan indicator of f ood security. Yetadequate energy intakedoesnot guarantee adequateintake of othernutrients. Forthisstudy,VitaminA intakewas alsoselectedasan indicator of f oodsecurity. Onein threechildren in SouthAf rica had marginalvitamin Astatus (SAVACG 1996) and the NFCSalsoshowed low dietary intake ofvitamin A(Labadarios,ed.2000).

Reference standardsforenergy and vitamin Aintakewere:Energy: RDAmale1-2yearolds: 4393kJlday,female1-2yearolds:4166kJlday,male3-8year olds: 7316kJlday, female3-8year olds:6896kJlday,male 9-13year olds:9572kJlday , female9-13yearolds:8698kJlday; VitaminA:RDA 1-3year olds: 300uglday,RDA 4-8yearolds:

400uglday,RDA 9-13year olds:600uglday (Institute ofMedicine(2001).Dietaryreferenceintakes:VitaminA;Institute of Medicine(2002). Dietaryreference intakes:Energy.).

SPECIFIC OBJECTIVES TO ADDRESS NFCS STUDY SPECIFIC STUDY VARIABLES CUT OFF POINT

RESEARCH AIMS DATA SOURCE

3. To determine (from Quantified Food Frequency data) Quantified Food

*

Energy (kJ/day) consumed by < 2/3 ofRDA for

the number and percentage of children not Frequency child child's requirement

receiving adequate nutrient intakes [< 2/3 ofRDA questionnaire

*

Vitamin A (ug RE / day) for energy and

for energy and vitamin A] consumed by child vitamin A

4.To determine the number and percentage of children 24 hour recall

*

Low dietary diversity <6food groups

with low dietary diversity [estimated by less than 6 consumed 11

food groups consumed by the child in the household in 1 day (24 Hour Recall data]

5. To determine from the age and height measurement Socio-demographic

*

child's height-far-age Z score Height-far-age Z taken, the number and percentage of children that questionnaire score<-2 SD from

are stunted [Height-far-ageZ score < -2 SD from medran' re erencef 12

median reference]

11 Dietarydiversity= numberofdifferent foods orfood groups consumedover agiven ref erenceperiod(Ruel 2002). Foodgroupdiversityisastronger predictorofnutrient adequacy than asimplecountoffood (Ruel2002).Cut ojJpoints above which betternutrientintakeswere seen (RueI2002): Vietnam:>8/11foodgroups, Kenya:>6/11food groups(-0 of food groups). Thedietarydiversityfoodgroup cutojJselectedfor thisstudywasexp loratory. The decision for thisstudy wasto usethe16food groups as in MRCfood composition tables (Saye d, Fransand Schonfeldt 1999,AppendixC) andexclude foodgroups "Baby foods" (mainlycommercial infant cerealand jar foods),

"Therapeutic/Special/Diet products " (mainly powderedenteralproducts)and"Miscellaneo us"(mainly alcoholicbeverages).Using therema ining 13food groups,the cut ofJpointfor lowdietarydiversitywas set at<6food groupoutofa possible13 (-0of foodgroup s).

12Thereference for median height-and weight-for-agewasthe 1977growth curvesfor children fromthe NationalCentre ofHealth Statistics ofthe US (as usedin the NFCS). ThecutojJp ointof<-2 SDwasselectedasitiscommonly usedtoreportanthropometric statisticsinSouthAfricae.g.s:4VACG report. Different countries use different cutofJpo intse.g.in India<-3SDsisused asacutofJpo int(ChungetaI1997).

SPECIFIC OBJECTIVES TOADDRESS NFCS STUDY SPECIFIC STUDY VARIABLES CUTOFF POINT

RESEARCH AIMS DATA SOURCE

6. To determine from the age andweight measurement Socio-demographic

*

child's weight-for-age Z score Weight-for-ageZ taken, the number and percentage of children that questionnaire score<-2 SDfrom

are underweight [Weight-for-age Z score<-2 SD med'ranre erenceD 12

from median reference]

7. To determine the number and percentage of households Hunger scale

*

Household and child hunger Yes responses to 5 that experience 'householdand child hunger' questionnaire expenence or morequestions^{. 13}

[qualitative questionnaire] (CCHIP)

8. To compare the estimates of the prevalence of - - ^-

household food security made by the different indicators above

9. Todetermine the overlap in the identification of food - Each indicator of householdfood - insecure households by the different indicators of insecurity as outlined in objectives 1 household food insecurity (i.e.how many of the to 7 preceding

same households are identified)

10. To investigate whether there is any correlation - - -

Two anthropometricindicatorswereused asstuntingmay bea betterindicationofchronic food insecurity ,while underweightmay betterreflecttransitory/acutefood security.

13AccordingtotheCCHIP questionnaire:5positiveresponses outof the 8questionsreflectedhouseholdhungerwith the child in thehouseholdalsobeing affected- childhoo d hunger was not analysedseparatelyin thisstudy.

14Itisassume d thatwith greaterhouseholdincome,householdmemberswill buyandconsume more food.

SPECIFIC OBJECTIVES TOADDRESS NFCS STUDY SPECIFIC STUDY VARIABLES CUT OFF POINT

RESEARCH AIMS DATA SOURCE

between the indicators selected Specifically,whether:

With increasing household income:14

Energyand vitamin A intakeincreases (from 24HR andQFFQ data)

Dietary diversityincreases Stunting decreases

Underweight decreases

Household and child hungerdecreases With increasing energy intake:¹⁵(24HR data)

Energyintake increases(QFFQ data) Vitamin A intake increases (24HR data) Dietar ydiversity increases

Stunting decreases Underweight decreases

Household and childhungerdecreases

15Increasing energyintakeis assumed tobeindicativeofconsuming more food.

16Increasing VitaminA intakeisassumed tobeindicative ofconsuming a betterquality diet.

17A greate rdietarydiversityis assumed tobe indicativeofconsuming abetter diet(bothintermsof qualityand quantity).

SPECIFIC OBJECTIVES TO ADDRESS NFCS STUDY SPECIFIC STUDY VARIABLES CUT OFF POINT

RESEARCH AIMS DATA SOURCE

With increasing energy intake:(QFFQ data) Vitamin A intake increases(QFFQ data) Dietarydiversity increases

Stunting decreases Underweight decreases

Household and child hunger decreases With increasingvitamin A intake:¹⁶(24HR data)

Vitamin Aintake increases(QFFQ data) Dietary diversity increases

Stunting decreases Underweightdecreases

Household and child hunger decreases With increasingvitamin A intake:(QFFQ data)

Dietary diversity increases Stuntingdecreases

Underweight decreases

Household and child hunger decreases With increasing dietarydiversity:17

Stunting decreases

SPECIFIC OBJECTIVES TO ADDRESS NFCS STUDY SPECIFIC STUDY VARIABLES CUT OFF POINT

RESEARCH AIMS DATA SOURCE

Underweight decreases

Household and child hunger decreases With increasing stunting

Underweight increases

Household and child hunger increases With increasing underweight

Household and child hunger increases