8/26/2013
1
Photo©HKI/KeangKhim
IMPACT OF HOMESTEAD FOOD PRODUCTION ON NUTRITIONAL STATUS OF CHILDREN AND WOMEN IN BAITADI DISTRICT, NEPAL
AkotoK.Osei,2013
• Today’sfocusisonresearchconductedinBaitadi districtaspartofthe“ActionAgainstHungerthroughAgriculture(AAMA)project”
• AAMAwasoneofHKI’sEnhancedHomesteadFoodProduction(EHFP)programs
• 4yearprojectfundedbyUSAID(2008‐2012)
• Targetedchildren<2yearsandpregnantandlactatingwomen
• Combinedhomegardenandpoultryraising,withbehaviorchangecommunication(BCC)strategiestopromoteEssentialNutritionActions(ENA)
• Implementedthroughthelocalhealthandagriculturesystems
ACTION AGAINST MALNUTRITION THROUGH AGRICULTURE (AAMA) PROJECT
8/26/2013
2
• Asof2008,thereweredemonstratedimpactofEHFPon:
• Improvedagriculturalpractices• Volumeanddiversityofvegetablesandfruitsproduced• Diversityoffoodsconsumed• Women’sinvolvementinhouseholddecisionmaking• Potentialreductionofanemiaamongchildrenandwomen
• Noimpactonchildanthropometry(stunting,wastingandunderweight)
• Noimpactonanthropometryofwomen(BMI)• Noimpactonoverallhouseholdincome,althoughsomepositiveimpactonincomefromsaleofEHFPproducts
WHY THE RESEARCH IN BAITADI DISTRICT WAS CONDUCTED
ANEMIA AMONG CHILDREN AGED 6-59 MONTHS FROM HFP PROGRAM AND CONTROL HOUSEHOLDS (2003-2006).
64
45
66
50
65
5764 65
7870 72 70
43
17
59
31
0
10
20
30
40
50
60
70
80
90
Target Contro l Target Contro l Target Contro l Target Control
Bangladesh Nepal Cambodia Philippines
Baseline
Endline
8/26/2013
3
• VeryfewevaluationsonimpactofEHFPinterventiononanthropometryofchildrenandwomen
• ScarcedataonimpactofEHFPonothermicronutrientindicatorsasideanemia
• LimiteddataonimpactofEHFPonoverallhouseholdincome
• EHFPevaluationdesignswerenotrigorousenough
• Lackofemphasisonprogramimpactpathways
Masset etal2012
OTHER REVIEW ARTICLES REACHED SIMILAR CONCLUSIONS AND PROVIDED THE FOLLOW REASONS FOR LACK OF IMPACT OF EHFP ON NUTRITIONAL STATUS
Baitadi andKailali Districts
Baitadi district:• Communitytrialtoaddressthelong‐standingchallenge:canEHFPimprovenutritionalstatusofchildrenandwomen
• Reducestunting,wasting,underweightandanemiainchildren
• Reduceunderweightandanemiainwomen
Kailali district:• Roll‐outEHFPthroughoutthedistricttoidentifyfactorsnecessaryforpotential“scaleup”
RESEARCH OBJECTIVES OF AAMA
8/26/2013
4
• Clusterrandomizedcontrolleddesign• Preandpostsurveyswithindependentsamples(“crosssectional”)
– BaselineAugust,2009– Follow‐upAugust,2012
RESEARCH DESIGN
Baitadi District
12 Ilakas (Sub‐Districts)
63 Village Development Committees (VDC)
12 Ilakas matched in pairs
4 Ilakas pairs included
(randomly selected)
Randomized
4 Ilakas (20 VDC) received EHFP
[Treament)
14 VDCs: Baseline and Follow‐up surveys
4 Ilakas (21 VDC) as Control
14 VDCs: Baseline and Follow‐up surveys
1 Ilaka pair not included
RADOMIZATION OF COMMUNITIES TO EHFP (TREATMENT) AND CONTROL
8/26/2013
5
• EHFPwasimplementedin6of~9wards(villages)pertargetVDC• Eachwardhad:
– 1VillageModelFarm(VMF),managedbyawoman– ~30householdswithchildren<2yearswhobenefitedfromtheEHFPinputs(homegarden,poultryandnutritioneducation)
– Beneficiarymotherswereorganizedinto‘mothersgroups’(~15‐ 20motherspergroup)
• FemaleCommunityHealthVolunteers(FCHVs)weregivenprioritytobeaVMF
• CascadeTrainingledbyCHD,HKI,NTAG,NNSWA,DADO&DLSO:– TrainingofTrainerswhothentrainedFCHVs,VMFandhouseholdbeneficiaries
ROLE-OUT OF EHFP INTERVENTION IN “TREATMENT” COMMUNITIES
8/26/2013
6
Eachsurveyexaminedfamilieswithchildrenchildren 12‐48months
BaselineSurvey:
‐ 2106families(treatment=1055;control=1051)
Follow‐up:
‐ 2614families(treatment=1307;control=1307)
• Qualitativeinterviewsatbaseline– 10FGDs,5‐8mothersofperFGD– 5IDIswithgrandmothers– 5IDIswithfathers
FAMILIES WHO PARTICIPATED IN BASELINE AND FOLLOW UP SURVEYS
• QuantitativeSurvey– Interviews– Heightandweightofmothersandchildren– Hemoglobinconcentrationofmothersandchildren
• QualitativeResearch– ToinformBehaviorChangeCommunicationstrategyofAAMA– FocusGroupDiscussions(FGD)withmothers– In‐depthinterviews(IDI)withgrandmothersandfathers
DATA COLLECTED
8/26/2013
7
• ProcessEvaluationinTreatmentcommunities– BasedonProgramImpactPathways(PIPS)– Monthlyhouseholdvisits– 4RoundsmonitoringdatacollectionbyLotQualityAssuranceSampling(LQAS)Technique
• (Aug2010;April2011;Aug2011;Mar2012)
• MidtermEvaluationSept2010(Onlyqualitativedatacollection)
MONITORING
Photo©HKI/Bartay
SECTION 4: RESULTS
8/26/2013
8
• EHFPwereimplementedin20VDCsand120wards
• 122VMFswereestablished
– ~30%oftheVMFswereownedbyFCHVs
• 3329Householdswithchildren<2yearsbenefitedfrominputsandtrainingsforhomegardens,poultryandnutritioneducation
COVERAGE OF EHFP INTERVENTION IN TREATMENT COMMUNITIES
Baseline Follow up
Treatment Control P value Treatment Control P value
Upper Caste, % 77.4 82.3 0.004 70.2 78.1 0.000
Household size 7.2 (4.0) 6.0 (3.0) 0.000 6.0 (3.0) 6.0 (3.0) 0.569
Male head, % 87.0 93.3 0.000 93.8 95.9 0.017
Wealth tercile, %
Lower 25.5 42.2 0.000 25.2 41.2 0.000
Middle 28.2 36.7 0.000 32.0 34.6 0.158
Upper 46.3 21.1 0.000 42.8 24.2 0.000
RESULTS FROM BASELINE AND FOLLOW-UP SURVEYS
CharacteristicsofHouseholdsinsample
8/26/2013
9
44.8
97.6
51.2
80.2
0
20
40
60
80
100
Baseline* Follow ‐ up*
Percent of households
Survey period
Treatment
Control
HOUSEHOLD WITH EITHER HOME GARDEN, POULTRY OR BOTH
* p<0.05
Treatment (T)%
Control (C)%
Difference(T – C)
P value
Garden only
Baseline 73.4 71.4 2.0 0.425
Follow up 69.8 79.2 ‐ 9.4 0.000
Poultry only
Baseline 15.9 18.0 ‐ 2.1 0.328
Follow up 1.1 4.2 ‐ 3.1 0.000
Garden + Poultry
Baseline 10.8 10.6 0.2 0.919
Follow up 29.1 16.6 12.5 0.000
HOUSEHOLDS WITH SPECIFIC TYPES OF EHFP ACTIVITY
8/26/2013
10
Treatment (T)%
Control (C)%
Difference(T‐C)
P value
Food insecure, %
Total Baseline 79.7 87.4 ‐ 7.7 0.000
Follow up 53.6 78.3 ‐ 24.7 0.000
Mild
Baseline 21.7 19.9 1.8 0.298
Follow up 17.4 14.5 2.9 0.041
Moderate
Baseline 8.8 15.4 ‐ 6.6 0.000
Follow up 23.3 35.8 ‐ 12.5 0.000
Severe
Baseline 49.2 52.2 ‐ 3.0 0.160
Follow up 12.9 27.9 ‐ 15.0 0.000
FOOD INSECURITY AMONG HOUSEHOLDS
13.223.3
16.9
39.2
0
20
40
60
80
100
Baseline* Follow ‐ up*
Percent of child
ren
Survey period
Treatment
Control
MORBIDITY AMONG CHILDREN (PAST 2 WEEKS RECALL)
* p<0.05
8/26/2013
11
62.5
76.7
61.6 63
0
10
20
30
40
50
60
70
80
90
100
Baseline Follow ‐ up*
Percent of child
ren
Survey period
Treatment
Control
EXCLUSIVE BREASTFEEDING PRACTICES
Onlymothersofsampledchildrenwithothersiblingbelow6monthsold
* p<0.05
35.6
79.4
37.744.1
0
10
20
30
40
50
60
70
80
90
100
Baseline Follow ‐ up*
Percent of child
ren
Survey period
Treatment
Control
COMPLEMENTARY FEEDING PRACTICES
Proportionofchildrenfedminimumacceptable diet(onlychildren12– 23months)
* p<0.05
8/26/2013
12
28.2 26.617.5
23
0
10
20
30
40
50
60
70
80
90
100
Baseline* Follow ‐ up*
Percent of mothers
Survey period
Treatment
Control
UNDERWEIGHT AMONG NON-PREGNANT MOTHERS
* p<0.05
Treatment(T)
mean ± SD
Control(C)
mean ± SD
Difference P value Ɨ P value ‡
Hemoglobin, g/L
Baseline 132.6 15.3 133.8 17.2 ‐ 1.2 0.077 0.028
Follow up 130.0 13.7 126.4 14.4 3.6 0.000 0.000
HEMOGLOBIN CONCENTRATION OF MOTHERS
Ɨ Adjustedforonlyclusterdesign‡ Adjustedforclusterdesign,mother’sageandpregnancystatus
8/26/2013
13
19.624.621.1
35.8
0
10
20
30
40
50
60
70
80
90
100
Baseline Follow ‐ up*
Percent of mothers
Survey period
Treatment
Control
ANEMIA AMONG MOTHERS
* p<0.05
Treatment(T) %
Control(C)%
Difference P value
Stunting (HAZ < ‐2)
Baseline 57.7 65.8 ‐ 8.1 0.000
Follow up 55.1 63.5 ‐ 8.4 0.000
Wasting (WHZ < ‐2)
Baseline 10.6 10.1 0.5 0.688
Follow up 10.5 9.7 0.8 0.462
Underweight (WAZ < ‐2)
Baseline 43.4 48.0 ‐ 4.6 0.033
Follow up 41.0 40.6 0.4 0.860
STUNTING, WASTING AND UNDERWEIGHT AMONG CHILDREN
8/26/2013
14
Treatment (T)
mean ± SD
Control (C)
mean ± SD
Difference(T – C)
P value Ɨ P value ‡
Hemoglobin, g/L
Baseline 118.5 12.1 118.1 12.6 0.4 0.446 0.751
Follow up 117.9 11.7 115.0 12.4 2.9 0.000 0.000
HEMOGLOBIN CONCENTRATION OF CHILDREN
Ɨ Adjustedforonlyclusterdesign‡ Adjustedforclusterdesign,child’sageandsex
28.2 30.831.6
42.5
0
20
40
60
80
100
Baseline Follow ‐ up*
Percent of child
ren
Survey period
Treatment
Control
ANEMIA AMONG CHILDREN
* p<0.05
8/26/2013
15
• IMPACT:ChildNutritionalStatus
• IMPACT:MaternalNutritionalStatus
#p=0.051*p<0.05
Multivariate Logistic regression analysis
Adjusted Odds Ratios (95% CI)
Binary outcome variable
Stunting Wasting Underweight Child anemia
Treatment X TIME 0.93 (0.73 – 1.18) 0.99 (0.68 – 1.45) 1.13 (0.89 – 1.43) 0.77 (0.60 – 1.00)#
Adjusted Odds Ratios (95% CI)
Binary outcome variable
Maternal underweight Maternal anemia
Treatment X Time 0.63 (0.47 – 0.84)* 0.59 (0.45 – 0.76)*
N Baseline%
Follow up%
Difference P value
(B) (F) (F ‐ B)
Anemia
EHFP + MNP 99 71.7 20.2 ‐ 51.5 0.000
EHFP 100 64.4 15.8 ‐ 48.6 0.000
Control 106 65.1 25.5 ‐ 39.6 0.000
SUB-STUDY WITHIN AAMA : EHFP + MNP VERSUS EHFP ALONE VERSUS CONTROL
8/26/2013
16
Photo©HKI/GeorgePigdor
SUMMARY OF FINDINGS
• ResultsshowsthatEnhanceHomesteadFoodProductionprogramhaspotentialofreducinganemiaamongchildrenandwomeninNepal
• ThefindingsalsodemonstratethatEHFPhaspotentialofreducingunderweightamongwomen
• Therewasnodemonstratedimpactonchildgrowth• Probablybecausefamilieswereexposedtotheinterventionforonly2.5years
• However,therewassignificantimprovementsinarangeofmaternalpracticesthatareknowntoimpactchildgrowth
SUMMARY OF FINDINGS
8/26/2013
17
• Strongpartnershipbetweengovernmentministriesandotherimplementingagencieswasbeneficialinrolling‐outoftheAAMAintervention
• Emphasisonprogramimpactpathwaysfacilitatedpropermonitoringoftheinterventionactivities
• Formativeresearchatbaselineallowedustoidentifyspecificmessages(“smalldoableactions”)topromoteinthedistrict
• Shortfundingcycle(4years)andlengthyprojectstartupprocesslimitedexposureoffamiliestotheintervention
• Multi‐sectoral planningandimplementationappearedeasieratthedistrictandwardlevelsthanatnationallevel
SOME LESSONS LEARNED
• GovernmentofNepal• USAID• NTAG• SMJK• MI• Alive&Thrive• UCDAVIS• HKI
THANKS TO
8/26/2013
18
THANK YOU.“Althoughtheworldisfullofsuffering,itisalsofulloftheovercomingofit.”‐HelenKeller