AAMA Operations Research Process: 2008-2012_Jennifer Nielsen_4.25.13
Download as PPTX, PDF1 like513 views
The AAMA operations research project from 2008-2012 combined HKI's homestead food production model with essential nutrition actions to improve nutrition among mothers and children under 2 in Baitadi, Nepal. A community randomized controlled trial examined the impact of the enhanced homestead food production intervention on nutritional outcomes. The intervention led to significant improvements across impact pathways including food production, assets, nutrition knowledge and health practices. However, the intervention did not find a significant reduction in child stunting, wasting or underweight, though it did significantly reduce maternal underweight and anemia. The results provide promising evidence for scaling up the model to address malnutrition in Nepal.
More Related Content
Similar to AAMA Operations Research Process: 2008-2012_Jennifer Nielsen_4.25.13 (20)
More from CORE Group (20)
AAMA Operations Research Process: 2008-2012_Jennifer Nielsen_4.25.13
- 1. AAMA Operations Research Process: 2008-2012 CORE Group Spring Meeting April 24, 2013
- 2. Photo © HKI/ Bartay INTRODUCTION & RESEARCH DESIGN HKI’s Enhanced Homestead Food Production Model
- 3. • The Action Against Malnutrition Through Agriculture (AAMA) project strategy combined HKI’s Homestead Food Production (HFP) strategy for improving year-round access to nutrient dense plant and animal-source foods with Essential Nutrition Actions to promote optimal nutrition practices (“EHFP”) • Targets mothers with children in the 1000 day nutrition “window of opportunity” • Operations research conducted under CSHGP grant in Baitadi hill district of Far Nepal’s Far Western Region • Community randomized control trial (RCT) to address long- standing question: can EHFP improve nutritional status? • Reduce stunting , wasting , underweight and anemia in children • Reduce underweight and anemia in women INTRODUCTION
- 4. • Cross-sectional Design – Examined children 12-48 months of age and their mothers in order to capture exposure during the (-9) to 23 month window – Baseline August 2009 – Endline August 2012 – Exposure to full treatment ~2.5 years • Randomization by sub-district (Ilaka) – 4 pairs matched by SES characteristics & population size – 1 in each pair randomly assigned to treatment arm; 1 to control – In treatment area 122 VMFs created reaching 3329 “HFPBs” • Baseline Survey Sampling: stratified 3-stage PPS – 1) random selection of 14 out of total 21 VDCs per study arm by PPS; 2) 6 wards randomly selected per VDC then 3) HH selected from center point of ward with spin of the pen – Sample size calculated to detect 10 percentage point difference in stunting rate: Baseline n=2106 Endline n=2614 Additional sample <6 mos: n=203 baseline; n=230 endline to assess exclusive breastfeeding RESEARCH DESIGN
- 5. Baitadi District INTERVENTION & CONTROL AREAS
- 6. Photo © HKI / Keang Khim Enhanced Homestead Food Production “Treatment”
- 8. • Implemented in 6 out of 9 wards (village) per targeted Village Development Committee (VDC) with hopes of spill over within VDC • Each ward developed a village model farm (VMF) that supported two mothers groups each with ~20 women in 1000 day window • Offered FCHVs opportunity to be VMF (~30%) • HFP Cascade Training led by HKI, NGO-NNSWA and Government partners from district agriculture & livestock offices: – Master trainers – VMF – Mothers’ groups • ENA Cascade Training by HKI, NNSWA, and district health office: – Master trainers – FCHV – Mothers’ groups TREATMENT DELIVERY: Enhanced Homestead Food Production
- 9. • Quantitative Survey (350 Qs) Baseline/Endline – HH SES - HFP practices - Food purchase/consumption - Food security (HFIAS) – WASH - Health - IYCN - IMCI – HH participation in community groups - Exposure to Ag training – ENA knowledge • Nutrition Measures (children & mothers) – Height, weight, age (WHO/FANTA Guidance) – Hemoglobin concentration (Hemocue ®) • Baseline Qualitative Formative Research – To inform Nutrition BCC strategy (barriers/facilitators IYCF) – 10 FGD with 5-8 mothers of children <24 mos, ages 18-44 yrs: 62 Qs – 5 IDIs with grandmothers; 5 IDIs with fathers: 48Qs DATA COLLECTION
- 10. • Ethics Review Board: Nepal Health Research Council – Submitted Feb 2009 – Approved June 2009 • Process Evaluation in Treatment Zones – 4 Rounds by LQAS (Aug 2010; April 2011; Aug 2011; Mar 2012) – Questionnaire and observations – Examined HFP practices, ENA knowledge and practice; Hygiene • Midterm Evaluation Sept 2010 • Final Evaluation Sept 2012 DATA COLLECTION
- 11. • Based on intent to treat (all HH in treatment zone considered part of study population) • Difference in Differences analysis (interaction TRT * Time) • Present unadjusted analyses of outputs & outcomes along pathways and adjusted analysis for nutritional impact in treatment compared to control HH. • Adjusted analysis controlled for differences at baseline that were independent of the impact pathway • Presentation highlights selected significant DID and descriptive analyses ANALYSIS
- 12. Photo © HKI / George Pigdor RESEARCH FINDINGS
- 13. • Significant differences detected in each stage of the program impact pathways (unadjusted analysis) • Significant reduction in maternal underweight and anemia (adjusted analysis) • No significant reduction in stunting, wasting or underweight • Borderline significant reduction in child anemia OVERVIEW OF FINDINGS
- 14. Quality of HFP Significantly Better in TRT ODDS RATIO: – Improved/”developed” HFP almost 40-fold higher – Variety of vegetables almost 60% higher – Quantity (kg) of vegetables over past 2 mos. about 80% greater DIMENSIONS OF CHANGE: – Treatment HH with improved/developed HFP increased from 1.3% to 60% – Median number of vegetables produced increased from 4 to 8 – Vegetable production over past 2 month period increased from 10 to 30 kg. FOOD PRODUCTION PATHWAY
- 15. Significantly Better Perceptions of Food Security in TRT as assessed by HFIAS (v. 3) ODDS RATIO: – Treatment group reported higher likelihood of feeling food secure and lower likelihood of insecurity DIMENSIONS OF CHANGE: – Reports of food security increased from 20% to 46% – Reports of severe food insecurity decreased from 49% to 13% FOOD PRODUCTION PATHWAY
- 16. Reported HH income significantly higher in TRT ODDS RATIO: – Lower likelihood of reporting no monthly income and higher likelihood of reporting income > 4000 Rs. – Odds of practicing HFP higher in treatment group; owning poultry more than double; owning fruit trees about 28% higher DIMENSIONS OF CHANGE – Poultry ownership increased from 12% to 30% – Any HFP increased from 49 % to 98% (but also increased in control group from 50% to 80%) ASSETS PATHWAY
- 17. Significantly Higher Levels of Nutrition Knowledge ODDS RATIO: – Significantly higher knowledge along a range of questions – Especially striking: treatment HH 9 times more likely to report eggs are good for children and 13 times more likely to report good for women – Women six-fold more likely to know women should eat more during pregnancy DIMENSIONS: OF CHANGE: – Proportion of treatment mothers citing eggs as nutritious for children increased from 29% to 80% – Proportion citing as nutritious for pregnant women from 32% to 85% – Women knowing to eat more during pregnancy increased from 62% to 93% NUTRITION KNOWLEDGE PATHWAY
- 18. Complementary Feeding Significantly Better (no change in Breastfeeding) ODDS RATIO: – Odds of child receiving Minimum Acceptable Diet and Minimum Dietary Diversity both five times greater in treatment – Odds of child receiving eggs, MN-rich plant source foods and legumes also significantly higher, although consumption of eggs & other ASF remained low. – Odds of pregnant woman eating more four times greater DIMENSIONS OF CHANGE: – Minimum Acceptable Diet increased in treatment group from 35% to 79% – Pregnant women’s reports of eating more increased from 15% to 40% ENA PRACTICES PATHWAY
- 19. Women’s health seeking significantly better (improvements in child health care both arms) ODDS RATIOS: – Maternal care (+4 ANC, skilled delivery) significantly higher in treatment – Treatment almost 3 times higher odds of consuming IFA – 4 times higher odds of receiving deworming treatment – 3 times greater odds of receiving PP VAS DIMENSIONS OF CHANGE: – Consumption of IFA in most recent pregnancy increased from 72% to 92% – Deworming treatment increased from 45% to 89% – PP VAS increased from 36% to 80% HEALTH PRACTICES PATHWAY
- 20. • IMPACT: Child Nutritional Status • IMPACT: Maternal Nutritional Status #p=0.051 *p<0.05 FINDINGS OF IMPACT Adjusted logistic regressions for assessing the effect of EHFP on child nutritional status Binary outcome variable Stunting Wasting Underweight Child anemia TRT 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 logistic regressions assessing effect on women’s nutrition Binary outcome variable Maternal underweight Maternal anemia Treatment X time 0.63 (0.47 – 0.84)* 0.59 (0.45 – 0.76)*
- 21. Photo © HKI/ Bartay DISCUSSION & CONCLUSION
- 22. • Lengthy project start up process and short project length limited exposure window • Results consistent with RCT literature on difficulty demonstrating impact on child growth in settings with extremely high prevalence of stunting (Baitadi ~60%) even with food supplements and ASF • Alternative determinants of child growth not addressed: hygiene, environmental enteropathy, high morbidity, geography • Poultry production challenges • Women’s benefits to be explored further. More sensitive to dietary improvements? Empowerment in decision-making? Benefits for future pregnancies? DISCUSSION & CONCLUSIONS
- 23. • Model being scaled up under Suahaara Project (25 of 75 districts nationwide) • Strengthened HFP model and poultry husbandry with focus on excluded groups • Increased emphasis on hygiene, sanitation and “toddler protection” • Strong promise of sustainability due to partnership with government and strong governance strategy • Research will continue on nutrition impact of EHFP FUTURE DIRECTIONS
- 24. THANK YOU. “Although the world is full of suffering, it is also full of the overcoming of it.” -Helen Keller
Editor's Notes
- #5: Baseline sample(n=1055 trt; n=1051 ctl) Endline sample (n=1307 trt; n=1307 ctl)
- #12: Variables in adjusted models included: number of children in household; sex of household head; caste (lower or upper); respondent working outside the home (yes/no); respondent pregnant; mother’s education (none vs. any formal); and mother primary caretaker (yes/no). Biological factors, including child’s sex and age and mother’s age, were also controlledUpper caste included: Brahmin, Chetri, Giri, or Thakuri; lower caste included: Dalit, disadvantaged Janajatis, relatively advantaged Janajatis, disadvantaged non-Dalit Terai caste, and religious minorities.
- #15: Odds of cultivating an improved (diversified) or developed (diversified & year round) garden significantly higher in treatment : OR 37.8 (8.38 – 170.2)*Variety of vegetables grown about 60% higher : 0.59 (0.52 – 0.65)* and increased from median of 4 to 8 (±4)Quantity (kg; past 2 mos) of vegetable production about 80% higher : 0.81 (0.68 – 0.93)* and increased from 10 to 30 (±30) kg.Number of eggs was about 40% greater in treatment: 0.39 (0.10 – 0.68)* and increased from 8 to 30 (±45)
- #16: Household Food Insecurity Access Scale (HFIAS) v3 2007Treatment group had higher likelihood of feeling food secure [OR 1.77 (1.31 – 2.37)]* Treatment group also had lower likelihood of feeling severe food insecurity [OR 0.43 (0.33 – 0.56)]*
- #17: Treatment HH 42% less likely to have no monthly income [OR 0.58 (0.45 – 0.73)*] and 39% more likely to have income > 4000 Rs. [1.39 (1.07 – 1.82)*]Significantly higher odds of being engaged in HFP [OR 13.13 (8.64 – 19.9)*], owning poultry [OR 2.64 (1.93 – 3.62)*], and owning fruit trees [OR 1.28 (0.79 – 2.08)*]Significantly higher odds of being engaged in HFP [OR 13.13 (8.64 – 19.9)*], owning poultry [OR 2.64 (1.93 – 3.62)*], and owning fruit trees [OR 1.28 (0.79 – 2.08)*]
- #18: eggs good for young children [OR 9.48 (7.32 – 12.27)]* and for women [OR 13.22 (10.18 – 17.15)]*women should eat more during pregnancy [OR 6.30 (4.67 – 8.49)]*
- #19: treatment significantly higher IYC dietary diversity [OR 5.33 (3.50 – 8.11)]*And minimum acceptable diet [OR 5.36 (3.53 – 8.14)]*Pregnant women in treatment group more likely to report eating more [OR 4.45 (1.56 – 12.63)]*
- #20: consuming IFA [OR 2.70 (1.95 – 3.73)]*PPVAS [OR 3.06 (2.38 – 3.92)]*
- #21: Covariates in the regression models included: number of children in household; sex of household head; caste (lower or upper); respondent working outside the home (yes/no); respondent pregnant; mother’s education (none vs. any formal); and mother primary caretaker (yes/no). Biological factors, including child’s sex and age and mother’s age, were also controlled for as the literature shows these can confound outcomes.Upper caste included: Brahmin, Chetri, Giri, or Thakuri; lower caste included: Dalit, disadvantaged Janajatis, relatively advantaged Janajatis, disadvantaged non-Dalit Terai caste, and religious minorities.