“Translating Adoption of Improved Varieties (Quality Protein Maize) into Nutritional Impact in Rural Ethiopia

The experimental design consists of two central treatment groups.

In the first treatment group (Adoption Encouragement), households receive a brief message about the nutritional benefits of QPM for young children in particular and an offer of free QPM seed (they may order a total of 3 bags of seed with 2kg of seed in each bag).

In the second treatment group (Consumption Encouragement), households receive a brief message about the nutritional benefits of QPM for young children in particular and an offer of free QPM seed (they may order a total of 3 bags of seed with 2kg of seed in each bag). At baseline, both the head of household and primary caregiver of the youngest child are further informed of the nutritional benefits of QPM specifically for young children. They are also informed that in order for their children to realize the maximum benefits of QPM consumption they must keep QPM grain separate from other grain. The primary caregiver is also given separate storage containers for QPM grain and flour that are labeled with the name of the youngest child.

2015-07-25

2016-02-29

Consumption
• Consumption of any QPM by the index child (measured both in the last 24 hours and in the last week)
• Amount of QPM consumed by the index child in the last 24 hours (grams)
• # of days in the last week that the index child consumed a QPM-based food
• Index child’s proportion of total maize consumption that was QPM in the last 24 hours
• Total caloric intake by index child in the last 24 hours
• Total utilizable protein intake (grams) by index child in the last 24 hours
Cooking
• Caregiver cooked a QPM-based food (measured both in the last 24 hours and in the last week)
• Amount of QPM used by caregiver for household cooking in the last 24 hours (grams)
• Amount of QPM used by caregiver for household cooking in the last 24 hours that was then consumed by children under 3 years of age (grams)
• Number of times the caregiver cooked a QPM based food (measured in the last 24 hours and in the last week)
• Proportion of QPM in maize used by caregiver for household cooking in the last 24 hours
• Caregiver cooked a QPM-based food that was consumed only by children under 3 years (measured in the last 24 hours and in the last week)
• # of days in the last week that caregiver cooked a QPM-based food that was consumed only by children under 3 years
• Caregiver cooked a food that was consumed only by children under 3 years (measured in the last 24 hours and in the last week)
• # of days in the last week that caregiver cooked a food that was consumed only by children under 3 years
Anthropometrics
• Height-for-age Z-score of index child
• Weight-for-height Z-score of index child
• Weight-for-age Z-score of index child
• Mid-upper arm circumference Z-score of index child
Biomarkers
• Serum transthyretin (continuous measurement)
• Insulin-like growth factor I (continuous measurement)
Agricultural outcomes
• Hectares of QPM planted
• Hectares of maize planted
• Kilograms of QPM produced
• Kilograms of maize produced
• Total income from agricultural season

The study design consists of three groups: a control group that receives no intervention, a treatment group that receives the Adoption Encouragement intervention and a treatment group that receives both Adoption Encouragement and Consumption Encouragement.

Households and clusters (nutrition groups) are pre-randomized from lists in the office by a computer.

The unit of randomization is both the cluster (women's groups) and the household. The cluster is a women's group formed for the purpose of health outreach by Health Extension workers in Ethiopia. The Adoption Encouragement intervention is randomized at the household level and stratified by kebele. The Consumption Encouragement intervention is randomized at the cluster level and stratified by kebele.

Yes

Across the two treatment arms, there are 562 clusters (women's groups). At the time of the intervention launch we do not have enough information to identify clusters in the control arm.

The experimental sample includes a total of 1424 households with 703 of the households included for the collection of biomarkers.

There are 400 households in the control group, 200 of the households will be included in the subsample for collection of biomarkers.
There are 511 households in the Adoption Encouragement group in 280 clusters. 252 of the households in 252 clusters will be included in the subsample for collection of biomarkers.
There are 513 households in the Adoption and Consumption Encouragement group in 282 clusters. 251 of the households in 251 clusters will be included in the subsample for collection of biomarkers.