Promoting Healthy Eating among Poor Children: The Roles of Information, Affordability, Accessibility, Gender, and Peers on Child-feeding in Ethiopia

There is a total of six intervention arms including control: 1) BCC for mothers, 2) BCC for mothers and fathers, 3) food vouchers, 4) BCC for mothers and food vouchers, 5) BCC for mothers and fathers and food vouchers, and 6) control. The duration of intervention is six months, but individuals will be followed up to 12 months to see whether the child-feeding habits continue to stay with the households. All interventions have been pilot tested in the project site to refine the BCC materials and voucher procedures, and to assess the feasibility and acceptability of the interventions. All interventions will be conducted and funded by AFF.

1. Mother BCC
A behavior change communication (BCC) intervention on complementary feeding practice for mothers has been developed based on the Alive and Thrive’s BCC program materials that have been successfully implemented in Ethiopia. BCC will be delivered in groups of six through weekly sessions for a duration of 6 months. Key gaps in knowledge and barriers identified through formative studies are also addressed in the BCC modules.

We also conduct individually-tailored trainings based on participant’s food recall survey results, youngest child’s age, and child caring behavior. The training informs the participants what to eat (local sources of protein and vitamins A and C), what not to eat (e.g., excessive oil, sugar, and processed food), how much and when to eat (e.g., ideal eating frequency by age groups, when to introduce solid food for infants, when to introduce non-breastmilk liquid for infants, etc.), ideal intra-household allocation of food, and short and long-term benefits of good nutrition. Counseling will consist of comparing the household's nutrition level in terms of micro and macronutrients with the ideal standard, and helping the mothers to set monthly concrete goals. Mothers are also encouraged to keep a food log for self-monitoring. Training, counseling, and food recording is aided by highly image-oriented pictures and brochures. The training and counseling materials will follow, as appropriate, WHO and UNICEF guidelines on child growth and development.

2. Father BCC
A BCC intervention on complementary feeding practice for fathers has been developed based on the Alive and Thrive’s resources. Materials to raise awareness on gender related constraints between spouses will be adapted from Nurturing Connections developed by Helen Keller International and USAID’s infant and young child nutrition project targeting male group. Key gaps in knowledge and barriers identified through formative studies will be addressed in the BCC modules.

As the gender-related modules from Helen Keller International and USAID have not been tested for acceptability in Ethiopia, we tested the developed modules through multiple father focus group interviews and modified the contents reflecting the comments from the fathers.

Materials will be delivered in groups of six through weekly sessions for the duration of three months. As most fathers are orthodox farmers who do not work during orthodox holidays (about seven days a month), BCC training will be conducted during orthodox holidays or seven to eight am in the morning before work begins. This schedule was found to be the most convenient time from the formative study of fathers. Motivational interview technique will be used to individually counsel the fathers.

3. Food vouchers
Food vouchers are transferred monthly at the Project Office in Holeta. Eligible participants will receive approximately 15% of their monthly income, 200 ETB, estimated around US$10. Transferors will verbally state that the purpose of the vouchers is to enable households to consume healthier food and that they are nontransferable, while making clear that nothing is required of them to receive the voucher and that no rule or regulation is tied to the receipt of the transfers. Food vouchers will be distributed monthly to mothers at Holeta Project Office, and they will be given in denominations of 20 ETB to facilitate small transactions. At the first disbursement, careful instructions will be provided as to how to use the vouchers. Each voucher will have a blue-colored official stamp to prevent fake vouchers and a household identification number to be matched with household IDs to prevent transfers.

Vouchers will be redeemable for a broad list of food groups including cereals, tubers, fruits, vegetables, legumes, meats, fish, milk products, and eggs. Vouchers can be used at selected shops and regular markets in which voucher stands will be available for sellers to cash out the vouchers. Shops or sellers will be required to match the vouchers with the household IDs and to keep a record of the food items that they bought and their monetary amount. Vouchers can be used over a series of visits per month and must be redeemed within the month it was received.

Below are interventions that were considered, but not implemented

4. Access to healthy food
We originally planned to send petty traders with diverse food items to random half of the local market in the study area to assess the impact of accessibility to healthy food on feeding practices. This was taking into consideration that some mothers especially rural residents, have access to a limited set of food items in the market that they usually visit, lacking dairy, meat, vitamin A rich vegetables and fruits. However, we decided not to implement this intervention due to the political turmoil that was ongoing during the study period. Our original plan was to implement the study in 400 village. However, due to the political turmoil that was ongoing during the study period, we reduced the sample size to 105 villages of randomly selected 6 kebele (district). This resulted in limited number of local markets to implement the study.

5. Price variation of healthy foods
We planned to introduce variation in the prices of healthy foods that are sold through petty traders to estimate the impact of price on demand. This intervention was not implemented due to the political turmoil that was ongoing during the study period. We were worried that price manipulation may cause possible conflicts, risks, or misunderstanding about the NGO’s activities in the field, especially given the political and civil unrest.

2017-07-10

2018-06-30

• Primary Outcomes (end points)
1) Mother’s and Father’s nutritional knowledge
2) Child’s dietary diversity scores (CDDS)
3) Food consumption scores (FCS)

The primary outcomes of interest are mean difference in parents’ nutritional knowledge and child dietary diversity score (CDDS). Mother and fathers’ IYCF knowledge are assessed using a separate survey module that contains 34 and 27 questions, respectively, on topics such as importance of the first two years in child growth, appropriate feeding frequency and dietary diversity, and nutrient rich foods. We count number of questions correct and divide the number by total number of questions. To calculate CDDS, we use a survey module that contains questions on 40 food items or food groups that are consumed by children in the study area and age group. For each question, the interviewer asks the mother "Yesterday, during the day or night, did [NAME] eat [FOOD ITEM]?” If the answer is yes, the interviewer asks how many times the food item was fed to the child the previous day. The food items are grouped into seven food groups, and the CDDS is calculated by summing up the number of food groups the child consumed in the past 24 hours. The Food Consumption Score (FCS), a food security indicator, represents households' dietary diversity and nutrient intake. The FCS is calculated using the frequency of consumption of different food groups consumed by a household during the 7 days before the survey.

Details on primary and secondary outcomes were initially registered in pre-analysis plan
1) AEA trial registry: July 11, 2017
2) Clinicaltrials.gov: June 25, 2017
We are updating the AEA RCT registry to reflect the details written on the initial pre-analysis plan that was submitted to the AEA trial registry and clinicaltrials.gov.

Secondary outcomes
Child feeding frequency
Child’s height-for-age Z scores (HAZ), weight-for-height Z scores (WHZ) and Mid-upper arm circumference (MUAC)
Social support actions of father

Other outcomes
Household hygiene environment and practice
Minimum dietary diversity
Minimum meal frequency
Minimum acceptable diet
Intra-household decision making between the husband and wife
Household labor allocation
Communication between partners
Gender norm attitude
Perceived social support of mothers
Change in mothers’ weight circumference
Household food and non-food consumption patterns
Food security measures

Secondary outcomes of interests are child anthropometry. Child’s height, weight, and MUAC are collected three times in units of cm, kg, and cm, respectively. Height-for-age Z scores (HAZ), weight-for-height Z scores (WHZ), stunting, and wasting are calculated for analysis using the WHO child growth standards. Other measures of IYCF practices include minimum dietary diversity, minimum meal frequency, and minimum acceptable diet standards (WHO, 2010). Minimum meal frequency is the proportion of children who consumed minimum number of meals recommended for the age, and minimum dietary diversity is proportion of children who receive food from 4 or more food groups. Minimum acceptable diet is calculated by combining minimum dietary diversity and minimum feeding frequency adjusting for child’s age.

This study is a community-based, clustered randomized controlled trial. The study was conducted in Ejere district located in the Oromia region of central Ethiopia, 52 km west of Addis Ababa. All villages in Ejere were eligible for clusters. We randomly selected three rural kebeles (equivalent to a ward) and all three urban kebeles within the Ejere district to include both urban and rural localities. Within the 6 selected kebeles, 105 garees (villages) were identified through Kebele leaders. We randomly assigned garees into treatment and control groups in a 1:3 (treatment: control) ratio: T1, maternal BCC only; T2, maternal BCC and paternal BCC; T3, food voucher only; T4, maternal BCC and food voucher; T5, maternal BCC, paternal BCC, and food voucher; and C, control. Randomization was stratified at the kebele level. Ownership of the food vouchers were individually randomized within the household between men and women. For female headed households, the women were given ownership of the food voucher.

The census data of 22,000 households living in Ejere district was collected during 2016 to identify eligible participants for the trial. For inclusion in this study, women had to be a resident of the study cluster, have children between 4-20 months, and consent to participate. Men were eligible if they lived with the participating women. Pregnant women or women who have children 0-3 months at the time of baseline was not eligible for the trial but was surveyed at baseline and follow-up to assess knowledge spillovers. The baseline survey was implemented from April to August 2017 before the start of the intervention, and the follow-up period was immediately after the completion of the BCC interventions (December 2017 to March 2018).

We found a total of 875 eligible mother-father-child group, all of which were included in the study for the treatment and control groups. There are 101 (15), 92 (11), 96 (14), 154 (13), 142 (15), and 290 (37) mother and child pairs (villages) randomly assigned to the treatment and control groups: T1, maternal BCC only; T2, maternal BCC and paternal BCC; T3, food voucher only; T4, maternal BCC and food voucher; T5, maternal BCC, paternal BCC, and food voucher; and C, control. We also surveyed 344 mothers who were either pregnant or had children between 0-3 months at baseline and follow-up.

The original planned sample size was 300 (50), 300 (50), 300 (50), 300 (50), 300 (50), and 900 (150) mother and child pairs (villages) randomly assigned to the treatment and control groups: T1, maternal BCC only; T2, maternal BCC and paternal BCC; T3, food voucher only; T4, maternal BCC and food voucher; T5, maternal BCC, paternal BCC, and food voucher; and C, control. However due to the political turmoil that was ongoing during the study period, we downscaled the study.

Randomization done in office by a computer.

o For the assignment of treatment groups, villages will be randomly assigned to one of the six groups including control.
o For the assignment of voucher recipients, all households within each voucher-assigned village will randomly assigned to either the mother or the father receiving the voucher.
o For market accessibility, a random half of all markets in the catchment area will be selected for the healthy food access intervention.

villlage (garee)

Yes

400 villages (planned)

2400 household (women-men-child pair)

50 villages in each treatment arm (planned)
150 villages for the control group (planned)