Improving Diets in Schools in Vietnam: Testing Innovative Methods of Influencing Demand

Treatment 1: The treated children will be given lunch time presentations (nutrition information communication) and encouraged to share their knowledge with their parents (child-parent communication about healthier diets).
Due to the limited time and facility available at schools, we will keep the nutrition education and communication activities for children short and relevant. The messages will focus on the 4-6 types of fruit and vegetables, some of which are served to children during their lunch times, and some are considered of exceptional nutritional value “superfood”. The teachers will be trained to provide this message to children in a 5-minute show-and-tell talk right before the lunch time. Complementing the presentation will be leaflets/posters with the same contents. The talk will cover the benefits of eating FAV for children, the risks of not eating enough FAV, and healthy methods for FAV preparation. The materials for this activity are developed by NIN.

Children will be encouraged (with age-appropriate in-kind rewards such as stickers bearing fruits and vegetables cartoons) to share what they have learned during the nutrition education sessions with their family members. The children will be given tips to communicate with their parents, as well as provided with supplementary communication materials such as leaflets to give their parents.

Through their children, the parents will also be provided with instructions on how to provide recommendations for the school meals to food contractors and schools. They can send text messages to a dedicated line, or return a filled survey to the schools voicing their opinion of the school food.

These treatment activities will last for 5 weeks.

Treatment 2: In addition to the Treatment 1, the treated will also be provided with health snacks during the afternoon breaks.

In Dong Anh, no snack is provided in the afternoon, which prompted children to either bring their own snacks (e.g. sweetened milk or cakes, hardly fruits or nuts) or buy processed snacks outside schools. We will recruit a contractor for supplying seasonal fruits as snacks to the schools, ensuring the selected company meets legal, safety and any school-based requirements to provide food for children at school. The fruit will be provided for children during the afternoon break. The snacks will serve as (i) outcome measurement: if the information session motivates the children to eat more FAV, and (ii) a default healthier option which can potentially replace their less healthy snacks.
This snack provision will also take place in 5 weeks on a sub-set of children who were given nutrition lessons.
Control: No intervention on the control group.

2019-03-11

2019-04-30

The experiment has two groups of outcome variables: one for the children and one for their parents.
Outcome variables for children:
- Knowledge about vegetables and fruits
- Attitude towards vegetable and fruit consumption
- Consumption of fruits and vegetables at school
- Consumption of fruits and vegetables at home
Outcome variables for parents:
- Knowledge about vegetables and fruits
- Attitude towards vegetables and fruit consumption
- Consumption of fruits and vegetables

The study will involve 3 data collection rounds: baseline, endline 1 (right after the intervention ends) and endline 2 (six months after endline 1). Baseline data will be used to validate the assumptions, adapt the theories, re-estimate the power, and adjust the pre-analysis plans accordingly. Besides, balance tests will be performed on baseline data to check if randomization has resulted in groups similar in relevant covariates. We will use standard data analysis techniques used in randomization studies to measure the treatment effects (Khandker, B. Koolwal, & Samad, 2009).
Outcome variables for children:
A KAP (knowledge, attitude and practice) questionnaires will be used to measure the knowledge of and attitude towards vegetable and fruit consumption. A FAV consumption frequency question will be included under “Practice”. The questionnaire will be carried out in a face-to-face interview with children.
To measure the consumption of fruits and vegetables at school, research assistants (RA) will observe the children during lunch time/afternoon snacks to fill in the school food diaries. The food diary will not only cover fruits and vegetables, but also other foods, to put fruit and vegetable consumption in the overall context of a diet.
Lunch time: The RAs will hand in the food trays (with stickers to identify the sampled children) with standardized portions. After children finish eating, they will collect the identified trays, measure the leftover (or extra food – the RA can see if he/she asks for extra food from the teacher/enumerator; and/or estimate the amount of food she/he was given by another student) and fill in the school food diaries for each children.
Afternoon break: The RAs will observe if the children pick up the provided food or take their home food and fill in the school food diary.
School food diaries will be done in 2 days: the day when the presentation is given, and 2 days afterwards.
To measure the consumption of fruits and vegetables at home, children will be instructed by RAs to fill in the home food diary at home themselves (with the parents’ help). The home food diaries will also be done in 2 days as per the school food diaries.
Outcome variables for parents:
A KAP survey for parents will be implemented to measure the knowledge and attitude towards fruits and vegetables consumption. The parents will be instructed by RAs to fill in the food diary themselves. The food diaries will also be done in 2 days as per the school food diaries.
The following variables/indicators will be used to assess the knowledge level on FAV:
- The number of incorrect/”do not know” answers to the questions on FAV
- The proportion of respondents who got all the answers on FAV correct
The following variables/indicators will be used to assess the attitude towards FAV:
- The scores on perceived barriers to FAV consumption
- The proportion of respondents who have X barriers to FAV consumption
- The score on perceived benefits of FAV consumption
- The proportion of respondents who recognize all the benefits of FAV consumption
- The score on perceived susceptibility to FAV insufficient consumption
- The proportion of respondents who considered themselves not eating enough FAV
- The score on food preference (“whether you like the taste” of FAV)
- The proportion of respondents who like the taste of FAV
The following variables/indicators will be used to assess consumption of fruits and vegetables
- Consumption (in grams) of the fruits and vegetables explicitly promoted in the activities – during lunch time
- Consumption (in grams) of the fruits and vegetables explicitly promoted in the activities – during afternoon break
- Total consumption (in grams) of fruits and vegetables consumed out of school in a day
- Total consumption (in grams) of fruits and vegetables in a day
- The share of FAV in total food consumption (in grams) (in %) in a day
- The share of FAV in total food consumption (in Kcal) (in %) in a day
- Daily frequency of FAV consumption
- Weekly frequency of FAV consumption

We've described the experimental design above in the intervention component.

Out of 12 schools, 6 of them will have 2 grades who are in the treatment group, the other grade in the control group, and half of them will have 1 grade who is in the treatment group, the other 2 grades in the control group. Whether a school has 1 or 2 grades in the treatment group will be determined in a public lottery where the school administrators participate and draw straws. This design has the following advantages: It is easier to convince the schools to treat one entire grade than some classes from different grade; besides, coming from different grades the children from the treatment group and control group are less likely to interact with each other, preventing diffusion of treatment.

In each school, which grade(s) to receive the treatment out of the 3 grades (3rd, 4th and 5th) will be randomly selected.
All the classes within that grade will receive treatment. Half of the classes will be randomly assigned to treatment 1 or treatment 2. Randomization will be done using the statistical package Stata.
Assuming each grade has the same number of classes (7 classes per grade), in 12 schools there will be 252 classes, half of which will be in the control group and the other in the treatment group.

Randomization of treatment is done at the classroom level.

Yes

Approximately 252 classrooms

10 children in each of the 252 classes, which corresponds to 2,520 children (and their parents) in total.

126 clusters (classrooms) in the control group, 52 clusters in the treatment group 1 (communication activities only), 52 clusters in the treatment group 2 (communication activities and snack provisions).

Power calculation is done using the clustersampsi command in Stata. The intra-cluster correlation was set at the high level of 0.4. The population mean was set at 119.5 grams of FAV per capita per day of consumption, standard deviation is set at 44.8 grams. These numbers were obtained from the Vietnam Household living Standard Survey 2014 data for observations in Dong Anh district. Significant level Alpha = 0.05 and Power Target = 0.8. The minimum detectable effect for the FAV consumption is estimated to be 10.8 grams (~0.24 SD).