Establishing a nutrition panel survey in Uganda

Nils Teufel
Nils  is an agricultural economist based at ILRI in Nairobi. Originally from Germany, he has worked extensively across South Asia and was based in ILRI’s office in New Delhi, India from 2006 to 2013.

Editor: In this month's blog article, we have asked Nils Teufel to share his work on a Nutrition research project in Uganda and Vietnam. The project utilises RHoMIS to look at food and nutrition security. A major focus of the research is to test diet recall and reporting accuracy over a 12 month period. The study in Uganda is beginning in August, while Vietnam began in March 2019.

The food and nutrition security modules of RHoMIS are critical for many evaluation studies as they provide data for the determination of major livelihood indicators. The core RHoMIS survey includes two main food insecurity indicators: the Food Insecurity Experience Scale (FIES) promoted by the FAO as an index of household exposure to food insecurity; and the Household Diet Diversity Score (HDDS), which is based on the concept of using the number of food groups consumed to evaluate a human diet. 

These indicators cluster foods of similar nutritional characteristics into groups, such as starchy, vegetable, fruit, dairy or eggs. Particular attention is given to foods providing vitamins and minerals, such as orange-coloured vegetables which are rich in vitamin A, which then form their own group. In the case of household diet diversity, the score is interpreted as a measure of food security: the ability of a household to access food of sufficient quantity and quality, while an individual’s diet diversity is interpreted to characterise this person’s quality of diet, or nutritional security.

Within RHoMIS surveys, it is standard for respondents to be asked to recall their food consumption experiences and practices over the past 12 months. In particular, the diet diversity questions focus on the household’s food consumption in the best and the most difficult months of the last 12 months. 

This is consistent with the overall RHoMIS approach for collecting farming system and household data and is common for many tools characterising farming systems. The approach allows the considerable seasonal variation in agricultural production and household livelihoods to be captured, all of which is typical for smallholder farming in developing countries. ​

A pig farmer in Vietnam. (Nils Teufel)

Seasonal scarcities and stresses in food availability can have considerable impacts on households’ well-being and are critical factors in determining food insecurity and malnutrition.

However, within nutrition research, recall periods of more than a week are generally avoided to ensure sufficient data quality, especially in regard to detailed diet information. It is assumed that remembering accurately which foods a household or an individual has consumed at a time several months past is too difficult and that the recall errors included in such data would bias any interpretation of results Therefore, the diet diversity results of RHoMIS data have often been questioned.

To test how far the recall error may compromise data collection on diet diversity with recall periods of up to 12 months a dedicated study has been initiated in Vietnam and Uganda including 200 households in each country. 
​
The objective of the study is to compare diet diversity data collected for the immediate past with data collected from the same household over a 12 month recall period. In addition, the concept of identifying months with the most favourable and the most difficult food security situation will be tested by comparing current diet diversity data with the RHoMIS identification of these months. Finally, it appears that diet diversity may be systematically underreported for the most difficult month. This will also be tested.

Thus, the RHoMIS Nutrition study has established two panels of households, one in Vietnam and one in Uganda. These panels are visited monthly over 12 months to record their household diet diversity. Subsequently, a final RHoMIS interview will collect diet diversity data for the entire study period by the standard 12 month recall process.

In Uganda, the study has been incorporated into an ongoing research project on pig genetics. This project has already established a longitudinal survey, collecting data at monthly intervals from 200 small-holder pig keepers in two districts, Hoima and Kamuli. Here, two field research assistants, one for each district, have established a strong relationship with local communities and in particular with the survey households. ​

RHoMIS survey practice  and enumerator training
​in Kamuli, Uganda, May 2019. (Nils Teufel)

So far, they have been collecting data mainly on pig production and pig breeding activities. In addition to the specific research objectives of testing diet diversity recall, the RHoMIS data will provide farming systems and household livelihood data to this project to put pig production and breeding results into a systems and social context.

From a research perspective, this study is a great opportunity to improve the understanding of how food security can be measured and what the trade-offs between data quality and capturing seasonal variation might be. Incorporating this into a production focused project will also allow various disciplines to collaborate more closely and gain greater appreciation of various research perspectives.

[Editor’s Note: soon we hope to have a blog article from the team in Vietnam, who are three month’s into their research]