28 March 2018

The consumption of animal sourced foods by infants in low-income settings: Does it need to increase? And, if so, what is holding it back?

I have just finished reading a very useful new IFPRI Discussion Paper by Derek HeadeyKalle Hirvonen and John Hoddinott.  The paper pulls together data on the food intake of 112,553 children 6-23 months old contained within Demographic and Health Surveys (DHS) across 46 low and middle income countries since 2006 (when more standardised and disaggregated food recall data began to be collected).
The paper asks three questions: (1) how does infant food intake vary across country? (2) what is the association between food intake components and infant linear growth? and (3) how sensitive is a food’s intake to the price of the food in question? There is a specific focus on animal sourced foods (ASF), given the density of high quality proteins and micronutrients in these foods.
The paper has its limitations, for example the data are 24 hour recall data and so are subject to recall error and the statistical analysis (i.e. beyond describing the data) cannot go to the subnational level (because the prices are national). Nevertheless, this is by far the best attempt to pull together and analyse this type of data that we have and, for organisations such as GAIN, which deliver programmes with partners, it is invaluable, and, in some respects, startling.
So what does the paper tell us?
  1. Half of all 6-23 month old children in Sub Saharan Africa did not consume even one ASF in the previous 24 hours, whereas 84% of kids in Latin America and the Caribbean consume at least one ASF.
  2. Over 10% of all children in Asia and Africa consumed no food in the previous 24 hours (although it is not clear if that means no breastmilk was also consumed).
  3. Fish, eggs and legumes/nuts are the least consumed foods in all regions at 19.9%, 21.9% and 25.9% of infants, respectively.
  4. The consumption of any ASF and any fruit is positively associated with decreased likelihood of stunting, especially for infants over 18 months of age, but, curiously, vegetable and pulse consumption were not, even when looking only at the subset of infants who do not meet the minimum diet diversity score.
  5. For the group of infants as a whole, dairy, meat and fish consumption had the strongest and most significant associations with reductions in stunting prevalence (although egg consumption less so) and more so in the upper age groups, with fish significant in all age groups.
  6. By region, for 18-23 month olds, vegetable consumption (either Vitamin A rich or leafy green) had very little association with stunting rates in any region. Dairy was strongly associated with stunting reduction in Latin American and Caribbean, south, east and central Asia and east and southern Africa, but not elsewhere. Fish consumption was only significantly associated with stunting reduction in west and central Africa and Latin American and Caribbean (and in Latin America and the Caribbean the authors could not separate out fish and meat consumption). Vitamin A rich fruit consumption was only associated with stunting reduction in Latin American and Caribbean and egg consumption only in west and central Africa. Consumption of  “other fruit” was only significantly associated with stunting reduction in south, central and eastern Asia.  Surprisingly meat consumption was not significantly associated in any region at 5%–with the exception of Latin American and Caribbean (again, note the authors could not separate meat and fish consumption for this region).
  7. How important are prices in deterring consumption? Normalising food prices in each country (and we also have high income country data available from this ICMP source) relative to that country’s cheapest staple cereal, we find that it costs between 6.0 and 11.5 times as much to buy a calorie from leafy green vegetables as it does to buy a calorie from the country’s cheapest cereal (even in high income countries the ratio is 9).
  8. Prices of other foods relative to the cheapest staple are highest in the two African regions, which probably reflects, in part, high costs of moving foods around in this geography. The highest prices relative to the cheapest cereal are for fresh cows milk, vegetables and, interestingly, fortified baby cereal (which is 16-24 times as expensive as the cheapest cereal in Africa and Asian regions).
  9. Egg and dairy consumption is most sensitive to changes in their own price (as prices go up, consumption goes down) but the consumption of meat and fish is less sensitive to own prices.
So what does this all mean? First, the region, country (and presumably subnational area) really matters: consumption of different diet components is highly context specific. Second, the relationship between diet components and stunting is also quite context specific. Third, the relationship between the price of a food and its consumption is consistent (negative) although the magnitude varies by food.
The paper points out that if the public nutrition goal is to increase the consumption of animal source foods for infants (or any age group), then we need to reduce the price of these foods for all and this means expanding the availability of that food (while maintaining profit so that farmers actually grow it).  This strategy would of course raise issues of concern around sustainability – how can we make ASF cheaper for infants in low-income countries without leaving those very same children with fewer natural assets to manage when they grow up?
Current strategies to convince parents to increase the consumption of ASF for infants are hampered by the high price of ASFs because families will sell the foods for income to buy other foods and other things – but the food that is purchased is not as rich in high quality protein and micronutrients as the food that is sold. In this high price context it is less easy to convince farming parents to forego income to feed high priced nutrient rich food to their infants.
This paper resonates with GAIN’s mission—to make nutritious foods more available, affordable and desirable.  Affordability of nutritious food is a real constraint to nutrition status and this paper presents data from 46 countries to show this. But the results in the paper also point to potential sustainability trade-offs.  We need more science to chart any trade-offs (or synergies).
In short, we need to get the price of nutritious foods down in a sustainable way—if we don’t these foods will not be eaten by the people that need them the most.