There is a lot we simply don’t know about how to address malnutrition effectively. If we knew more, we could save millions of lives every year, since malnutrition is connected in one way or another to almost half of all child deaths. One reason for the gaps in our understanding about malnutrition is that the data to which we have had access has been better suited for observing general trends, like United Nations estimates of under 5 mortality trends, which does not highlight the remaining work that needs to be done. Because Ki partners have teamed up, we have individual level data on stunting and wasting in this age range, and we can investigate look more closely at the relationship between stunting and wasting and its role in mortality in early childhood. Our findings have refocused efforts and resources to accelerate progress towards eliminating childhood mortality.
Consider two manifestations of malnutrition: stunting and wasting. Stunting is a measure of chronic malnutrition. Stunted children, who are short for their age, are unlikely to ever reach their full cognitive or physical potential. Wasting is a measure of acute malnutrition. Wasted children, who weigh less than they should based on their height, are at imminent risk of dying.
By necessity, the field has relied on cross-sectional data, or data that looks at entire populations at a single point in time, to analyze both stunting and wasting. But cross-sectional data can only answer certain kinds of questions. Longitudinal data that tracks the same people over time can answer more questions, but it is harder to get.
Longitudinal data that tracks the same people over time can answer more questions, but it is harder to get.
For example, using cross-sectional data, it is possible to estimate the prevalence of malnutrition, or how many people are suffering from stunting or wasting at one point in time. It is not, however, possible to estimate incidence, or the proportion of people who start suffering from stunting or wasting over a given period of time. Longitudinal data, on the other hand, can help us learn when children become stunted or wasted, how these conditions progress, how long they last, and how children recover (or not).
This is where the Ki approach comes in. By combining data from multiple data sets, it is possible to piece together enough observed longitudinal changes to address some of these open questions. For wasting, we used data from 43 cohorts representing approximately 123,000 children from 35 countries. For stunting, we used data from 41 cohorts representing approximately 86,000 children from more than 30 countries. These large sample sizes allowed for dramatically more sophisticated analyses than had been possible in the past, including incidence analyses. These analyses, in turn, changed our understanding of stunting and wasting—and how countries are recommending they be addressed.
Based on prevalence studies, the common assumption was that children were most likely to become stunted or wasted at the age of 12 months or older, after they stopped exclusively breastfeeding. That is because prevalence was highest among that age group. However, incidence for wasting turned out to be highest in the first three months of life, and about half of stunted children became stunted in their first six months. In other words, although more children who are stunted or wasted fall within the 12-18 month age group than any other age group, most children become stunted or wasted earlier (and remain stunted as they age).
This finding is consequential, because many experts were focused on interventions that would come too late to help most children. Now, we know that we need to concentrate on interventions during the prenatal and early postnatal periods, too. We are now working with many partners to conduct more research and consider updating guidelines related to stunting and wasting based on the strength of these findings.
Now, we know that we need to concentrate on interventions during the prenatal and early postnatal periods.
Continue reading the published case study