Why Children in Northern Rwanda Are Still Stunted Despite Green Fields

A new study using machine learning finds that child stunting in Northern Rwanda is driven not only by food insecurity and caregiving practices, but also by rainfall, vegetation, elevation, and geographic location. The research highlights the need for climate-aware, locally targeted interventions to effectively tackle chronic undernutrition.

Northern Rwanda is known for its rolling green hills and fertile farmland. Crops grow well, rainfall is plentiful, and agriculture is the backbone of daily life. Yet behind this picture of abundance lies a troubling reality: many children in the region are too short for their age, a condition known as stunting.

Stunting is more than just being small. It is a sign of long-term undernutrition that can affect brain development, school performance, and future earnings. Although Rwanda has made progress in reducing malnutrition, the Northern Province continues to record some of the highest rates in the country. Researchers wanted to understand why this problem persists in an area that appears agriculturally productive.

Turning to Artificial Intelligence for Answers

A team of researchers from Umea University in Sweden, the University of Rwanda, and Lund University decided to look beyond traditional explanations. Instead of relying only on standard statistical methods, they used machine learning, a form of artificial intelligence that can detect complex patterns in data.

Between October and December 2021, the team collected information from 615 households across five districts in Northern Rwanda: Musanze, Burera, Gicumbi, Rulindo, and Gakenke. They measured children’s height and weight, asked mothers about food security and childcare practices, and recorded the exact geographic location of each home.

They also gathered environmental data from satellite sources, including rainfall, temperature, vegetation cover, soil fertility, elevation, and how urban or rural each area was. This allowed them to study not only family conditions but also the surrounding environment.

Rainfall, Food, and Caregiving Matter

The results showed that stunting is shaped by many factors working together. Rainfall turned out to be one of the strongest influences. In some areas, good rainfall supports farming and improves food availability. In others, too much rain may cause flooding, damage crops, or increase waterborne diseases, which can harm children’s health.

Food insecurity was another major factor. When families struggled to access enough diverse and nutritious food, children’s growth dropped significantly. Even in farming communities, families may sell their best produce for income and eat less varied diets at home.

The number of days a child was left alone also made a difference. In farming households, caregivers often spend long hours in the fields. When young children are left without proper supervision, they may not be fed regularly or may fall ill more easily. The study found that children who were left alone more often tended to have lower height-for-age scores.

The Power of Place

One of the most important findings was that location itself matters. The researchers used a special method that allowed them to see how geography interacts with other factors. They discovered that stunting is not evenly spread across the province. It clusters in certain areas where environmental and social conditions combine in harmful ways.

Elevation, for example, showed a mixed effect. Children living at moderate altitudes tended to grow better than those at very high elevations, where access to healthcare and dietary diversity may be more limited.

Vegetation health, measured using satellite images, was linked to better growth when it reached certain levels, likely reflecting stronger agricultural potential. Urban areas generally showed better child growth than rural ones, highlighting differences in services, sanitation, and food access.

Rethinking Solutions to Malnutrition

The study found that machine learning explained much more of the variation in children’s growth than traditional methods. This means that the problem of stunting is more complex than previously understood. It is not only about poverty or food shortages. It is also about climate patterns, farming systems, caregiving practices, and geography.

The researchers suggest that solutions must be targeted to specific locations. Climate-smart agriculture, better support for farming families, improved childcare practices, and stronger food security programs should work together. Policies designed at the national level may need to be adapted for local conditions.

In Northern Rwanda, child growth is shaped by the land, the weather, and daily life. Understanding these connections could help policymakers design smarter interventions and give children a better chance to grow, learn, and thrive.