Food insecurity, caloric intake and nutritional status among children under 5 years old: a predictive modelling analysis of the MAL-ED multi-country cohort

The study examined whether statistical models could accurately predict caloric intake or changes in nutritional status (measured by weight-for-height ratio) in children under 5 years based on household food insecurity and other factors. The analysis used data from the MAL-ED cohort from 2009-2014 across eight countries (Bangladesh, Brazil, India, Nepal, Pakistan, Peru, South Africa, Tanzania) with 2,957 to 23,651 child observations. Researchers tested three models: predicting nutritional status changes based on food insecurity, predicting nutritional status changes based on caloric intake, and predicting caloric intake based on food insecurity. All three models showed low performance and significant prediction errors. The results suggest that statistical prediction of key variables leading to childhood malnutrition may not be feasible, which has implications for developing models that depend on empirical data rather than statistical shortcuts.