Researchers model famine as a crisis that can still be interrupted

A community does not cross into famine in a single instant. Crops and livelihoods may be damaged by conflict, drought or an economic shock; then a second force can keep people trapped, such as insecurity or a policy that blocks humanitarian aid. More than 30 researchers suggest that mapping this chain could show where a crisis is building—and where it might still be interrupted.

The work appears in a special issue of the Journal of Public Health Policy, edited by Elena Naumova and Paul Howe of Tufts University. Its central idea is to treat famine as a dynamic system rather than either a sudden wave of starvation or a long, unstructured socioeconomic decline. Models can track interacting pressures, escalation and the point at which a crisis shifts into a state of very high mortality.

Different models answer different questions. Climate and statistical models examine how drought, extreme weather and food insecurity may increase future risk, and how planting alternative crops may help prevent conditions from worsening over time. Another study follows events in Yemen from climate shocks to food insecurity and then rising malnutrition, looking for warning signs in the sequence. A separate team studied the famine that struck southern Sudan in 1998.

The mechanism matters because pressure does not always lead to catastrophe. Some forces amplify one another, while others can counteract them. The researchers compare the task partly to forecasting a hurricane: understanding how the conditions combine can reveal which signals to watch. The comparison has a limit, though. Famines almost always involve human action or inaction, including choices that determine whether aid can reach people.

And so, concretely? Better models could help humanitarian organizations and governments recognize intensifying conditions before they become deadly, but only if they can obtain reliable data and act early. Crisis zones often have damaged infrastructure, incomplete information and serious political and ethical constraints on data sharing. The studies point to a path forward, while making clear that prediction alone does not feed anyone.