Major cholera outbreak in Yemen controlled by computer predictions
Cholera cases in Yemen have been slashed by a new system that predicts where outbreaks will occur.
Last year, there were more than 50,000 new cases in just one week – this year, the numbers plummeted to about 2,500.
The system has enabled aid workers to focus efforts on prevention several weeks in advance of an outbreak – by monitoring rainfall.
It comes as the UN says it is concerned about a possible “third wave” of the epidemic.
The deployment of the technology has been coordinated by the UK’s Department for International Development.
Prof Charlotte Watts, the department’s chief scientific adviser, said that the system had helped aid workers bring a rampant epidemic under control.
“We have thousands of people around the world that died from cholera each year,” he said.
“And I think this approach could really help put a dent into that figure.
“What this technology enables us to do is really home in to where we’re going to get new outbreaks, and respond really effectively.”
Last year, there were a million cases of the waterborne disease in Yemen. More than 2,000 people died and many of them were children.
It was the largest and fastest-spreading epidemic on record – and its rapid spread was caused by the destruction of sewerage and sanitation systems during the country’s civil war.
Although cases have reduced dramatically in 2018, the UN says it is concerned about a possible “third wave” of the epidemic.
The UK’s overseas aid department has worked with the Met Office to develop a system that predicts where cholera will occur four weeks ahead of time.
How does it work?
The Met Office produces a rainfall forecast for Yemen. Using its supercomputers, it is to determine the specific amount of rain that will fall and pinpoint the areas it will hit within a 10km (six-mile) radius.
These are important because downpours overwhelm the sewerage system and spread the infection.
The forecasts are used in conjunction with a computer model developed by Prof Rita Colwell, at the University of Maryland, and Dr Antar Jutla, at West Virginia University.
It draws on a variety of local information such as:
- population density
- access to clean water
- seasonal temperature
Together, this information enables scientists to predict the areas most likely to experience an outbreak, up to four weeks in advance.
