In the poorest parts of the planet, being bitten by a venomous snake is a daily reality and every second that passes before treatment with antivenom is a matter of life and death. “We call the time between bite and treatment the golden hour“, explains Gabriel Alcoba, a snakebite expert doctor from Médecins Sans Frontières (MSF) who has worked for years in countries such as South Sudan, Cameroon and Nepal.
During his long experience, this doctor has seen all types of cases and knows that information about the snake is vital, whether it is a description of witnesses, a photograph or the corpse of the animal , even if the authorities recommend not trying to kill him. “I remember the case of a 7-year-old boy who put his hand in a burrow and was bitten by a cobra,” he tells elDiario.es. “The father reacted and the snake also bit him. “We were only able to save the child.”
According to data from the World Health Organization (WHO), snake venom kills between 81,000 and 138,000 people worldwide each year and leaves 400,000 people disabled for life. Many of the 2.7 million people who are bitten by venomous snakes each year are usually plunged, at the same time, into forced migration due to conflict or natural disasters, increasing the risks of crossing flooded areas and receive a stroke.
Some patients arrive after 24 hours in very serious condition, with gangrene which ends in amputations.
Gabriel Alcoba
— Snake bite expert doctor from Doctors Without Borders (MSF)
Lacking access to medical care, victims’ first recourse is generally a healer, which increases the risk of a fatal outcome. “They come to us with tourniquets and incisions, sometimes they have rubbed urine or cow feces on them, things that can be very toxic,” Alcoba explains. “Some patients arrive after 24 hours in very serious condition, with gangrene which ends in amputations. » Knowing whether it is a venomous snake or an imitator is also important to use scarce resources wisely and avoid the risk of anaphylactic shock, because antivenoms, made from horse serum, can trigger a allergic reaction.
AI against the black mamba
This harsh reality led Alcoba and Rafael Ruiz de Castañeda of the University of Geneva to develop a program based on artificial intelligence (AI) that helps local teams make quick decisions. The system, presented in the magazine Tropical diseases neglected by PLOSis already underway in an MSF pilot project in South Sudan. Its basis is a machine learning program that has been trained on more than 380,000 images of snakes from around the world and is used to identify the species and its venom type.
“Even if in some places there is very little information, we manage to improve the results and the algorithm reaches success rates of 90%,” Ruiz de Castañeda explains to elDiario.es. “Ultimately, it lets us know what’s going on there, because we have a big lack of information; “We know people are dying, but we don’t know what bit them. »
We have a great lack of information; We know there are people dying, but we don’t know what bit them.
Rafael Ruiz de Castaneda
— Researcher at the University of Geneva
Researchers are aware that there is a very complex reality and different levels of information and access to treatments depending on each country and region. If patients arrive at a health center in time, doctors can try to find out which snake it is from the symptoms: swelling and bleeding can indicate vipers, while in cases of muscle paralysis, it could be a mamba or a cobra. All this conditions the treatment which, in the case of the black mamba, can include between 6 and 9 doses, but in the event of an error, a large quantity of antivenom can be wasted, which constitutes a very rare and expensive treatment.
A growing problem
“When I worked in South Sudan in 2014, we were seeing about three patients a day, and ten years later the situation has gotten worse,” says Dr. Alcoba. Between January and July this year alone, MSF treated more than 300 snake-bite patients in health centers across the country. Doctors chose this region because it has one of the lowest numbers of snake studies but has one of the highest rates of bite admissions.
“Some snakes look very similar, they camouflage themselves as poisonous and it is very difficult to identify them, there is a fairly high level of error,” he underlines. Another case that shocked him was that of a 4-year-old boy he treated in 2016 who was bitten by a snake on the right cheek, just below the eye. “We didn’t know which snake had bitten him, we only saw that it was swollen and that’s dangerous because the inflammation can affect the respiratory tract,” he recalls. “The father had seen a snake in the house, small and black, we didn’t know if it was a cobra or if it was a Atractaspisin which case there is no antivenom. It is in this type of case that the tool can prove vital, he insists.
“Many situations can arise, we want to use the tool to map and know where the most dangerous snakes are to react immediately,” explains Dr Alcoba. “We now understand which species affect which populations and which parts of the world,” adds Ruiz de Castañeda. “And now we also know what treatments need to be implemented in this part of the planet, because the problem is that the ministries of health buy the products, but they don’t know very well how to distribute them, because they don’t know not what’s there.
The new tool will not only be used to react to cases that arrive, but also to prepare. “Knowing that there are venomous snakes in an area, thanks to the collaboration of users, will allow us to anticipate the arrival of possible victims,” concludes Ruiz de Castañeda. “Identify their presence using AI and alert the area doctor to prepare to receive cases. “It’s just one more grain of sand, one more contribution to what we call zoological diagnosis, which in certain regions of the world can make a difference.”
