Snakebite poisoning is a global health problem. Each year, more than five million people are bitten by snakes; Of them, two million suffer serious effects, and between 81,000 and 138,000 die as a result of them, according to the World Health Organization. Snake venom can cause skin inflammation, muscle destruction, shock hemorrhagic or paralysis of the respiratory muscles.
Snakebite poisoning is classified as a neglected tropical disease (ETD) and primarily affects people in remote, flooded or conflict-affected areas of low- and middle-income countries. The majority of victims – children, rural residents and agricultural workers in low- and middle-income countries – tend to be the poorest of the poor. It is a disease, in short, of barefoot people.
It is not a disease that can travel quickly to more developed countries, although it is likely that climate change causes migration of venomous snake species into new regions and unprepared countries.
The majority of victims – children, rural residents and agricultural workers in low- and middle-income countries – tend to be the poorest of the poor. It is a disease, in short, of barefoot people
Consequently, until now, it has not been a lucrative market for large pharmaceutical companies, nor does it occupy a prominent place on the public health agenda, either at the national or international level of policy makers, despite the millions of people affected. . In fact, research into effective, multipurpose, less complex, safer and more affordable antivenoms has not been prioritized.
I remember a time when we used photo albums to identify snakes in hospitals. Doctors Without Borders (MSF). The medical staff flipped through the photos with the patient to find out which snake had bitten him. I witnessed this firsthand in South Sudan and other countries while working there as a referral doctor for snake poisoning (bite from snakes that contain toxic substances that damage tissue) and other cases. neglected tropical diseases.
Today, experts have created a snake identification database with hundreds of thousands of photos of snakes from different countries, using artificial intelligence (AI) and machine learning. We have begun to apply this innovative approach on a pilot basis in two MSF-supported hospitals in Twic and Abyei (South Sudan).
We are improving our snake database together with the One Health Unit (One health) from the University of Geneva and the support of MSF teams on the ground in South Sudan and the association’s medical and innovation departments in Switzerland. Together we are developing an application for mobile phones with software that uses AI to help identify snake species in the field, distinguishing between poisonous and harmless snakes. This software can recommend the best action for a person bitten by a snake before they reach the hospital.
For the AI to identify a specific snake, we need a photograph of the animal (…). This is analyzed with software that compares it with thousands of images to identify the snake or add it as a new entry with GPS coordinates.
Currently, we are working on collecting quality photographs to enter them into this computer program. South Sudan is one of the countries with the fewest ecological studies on snakes and, at the same time, the country with the highest rates of admission for snake bites in MSF hospitals, especially from May to October due to rains and floods. MSF has treated more than 500 snakebite patients in just two hospitals in less than a year. The rainy season, which lasts four months, causes floods in this area of the country located in the Nile basin that forces snakes out of their burrows. In Twic County alone, my colleagues saw between 20 and 25 snakebite patients a week in the rainy season.
For the AI to identify a specific snake, we need a photograph of the animal, either when it is still alive (taking all precautions, with protective material, at a distance of more than two meters and using the zoom on the camera) or, when it has died (even so you have to continue being very careful since it could still bite). If someone has been bitten, the bite victim or someone nearby can also try to take a photo of the snake after the attack occurs, but this should be done with extreme caution. If it was not possible to take a photo at the time of the bite, our staff can return to the place where it occurred to take a snapshot of the snake, always with the greatest of precautions.
Once we have a photo, it is analyzed with software which compares it with thousands of images to identify the snake or add it as a new entry with GPS coordinates.
The first results are promising: sometimes the AI identifies snakes even better than experts. For example, it can differentiate between poisonous species such as the blow viper, the Egyptian cobra or the black mamba and other harmless ones such as various African snakes, but some “camouflage” themselves as dangerous snakes by imitating very similar cobras or vipers.
At Doctors Without Borders we treat more than 7,000 snakebite patients every year. Knowing at least the gender of the snake is essential when selecting the antidote
With better quality photos, funding, and more research, this app could help patients in real time, from identifying the snake to choosing the right antivenin.
At MSF we treat more than 7,000 snakebite patients every year. Knowing at least the gender of the snake is essential when selecting the antivenin to administer to prevent death and disability.
If the patient has been bitten but still shows no signs of reaction, we treat the wound and monitor him or her closely for 24 hours. Venom is injected in only half of snake bites and antivenin costs about $90 to $180 per dose (between €83 and €166). At a hospital like Mayen Abun (in Twic County), $90 is a huge amount, and the remote location means the team has to carefully judge the use of all medications to make sure we don’t run out. We have to make sure that all patients who receive a dose of antidote really need it because, in another vein, antivenom serums can have adverse side effects.
More research remains necessary. Often, patients receive the wrong treatment because the snake is not correctly identified, or valuable antivenin is wasted on snake bites that are painful for a few hours but not venomous, products that can also cause serious side effects (anaphylaxis or serum sickness). The antidote is scarce and extremely expensive, and can cost the patient between a month and a year of salary.
We hope that the use of AI will help us improve, both ourselves and the rest of the health actors in the most affected countries, in our knowledge of the different species of this animal and optimize the use of antivenoms.
