Bart Weetjens – TEDx Talk TRANSCRIPT
All right. So we get a lot of crazy ideas here and I’m coming with maybe even a more crazy idea.
We train rats, to save human lives, by detecting landmines and tuberculosis. I brought a live rat here. It’s not really a live one but it’s actually an example of a giant rat. And that’s what we train. We train giant rats.
And this may seem really crazy to you but actually it’s an example of what the professor in the school earlier expressed, it’s actually a good example of a design process.
Because this may sound crazy to you at first glance but in practice, if you look at the situation, the context in which these rats are used, then it may make a lot of sense.
All of this approach is actually based on the empathy, based on putting ourselves in the situation of subsistance farmers in Africa. Families like these who actually cannot access their farmland, they cannot fetch water, or it’s perilous to fetch water, it’s perilous to fetch firewood, because of the danger of exploding landmines.
And really, this brings us to the problem.
Most of these communities are at the bottom of the pyramid, living in Africa, Latin America, Southeast Asia. They are actually affected by undetected threats. And tuberculosis is one of those threats, landmines is another one. And these two particular threats, this organization I founded, Apopo, is working on.
Because we believe, that actually it is possible, if we utilize resources in our own environment, that we can come to solutions which are much more sustainable than what is often imposed from outside. And, let me go at once to how this system works.
Actually, here you see one of our trainers, he is in safe area, on the side of the road and there’s a landmine a bit here on the side, actually animal smells it and it wants to get there but it has to follow the search pattern, to systematically search the field without leaving any blind pockets behind.
And you’ll see in the next row actually the animal starts scratching on the soil, I’m not sure there is sound but what you should hear now is this sound, which is a click sound and the animal actually comes back for the food reward which is banana or peanuts. Very simple, these animals work for peanuts.
So, why using rats. Well, first of all, these animals are endemic to Africa. They are giant African pouched rats and you find them all over Sub-Saharan Africa. They have more genetic material allocated to olfaction than any other mammal species. They don’t rely so much on audition or sight, they actually mostly get their perceptory input through smelling and it’s something they start immediately at birth.
This is what the animals actually look like when they’re 28 days old. Then they open the eyes and then we immediately wean them from the litter. We socialize them, we bring them in all kinds of environments, expose them to different sounds, different smells, different people.
They drive around in a car, they go with their trainers wherever and they learn that humans are friends. Really important, because, if the animals are well-trained, they can be very sociable, they can be very intelligent, they can be very reliable pets, they can be even very clean pets.
In the next phase, when they’re well-socialized, we actually start clicker training. We call this, this click sound, this is a simple metal clicker and we pair this click sound with food reinforcer. In this case you see mashed banana in a syringe but we also use pellets, we use all kinds of primary reinforcers, the foods.
And these animals love to perform repetitive tasks in return for a simple food reward. I think all of you have seen already a hamster, in a cage, that is happy by running around in that wheel for hours.
And in a way, well, we should be grateful to the animal that they want to do this work, because it’s… landmine detection is a task that requires high concentration, most of the problem occurs in very hot areas, you have to wear a lot of protective gear, it is really a difficult, dangerous task and rats can help us enhance it a lot.
So back to the training. Once this click and foods, click and foods is associated with each other, we can start now shaping the behavior of the animal. And you can’t see it well from the picture, actually on this one you can actually.
There are three holes in the cage. The animal actually learns to put its hole in the…its nose in the hole for three seconds which is really long for a rat. And every time it does that, it gets again a click sound and a food reinforcer.
And under that hole we can put a target scent. In case of landmines, the target scent is explosives, in case of tuberculosis, they’re positive sputum samples. Gradually, we make it more difficult, the animals work in a harness, they learn to detect sources of explosives in the open and later on they detect real landmines in a minefield.
At the end, before we ship the animals to operational sites, the animals undergo an accreditation test. They, in two consecutive days, they have to cover eight boxes of 100 square meters each, and have to find all the mines in all of these boxes, which is really difficult.
Once they pass this test, they are shipped abroad, and again, in the country, tested according to the same international mine action standards by an external party. It’s really wonderful to see this interaction between the animals and the trainers.
After the animals are ready, they actually follow their trainers back to the cage, just like a dog would do. This is, actually our capacity of… a mine action capacity in Mozambique, three years ago. And this is our current capacity. I don’t show this to actually blow the trumpets, but it is to demonstrate that the small rat capacity, you see these people here on the side, a small rat capacity can enhance the output of a mine action operation considerably.
And you have mechanical tools, you have manual operators with metal detectors, and all these are combined for a relatively slow, tedious process. Where the rats come in, they can speed up this process considerably and make it a lot cheaper.
To give you an example, where rat and a trainer combined can do 200 square meters in half an hour. For the same 200 square meters, you’d need actually four people to work an entire day with metal detectors to get that done. In combination, rats supporting operation with mechanical tools and manual operators utilizing an innovative land release approach, we can actually enhance mine action and provided at only 60% of the cost of the traditional mine action operation.
We’ve been bringing this mold now in several countries. Our headquarters are in Tanzania. Over the last 15 years we have developed this technology to what it is now.
We talk about ‘detection rats technology’ and really it is a technology, because it’s very standardized. We can train a rat in this place, send it abroad and is there used by somebody else according to same procedures.
So we have operations in Tanzania. We have mine action operations in Mozambique and Angola. And we’re also expanding into Southeast Asia, in Thailand, Cambodia and also in Myanmar, where there’s a lot of residual landmines left.
Let me talk a bit about tuberculosis problem as well. And for the TB detection application the animals don’t get out of setup. So they’re not immediately in contact with patients.
What we do, is we collect samples from hospitals, put them in these cassettes ten by ten, and insert this cassette in this setup, and watch now carefully, because this goes really very fast, it’s already at the third sample. So the animal only needs a fraction of a second to keep the nose in the hole, the trainer is now providing a click sound and the animal comes for a food reward.
Again very, extremely simple. And then if you know that these folks have like a secondary level of education, then we provide them with lot of skills, expertise, with which they can enhance their development and tackle those difficult, expensive detection challenges more independently.
Look how the animal is actually continuing here, so it takes, it continues where it had been rewarded. Here’s an indication on another sample, for which it will not be rewarded. Actually, in practice, this is an operational sample, we only reward them on known samples.
This sample, after rat indication will be reconfirmed by other techniques. But by using this method, we can actually enhance case detection by over 30% in the 17 hospitals that we work together with currently in Dar es Salaam, Tanzania. And maybe we just go over the system’s approach.
So a patient with pulmonary complaints goes to a hospital for a microscopic test, microscopy, which is not very reliable. We collect all the samples from these hospitals and bring them to a central facility unit, both the positives and negatives. We inactivate all these samples by heating them at 90 degrees for 30 minutes. We train the animals on indicating positive, known positive TB samples, then the animals run through the setup with operational samples and pick out those that have been missed by microscopes.
We then reconfirm these with a higher level of technology, with a fluorescence microscope or a gene expert, and this algorithm together jointly increases case detection rate by over 30%.
Now the challenge we face is to find these people back. We’re talking African urban slums. People all have a mobile phone but they don’t necessarily have a fixed address. So we trace people by means of mobile phones.
And currently we manage to reach some 70% of those people that have been detected by the rats, confirmed, to enroll into treatment.
So if you compare rats and microscopes, who is then better? Well, simply, microscopy on a routine basis, and I’m talking simple light microscopy like they’re using in dispensaries in Africa, has less than 50% case detection rate. That’s in terms of sensitivity.
Traditionally, it’s a highly specific, because it’s a visualization technique, so it’s a highly specific test. In combination with rats, rats are… we tested ten rats on a number of samples and the average per rat was 67% in terms of sensitivity, still with a relatively good specificity of 94%.
So using several rats together, we can actually have a highly sensitive test and then using microscopy to confirm, we have both highly sensitive and highly specific test together.
And really, I wanted to conclude this talk, because it actually shows that, well, these 3600 people that have been found by rats, the millions of square meters that we have been returning to populations, the thousands of land mines that we found in this process and surprisingly the tens of thousands of unexploded bombs and small arms and ammunition, if you look at the impact created by this purely, well, out of the box thinking, let it be.
But in essence, it is based on empathy, and on respectfully approaching those natural resources and harmonizing with them in order to foster a better world, I think this can be an inspiring story for all.
Thank you very much.
Download This Transcript as PDF here: Rats That Sniff Out Landmines and TB_ Bart Weetjens at TEDxBratislava (Transcript)