Sept. 20, 2021 — Cobra the dog has been hard at work at the Miami International Airport, sniffing masks proffered by American Airlines employees making their way through a security checkpoint. If she identifies a specific scent, she’ll let her handler know simply by sitting down. When this good girl sits, that means Cobra has detected an olfactory signal of the coronavirus, the virus that causes COVID-19.
Cobra, a Belgian Malinois, is one of two canines — her partner is One Betta, a Dutch shepherd — working this checkpoint at Miami International. They are part of a pilot program with the Global Forensic and Justice Center at Florida International University, using the detection dogs as a quick screen for people who have COVID-19.
Their detection rate is high, at more than 98%, and the program has been such a success that it’s being extended for another month at the airport.
If these two dogs continue to accurately detect COVID-19, they and other canines with similar training could be deployed to other places with lots of people coming and going at once, including other airports or even schools. In fact, COVID-sniffing dogs are in use in some university classrooms already.
But building up a big brigade of live animals as disease detectors involves some thorny issues, including where the animals retire once their careers are complete.
“When COVID first arose, we said let’s see if we can train these two dogs on either the virus or the odor of COVID-19,” says Kenneth Furton, PhD, a professor of chemistry and biochemistry, provost, and executive vice president at Florida International University.
His team had completed a study with what he calls “medical detector dogs,” animals that might be able to detect the odor of someone having a seizure. That led them to see how well the animals could detect other kinds of disorders.
Training a dog to sniff out specific odors starts with getting them to understand the task in general. Furton says that the animals first are trained to grasp that their job is to detect one odor among many. Once the dogs grasp that, they can be trained on just about any specific odor.
In fact, in addition to detecting seizures, dogs reportedly have been able to identify diabetes and even some cancers, such as ovarian cancer.
Furton says he’s not aware of any previous use of dogs to screen for infectious disease. That may simply be because nothing recently has struck with the global ferocity of COVID, driving humans to turn to their best friends for help.
Cobra and One Betta got their start learning to identify the presence of laurel wilt, a fungus that attacks avocado trees and kills them, costing Florida growers millions. With that expertise under their collars, the two dogs need only a few weeks to get good at detecting other smells assigned to them.
Training the Dogs, Safely
To train Cobra and One Betta on COVID-19 odors, Furton’s team first acquired mask samples from people hospitalized with COVID and people who did not have the disease. In battling the viruses, people produce certain chemicals that they exhale every time they breathe. When Furton and his colleagues compared the exhaled components trapped in the masks, they found differences between masks from people with COVID and those without.
Having confirmed that exhalations can be COVID-specific, the research team trained four dogs — Cobra, One Betta, Hubble, and Max — to detect masks from people with COVID among an assortment of mask choices. Before this step, though, the researchers made sure that any trace of active virus was destroyed by ultraviolet light so that the dogs would not be infected.
Each time the dogs accurately selected a mask from a COVID patient, their reward was access to a favorite toy: a red ball to chew on. Although all four dogs performed very well, yes, they did, Cobra and One Betta showed the most accuracy, outperforming their training colleagues. From their training scores, Cobra ranked first, with 99.45% accuracy. Despite her name, says Furton, One Betta was “not one better,” coming in second at 98.1%, which is still quite high.
Both dogs are good at their airport screening duties. If one of them sits after sniffing a mask at the checkpoint, the next step is for the mask owner to be tested.
From Aug. 23 to Sept. 8, the two canines screened 1,093 people during 8 working days, alerting on only one case, according to Greg Chin, communications director for the Miami-Dade Aviation Department. That person had tested positive for COVID 2 weeks earlier and was returning to work after quarantine, and their rapid test after the dog alerted was negative.
Furton says that there are some reports of dogs also alerting before tests can show a positive result, suggesting the dogs’ odor detection can be more precise. They hope to expand their study to see how tight the window of dog-based detection is.
For now, the detector dogs are doing so well that the program has been extended for 30 more days, Chin says.
As promising as this seems, using dogs for screening carries some logistical and ethical tangles. Training a canine army to deploy for high-volume detection points means that once the work is done, a whole lot of dogs will need a safe place to retire. In addition, the initial training takes several months, says Furton, whereas if a device were developed for screening, manufacturing could likely be ramped up quickly to meet demand.
The dogs might not need to retire right away, though.
“We envision that they could be redeployed to another type of detection for another infectious disease” if the need arises, Furton says. But in the end, when working with dogs, he says, there is “a moral connection that you don’t have to deal with using instruments.”
Although the pilot screening at Miami International is the first airport test, the dogs have also done this work in other venues, including at a state emergency operations center in Florida and in some university classrooms, says Furton.