Sniffer dogs show 88% accuracy in detecting COVID-19 positive at airports and train stations, says study
- 24 May 2021
Sniffer dogs can be trained to detect more than 90 per cent of COVID-19 infections even when patients are asymptomatic, according to research published on Monday, which authors hope could help replace the need to quarantine new arrivals.
Using their remarkable sense of smell — which can pick up the equivalent of half a teaspoon of sugar in an olympic-sized swimming pool — dogs have already shown that they can sniff out maladies such as cancer, malaria and epilepsy.
Several previous studies have shown proof-of-concept that dogs can detect SARS-CoV-2. While even the quickest COVID-19 tests take 15 minutes to show a result, the dogs can sniff out the disease in seconds. This means that two dogs can together screen 300 people in half an hour, researchers said.
This could make the sniff test "a suitable method for mass screening," Professor Logan at the London School of Hygiene & Tropical Medicine, was quoted as saying. Mass screening can be done at airports or at busy train stations to help prevent a super-spreading event.
The dogs picked up roughly 88 per cent of positive cases — meaning, for every 100 cases, the dogs failed to recognise just 12 infected people. But out of 100 people who did not have COVID-19, the dogs wrongly suggested — via the sniff test — that 14 of them were infected.
The research team thus does not recommend dogs alone to sniff out positive cases, but use them as an additional screening tool alongside more conventional tests. Researchers from the London School of Tropical Medicine wanted to see if dogs could detect a distinctive odour given off from chemical compounds associated with someone who is COVID-19 positive but doesn't show symptoms.
They gathered samples of clothing and face masks from people who had tested positive for mild or symptomatic SARS-CoV-2. Samples of the socks of 200 COVID-19 cases were collected and arranged in lab tests for six dogs that had been trained to indicate either a presence or absence of the chemical compound.
The dogs needed to be trained not to identify "false positives" in a bid to hack their reward system and obtain treats even if there were no COVID-19 samples in a given test. "This means that the dog fully understands and gets a reward for a correct negative as well as a correct positive," said Claire Guest, from the school's Faculty of Infectious and Tropical Diseases.
Overall, the dogs were successfully able to identify between 94 and 82 per cent of SARS-CoV-2 samples. The researchers then modelled how effectively these success rates, combined with traditional PCR tests, could help detect mild or asymptomatic COVID-19 cases. They found that using dogs to screen arrivals at terminuses such as airports could detect 91 per cent of cases, resulting in a 2.24 times lower rate of transmission than with PCR tests alone.
'Important start'
Authors of the research, which has yet to be peer-reviewed, said they hoped it could eventually replace the need for travellers to quarantine — which necessarily disrupts every arrival even though the vast majority are not COVID-19 positive. "The key thing is that dogs are significantly quicker than other tests," said co-author James Logan. "What we're suggesting is that dogs would give the first initial screening, and then those (arrivals) that were indicated as positive would then receive a complimentary PCR test."
The team said that out of a plane full of arrivals — around 300 people — less than one per cent were statistically likely to be carrying SARS-CoV-2.
Under current quarantine regulations employed by some countries, all 300 would need to isolate, causing significant inconvenience. But given the sensitivity of trained dogs, a maximum of 35 people on board would be indicated as positive, the paper said. Of these, only around 3 would be expected to return a positive PCR test.
"This is a really important start and could lead to a useful, usable system," said Mick Bailey, professor of Comparative Immunology at the University of Bristol, who was not involved in the research. "But there's a lot more validation needs to be done before we could be confident that the dogs can reliably and specifically detect asymptomatic SARS-CoV-2 infection in people in airports and train stations."
