Research suggesting increased Covid-19 risk for dog-owners may be barking up the wrong tree

“They’re known as ‘man’s best friend’, but if you live with a dog, you may unknowingly be increasing your risk of coronavirus.”

“Walking your pet dog may raise your risk of catching coronavirus by 78 per cent”

Recent research from the University of Granada has claimed that people who walk their dog are at higher risk from contracting Covid-19 than people who either don’t have a dog, or have a dog they don’t walk. It goes on to suggest that this may be because dogs themselves are vectors for the virus.

There are various issues with how the research was conducted which means we should be cautious when drawing any conclusions from it. 

Perhaps most importantly, it remains unclear whether owning a dog specifically had anything to do with the finding. The results, even if they are accurate, could just indicate that people who left their home on a regular basis are more at risk from getting Covid-19. 

What did the survey do?

The study surveyed people in Spain from 4 April to 5 May during which there were restrictions on movement, similar to the period of lockdown in the UK. 

It asked whether they had or suspected they had contracted Covid-19 and for information on the respondent’s demographics and behaviours.

The survey was circulated using a link shared with people on University of Granada mailing lists. This is the first issue. No effort was made to obtain a sample representative of the Spanish population, and so the findings are not necessarily applicable at a national, let alone international, level. 

A bigger problem is to do with just how many variables the researchers analysed as potentially having a link with suspected or confirmed Covid-19 infection. 

Paul Hunter, Professor in medicine at the University of East Anglia, told us: ”In epidemiology one of the problems is with what we call type one errors.”

This is essentially a “false positive”—finding an association between two variables where none exists.

Professor Hunter added: “When a study investigates multiple risk factors the type 1 errors are more common. 

“Using a p<0.05 cut off for significance we would expect that for every 20 risk factors tested one (on average) would appear to be a significant risk factor. 

This refers to the fact that the threshold for determining statistical significance (in this case the “p-value”) is often set at 5%. This is the chance of finding this association, or a stronger one, by chance, if none actually exists.

However, if you test lots of variables as potential risk factors, you increase the likelihood of a result which appears to be statistically significant, but is not actually real and is in fact down to chance.

This research appeared to test around 40 variables to see whether they were associated with an increased likelihood to report having contracted Covid-19. 

Is it the dog, or the walking?

While any link between dog-walking and risk of Covid infection is unclear for these reasons, it’s worth noting that it isn’t perhaps that unusual to find a relationship between dog-walking and an increased likelihood to contract Covid-19. 

People who are walking their dog are, by nature, leaving their house—possibly many times a day—and so may be more likely to come into contact with other people outside of their household.

What is odd is that the research did not actually ask people how often they spent outside the home during the period of confinement in Spain, so the researchers couldn’t isolate any increased risk of having a dog specifically from any increased risk from being outside the home for however long each day.

The survey asked whether people went to certain shops, travelled to work, and walked their dog among other things, but not whether they just left their house for any reason, or how long they were spending outside the home. 

What has the potential to mislead (and perhaps cause unnecessary worry) is that the paper then moves from this uncertain association between dog-walking and the risk of contracting Covid-19 to theories which focus on the role of dogs specifically in spreading Covid-19. 

The university’s press release says: “The authors warn of the need among dog-lovers to take extreme hygiene measures regarding their pets, as it is not yet clear whether the owners were infected because the animal acted as the host for the virus and transmitted it directly, or whether they picked it up indirectly due to the increased exposure of the dog to vehicles of the virus (that is,objects or surfaces where the virus lies).”

This ignores the possibility that the association may simply be because people going outdoors more often are more likely to have reported contracting Covid-19.  

While it’s certainly true that there have been some cases of dogs contracting the novel coronavirus, it’s not clear that they can also pass it on. The US Food and Drug Administration says: “Based on the limited information available to date, the risk of pets spreading the virus that causes COVID-19 in people is considered to be low. At this time, there is no evidence that animals play a significant role in spreading the virus that causes COVID-19.”