Well it’s not even Thanksgiving yet and COVID-19 cases are once again skyrocketing around the United States. Admittedly cases were way down in the Bay Area so I’ve been a bit lax about quarantine, in that I’ve been going to grocery stores to shop (masked obviously) instead of ordering delivery or pickup, and I’ve been having more socially distant outdoor hangouts with couples in the area. But it’s time to go back to being a shut-in freak because Kyle decided to have a Halloween party without masks. A Halloween party. Without masks. Like, it’s the easiest kind of party to have masks at, Kyle. What the fuck is wrong with you?
Anyway I was skimming science news headlines this week when I noticed this quite alarming statement: Spanish study finds dog owners have 78% higher risk of catching Covid. Oh, no! All my careful quarantine planning has gone to waste because of this diseased beast. And clicking through I found even worse news: “Activities that significantly increased someone’s risk of testing positive included accepting supermarket deliveries at home, which raised the risk by 94 per cent and was found to be more dangerous than actually going to the shop.” Oh no, not the grocery deliveries, too! What the frick!
Okay, let’s stick a pin in this balloon: I was pretty skeptical but at the end of the article I learned it was just copy and pasted from the Daily Mail. Yeah. Oops.
This is all based on a real study, which was published in early September and which you can read in full online. Spanish researchers surveyed 2,000 people in April of this year to ask if they’d knowingly contracted COVID-19 or thought they had and then a long list of their demographics and activities. Then the researchers went through and highlighted anything that looked like it had a strong correlation.
Okay, time for a quiz! Just based on that description, what problems do you think this study might suffer from? I’ll give you a minute to think it over.
Did you get it? First of all, a survey like this can’t tell you causation. Only correlation. They found that people who contracted COVID-19 were more likely to have groceries delivered than to go shopping in the store. Is it possible that they didn’t go into the store because THEY HAD A DEADLY VIRUS AND WERE SUPPOSED TO BE UNDER QUARANTINE? I’d bet a lot of those COVID-19 patients were also more likely to buy tea and cold medicine, but buying tea and cold medicine don’t increase your risk of contracting COVID-19. In fact, 6.9% of positive respondents said they got grocery delivery. 8.1% had been in a medical establishment. Hmm.
The other issue with this study is that it’s classic p-hacking. I’ve talked about it before — you can’t build a solid study by collecting a thousand different data points with no real hypothesis and then say “Look at this one data point that’s significant!” If you look at enough different attributes, you are more and more likely to find something interesting. That’s why scientists consider that kind of study preliminary. “We looked at a thousand different demographics and found that owning a dog correlated to COVID-19. Now someone else needs to follow up on that by specifically studying dog-ownership and COVID-19.”
This survey asked people about their sex, age, weight, physical activity, residence type, family size, number of children, pets, usage of masks, usage of gloves, usage of disinfectant on hands, surfaces, shoes, and products, clothes washing habits, use of and type of public transportation, travel to grocery stores, pharmacies, tobacco shops, banks and doctors’ offices, working at home or on site, traveling abroad, and general comorbidities. With all that, is it any shock that they found some correlation somewhere?
59% of the COVID-positive respondents reported that they lived with someone who tested positive. That’s a pretty solid correlation even amongst dozens of attributes. Compare that to the 6.9% of COVID-positive respondents, or 28 people, who said they walked their dogs.
Oh and by the way, the headline just says “dog owners” but in fact the study specifically highlighted people who walk their dogs. There is a difference.
Anyway, it’s not a bad study per se if it’s appropriately considered as a preliminary paper. But I stop wanting to give the authors the benefit of the doubt when they write in their discussion, “These results point to living with dogs as a strong risk factor for COVID-19 infection.” They…they really don’t. They conclude, “The results of this study demonstrate that living with dogs, working on-site, purchasing essential commodities by using home delivery service, and especially, living with a COVID-19 patient, have been the main routes of transmission of SARS-CoV-2 during the most restrictive period of confinement in Spain.” That is a completely ridiculous thing to say, especially considering that the study found absolutely no correlation between COVID-19 positivity and home delivery of ready meals or nonessential products. There’s no discussion of it, and no suggested explanation for why that is except for the single sentence, “These results would also suggest certain weaknesses in maintaining the hygiene chain in home delivery transportation systems.” Which is an absolutely galling thing to publish in September of 2020, months after scientists have established conclusively that the primary spread of the virus is through person to person contact and not through surfaces. Like, that is actively unhelpful and it should probably be retracted.
The simple fact is that we don’t know much about COVID in animals because it appears to be extremely rare, and the only instances we have of dogs and cats catching it is when their human already had it. Any other method of transmission would have to be through contact the dog has with someone positive, which they then bring to you, which would be a surface transmission, which scientists now think is very rare.
So please do not be afraid to walk your dog, regardless of what the Daily Mail or these study authors say. Wear a mask, wash your hands, and don’t take your dog to an indoor party thrown by COVID-deniers, for god’s sake.