Can someone more knowledgeable than me about epidemiology please help debunk this?

(30 Posts)
TinyGarden Tue 17-Nov-20 17:56:44

https://scitechdaily.com/are-dogs-spreading-sars-cov-2-study-finds-living-with-a-dog-increases-risk-of-contracting-covid-19/

So far this study has only been reported in such illustrious tomes as The Daily Mail and the Metro.

Actual study here:

https://www.sciencedirect.com/science/article/pii/S0013935120311208?via%3Dihu

I know correlation is not causation and it looks like it's an online self report study. But would very much appreciate further scepticism from people who (unlike me) really know what they're talking about.

Not only cos I don't want to have to worry about my parents interacting with my dog (when it's allowed again). But also because it tries to paint an association between supermarket home deliveries and Covid. confused

Took me months to persuade my parents to go for online deliveries rather than pop to supermarkets!

OP’s posts: |
TinyGarden Tue 17-Nov-20 19:24:37

Anyone please? Hoping the data around dog walking (and supermarket deliveries!) is not strong.

OP’s posts: |
damnthisvirusandmarriage Tue 17-Nov-20 19:41:54

Following with interest. I live with three dogs.
I also walk dogs of 20 odd people. Abs train many more....

amicissimma Tue 17-Nov-20 19:44:11

I don't have a dog and I don't have supermarket deliveries and I haven't had Covid.

There you are - conclusive!

(I thought the washing groceries thing had been discredited.)

Gloschick Tue 17-Nov-20 19:48:09

I've only had a skim read but the article doesn't prove anything. There is a decent p value for living with a dog, but there are several issues. Firstly it isn't even definite covid, just symptoms suspicious of covid, so it is unclear how many people actually had covid. Also, it was done in the height of covid in Spain when we saw all those images of Spanish people sitting at windows, so any increase could just reflect the fact that dog walkers were leaving the house a lot more. There could be any number of associated variables eg dog owners more likely to live with other people therefore more likely to be exposed etc. I wouldn't be rehoming my dog based on the article!

TinyGarden Tue 17-Nov-20 19:51:57

Thanks @Gloschick for taking a look.

Yes, deffo no danger of rehoming! Just bit worried about my elderly parents.

It's strange as at the start of the summer there were loads of studies (or so I thought), that said pet transmission (whether via fur or other means) not a worry.

The supermarket deliveries thing is odd too.

OP’s posts: |
TinyGarden Tue 17-Nov-20 19:54:22

I like your science @amicissimma grin

OP’s posts: |

Advertisement

Gloschick Tue 17-Nov-20 20:26:12

The confidence interval crosses 1 for home delivery, which basically means it isn't statistically significant.

TunMahla Tue 17-Nov-20 20:30:27

Also, are not 'dog' people supposed to more often have an extroverted personality type and so socialise more? Speaking as an introverted 'cat' person ;).

SexTrainGlue Tue 17-Nov-20 20:46:32

I've only skimmed it, but they appear to have omitted description of the population reguiations at the time they collected data from.

Spain was pretty strict, wasn't it? Not even exercise, but with an exception for talking dogs out for toilet walks. So it's possible that what it means is simply that 'those who go out in addition to necessary shopping, medical attention and permitted essential occupations had higher infection rates'

Or am I getting muddled with French/Italian restrictions? If so ignore me!!

orangenasturtium Tue 17-Nov-20 20:50:21

As Gloschick said, the cases aren't confirmed as COVID 19.

It could be that dog walkers were leaving the house far more often than non dog owners. I believe exercise was not allowed during the lockdown. Assuming dog owners walk their dog at least once a day, they would be going out 7 times a week, meeting more people, spending more time in confined communal indoor spaces eg apartment lobbies and lifts, compared to once a week or never for many households.

However, it could also be that the dogs are acting as fomites, bringing in the virus from the pavement on their paws and spreading it through the home, on furniture and floors, touching people. Given that there is at least one confirmed case of a dog being infected with SARS CoV 19, it could be possible that there is dog to human transmission.

There is no way of knowing without further investigation.

It's not that surprising that home deliveries could be a greater risk than shopping in the supermarket. The number of contacts will be greater in a supermarket. However, the number of people who will have handled the food in the past 24 hours will probably be a lot greater for home deliveries (unpacking the delivery, picking the delivery, packing the delivery, and the driver), whereas many items in the supermarket will have been untouched for 24 hours or more. Also, all of the food will have been kept in a cold warehouse, then a refrigerated van, better conditions for spreading and preserving the virus than a supermarket. If my delivery drivers are anything to go by, they are often out of breath from unloading the shopping, so although you might spend less time in close contact with them than a check out assistant, it is a high risk contact, particularly if they are delivering to an apartment front door, so indoors and less well ventilated than a supermarket.

MedSchoolRat Tue 17-Nov-20 21:27:10

There's a lot of bad covid research.

I'd like to leave it there, but here goes the longer version:
It's observational, based on self-diagnosis (!!) and self-reported behaviour. 44% of the respondents had post-grad degrees. Does that sound you like the average Spanish person?

I dunno what an elastic net model is, it seems to be a method for choosing covariates, they don't even bother to give a reference for this. Their significance threshold seems to be p < 0.1 (most studies like p <0.05 instead).

And even if walking the dog was hazardous.... was it walking the dog or stopping to chat and share a smoke with the neighbours that created a problem. Do they know?

Case-control is a type of fairly robust observational study design to look for genuine risk factors. There's something called the hierarchy of evidence and this design (cross sectional survey) is at the very bottom of the hierarchy.

"In this study, no different susceptibility to the virus was found between men and women. ... The present study did not show effect of age on the estimated prevalence of COVID-19."

and yet we know that older people & males get more symptomatic covid illness. They weren't detecting covid at all so no wonder they found no difference for age/sex. That explains that.

Sure. Whatevs. Follow #epitwitter for potential evisceration.

TinyGarden Tue 17-Nov-20 21:47:03

Thanks @MedSchoolRat @orangenasturtium @SexTrainGlue

Love the divorce rate in Maine correlates with per capita margarine consumption smile

Thanks all for the critiquing.....this has cheered me up a little.

Does make you wonder how responsible the authors are in publishing - especially when you consider their bar for 'statistically significant' seems so low...

OP’s posts: |
Sb2012 Tue 17-Nov-20 22:56:48

amicissimma

I don't have a dog and I don't have supermarket deliveries and I haven't had Covid.

There you are - conclusive!

(I thought the washing groceries thing had been discredited.)

2 dogs and we do a lot of online supermarket deliveries (even before covid)
Also no covid here.

QueenStromba Wed 18-Nov-20 09:50:49

orangenasturtium

As Gloschick said, the cases aren't confirmed as COVID 19.

It could be that dog walkers were leaving the house far more often than non dog owners. I believe exercise was not allowed during the lockdown. Assuming dog owners walk their dog at least once a day, they would be going out 7 times a week, meeting more people, spending more time in confined communal indoor spaces eg apartment lobbies and lifts, compared to once a week or never for many households.

However, it could also be that the dogs are acting as fomites, bringing in the virus from the pavement on their paws and spreading it through the home, on furniture and floors, touching people. Given that there is at least one confirmed case of a dog being infected with SARS CoV 19, it could be possible that there is dog to human transmission.

There is no way of knowing without further investigation.

It's not that surprising that home deliveries could be a greater risk than shopping in the supermarket. The number of contacts will be greater in a supermarket. However, the number of people who will have handled the food in the past 24 hours will probably be a lot greater for home deliveries (unpacking the delivery, picking the delivery, packing the delivery, and the driver), whereas many items in the supermarket will have been untouched for 24 hours or more. Also, all of the food will have been kept in a cold warehouse, then a refrigerated van, better conditions for spreading and preserving the virus than a supermarket. If my delivery drivers are anything to go by, they are often out of breath from unloading the shopping, so although you might spend less time in close contact with them than a check out assistant, it is a high risk contact, particularly if they are delivering to an apartment front door, so indoors and less well ventilated than a supermarket.

Or the people most in need of a supermarket delivery are people who are self isolating due to covid or suspected covid.

GreyishDays Wed 18-Nov-20 10:01:01

I also think it looks like a reasonable study, but being an environmental cross section type of thing they can’t account for factors separately. Eg people who was their groceries might also do something else protective.

Also yes to the dates of the study, they weren’t allowed to do anything else in that time I think. Wasn’t it Spain who had funny tales of hiring out dogs so people had an excuse to walk? So that could be people leaving the house, touching gates, speaking to other people etc.

QueenStromba Wed 18-Nov-20 10:01:58

amicissimma

I don't have a dog and I don't have supermarket deliveries and I haven't had Covid.

There you are - conclusive!

(I thought the washing groceries thing had been discredited.)

Advice seems to be that you are unlikely to catch covid off your shopping but that giving it a wash won't hurt. Whether you wash your shopping is probably a decent proxy for how careful you're being in general.

PowerslidePanda Wed 18-Nov-20 10:18:29

Before everyone starts panicking about home deliveries, can I just point out that the way it works in Spain/in that study is not the same as in this country? They're talking about people who went to the supermarket themselves, picked up and paid for the items themselves and then paid the supermarket to deliver them, rather than taking them home themselves. And concluding that it's higher risk than taking your own shopping home (obviously, because it's the same process but with extra people involved!) It can't be compared to the risk of avoiding the supermarket altogether.

PowerslidePanda Wed 18-Nov-20 10:22:30

A higher prevalence of the disease was also detected among those surveyed who had purchased their basic products at a supermarket and then used the home delivery service, compared to those who brought their shopping home themselves (the risk increased by 94% among the former group).

MereDintofPandiculation Wed 18-Nov-20 10:51:00

Their significance threshold seems to be p < 0.1 (most studies like p <0.05 instead). What the p value means is that the probability of getting the observed result or one even more extreme by random chance alone is p. So if they really are using p<0.1 (I can't see the article, only the abstract), and they are looking at 30 different factors, then chance alone would mean that on average 3 of them would meet the p<0.1 criterion.

TinyGarden Wed 18-Nov-20 13:29:44

@PowerslidePanda Gosh that's a huge difference in context isn't it, if in Spain 'home deliveries' often meant going to the store, picking products off the shelves then getting them delivered.

Reassuring for the U.K. context (where I'm not aware of that being a service - maybe it is for people without cars etc?).

OP’s posts: |
TinyGarden Wed 18-Nov-20 13:30:41

PS. Sorry - actually just seen on another thread using the @ symbol is rude....didn't realise. Will try to bold / quote instead!

OP’s posts: |
MissMatchedClaws Wed 18-Nov-20 13:49:29

Think others have pretty much said it, but cross sectional survey:
do people who remember walking their dog believe they took a risk, and therefore report that "yes, they 'think' they had COVID" more than people who know full well they were indoors the whole time? Without a decent outcome measure of actual infection, you can't tell.

Also, the confidence interval around the OR - the one for home delivery crosses 1. This tells you that the true difference between home delivery and not home delivery people is either less risk or more risk. So no difference. With walking the dog, the true difference in self-assessed COVID is from virtually none (odds of 1.03) to triple (3.07)

PowerslidePanda Wed 18-Nov-20 13:49:35

Yes - I don't think many countries have the order-online option that we do. And yes - if you think about the kind of people who would be using the kind of service they have in Spain, that probably explains the observed difference in risk - people who have to go to the supermarket via public transport rather than in their own car, people too frail to carry heavy bags, etc. The groceries themselves may be nothing to do with it.

No worries about the @ tagging - I do it all the time, I wasn't aware of the etiquette either!

QueenStromba Wed 18-Nov-20 14:00:38

TinyGarden

*@PowerslidePanda* Gosh that's a huge difference in context isn't it, if in Spain 'home deliveries' often meant going to the store, picking products off the shelves then getting them delivered.

Reassuring for the U.K. context (where I'm not aware of that being a service - maybe it is for people without cars etc?).

The only place I can think of the does it is Iceland.

Join the discussion

To comment on this thread you need to create a Mumsnet account.

Join Mumsnet

Already have a Mumsnet account? Log in