Does this prove masks work?

[lip09am]

This meta anlysis of several RCTs seems to show thta mask use works within a community setting? can anyone who actually knows their stuff confirm or deny this? Thanks

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7191274/

[miahoneybee]

I dont read that from the article. Lots of words like " may" ..respirators better than medical masks..cloth masks pretty useless from what I can read of it.
Personally I think the mask wearing is symbolic rather than for protection for the masses.
I think holding a lucky rock in your pocket gives as much protection as a mask.

[Splatt]

I've (briefly) gone through some of the studies they used.
First up, none specifically did any sort of randomised, control group.
They failed to filter out variables (lumping masks in with hand washing and distancing measures to varying degrees etc).

Lets also forget respirators and healthcare workers. Not relevant to community.

Im not a fan of meta-studies. Especially meta-studies that look at a small number of really poor quality studies all looking for and testing different things.

Its more interesting that they go back to 2008 for some studies to "support" their claim whilst ignoring others from the same period contradicting it. Cherry picking data perhaps.

Here are some of the studies from that period they chose to NOT look at:-


This is from a World Health Organisation summary of existing RCTs for masks in 2019.


ECDC issued a mask technical report this week. They didn't find much at all.

https://www.ecdc.europa.eu/en/publicati ... ansmission

Basically they summarised this:

The evidence regarding the effectiveness of medical face masks for the prevention of COVID-19 in the community is compatible with a small to moderate protective effect, but there are still significant uncertainties about the size of this effect. Evidence for the effectiveness of non-medical face masks, face shields/visors and respirators in the community is scarce and of very low certainty.

Additional high-quality studies are needed to assess the relevance of the use of medical face masks in the COVID-19 pandemic.

Certainly nothing there to scream "masks work" and its very important to realise they we're talking about *non-clinical* (non-medical) masks in public here. For that they say very scarce, poor quality data to say they do anything.

So basically this is unchanged since Feb 2020. The data hasnt improved or changed, no strong signals appear.

[lip09am]

That's what I thought but it seems to be from an RTC - which are considered the gold standard for this type of research.

[Occamsrazor]

Thanks Splatt, that's really useful.
Hands up I can't face reading another mask study but the problems with all these studies are fairly basic ones which all good quality scientific work tries to minimise but which are harder to minimise during a crisis:
1. Choosing sample groups: how to ensure that what is being shown is to do with mask efficacy rather than with something else (eg health, lifestyle, other medical or non medical interventions etc) or with a mix.
2. Use of control groups. Because there are so many variables, control groups may need to be numerous.
3. The time at which research was undertaken. Because of the difficulties of 1. And 2., the evidence of pre-pandemic mask studies is likely to be stronger, especially in terms of 'real life' effects rather than laboratory or other controlled- environment effects. This is because they took place over a longer period of time and not in a panicked response to a particular issue.
4. Number of participants. Because of the huge range of individual differences that can affect the outcome, researchers would have to use a lot of people in order to minimise the effect of those differences - a bit like the bloody R number: if you start with 1 case and end with 2 cases, that skews the R number towards looking terrifying. You need a lot more cases for the R number to be statistically meaningful in real life. I don't know how many because I'm not a mask researcher!... (Wouldn't want to be accused of misinformation!)
5.taking all the above, any conclusion has to try to ensure causation rather than correlation ideally, or if the latter, state that openly.
The other point is that masks can't protect the wearer - that's why if you're dealing with Ebola you really do have to wear a hazmat suit!
In terms of protecting others, some evidence suggests that masks can have a very minimal effect when in close, not fleeting contact with symptomatic people in unventilated indoor settings.
But frankly, just as I never visited my aged parents if I had a cold, and always covered my mouth if I coughed or sneezed and never spat in the tube or whatever... , just normal, polite, respectful behaviour that most people have always undertaken towards others is going to have the same effect as masks on transmission rates.
(and of course there are all sorts of issues around masks which mean that any weak evidence for them is overwhelmed by the negatives)

[lip09am]

I've (briefly) gone through some of the studies they used.
First up, none specifically did any sort of randomised, control group.
They failed to filter out variables (lumping masks in with hand washing and distancing measures to varying degrees etc).

Lets also forget respirators and healthcare workers. Not relevant to community.

Im not a fan of meta-studies. Especially meta-studies that look at a small number of really poor quality studies all looking for and testing different things.

Its more interesting that they go back to 2008 for some studies to "support" their claim whilst ignoring others from the same period contradicting it. Cherry picking data perhaps.

Here are some of the studies from that period they chose to NOT look at:-


This is from a World Health Organisation summary of existing RCTs for masks in 2019.


ECDC issued a mask technical report this week. They didn't find much at all.

https://www.ecdc.europa.eu/en/publicati ... ansmission

Basically they summarised this:

The evidence regarding the effectiveness of medical face masks for the prevention of COVID-19 in the community is compatible with a small to moderate protective effect, but there are still significant uncertainties about the size of this effect. Evidence for the effectiveness of non-medical face masks, face shields/visors and respirators in the community is scarce and of very low certainty.

Additional high-quality studies are needed to assess the relevance of the use of medical face masks in the COVID-19 pandemic.

Certainly nothing there to scream "masks work" and its very important to realise they we're talking about *non-clinical* (non-medical) masks in public here. For that they say very scarce, poor quality data to say they do anything.

So basically this is unchanged since Feb 2020. The data hasnt improved or changed, no strong signals appear.

thanks for the analysis - it backs up pretty much waht I thought (but i have no scientific standing to back it up)

[Splatt]

That's what I thought but it seems to be from an RTC - which are considered the gold standard for this type of research.

Its not, read the list of studies they used - they're RCTs but using grouped variables (mask with sanitiser, with distance in some cases). They also fail to take into account any other NPIs that might or might not be the case between them.
On top of that they ignore other studies in the same timeframe that don't support their results. No explanation as to why they used some studies and not others.

An RTC is only useful if you can separate out all the variables and normalise the data. Otherwise the best you can get is a result that "all of these together did X" which is what the cited papers show.

Again, the WHO themselves host other studies from the same period that completely contradict the ones they've chosen (for no clear reasons) to use.

Cherry picking studies with group variables and doing meta analysis isn't going to produce a useful result.

So again i'd be asking for their reasoning for picking the specific studies they did (why not the WHO list, why not DanMask etc) and going from there.
Meta studies can be handy but only if they've eliminated bias in the choice of papers used.

[MyHomeIsMyCastle]

Surely the bottom line with masks is they can't prevent/reduce transmission from someone who is not infected in the first place, so requiring healthy people to wear them is achieving absolutely no benefit whatsoever?

If we started regarding people who are "asymptomatic" as healthy rather than as potential plague carriers, the case for masks just disappears.

[miahoneybee]

Myhomeismycastle..that's two useful conversation starters I have picked up on here today to discuss with some sheep so thanks..

[Splatt]

If we started regarding people who are "asymptomatic" as healthy rather than as potential plague carriers, the case for masks just disappears.

However, pre-symptomatic spread is a real thing. That person isn't "healthy" as they are sick and will become sick but just dont know if for a day or 2.
Thats not the same as asymptomatic.

So if masks worked there'd be some logic to it. But as above, even the ECDC are saying the evidence for masks working is scarce and of very low quality.