Recent review of the studies on using masks to prevent the transmission of Covid-19.

[PeterNSteinmetz]

One problem seems to be private gatherings. Some workplaces don't require masks, or require them only in limited circumstances. I suspect that people, most of the time, are in situations where they aren't compelled to wear masks. Some choose to wear them, some don't need to wear them, and some should wear them but choose not to in those situations. No one really knows what people do out of the public eye.

I agree with you and the authors of the review that there are certain higher risk situations where using a mask that really works, like an N95, is likely wise. That is one of the reasons that my institute developed the adaptor for the 3M 6200 - so more people would be able to access that technology.

I notice that almost no mechanics that work on planes outside wear them, which is likely lower risk. Airplane cockpits are of course a different matter and higher risk in my view.

My primary concern, as may be obvious, had been with mask mandates. Both because of the insidious rights implications as well as the potential enablement of genuinely risky behavior where people over-estimate the value of a cloth mask.

 

[PeterNSteinmetz]

I say the medical community consensus supports mask wearing.

I will have to disagree. If you looked at the actual scientific reviews (not commentaries or lay articles), the evidence is well described as mixed on whether asking the general public to wear cloth masks inhibits the spread of COVID-19.

I thought the review cited makes a pretty fair assessment based on the more recent evidence. Prior to that, the reviews, which are listed on my medical interest page at http://steinmetz.org/peter/Medical were about 50/50 split.

If you want we can discuss those reviews and the articles underlying them in detail. But I gather you don’t really feel this is the appropriate forum.

But I really do not think it accurate to say there is a clear scientific consensus on this point.

It may be uncomfortable, but the evidence is just not that clear presently on this specific question.

 

[PeterNSteinmetz]

The US was never locked down. There were so many "essential businesses" I doubt we reached 50% lockdown.

There was about a 30% drop in mobility measures per both Apple and Google which preceded the initial phase of lockdown orders by about 3 weeks.

The initial lockdown orders were significantly correlated with a rise in mortality when adjusting for population and the timing of both the orders and the initial cases on a per state basis in the US. But not in the way one would expect - the states with lockdown orders had HIGHER rates of mortality attributed to COVID-19.

One could interpret that to mean lockdowns caused more deaths, though I am inclined to think there is some other uncontrolled independent variable.

 

[PeterNSteinmetz]

Why has this gotten so politicized?!?

We should probably limit this political aspect of the discussion as that is not permitted on PoA and will result in a lock. PM me if you would like to discuss further why I take it as a political issue. I appreciate your insights and opinion.

 

[PeterNSteinmetz]

From that paper's "conclusions" section:

"Evidence that the virus can be airborne (and therefore be inhaled) and that masking policies, when effectively delivered, save lives is now strong."

"[W]e recommend that the public wear masks or face coverings..."

The most fascinating part of the story is how you can continually take papers and draw conclusions from them that are totally opposite to the paper's own authors.

I don't know if I am disagreeing with the authors much here. For example, please see the part after the ellipsis which you omitted and which limits the conditions under which they recommend the public use masks. Pretty much what I said above. Use them in high risk situations.

Please also see the key points summary which disagrees with the one statement from the discussion which you quoted above. Sometimes the authors of papers make a variety of different statements, some of which, if taken in isolation, will appear to have a different meaning than what they intended overall.

As I noted above, rather than dithering about the specific meaning in one sentence or word or another, I usually find it best to read the whole paper to understand the authors’ whole meaning. Considered as a whole, I think it is well described by their 5 key points at the top of paper in terms of the level of scientific certainty.

 

[PeterNSteinmetz]

This is nonsense. The premise is the number virus particles you are exposed to is a finite quantity. As if the virus does not reproduce in the host. Find me a virologist who agrees with and I'll eat that hat I have left over from the other thread. Sure would be nice if it were true, though.

Agreed that the initial viral load being correlated with the likelihood of getting sick is better established. But I thought I had read some reports on both sides of this issue of correlation with disease severity. This article may be of interest: https://www.sciencealert.com/does-t...-are-exposed-to-determine-how-sick-you-ll-get.

It states that basically we don't know if initial viral load affects disease severity. Please note this is a secondary source and from early on. A quick search of scholar.google.com did not reveal any more detailed recent studies, so perhaps it is not a viable current idea.

 

[PeterNSteinmetz]

... My premise is the masks reduce the number of particles I am exposed to when I’m wearing it (by blocking many of them from getting in my mouth/nose) and reduces the number I expose YOU to if I’m infected/contagious (by blocking many of them from getting OUT of my mouth/nose). ... That said, I probably should have said it gives me a better CHANCE of a less severe infection at the start.

Speculating on this a bit more it strikes me both things could essentially be true. It could be that the initial number of virions you are exposed to determines primarily the likelihood you will become ill and also have an effect on disease severity. This would be consistent with what we know about immune response.

If you receive less than a certain threshold number of virions, your immune system is able to mount a completely effective response and wipe out the virus. If you receive just a bit over that number, perhaps you mount a response and in some sense were infected, but it is tamped down quickly without much ill effect. And if you receive more than even another higher threshold, the virus is able to start geometric growth and you become fully ill, with the severity now determined by a number of other genetic and acquired health factors.

Would be a hard thing to prove in some sense without a lot of virus challenge experiments, which are difficult ethically. This is just my musings - I don't know of any hard papers to support this specifically in the case of Covid-19.

 

[PeterNSteinmetz]

California did "flatten the curve" from from August through October, while having limited indoor and outdoor dining and limited barber/salon services, open retail with masks and worker precautions, etc. It was largely private gatherings that did us in, then those folks spread it to their workplaces.

I don’t know that one can clearly attribute the cause of the recent run-up like that. Most flu seasons follow this sort of course.

The reasons for the seasonality are presently debated but may include more time inside, colder temperatures, kids in school (though presumably not as big a factor this year).

Certainly gatherings is one reasonable hypothesis, but there may be other factors that are substantial contributions as well.

 

[PeterNSteinmetz]

Edit: mask mandates also save lives (or at least have a 50/50 chance according to the OP). Which is worse? Totally valid question. It is like the classic choice of crashing the school bus to avoid a single child on the road. Which do you choose?

Exactly, whether to use a coercive government mandate in the presence of such mixed evidence is the key policy question.

And a political one which I won’t personally go into here per PoA policy.

 

[PeterNSteinmetz]

Flattening the curve worked — until it didn’t
The US started 2020 by “flattening the curve” — and never came up with a plan for what comes next.

https://www.vox.com/22180261/covid-19-coronavirus-social-distancing-lockdowns-flatten-the-curve

Excerpt:

The US did succeed at flattening the curve — at least at first. Businesses closed and most states issued stay-at-home orders, and later research concluded those lockdown measures helped prevent tens of millions of Covid-19 cases.

I don’t think the peer-reviewed literature is so crystal clear as Vox would represent here. Particularly regarding the effect of coercive government lockdowns.

Two things were going on initially. People began socially distancing voluntarily about 2-3 weeks before any lockdown orders, as shown by social mobility data. After the orders, they started resuming more normal travel, though I suspect that is just fatigue rather than caused by the orders.

Also at first testing was very limited and there was a strong correlation between number of tests and number of positive cases reported, particularly early on. (The lack of random population sampling really hurt us in tracking this.)

While the question of cases is quite confounded by increased testing capacity in that timeframe, what is clearer is that lockdowns initially had no significant effect on deaths attributed to COVID-19 in the US.

There is good data that things like staying away from people and using N95 masks work to prevent the spread of Covid-19. The effectiveness of coercive lockdowns and mask mandates are a different matter.

 

[PeterNSteinmetz]

That's the whole point of why you brought up the subject. You have been there before. How many times have you written, "coercive government mandate" on this board? You want people to agree with you on policy, which is not able to be discussed here.

I understand that the policy can’t be discussed here. Regardless of whatever one wants to make of policy in the presence of mixed data, I think it is important for people to understand the nature of the data and analyses and the levels of scientific certainty or uncertainty involved. Then people can fit that into their own value framework.

Much of the current acrimony surrounding this issue is caused by people assuming the evidence is clearer than it is. Hopefully by recognizing the actual level of scientific certainty or lack thereof, people can come to greater peace with the idea that we live in uncertain times and that different people may reasonably make different choices about this in their lives.

 

[PeterNSteinmetz]

Every study has its limitations and caveats. You keep taking those limitations and caveats as the main conclusions of the paper while ignoring the actual conclusions.

Ok. So which of the main key 5 points made by the authors do you think I have disagreed with?

I think I would only probably state #2 slightly differently. They say “The benefits of mask wearing seem to outweigh the harms when COVID-19 is spreading in a population.”

“Seem to” is already fairly weak but I would probably be a bit more cautious or say this with more qualification. Such as “Considering only direct health outcomes, the benefits of mask wearing seem to”. Since they did not consider secondary health effects such as people deferring other medical treatments.

I would also not write the one sentence in the conclusion as they did “Evidence that the virus can be airborne (and therefore be inhaled) and that masking policies, when effectively delivered, save lives is now strong”

This is compound and while I think the evidence for part of it has strong evidence, such as the virus can be airborne, I think the phrasing and qualifier “when effectively delivered” is not clear enough. What does “effectively delivered” mean? Does that mean when healthcare workers are wearing N95s? Certainly the evidence there is fairly strong.

But it can easily be confused by people to think it means that cloth mask wearing by the general public saves lives has strong evidence. But that is not really what they are showing in the rest of the review.

These are the sort of somewhat vague statements that often appear in discussions at the end of papers and the reason that it really is best to read the whole paper and objectively consider all of it.

Those nits aside, I think they did a good job with their five key points and overall. I think it is good that the limitations of our certainty on these questions are being more clearly discussed.

 

[PeterNSteinmetz]

When you look specifically at Southern California, you can't attribute the changes to people spending more times indoors. If anything, it would be flipped with people spending more time inside in the summer on the hottest days than in the more mild weather we've had Sept-Dec.

Occam's Razor approach to looking at the California data is that the inflection point is Halloween/Dia de Muertos (Nov. 1).

Interesting question. What does the mobility data tell us about those time points?

It is of course going to be difficult to de-confound the normal tendency for the flu to increase in the fall from any specific factor.

And it may be, as you say, that there are more indoor personal gatherings in the fall which has generally driven this. I am not expert on that literature but understand that the causes of seasonal flu variation are actively debated in the epidemiological community.

 

[PeterNSteinmetz]

Mobility data aren't going to tell you when there was a block party with 50-100 people in a small neighborhood.

Crowds are definitely one of those high risk situations best avoided. Or use a good mask like an N95 if you have to be in one.

 

[PeterNSteinmetz]

You've used this paper to conclude that there is "mixed" evidence to support mask wearing.

You use the word "mixed" while the authors use the word "strongly":

"...fluid dynamics of aerosol spread and international epidemiologic data summarized in this review already strongly support the hypothesis that masks are likely to be effective..."

They're not vague statements, they are being scientifically precise by noting the limitations involved.

That's right. You need to read the whole paper so you can pick up on a few words such as "seem" and use them as an excuse to flip the meaning of the entire paper and make it "seem" like it's concluding the opposite of what it actually is.

I think the authors are being precise by using "may substantially reduce" and "seem to outweigh" in their summary points. This is much weaker than what they could have said. The reason they may have chosen that wording rather than saying simply "substantially reduce" and "outweigh" is that they may recognize that the results of all the randomized trials that are available do not support this idea. And they recognize that observational studies are subject to potential large confounds, especially when looking at small effects. And in-vitro studies on particle dynamics do not account for things like fickle human behavior which occurs in actual populations.

So let's look at the actual evidence which they are dealing with, and as summarized in my initial OP:

In-vitro studies of droplet dynamics: Some fairly good reasons to think that masks interfere with transmission of respiratory droplets of the appropriate size if worn properly. Less clear about droplets when actually worn by people.

Observational studies: Some, such as Lyu & Wehby, have shown an apparent small reduction in R0, which if due to masking, could add up to a sizeable reduction in cases. But attributing causation to such small effect sizes is always tricky and one is often detecting an effect, but due to some other confound.

Randomized trials: The Danmask-19 study shows no significant effect. Interpreted as point estimate, it suggests a modest reduction of 15% however there is also a 25% chance it may have resulted in an increase in cases. Older studies of flu transmission also failed to show a significant reduction in transmission when people wore surgical masks.

It is hard to reconcile all this. It does not point to a consistent story. The randomized trials are the gold standard in clinical work for a reason -- because they avoid the problems with confounds of the observational studies and account for actual human behavior. They should be weighted more heavily in an objective evaluation than other evidence.

It is possible that given droplet dynamics that there is a source effect even though there is no noticeable effect for the wearer. Seems a stretch to me, but it is possible.

I will also note that a similar situation, some nice theory and some strongly suggestive observational work but a failure in actual randomized trials is a common outcome for proposed new treatments in medicine. Happens all the time.

Given that overall, I think "mixed" is a good description. 1/3 of the main categories of data, and the more reliable third, does not support the hypothesis that the general public wearing cloth masks slows the spread of Covid-19. The use of the word "strongly" by the authors pertained only to the other less reliable 2/3. "mixed" is also a good description of the conclusions of prior reviews looking at the effect of masks on transmission (listed on my medical interest page - they were basically 50/50 on whether a recommendation for the general public to wear cloth masks was merited), though these did not include Lyu & Wehby or Bundgaard et al. as they were not yet published.

Now if you want, we can go through and look at the specific randomized and observational studies and evaluate the strength of evidence of each, to see if "mixed" is a good description overall. That could be interesting.

Quibbling over the meaning of "may substantially reduce" versus "substantially reduce" and "seem to outweigh" versus "outweigh" and "mixed" in a brief summary versus one statement containing "clearly" in an entire paper is a perhaps interesting semantic exercise (and I understand you are an attorney so may enjoy that sort of thing) but is considerably less informative of the science than really looking at the data and analyses.

If you are truly interested in the scientific issue, let's examine those studies. You may not like my word choices in one sentence of a brief summary -- so be it. I will not be commenting further on these word choices here.

 

[PeterNSteinmetz]

That's all despite the fact that every single model relied on initially to justify all the "NPIs" showed that there's be spikes once restrictions were limited. But there's very few elected "leaders" with the balls to stand behind a podium and say, "People are going to die, avs the consequences of effectively preventing that would be worse than letting it happen."

Interestingly, one of the early studies of the likely outcome of lockdowns predicted that they might actually make things worse in the fall.

 

[PeterNSteinmetz]

I have to ride a transcontinental flight twice a month for work. I can’t afford not to ride the airplane. My flight back to California earlier this week only had two empty seats. Sometimes there just isn’t the option to avoid a crowd.

Wear an N95 or P100. If you can’t get the disposables, get a 3M 6200 and it’s N95s, which are more available. You can then 3D print the adaptors to add cloth over the outflow valve.

 

[PeterNSteinmetz]

And now the Jefferson et al review updated and is rather explicit.

“ Wearing masks in the community probably makes little or no difference to the outcome of influenza-like illness (ILI)/COVID-19 like illness compared to not wearing masks (risk ratio (RR) 0.95, 95% confidence interval (CI) 0.84 to 1.09; 9 trials, 276,917 participants; moderate-certainty evidence. Wearing masks in the community probably makes little or no difference to the outcome of laboratory-confirmed influenza/SARS-CoV-2 compared to not wearing masks (RR 1.01, 95% CI 0.72 to 1.42; 6 trials, 13,919 participants; moderate-certainty evidence)”

Do physical measures such as hand-washing or wearing masks stop or slow down the spread of respiratory viruses?

www.cochrane.org

 

[Bob Noel]

well, given how masks were worn (as chin straps or just covering the mouth) and the performance of the masks (e.g., homemade vs real masks).... what would anyone expect?

 

[PeterNSteinmetz]

Certainly that does apply to what was happening recently. Jefferson et al's review covers pretty much all the studies that have been done, some under better circumstances. But even then the leakage around the masks etc likely prevents them from having any significant effect on this type of illness.

A lot of what people were doing is going to look even more ridiculous in 10 or 20 years.

Nonetheless, let us not forget all the mask Nazis.