[BostonCard]

I had said for a while that while I believed in masks to prevent COVID-19 transmission, the evidence in support of them was underwhelming... until now.  Big new meta analysis in the Lancet covering 172 observational studies across 16 countries and six continents and 44 relevant comparative studies in health-care and non-health-care settings and a whopping 25,697 people.  They mix SARS, MERS, and COVID-19 but I think that's ok; I would expect the transmission dynamics of the coronaviridae to be similar.

https://www.thelancet.com/journals/lance...9/fulltext (pdf here)


Three items come from this:
1) Physical distancing makes a very big difference


Going from less than a meter to over a meter (six feet) of distancing cuts the risk by about 70%, and more distancing is better.  According to the authors:

A physical distance of more than 1 m probably results in a large reduction in virus infection; for every 1 m further away in distancing, the relative effect might increase 2·02 times

Second, masks do work, both in a healthcare setting and in a community setting, reducing the risk of transmission by about 80%.

Third, N95's work better than surgical masks or multilayer cotton masks.  The reduction in the odds of transmission is about 96% for N95's, versus about 2/3 for surgical and cotton masks.

Fourth, eye protection may also help cut the risk down (by about 3/4).

Fifth, to quote to quote the investigators "However, no intervention, even when properly used, was associated with complete protection from infection."

Still, that is a substantial risk reduction.

BC

[BostonCard]

I had said for a while that while I believed in masks to prevent COVID-19 transmission, the evidence in support of them was underwhelming... until now.  Big new meta analysis in the Lancet covering 172 observational studies across 16 countries and six continents and 44 relevant comparative studies in health-care and non-health-care settings and a whopping 25,697 people.  They mix SARS, MERS, and COVID-19 but I think that's ok; I would expect the transmission dynamics of the coronaviridae to be similar.

https://www.thelancet.com/journals/lance...9/fulltext (pdf here)


Three items come from this:
1) Physical distancing makes a very big difference


Going from less than a meter to over a meter (six feet) of distancing cuts the risk by about 70%, and more distancing is better.  According to the authors:

A physical distance of more than 1 m probably results in a large reduction in virus infection; for every 1 m further away in distancing, the relative effect might increase 2·02 times

Second, masks do work, both in a healthcare setting and in a community setting, reducing the risk of transmission by about 80%.

Third, N95's work better than surgical masks or multilayer cotton masks.  The reduction in the odds of transmission is about 96% for N95's, versus about 2/3 for surgical and cotton masks.

Fourth, eye protection may also help cut the risk down (by about 3/4).

Fifth, to quote to quote the investigators "However, no intervention, even when properly used, was associated with complete protection from infection."

Still, that is a substantial risk reduction.

BC

[M T]

Before looking at some detail, I will say I find this study as muddy to me.  It asserts lots of things and gives a little evidence of some of those, assuring the reader that they've done all sorts of stuff but not giving the details.  I often wonder when I hear presentations like this, if the idea is that by the time you have a thought that something might be not solid, they've thrown a lot of other stuff at you.  Just because my pre-existing notions may generally line up with their conclusions doesn't mean that they've done what is needed to get to the conclusions.  (Nor does it mean they haven't.)
 
I also want to point out that their study regarding masks seems always to determine whether the wearer of the mask is protected against infection, not whether the wearing of the mask protects others from being infected by the wearer.   (I don't recall seeing a study yet of HCW's masks protecting patients from COVID-19 infection by the HCW.  Someone probably has done such a study; I just haven't seen it.)   I believe that each of the general public needs to be convinced that each is protected by their own mask, in order to get the level of mask usage that also reduces the spread by an asymptomatic person because he is already wearing a mask for his self-protection.  (Some seem to think that masks are primarily to prevent the wearer from spreading it.)

Third, N95's work better than surgical masks or multilayer cotton masks.  The reduction in the odds of transmission is about 96% for N95's, versus about 2/3 for surgical and cotton masks.

Just taking this one for an example...
In the section dealing with N95 masks versus surgical or 12+ layer cotton masks, all the N95 masks in the studies were used by healthcare workers.  Moreover, at least for one set of results they used that I looked at, only the usage of N95 masks was also associated with (say) PPE usage.   I would say there are confounding factors that aren't separated out (and can't be with this amalgam of studies).

  Doesn't the study only associate the risk for situations in which N95 masks are used (ie, the situation included fitted N95, maybe in a negative pressure room and full PPE, and used by HCW used to working in infectious situations) versus those in which no mask is used (perhaps a waiting room, no PPE)?  
  Then it separately associates the risk for situations in which surgical or 12+ layer cotton masks are used (I may be wrong, but these do not seem to include the homemade masks) versus no masks.

So, when one suggests that N95 masks are better, should that be "properly fitted N95 masks, as used by medically trained staff familiar with working in infectious situations, and doing all the other procedures they would do in such situations (shoe coverings, whatever) is less likely to result in a transmission than when using a surgical mask, with the procedures one might do normally when wearing a surgical mask'?

I don't think this study is sufficient to say that if you hand Joe Public a N95 mask that he would be less likely to get (or give!) an infection than if you handed him a surgical mask.

I also suspect this study will be used to say that homemade masks as worn by a lay person provides the level of protection suggested for the non-N95 masks here.  (They do mention "single-layer masks" but only in a context of referring to what some of the studies report, not what they evaluated.)  I think that would be a misuse of the results.

[M T]

Before looking at some detail, I will say I find this study as muddy to me.  It asserts lots of things and gives a little evidence of some of those, assuring the reader that they've done all sorts of stuff but not giving the details.  I often wonder when I hear presentations like this, if the idea is that by the time you have a thought that something might be not solid, they've thrown a lot of other stuff at you.  Just because my pre-existing notions may generally line up with their conclusions doesn't mean that they've done what is needed to get to the conclusions.  (Nor does it mean they haven't.)
 
I also want to point out that their study regarding masks seems always to determine whether the wearer of the mask is protected against infection, not whether the wearing of the mask protects others from being infected by the wearer.   (I don't recall seeing a study yet of HCW's masks protecting patients from COVID-19 infection by the HCW.  Someone probably has done such a study; I just haven't seen it.)   I believe that each of the general public needs to be convinced that each is protected by their own mask, in order to get the level of mask usage that also reduces the spread by an asymptomatic person because he is already wearing a mask for his self-protection.  (Some seem to think that masks are primarily to prevent the wearer from spreading it.)

Third, N95's work better than surgical masks or multilayer cotton masks.  The reduction in the odds of transmission is about 96% for N95's, versus about 2/3 for surgical and cotton masks.

Just taking this one for an example...
In the section dealing with N95 masks versus surgical or 12+ layer cotton masks, all the N95 masks in the studies were used by healthcare workers.  Moreover, at least for one set of results they used that I looked at, only the usage of N95 masks was also associated with (say) PPE usage.   I would say there are confounding factors that aren't separated out (and can't be with this amalgam of studies).

  Doesn't the study only associate the risk for situations in which N95 masks are used (ie, the situation included fitted N95, maybe in a negative pressure room and full PPE, and used by HCW used to working in infectious situations) versus those in which no mask is used (perhaps a waiting room, no PPE)?  
  Then it separately associates the risk for situations in which surgical or 12+ layer cotton masks are used (I may be wrong, but these do not seem to include the homemade masks) versus no masks.

So, when one suggests that N95 masks are better, should that be "properly fitted N95 masks, as used by medically trained staff familiar with working in infectious situations, and doing all the other procedures they would do in such situations (shoe coverings, whatever) is less likely to result in a transmission than when using a surgical mask, with the procedures one might do normally when wearing a surgical mask'?

I don't think this study is sufficient to say that if you hand Joe Public a N95 mask that he would be less likely to get (or give!) an infection than if you handed him a surgical mask.

I also suspect this study will be used to say that homemade masks as worn by a lay person provides the level of protection suggested for the non-N95 masks here.  (They do mention "single-layer masks" but only in a context of referring to what some of the studies report, not what they evaluated.)  I think that would be a misuse of the results.