"Munro points out that the question of the study is not: do masks work? We know that in a physics lab, the N95 filters more virus than a medical mask. Heck, you don’t even need a physics lab; just look at the profile of someone wearing a medical mask. Masks aren’t used in physics labs; they are used in the messy real world."
The idea that testing masks glued to mannequins tells us anything about real people is so prevalent among "scientists" who should know better one wonders if they are speaking in good faith.
Thanks for explaining this study. You're substacks are really interesting and help to keep me updated. I don't work on a covid ward any longer but it's good to know we did the best for our staff when I did!
Given that surgical masks leave large gaps that allow ingress of entirely unfiltered air, and that everyone from patients to hospital staff may be infected and transmitting (staff and patients both may be asymptomatic), it seems clear that in a period when prevalence is high, even if an N95 respirator were to afford 100% protection from infection, the remaining risk of infection from the long periods wearing only a surgical mask while in the hospital, coupled with the risk of infection outside of work (where some workers may mix widely while unmasked), will be very high. As such, it is hardly surprising that wearing an N95 respirator only for aerosol-generating procedures and when directly caring for a Covid-positive patient doesn't significantly reduce overall infection risk. There's simply too much risk that remains from the (majority) of the time when workers aren't wearing an N95 respirator.
My partner intubated Covid patients on 12-hour shifts in a London hospital's ICU during the first year of the pandemic, most of which was during the pre-vaccine era. She wore an elastomeric respirator at all times while in the hospital for several months, and switched to FFP3 disposable respirators thereafter, once the hospital had inventory of them. Her hospital conducted N antibody surveillance; she never contracted Covid. Nor did many of her colleagues who also intubated Covid patients and wore elastomeric respirators and FFP3 disposable respirators at all times while in the hospital during the pre-vaccine era.
A sensible response to this study from a person who wishes not to contract Covid while working in a hospital is therefore that the study simply serves to underscore that a fit-tested respirator should be worn at all times while in the hospital--clearly, wearing a surgical mask for much of the time is not sufficient protection. While wearing an FFP3 respirator on the hospital premises does not address the risk of becoming infected while not at work, it is up to the worker to decide what precautions they take outside of the workplace. Generally modern societies hold an employer responsible for ensuring the safety of workers on the employer's premises--we don't say that risks an employee takes off-premises obviate protections for while the employee is on-premises.
Exactly. If you wear a respirator only for a relatively small amount of time in a hospital setting, you might as well go maskless, because you'll just get infected when you're not wearing a respirator (because non-respirator masks don't work well at preventing covid due to poor face seals, inferior filter media, etcetera). Also, compared to elastomeric respirators, N95s aren't that great either (https://www.lesswrong.com/posts/Nf8EbepFNqDqkJnb9/?commentId=rraNx5Wq24vNZrmA3).
It looks like Egypt had no requirement for eye protection given the supplementary materials shows that ~25% always wore eye protection versus ~88% in Canada. Canada had a HR of 2.85 (albeit with wide CIs) whereas for Egypt it was 0.95.
So based on this I don't think your summary is accurate, the risk of bias for Egypt and the underpowered nature of the other countries means that this study doesn't add all that much to our knowledge on mask wearing for HCWs.
I'm not sure why eye protection matters. Given the intervention would still be randomised, the use of eye protection cannot bias the result. It shouldn't matter if it is different in different centres. I would suggest ignoring the differences by region (as discussed) as this is a post-hoc analysis with many, many problems.
Per @gidmk's comments on twitter, given Egypt is >50% of the trial participants any issues in that country would bias the overall study.
The ocular route is an important one for Covid (Coroneo & Collignon, Lancet 2021), so because eye protection wasn't worn, both study groups in that country would be more likely to be infected at work and hence the HR would be reduced.
As the final line in the study says: "The subgroup results varied by country, and the overall estimates may not be applicable to individual countries because of treatment effect heterogeneity."
It's disappointing that it's been almost 3 years and this not great study is that the best that has been done on such an important topic of the OHS requirements of Healthcare Workers.
Unless the difference in eye protection impacts N95 or medical mask wearers differentially, this does not bias the study result and should not impact the HR.
Rather than assume like in my previous post, I ran the numbers to see what would happen to the HR/CIs regarding the hypothesis that ocular transmission was a feature in Egypt due to lack of eye protection.
If the number of positive cases in Egypt in both arms of the study drop by the exact same amount (eg. 20 - because these 20 were infected by the ocular route), then this changes Egypt from (HR, 0.95 [CI, 0.60 to 1.50]) to (HR, 0.83 [CI, 0.42 to 1.65]). Overall this would change the study from (HR, 1.14 [95% CI, 0.77 to 1.69]) to (HR, 1.19 [95% CI, 0.71 to 1.98]) which means the CI approaches the cut-off set for non-inferiority.
Here all you have done is reduced the total number of events from the study and made it underpowered. You’ve simply removed informative information. The more infections you remove, the wider the CI will get because you’re just increasing uncertainty. This is the worst thing to do
I believe an important difference is also omicron versus previous variants. Omicron is known to be very explosive in spreading.
An interpretation is that n95 did work much better against earlier variants in the setting, but were no longer useful with omicron.
Note also the Egypt part was an extension to the trial. If they had ran the originally planned trial, they would have had more than 2x reduction as point estimate!
Also, good to check what John Conly, one of the authors, has been saying about N95. In short, he has been vocally against them, calling them harmful, while the study has been ongoing.
I would take great care in inferring much from the subgroup analysis, for all the reasons stated in the blog. Without Egypt the study is underpowered to really detect anything, and would have been unethical. Also, unless you think the authors deliberately sabotaged the study, I don't think it's too relevant what one of the authors said. Best to focus on the details of the trial.
"At the most basic level, we know masks designed to filter particulates work because they are rigorously tested to confirm they do so."
Physicist here. Mask limitations are stated by manufacturers. Not suited for protection from virus aerosols. MAY protect from viruses in resp droplets, but the physics to study this is in its infancy. So any claim of protection at this point is simply based on faith.
Masking is a _physical_ intervention and the mechanism of masks/droplets/virus should be studied by _physicists_. Or do you believe that you can have science without understanding the mechanisms?
I don't know of any physicists who advocate public masking. I am agnostic about masking HCWs. There would need to be studies of carrying capacities of N95 masks before recommendations are issued for HCWs. Training is also necessary and is generally missing from hospital education, per a physician who has worked at many hospitals during the covid era.
Thank you for the analysis. As a pediatrician taking care of a ton of patients with Covid everyday and I don’t wear a mask, I suspect that if the study had a third arm of no masks, it would have been the same as n95 or surgical masks. It’s predominantly aerosol spread. Surgical masks only protect against droplets. So why would they be equal to n95’s? They couldn’t be if masks were actually doing something. Your nose is the best filter for aerosols and it has peroxidases that inactivate viruses. I’m a pediatrician and have practiced for over 20 years. I don’t get sick from my patients. I get sick from my husband or my kids.
It is a shame that personalisation of masks isn't a bigger thing. Wearing a vlfex n95 has less breathing resistance than a surgical mask. It's even the same thickness. For the comfort of a surgical mask you can have n95 protection.
Just prior to the pandemic we purchased pollution masks to prevent inhalation of breakdust on the London underground as PM2.5 levels are high and we used to spend about 1.5 hours a day down there. The mask we purchased was deeply uncomfortable and I hated wearing it. Over the course of the pandemic I settled on two masks that work perfectly for us and my wife has her favorite and I have mine. Just like with headphones or anything that fits on a unique head you need to encourage experimentation for personal preference.
You wrote that the question the study is designed to answer was: "Does requiring health care workers (who are already wearing medical masks all day and N95 masks for aerosol generating procedures) to wear N95 masks for routine care of Covid-19 patients, result in a less then doubling of the hazard of acquiring Covid-19".
So, they were concerned that additional time wearing N95 masks might *increase* the hazard? If that's all they're trying to answer, I'm losing interest in this study.
... but this is just one study. A recent systematic review and network meta-analysis published in Rev Med Virol found N95 or equivalent masks (FFP2) were the most effective in providing protection against coronavirus infections (OR, 0.30; CI, 0.20-0.44) consistently across subgroup analyses of causative viruses (influenza virus, SARS-CoV, MERS-CoV and SARS-CoV-2) and clinical settings.
Thanks Alasdair -- a strong analysis so clearly written.
I liked your piece as well, especially this:
"Munro points out that the question of the study is not: do masks work? We know that in a physics lab, the N95 filters more virus than a medical mask. Heck, you don’t even need a physics lab; just look at the profile of someone wearing a medical mask. Masks aren’t used in physics labs; they are used in the messy real world."
The idea that testing masks glued to mannequins tells us anything about real people is so prevalent among "scientists" who should know better one wonders if they are speaking in good faith.
Thanks for explaining this study. You're substacks are really interesting and help to keep me updated. I don't work on a covid ward any longer but it's good to know we did the best for our staff when I did!
Excellent, clear analysis. Thank-you.
Given that surgical masks leave large gaps that allow ingress of entirely unfiltered air, and that everyone from patients to hospital staff may be infected and transmitting (staff and patients both may be asymptomatic), it seems clear that in a period when prevalence is high, even if an N95 respirator were to afford 100% protection from infection, the remaining risk of infection from the long periods wearing only a surgical mask while in the hospital, coupled with the risk of infection outside of work (where some workers may mix widely while unmasked), will be very high. As such, it is hardly surprising that wearing an N95 respirator only for aerosol-generating procedures and when directly caring for a Covid-positive patient doesn't significantly reduce overall infection risk. There's simply too much risk that remains from the (majority) of the time when workers aren't wearing an N95 respirator.
My partner intubated Covid patients on 12-hour shifts in a London hospital's ICU during the first year of the pandemic, most of which was during the pre-vaccine era. She wore an elastomeric respirator at all times while in the hospital for several months, and switched to FFP3 disposable respirators thereafter, once the hospital had inventory of them. Her hospital conducted N antibody surveillance; she never contracted Covid. Nor did many of her colleagues who also intubated Covid patients and wore elastomeric respirators and FFP3 disposable respirators at all times while in the hospital during the pre-vaccine era.
A sensible response to this study from a person who wishes not to contract Covid while working in a hospital is therefore that the study simply serves to underscore that a fit-tested respirator should be worn at all times while in the hospital--clearly, wearing a surgical mask for much of the time is not sufficient protection. While wearing an FFP3 respirator on the hospital premises does not address the risk of becoming infected while not at work, it is up to the worker to decide what precautions they take outside of the workplace. Generally modern societies hold an employer responsible for ensuring the safety of workers on the employer's premises--we don't say that risks an employee takes off-premises obviate protections for while the employee is on-premises.
Exactly. If you wear a respirator only for a relatively small amount of time in a hospital setting, you might as well go maskless, because you'll just get infected when you're not wearing a respirator (because non-respirator masks don't work well at preventing covid due to poor face seals, inferior filter media, etcetera). Also, compared to elastomeric respirators, N95s aren't that great either (https://www.lesswrong.com/posts/Nf8EbepFNqDqkJnb9/?commentId=rraNx5Wq24vNZrmA3).
It looks like Egypt had no requirement for eye protection given the supplementary materials shows that ~25% always wore eye protection versus ~88% in Canada. Canada had a HR of 2.85 (albeit with wide CIs) whereas for Egypt it was 0.95.
So based on this I don't think your summary is accurate, the risk of bias for Egypt and the underpowered nature of the other countries means that this study doesn't add all that much to our knowledge on mask wearing for HCWs.
I'm not sure why eye protection matters. Given the intervention would still be randomised, the use of eye protection cannot bias the result. It shouldn't matter if it is different in different centres. I would suggest ignoring the differences by region (as discussed) as this is a post-hoc analysis with many, many problems.
Per @gidmk's comments on twitter, given Egypt is >50% of the trial participants any issues in that country would bias the overall study.
The ocular route is an important one for Covid (Coroneo & Collignon, Lancet 2021), so because eye protection wasn't worn, both study groups in that country would be more likely to be infected at work and hence the HR would be reduced.
As the final line in the study says: "The subgroup results varied by country, and the overall estimates may not be applicable to individual countries because of treatment effect heterogeneity."
It's disappointing that it's been almost 3 years and this not great study is that the best that has been done on such an important topic of the OHS requirements of Healthcare Workers.
Unless the difference in eye protection impacts N95 or medical mask wearers differentially, this does not bias the study result and should not impact the HR.
Rather than assume like in my previous post, I ran the numbers to see what would happen to the HR/CIs regarding the hypothesis that ocular transmission was a feature in Egypt due to lack of eye protection.
If the number of positive cases in Egypt in both arms of the study drop by the exact same amount (eg. 20 - because these 20 were infected by the ocular route), then this changes Egypt from (HR, 0.95 [CI, 0.60 to 1.50]) to (HR, 0.83 [CI, 0.42 to 1.65]). Overall this would change the study from (HR, 1.14 [95% CI, 0.77 to 1.69]) to (HR, 1.19 [95% CI, 0.71 to 1.98]) which means the CI approaches the cut-off set for non-inferiority.
Here all you have done is reduced the total number of events from the study and made it underpowered. You’ve simply removed informative information. The more infections you remove, the wider the CI will get because you’re just increasing uncertainty. This is the worst thing to do
I believe an important difference is also omicron versus previous variants. Omicron is known to be very explosive in spreading.
An interpretation is that n95 did work much better against earlier variants in the setting, but were no longer useful with omicron.
Note also the Egypt part was an extension to the trial. If they had ran the originally planned trial, they would have had more than 2x reduction as point estimate!
Also, good to check what John Conly, one of the authors, has been saying about N95. In short, he has been vocally against them, calling them harmful, while the study has been ongoing.
I would take great care in inferring much from the subgroup analysis, for all the reasons stated in the blog. Without Egypt the study is underpowered to really detect anything, and would have been unethical. Also, unless you think the authors deliberately sabotaged the study, I don't think it's too relevant what one of the authors said. Best to focus on the details of the trial.
"At the most basic level, we know masks designed to filter particulates work because they are rigorously tested to confirm they do so."
Physicist here. Mask limitations are stated by manufacturers. Not suited for protection from virus aerosols. MAY protect from viruses in resp droplets, but the physics to study this is in its infancy. So any claim of protection at this point is simply based on faith.
Masking is a _physical_ intervention and the mechanism of masks/droplets/virus should be studied by _physicists_. Or do you believe that you can have science without understanding the mechanisms?
I don't know of any physicists who advocate public masking. I am agnostic about masking HCWs. There would need to be studies of carrying capacities of N95 masks before recommendations are issued for HCWs. Training is also necessary and is generally missing from hospital education, per a physician who has worked at many hospitals during the covid era.
Thank you for the analysis. As a pediatrician taking care of a ton of patients with Covid everyday and I don’t wear a mask, I suspect that if the study had a third arm of no masks, it would have been the same as n95 or surgical masks. It’s predominantly aerosol spread. Surgical masks only protect against droplets. So why would they be equal to n95’s? They couldn’t be if masks were actually doing something. Your nose is the best filter for aerosols and it has peroxidases that inactivate viruses. I’m a pediatrician and have practiced for over 20 years. I don’t get sick from my patients. I get sick from my husband or my kids.
Thank you for the well written article.
It is a shame that personalisation of masks isn't a bigger thing. Wearing a vlfex n95 has less breathing resistance than a surgical mask. It's even the same thickness. For the comfort of a surgical mask you can have n95 protection.
Just prior to the pandemic we purchased pollution masks to prevent inhalation of breakdust on the London underground as PM2.5 levels are high and we used to spend about 1.5 hours a day down there. The mask we purchased was deeply uncomfortable and I hated wearing it. Over the course of the pandemic I settled on two masks that work perfectly for us and my wife has her favorite and I have mine. Just like with headphones or anything that fits on a unique head you need to encourage experimentation for personal preference.
Wouldn't another limitation be that a large portion of their covid exposure probably happened outside of work?
You wrote that the question the study is designed to answer was: "Does requiring health care workers (who are already wearing medical masks all day and N95 masks for aerosol generating procedures) to wear N95 masks for routine care of Covid-19 patients, result in a less then doubling of the hazard of acquiring Covid-19".
So, they were concerned that additional time wearing N95 masks might *increase* the hazard? If that's all they're trying to answer, I'm losing interest in this study.
Great minds think alike. Sometimes. 😉
Non-scientist, here. Can someone explain to me what Hazard Ratio (HR) means?
Here, HR means the relative risk between medical masks and N95s regarding catching covid.
... but this is just one study. A recent systematic review and network meta-analysis published in Rev Med Virol found N95 or equivalent masks (FFP2) were the most effective in providing protection against coronavirus infections (OR, 0.30; CI, 0.20-0.44) consistently across subgroup analyses of causative viruses (influenza virus, SARS-CoV, MERS-CoV and SARS-CoV-2) and clinical settings.
Can we just stop?!??
Masking the population is NEVER warranted.
Disease happens.
Some die.
End of.
Do you want to stop talking about masks or medicine in general?
I'm fed up with people who stand for anti-evolutionary actions paid for by the public.
This research isn't about masks for the whole population. It is about do N95 masks work well enough in certain hospital settings.
That's irrelevant. People who work in health don't get brain Severance when they go home after work. The ideas in one place bleed into the other.
The only way to make sense is to ban mandates, all of them.
Tracy's seems to object to any effort to keep people from dying...