Read on Twitter
It’s challenging finding the signal through the noise when looking for evidence around respiratory protection in reducing transmission risk for SARS-CoV-2. There are many confounding variables related to the fact that transmission can still be via large droplets or fomites. 
There are several questions and concepts that need to be considered and understood when examining the evidence. Several layers of intervention are required to cover all routes of transmission and this greatly confuses the analysis of data.
1. Was there adequate droplet precautions on top of respiratory protection?
A drop of saliva in the eye or on the face can still result in transmission while wearing adequate respiratory protection.
A drop of saliva in the eye or on the face can still result in transmission while wearing adequate respiratory protection.
2. What were the environmental attributes? Absence of risk reduction for respirators in one context (well ventilated OR with a silent patient) where droplet transmission is likely the higher residual risk, cannot be extrapolated to a poorly ventilated room with a coughing patient https://twitter.com/linseymarr/status/1326940468953960448
3. Was respirator use undertaken as part of a respiratory protection program where staff are fit tested & educated on their appropriate use?
Relying only on an inadequately sensitive “fit check” will miss a significant number of leaking respirators & muddy data on efficacy. https://twitter.com/drpieterpeach/status/1306414813392715778
Relying only on an inadequately sensitive “fit check” will miss a significant number of leaking respirators & muddy data on efficacy. https://twitter.com/drpieterpeach/status/1306414813392715778
4. Were patients infectious at the time the data was collected?
Eg We may have falsely concluded droplet precautions were adequate to prevent transmission in ICU HCWs looking after a patient that potentially may have been in their post-infectious period. https://twitter.com/drpieterpeach/status/1322095526876114950
Eg We may have falsely concluded droplet precautions were adequate to prevent transmission in ICU HCWs looking after a patient that potentially may have been in their post-infectious period. https://twitter.com/drpieterpeach/status/1322095526876114950
Conclusions from data like this may introduce survivorship bias.
As well as looking at evidence where no transmission occurred, we should be looking at data where transmission *did occur* and use it to delineate the holes in the processes. This is standard root cause analysis. https://twitter.com/maxcroser/status/871122567934783488
As well as looking at evidence where no transmission occurred, we should be looking at data where transmission *did occur* and use it to delineate the holes in the processes. This is standard root cause analysis. https://twitter.com/maxcroser/status/871122567934783488
It leaves me with this question.
At what point does the evidence of a series of suggestive cases of transmission whilst wearing adequate droplet precautions begin to overwhelm any uncertainty inherent in the currently noisy data of observational analyses with these confounders?
At what point does the evidence of a series of suggestive cases of transmission whilst wearing adequate droplet precautions begin to overwhelm any uncertainty inherent in the currently noisy data of observational analyses with these confounders?
The case reports of transmission are often dismissed as anecdotal evidence, or the cause of the transmission events are put down to inadequately executed droplet precautions. Are we looking at the evidence in the right way? https://twitter.com/drpieterpeach/status/1298958549511356417
If we take a step back to consider our priors, what other data should we look at? SARS, MERV, Flu? Basic science on the production & physics of aerosols such as the the work by @Don_Milton and @linseymarr? Should we use all the relevant data available to us? https://twitter.com/drpieterpeach/status/1331406462782881792
This is not like extrapolating invitro studies of drug out to clinical efficacy, nor is it like extrapolating a chance anecdotal outcome of a drug in a friend to a wider population.
It’s also not about forgetting that a respirator is part of a bigger IPC strategy.
It’s also not about forgetting that a respirator is part of a bigger IPC strategy.
In anaesthesia when we assess adverse events, we undertake root cause analyses & make process improvements where an intervention is likely to reduce even uncommon risks where the consequences are serious and the intervention is cost effective.
This is probably why so many anaesthetists here are publicly uncomfortable with the approach to analysing evidence and guidance on this issue. We see a low probability high consequence event with a low cost intervention with plausible benefit that is readily available.
Transmission is not only via droplets, or fomites, or short range aerosols, it’s all of them depending on context, *sometimes at the same time*, and it seems clear we can’t predict that context reliably enough to not try and cover all the bases when protecting our staff.
