|
|
|
|
|
|
by sachinag
2051 days ago
|
|
|
The "non-N95 masks don't protect yourself" was what we thought in March; it's no longer what we think after the Starbucks contact tracing and other studies. Regular masks absolutely provide substantial - admittedly, not perfect, but substantial - personal protection. |
|
|
There is no evidence for this, and if you read the CDC guidance document that suggested it this week, you'll see that every citation in the newly added section refers to studies about filtration efficiency, which is a very different question:
https://www.cdc.gov/coronavirus/2019-ncov/more/masking-scien...
The high-quality evidence for masks has not changed substantially since March. The few lines of direct evidence suggest that it's possible they help a tiny amount, but we haven't been able to detect much evidence of this, either way. Only by combining multiple (poor) studies together, pooling the data and statistically re-weighting the pooled results was the WHO meta-analysis able to find any signal at all, and the signal wasn't strong:
http://www.economicsfaq.com/retract-the-lancets-and-who-fund...
In 2020, we've had a number of studies that attempt to use statistical analysis of infection data from late spring (when cases were declining in the northern hemisphere) to draw conclusions about masks, and a handful of anecdotes that are repeated frequently by the media -- the Missouri hairdresser, the Korean Starbucks -- that were neither controlled, nor subjected to statistical analysis of any sort.
The only randomized controlled trial conducted on masks for personal protection against Covid -- the highest quality form of evidence for a health intervention -- has been completed since June, and rejected by three major publications:
https://clinicaltrials.gov/ct2/show/NCT04337541
https://www.berlingske.dk/videnskab/professor-stort-dansk-ma...
Moreover, as we have now seen, most of the places around the world with mask mandates and high rates of self-reported compliance are experiencing spikes in infection. At the very least, these call into question the publications that attempted to make claims based on data from the spring. We're starting to see retractions related to this:
https://www.medrxiv.org/content/10.1101/2020.10.21.20208728v...