4 ms·

There's at least one RCT on masking w.r.t. covid that finds no statistical benefit. I don't think it's appropriate to reference modeling as evidence after we h

by toolz 5y ago

There's at least one RCT on masking w.r.t. covid that finds no statistical benefit. I don't think it's appropriate to reference modeling as evidence after we have thorough studies published. https://www.acpjournals.org/doi/10.7326/m20-6817 https://www.acpjournals.org/doi/10.7326/m20-6817

gjulianm 5y ago
Inconclusive results != no statistical benefit. A 95% of CI that ranges between 46% decrease and 23% increase is pretty much inconclusive, you can't say masks are better with that study but you wouldn't be justified either in saying they don't do anything.
- toolz 5y ago
  Quoting the results section. > Although the difference observed was not statistically significant I'm not sure I understand your objection, but you most definitely can say there was no statistically significant benefit.
  - gjulianm 5y ago
    > but you most definitely can say there was no statistically significant benefit. Yes, but this study also doesn't point too conclusively that they don't have any effect. If the CI was something like 1% increase in infections - 1% decrease, you could say with a lot of confidence that it doesn't seem that masks did anything at all. But -23% to 46% is a gigantic range to say anything at all about effectiveness. So while this study is, as you say, not supporting the hypothesis that masks are effective, it doesn't support either the hypothesis that masks are not effective.
    - toolz 5y ago
      No, it very specifically supports the hypothesis that masks are not effective. Maybe not to the rigor you would hope for, but to my knowledge this is the most rigorous covid/mask study to date, so it's the best we've got. That isn't to say we should ignore all other studies, but the body of literature on mask effectiveness is very split. Maybe masks do work, I wouldn't be surprised if that were true, but given the body of evidence I'm of the opinion if they do work their effect is marginal.
      - gjulianm 5y ago
        > No, it very specifically supports the hypothesis that masks are not effective. It very clearly doesn't. No support for "masks are effective" is not the same that "support for masks are ineffective". To put an extreme example, suppose I make an study with just two participants, one with mask and another one without, and neither of them get infected. The results wouldn't support the hypothesis that masks are effective, but neither could support the hypothesis that masks are ineffective. Which is the point: the study you linked to does not have enough power to actually check effectiveness of the measure. The 95% confidence interval is way too large to draw any conclusion. Another issue for the study is that it's evaluating surgical masks, which are less effective than N95 masks, and only for infection of the wearer, when we know they're much more effective at avoiding others from getting infected. That's why population-level studies are important, and you'll find that those support the effectiveness of masks.
md_ 5y ago
That study is about the benefit of wearing surgical masks "in a setting where masks were uncommon and not among recommended public health measures"--thus answering the question "do surgical masks protect the wearer." Surgical masks are thought to have a relatively limited benefit for the wearer, so this isn't surprising. However, https://www.nature.com/articles/d41586-021-02457-y https://www.nature.com/articles/d41586-021-02457-y demonstrates clearly that in a randomized controlled trial where surgical masks are distributed to a whole community, there was a statistically significant reduction in cases: "The decrease was a modest 9%, but the researchers suggest that the true risk reduction is probably much greater, in part because they did no SARS-CoV-2 testing of people without symptoms or whose symptoms did not meet the World Health Organization’s definition of the disease." There's certainly more to learn here, of course.

by toolz 5y ago

gjulianm 5y ago

Inconclusive results != no statistical benefit. A 95% of CI that ranges between 46% decrease and 23% increase is pretty much inconclusive, you can't say masks are better with that study but you wouldn't be justified either in saying they don't do anything.

toolz 5y ago
Quoting the results section. > Although the difference observed was not statistically significant I'm not sure I understand your objection, but you most definitely can say there was no statistically significant benefit.
- gjulianm 5y ago
  > but you most definitely can say there was no statistically significant benefit. Yes, but this study also doesn't point too conclusively that they don't have any effect. If the CI was something like 1% increase in infections - 1% decrease, you could say with a lot of confidence that it doesn't seem that masks did anything at all. But -23% to 46% is a gigantic range to say anything at all about effectiveness. So while this study is, as you say, not supporting the hypothesis that masks are effective, it doesn't support either the hypothesis that masks are not effective.
  - toolz 5y ago
    No, it very specifically supports the hypothesis that masks are not effective. Maybe not to the rigor you would hope for, but to my knowledge this is the most rigorous covid/mask study to date, so it's the best we've got. That isn't to say we should ignore all other studies, but the body of literature on mask effectiveness is very split. Maybe masks do work, I wouldn't be surprised if that were true, but given the body of evidence I'm of the opinion if they do work their effect is marginal.
    - gjulianm 5y ago
      > No, it very specifically supports the hypothesis that masks are not effective. It very clearly doesn't. No support for "masks are effective" is not the same that "support for masks are ineffective". To put an extreme example, suppose I make an study with just two participants, one with mask and another one without, and neither of them get infected. The results wouldn't support the hypothesis that masks are effective, but neither could support the hypothesis that masks are ineffective. Which is the point: the study you linked to does not have enough power to actually check effectiveness of the measure. The 95% confidence interval is way too large to draw any conclusion. Another issue for the study is that it's evaluating surgical masks, which are less effective than N95 masks, and only for infection of the wearer, when we know they're much more effective at avoiding others from getting infected. That's why population-level studies are important, and you'll find that those support the effectiveness of masks.

md_ 5y ago

That study is about the benefit of wearing surgical masks "in a setting where masks were uncommon and not among recommended public health measures"--thus answering the question "do surgical masks protect the wearer." Surgical masks are thought to have a relatively limited benefit for the wearer, so this isn't surprising. However, https://www.nature.com/articles/d41586-021-02457-y https://www.nature.com/articles/d41586-021-02457-y demonstrates clearly that in a randomized controlled trial where surgical masks are distributed to a whole community, there was a statistically significant reduction in cases: "The decrease was a modest 9%, but the researchers suggest that the true risk reduction is probably much greater, in part because they did no SARS-CoV-2 testing of people without symptoms or whose symptoms did not meet the World Health Organization’s definition of the disease." There's certainly more to learn here, of course.