Mask-wearing cuts Covid incidence by 53%, says global study | Coronavirus | The Guardian This is all over my timeline today. I’m a fan of masks, and it is AMAZING progress to see a meta-analysis of public health interventions for covid. But….. theguardian.com/world/2021/nov…
..we need to look past headlines. This study looked at a range of measures, including masks, handwashing+disinfection. It’s well written and methodologically fine. But a meta-analysis can only analyse primary research and it’s conclusions are only as good as those studies…. 
…While people have been screaming at public health to “follow the science”, public health has been saying the science is blighted by confounders. So we need to be careful about interpreting the finding around masks, as the authors themselves point out. 6 studies. Big bias risk…
…Authors also concluded that handwashing is useful. But wide confidence intervals make it hard to determine effect size. Reassuring to see that handwashing *seems* to be beneficial. Important to note they’re doing a meta-analysis of mostly retrospective observational studies…
…They also looked at social distancing. 25% risk reduction looks about right, but is clearly dependent on the actual distance in the real world. Again, these studies weren’t particularly well designed for the outcome in question, so no surprise that the risk of bias was high…
..Disinfection of surfaces in households got a look-in too. Interesting to see very high reduction in odds of transmission, but mega wide confidence intervals and only one study made the grade. So, again, this shows the uncertainty we work with every day in managing this virus…
…I think it’s important for people not to pull the pieces that they’re ideologically tethered to out of this study to increase the “follow the science” noise in the pandemic space. Let’s just acknowledge uncertainty. The accompanying editorial to this paper says it best…
…Here’s what the editorial says about the section on masks, as this is pretty measured and I suspect the headlines generated from this paper will mostly be about masks. Most likely a small but welcome reduction in risk…
..A bit more context on this from someone who knows what he’s talking about.
https://twitter.com/bobepidemiology/status/1461212866971717633
• • •
Missing some Tweet in this thread? You can try to
force a refresh
