>>13847092>>13852695Table 3:
The three studies - in green, blue, and purple - are significantly better studies than the other two. They are randomized and compare to proper controls. They aren't phone surveys. Even while being decent studies, the meta-analysis fucked it all up and each one shows critical judgement failures in its design. Ailello had a study printed two years prior that was apart of the same series used by Liang. They failed to include it because the data was not in a nice chart nor were the results supportive of their conclusion.
They included the 2009 Cowling et al. even though it does not match the meta-analysis criteria. Instead of comparing a control group and a masked group, the study is looking at applying proper hand hygiene, with or without facemasks. This had no legitimate reason to be included in the meta-analysis. As for their 2008 study, which is looking at face mask effectiveness, these researchers still tripped over themselves and failed to include data which matched their search criteria.
Finally, the Suess et al. study highlights the meta-analysis could not correctly parse data in from studies.
It is pretty obvious that this meta-analysis has bungled up good studies with bad studies while narrowly focusing on the desired outcome in the non healthcare worker case. There is more than enough reason to not venture further into this thing and just throw it away, but I find enjoyment in how every study so far has highlighted some unique flaw in the construction of the meta-analysis. Is this the science that should be trusted?
An interesting point is that there are plenty of good studies here, but they were skewing data and not using their own criteria for the analysis. They could have made a far better analysis, but the outcome would have been even less convincing, and perhaps, this is because surgical masks do fuck all even when carefully used to prevent immediate transmission.
