As you just said, this meta analysis is a study of studies.
1)The studies that this meta analysis evaluated did not evaluate the drug in the dose/duration in question. Th average doses were about twice the recommended dose for corona tox. This is particularly important wrt evaluating side effects.
2)Many studies are complete trash. A study that studies studies is thus, by its very nature much more subject to incorporating trash into its own findings. It takes a great deal of time and effort to do a meta analysis properly to filter out all of the bad studies, and it has to be done in person, study by study. A good meta analysis usually takes a long time.
This study was, by its own admission, rushed out because of the “hotness” of the topic. Any professional reviewer would be very skeptical of this study based on this fact alone. Still, I trust the Lancet, so I will take the at their word on this - though the Lancet never presented this study in the light that the WaPo did.
3)Meta studies also compare a lot of apples and oranges by their very nature. Thus, they are best used when evaluating studies that are similar (example indomethicin as a treatment for gout). In this case, there has never even been a direct prospective study because the infection is brand new. Thus, by its design, the study is not ideal type of study to evaluate the question, and is potentially inaccurate.
Go back and reread my posts on this, and my original thread down the board. You may have misunderstood me.