4 Comments

I think you need to be highly sus of meta studies. Very difficult to understand actual inclusion/exclusion and so subject to bias that in my experience most are worthless other than as sources of cites got dive deeper into.

The larger the effort, the bigger the meta study, the worse this effect is because larger meta analyses are more subject to statistical sleight of hand and bias.

Expand full comment

Do "patient-level data meta-analysis" constitute higher level of evidence? Couldn't it lead to more "statistical magic" through data manipulation and overestimations? Genuinely asking.

Expand full comment
author

Great point! That's certainly possible. It depends how the data were standardized and the analysis performed. But patient-level meta-analyses generally remain higher level of evidence compared to trial-level meta-analyses that are commonly used nowadays; this is what we meant with our statement.

Expand full comment

Thank you for the answer!

By the way, I wanted to ask: even if the idea of presenting the themes gradually through posts and podcasts is a very smart one, I think I would be keen to have all the information in a book (even if pdf / ebook, to facilitate) at the end to facilitate future consultations. I would pay for that and I think I wouldn't be the only one!

Cheers!

Expand full comment