Thread by @GidMK

The ivermectin literature is full of fraud

What can we do about it?

Our new paper in @NatureMedicine argues that we need a systematic change to counter this issue

nature.com/articles/s4159…

The basic issue is that science works on trust. We assume that no one would ever fake a study, because it's ethically and morally indefensible, and work with that

Which makes it very easy for people to fake studies

However, people generally aren't very good at faking things. There are dozens of very simple checks you can run on data to see if it's real

Problem is, you need the data to do that

So, the solution: if you are going to meta-analyze clinical research, and particularly clinical research on Covid-19 drugs, you should ask authors to provide their anonymized, individual patient data

If people refuse or fail to send through minimal datasets, it should be a red flag and the study rated as high risk of bias

It takes more work, but the alternative is accepting that our opinions may be based on fraud