Never ever interpret clinical trial data in medicine in isolation. It is always best interpreted in context of biologic plausibility and pre test probability.
Pre test probability is affected by how much we know, especially results of other similar studies.
Not everyone puts in the effort to learn what's out there to calculate the best pre test probability.
Not everyone puts in the effort to learn what's out there to calculate the best pre test probability.
Similarly not everyone may have studied enough to understand the underlying biology.
Just interpreting a trial based on the numbers absent these two considerations is unnecessarily blinding oneself and trying to describe the elephant.
The lower the pretest probability and bio-plausibility, the stronger the data need to be to be believable. Successes do exist in such circumstances, but often the result is due to chance or bias.
Conversely the higher the pre test probability and bio-plausibility, the more likely you are dealing with a true result. In those cases you could even live with higher P values and some potential biases/limitations.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
