The pressure for medical treatment for COVID generated lots and lots of studies. Some good, some awful, few peer-reviewed before being widely and wildly disseminated. A new study looks at how we might separate the good from the bad and ugly.
p-values
Physicians have to apply population-based guidelines to individuals. How do they know which to use? P-values don't work and eminence is not evidence. Can a measure of a studies "fragility" be an answer?
The proverbial searching for "the needle in a haystack" can help us understand science's problem with p-values, and why so many studies find contrary things.