Problems of statistical analysis in clinical practice

#1
Sorry for my english, it's terrible.
Friends, I would like to know your opinion, why most medical research is perfected by analysing the confidence interval. Draw conclusions only from a comparison of the mean values and the value of P.
If the CI strives for 1, or the value of p is too small (for example, p = 0.008), then clinically such a difference is either not significant, due to the rarity of the event or its insignificance, or based on a CI close to 1, such a difference or such RR is not is clinically significant.

What do you think?
 
Last edited:

hlsmith

Not a robit
#2
"Clinically" significant is up to the researcher or reader based on substantial impact, etc. Statistical significant is a different thing and is based on a priori objective rules set out, but may not be clinically significant.
 
#3
"Clinically" significant is up to the researcher or reader based on substantial impact, etc. Statistical significant is a different thing and is based on a priori objective rules set out, but may not be clinically significant.
Absolute you are right. But reseachers sometimes, or last time always, make conclusion about there statistical significant like clinical. It is a big problem! It is very bad, that journals like bmj or a&a, critical care, posted articles whit poor statistical significance and very poor cilinical significant.
 

hlsmith

Not a robit
#4
Yes. The big journals require more rigor, but when there are 10,000+ journals out there - lots of garbage makes it to print.