DP data interpretation

Dear all,

I am doing the dose proportionality analysis with power model, and I have a question regarding the result interpretation. I know the Smith criterion, in which the 90% CI of β need to fall in the (1+ln0.8/lnr,1+ln1.25/lnr) range. However, this criterion is very strict to most of the studies.

Is there any other widely accepted criterion?

Some people says as long as the 90% CI across 1, it can be concluded as DP. However, I could not see any regulations about this. Is there anyone use this criterion?

Thanks

MDMD

Hi MDMD,

some fruitful discussion could be found here (regarding criteria too)

Mittyright

Yes; mittyright above points you to a useful discussion for the calculation of dose proportionality along the lines discussed in Brian Smith’s paper. However you will note that Helmut has
emphasized a transformation to be used for calculation of the final confidence intervals needed.

Angus