Csm sensitivity to ground stiffness

[orfilos1]

I am using AASHTO LFRD to design the pile group foundations of a viaduct. I carried out a sensitivity analyses for the elastic Seismic Response Coeffiecient Csm against stiffness profile and it seems very insensitive. I use REPUTE (a BEM pile group code based on PGroupN) to calculate the rotational stiffness of the pile group and I put this as an input to LUSAS to get the natural period of the viaduct which I analyse as a SDOF. I define with this the Csm. What I get is essentially no variation of Csm with ground stiffnesss which does not make sense.
Has anybody experienced similar patterns? Any tips to identify what I should be checking to get to the bottom of this?

Orfilos

PS My regards to our Geo-colleagues from the States.