We've run seismic analyses where we have included in the model displacements to account for building drift. These displacements were "large" - say 4" to 6" - so it's easy to accept the fact that they need to be accounted for in CAESAR. However, we now have a structural steel pipe bridge which the structural folks say will deflect 1-1/2" down under load. So our usual assumption that the support steel is "fixed" is not valid. We either stiffen the steel to get to something acceptable, whatever that is, or model the steel in CAESAR and see what happens. Are there any guidelines to indicate when the support steel can be assumed fixed and when its deflection needs to be accounted for?