I have tackled this problem before. A mentor of mine recommended doing a wedge analysis over a perscribed "failure surface" using a peak phi angle and ignoring cohesion, assume there is no earth downslope of the piers, and find the unbalance force that corresponds to a factor of safety of 1.2. The rational is that creep occurs when the driving stresses get within about 80 percent of the resisting stresses, so the calculated unbalanced force should prevent creep if it is applied to the earth (in this case, structurally). Then, I applied a creep load on each pile equal to the calculated unbalanced force times the pile spacing, and applied it to the pile at a depth below the ground surface equal to 2/3 the depth to top of support (2/3 the depth to top of rock in my case). Then, I recommneded that the structural engineer design the pile to resist that force and any other lateral loads (seismic and wind) using the UBC pole formula. Nowadays, I use L-pile, do the analysis myself.
In my case, prescribing the failure surface, or a surface over which creep occurs, could be reasonably estimated, because it was coluvium creeping over top of rock, and the rock was thestable top of support. May be difficult for your problem if there is no rock.
Good luck - not an easy problem. My gut feel is that the approach I describe is reasonably conservative and captures all the soil mechanics that influence the problem. Incidentally, the projects I worked on held up fine.