I guess part of my question is how you deal with variation. Every component you design with is simulated over corners and process variation, how do inductors fit into this? Maybe it's possible to pull out the parameters from the monte Carlo spice files but I feel this would a lot of "I think..." rather than "I know...". Especially as thick metals are not usually considered as process steps requiring high control like anything down in the bulk.