Fred,
There is a series of formulae that are followed for this in the ASD manual. You have to combine all of the stresses: compression, strong axis bending, weak axis bending.  I don't know them off the top of my head but I do know that if fa/Fa is greater than 0.15 it gets more complex.  Well, actually, I do know the formula for that off the top of my head:
fa/Fa + fbx/Fbx + fby/Fby < 1.0 under normal loading
fa/Fa + fbx/Fbx + fby/Fby < 1.33 under transient loading
but that is only if fa/Fa < 0.15
It's under combined axial and bending in the steel manual

ChipB,

yopu neglected to include the Eulers reduction portion in the equations.

The correct equation is:
fa/Fa + fbx/(1-fa/Fex)Fbx + fby/(1-fa/Fey)Fby < 1.0

Also, I believe that the 1/3 increase has been removed from ASCE 7 and subsequently from the interaction equations.  

Fred52, you would design the moment in the weak direction of the coum in a similar manner as you would in the strong direction.  The difference is that the top and bottom horizontal forces must be transferrd into the column flanges which will entail full penetration welded horizontal top and bottom stiffener plates.  The plates will need to be designed for the tension and/or compression force caused by the moment.  

These plates are not the same as what is required for web buckling.  Be sure to check that the web transfer plates don't interfere wit the web stiffener plates

The best thing to do if you can is to eliminate the weak axis moment connections.