Good question, to which I do not have an answer but do have a suggestion: suck it and see.
Take images of a field rich in stars with accurately measured BVR magnitudes (a Landolt field would be ideal, otherwise find a VS at the AAVSO) and see whether your R-band measurements show a systematic dependency on the tabulated B-V and V-R values. If you have Johnson B and/or V filters, so much the better.
Regular flats should take out spatial variation.