I don't think this is a matter of concentration. Rheumatoid factor assays measure binding and that depends both on concentration and a profile of affinity going from very high to not so high. One patient may have a small amount of high affinity RF and another a large amount of medium affinity RF and the same amount may bind in the assay. Binding may vary with lab conditions, so you need to calibrate against a standard. It's a bit like fishing for crabs with bacon. If the crabs are hungry you will get a good catch when there are only a few. If they are not so hungry, you need more in the sea to get the same catch. 

If you had a standardised monoclonal rheumatoid factor you could probably work out the relation of IU to concentration, but not for polyclonal RF in sera.