How do you treat a patient with warfarin failure, with therapeutic INR 2-3 at the time of DVT, and no underlying malignancy or hypercoagulable state?

I would give DOACs a shot in this case. The INR of 2-3 at the time of DVT "Diagnosis" might have been <2 at the time of DVT "development/occurrence" depending on how frequently the INR had been checked. I would, of course, maximize risk factors control as well.

Let INR drop to 1.5 and then start heparin.