We have previously developed and evaluated a new in vivo method for measuring protein binding (PB) of drugs. In this method, concentrations of drug and protein are determined prior to and during the 15 min after inflation of a blood pressure cuff on one arm. The free concentration of drug is calculated from the regression line of drug concentration against protein concentration. The robustness of the method was tested by a series of simulations to explore: 1) Effect of random errors of the assays of protein and drug. 2) Performance of the method over the range of PB (0-100%) with fixed assay error of ±5%. The coefficient of variation of estimated PB from the method increased (heteroscedastically) with increase in assay error. Variability was acceptable (less than 12%) when the assay error was less than 5%. The simulations show that the variation is constant down to 30% protein binding below which it increases rapidly. In practice this should not invalidate the method as PB lower than 50% is of lesser pharmacokinetic importance. For measuring PB of a drug with PB=90%, the new method requires an assay 33 times less sensitive than that required for standard in vitro methods. In conclusion, the new method offers acceptably low variability without the need for highly sensitive assays and can be used with readily available equipment.