Comments on Nedelman and Jia's extension of Satterthwaite's approximation applied to pharmacokinetics.

In this letter, I have pointed out how the estimator given by Nedelman and Jia can be made unbiased with only a small modification. Moreover, I report on a new simulation study to show that the unbiased estimator of psi has favorable statistical properties in terms of confidence interval width and coverage probability. Further, the simulation study shows the effect of enforcement of the C-S inequality on the performance of both the biased and unbiased estimators.