
N^{b}

Observed
Median

Median cutoff to pass validation

Observed 25^{th} percentile

Observed 75^{th} percentile


Prediction error percents (PE%)

5382

0.72

± 15

9.37

10.47

Absolute prediction error percents (PE%)

5382

9.96

30

4.55

17.65


^{a} The prediction error percents (PE%) were computed for each concentration value using the equation below and the absolute prediction error percents (PE%) was computed as the absolute value of the PE%. PE%
_{
ij
} = (DV
_{
ij
}
 PRED
_{
ij
}
)/PRED
_{
ij
}* 100, where, PE%
_{
ij
} is the percent prediction error between the i
^{th} value of the dependent variable in the j
^{th} subject; and the population prediction of the i
^{th} observation in the j
^{th} subject; DV
_{
ij
} is the value of the i
^{th} observation (i.e. the dependent variable) in the j
^{th} subject; and PRED_{ij} is the population prediction of the i
^{th} observation in the j
^{th} subject.

^{b}N refers to the number of pharmacokinetic samples used in the calculation of the PE% and PE% statistics. The 5382 pharmacokinetic samples were derived from 211 patients.