Skip to main content
. 2020 May 27;11:794. doi: 10.3389/fphar.2020.00794

Table 6.

Bias [mean absolute percentage error (MAPE)] in individually predicted doses at the next occasion, taking information from one (day 1), two (day 1, 7) or three (day 1, 7, 14) sampling occasions using three samples (pre-dose, and 2 and 4 h post-dose) per occasion into account for Empirical Bayes Estimates (EBE) estimation.

True dose MAPE (%) ΔMAPE (%) MAPE (%) ΔMAPE (%) MAPE (%) N
One occasion (One/two occasions) Two occasions (Two/three occasions) Three occasions
(95% CI) (95% CI)
All doses 13.3 [−0.09 to −0.07]* 5.6 [−0.01 to 0.00] 5.5 1,000
600 50.0 0.0 0.0 1
900 46.1 [−0.53 to −0.33]* 2.9 [−0.02 to 0.04] 3.9 34
1200 21.1 [−0.20 to −0.14]* 4.1 [−0.0 to 0.02] 4.7 133
1,500 11.9 [−0.08 to −0.05]* 5.5 [−0.01 to 0.01] 5.5 325
1,800 6.5 [−0.03 to −0.01]* 4.1 [0.00 to 0.01] 4.3 292
2100 12.6 [−0.06 to −0.04]* 7.8 [−0.02 to 0.00] 7.1 162
2400 18.6 [−0.08 to −0.04]* 12.8 [−0.05 to −0.02]* 9.3 47
2700 24.1 [−0.13 to −0.06]* 14.8 [0.00 to 0.00] 14.8 6

MAPE, mean absolute percentage error; CI, confidence interval.

*Statistically significant difference (p-value < 0.05) in MAPE between taking information from one or two occasions, or between taking two or three occasions, into account to predict individual doses at the next occasion.