Skip to main content
. 2020 Aug 8;47(6):527–542. doi: 10.1007/s10928-020-09704-1

Table 4.

Predicted versus actual treatment category based on midazolam clearance (Qmet) cut-points

Predicted category and midazolam clearance cut-point Inhibition No modulation Induction
(4.82–16.4 L/h) (16.4–41.8 L/h) (41.8–88.9 L/h)
Actual treatment—original dataset
Inhibition [N (%)] 59 (93.7) 4 (6.35) 0 (0)
No modulation [N (%)] 2 (1.35) 146 (98.6) 0 (0)
Induction [N (%)] 0 (0) 0 (0) 12 (100)
Actual treatment—external validation set
Inhibition [N (%)] 72 (83.7) 14 (16.3) 0 (0)
No modulation [N (%)] 1 (5.00) 19 (95.0) 0 (0)
Induction [N (%)] 0 (0) 15 (62.5) 9 (37.5)
Actual treatment—external validation set, no weak induction
Inhibition [N (%)] 72 (83.7) 14 (16.3) 0 (0)
No modulation [N (%)] 1 (5.00) 19 (95.0) 0 (0)
Induction [N (%)] 0 (0) 4 (33.3) 8 (66.7)
Actual treatment—both sets
Inhibition [N (%)] 131 (87.9) 18 (12.1) 0 (0)
No modulation [N (%)] 3 (1.79) 165 (98.2) 0 (0)
Induction [N (%)] 0 (0) 15 (41.7) 21 (58.3)
Actual treatment—both sets, no weak induction
Inhibition [N (%)] 131 (87.9) 18 (12.1) 0 (0)
No modulation [N (%)] 3 (1.79) 165 (98.2) 0 (0)
Induction [N (%)] 0 (0) 4 (16.7) 20 (83.3)