Table 1.
Simulation scenarios for the unconditional and conditional settings.
Figure 1.
Probability of imbalance using absolute measure (D1) with different trial sizes.
Lines correspond to Pr (D1≥d1), where d1 = 0.005 (hollow circle), 0.01 (triangle), 0.025 (cross), 0.05 (X), 0.10 (diamond), 0.15 (inverted triangle), and 0.20 (filled circle), from the top to the bottom, respectively. Top left: 25/arm, top right: 50/arm, middle left: 125/arm, middle right: 500/arm, bottom left: 1000/arm, bottom right: 2000/arm.
Figure 2.
Bias, simulation standard deviation (SD), coverage proportion and statistical power for the unadjusted and adjusted logOR, in scenario 1, the unconditional setting, with 125 patients per arm.
The unadjusted model is indicated by the dotted line with hollow circles, and the adjusted model is indicated by the solid line with filled circles.
Figure 3.
Bias, simulation standard deviation (SD), coverage proportion and statistical power for the unadjusted and adjusted logOR, in scenario 1, the unconditional setting, with 2000 patients per arm.
The unadjusted model is indicated by the dotted line with hollow circles, and the adjusted model is indicated by the solid line with filled circles.
Figure 4.
Probability of difference between the estimated and true ORR (deviation in either direction) in scenario 1, the unconditional setting, with 125 patients per arm.
Within each graph, lines correspond to Pr (|DORR|≥d2), where d2 = 0 (solid circle), 0.05 (bullet), 0.10 (little circle), 0.15 (square), 0.2 (diamond) and 0.25 (triangle), from top to bottom, respectively.
Figure 5.
Probability of difference between the estimated and true ORR (deviation in either direction) in scenario 1, the unconditional setting, with 2000 patients per arm.
Within each graph, lines correspond to Pr (|DORR|≥d2), where d2 = 0 (solid circle), 0.05 (bullet), 0.10 (little circle), 0.15 (square), 0.2 (diamond) and 0.25 (triangle), from top to bottom, respectively.
Table 2.
Probability of difference between the estimated and true ORR≥0.05 in the unconditional setting scenario 1.
Table 3.
Probability of difference between the estimated and true ORR≥0.10 in the unconditional setting scenario 1.
Table 4.
Probability of difference between the estimated and true ORR≥0.05 in the unconditional setting scenario 3.
Table 5.
Probability of difference between the estimated and true ORR≥0.10 in the unconditional setting scenario 3.