Skip to main content
. Author manuscript; available in PMC: 2017 Jan 1.
Published in final edited form as: J Biopharm Stat. 2015 Jun 22;26(3):475–498. doi: 10.1080/10543406.2015.1052492

Table 3.

Parameter estimates from 500 simulated datasets containing N = 10 and N = 30 patients distributed equally to receive one of five dose groups d1 . . . d5 over six cycles comparing comparing estimates based on Model 3 (α, β, ρ) that use multiple toxicity responses to estimates based on Models 6 (γ) and 7 (δ) that use a single binary toxicity response.

Prior Mean (SD) Estimate(Bias) MSD1 ESD2 Coverage
N=10 patients
        α 1 (2) 0.945 (0.055) 0.682 0.585 99.6
        β 0.5 (1) 0.515 (−0.015) 0.464 0.308 100
        ρ 0.83 (0.14) 0.820 (−0.020) 0.137 0.040 100
        γ 1 (2) 0.916 (0.084) 0.703 0.630 99.8
        δ 1 (2) 0.976 (0.024) 0.548 0.536 92.6
N=30 patients
        α 1 (2) 0.989 (0.011) 0.466 0.408 96.0
        β 0.5 (1) 0.507 (−0.007) 0.343 0.285 98.8
        ρ 0.83 (0.14) 0.802 (−0.002) 0.133 0.058 99.8
        γ 1 (2) 0.940 (0.060) 0.474 0.428 96.6
        δ 1 (2) 0.998 (0.002) 0.333 0.331 94
1

MSD is mean of the SD from 500 estimates

2

ESD is empirical SD of the 500 estimates