Skip to main content

Table 3 Model parameter estimates for different fractional polynomial network meta-analysis models.

From: Network meta-analysis of survival data with fractional polynomials

   Fixed effects model, first order fractional polynomial Fixed effects model, second order fractional polynomial Random effects model (d0), second order fractional polynomial
   Median of posterior distribution 95% Credible Interval Median of posterior distribution 95% Credible Interval Median of posterior distribution 95% Credible Interval
p1 power 1 -2   -2   -2  
p2 power 2    1   1  
Pooled estimate for difference in β 0
d 0AB gefitinib vs. docetaxel -0.003 (-0.111; 0.112) 0.158 (-0.065; 0.409) 0.144 (-0.156; 0.411)
d 0AC BSC vs. docetaxel 0.770 (0.598; 0.939) 1.674 (1.108; 2.193) 1.653 (1.039; 2.167)
d 0AD pemetrexed vs. docetaxel 0.020 (-0.189; 0.243) 0.141 (-0.483; 0.744) 0.142 (-0.540; 0.761)
Pooled estimate for difference in β 1
d 1AB gefitinib vs. docetaxel 0.634 (-0.429; 1.726) -0.003 (-1.470; 1.362) 0.070 (-1.330; 1.542)
d 1AC BSC vs. docetaxel -2.507 (-4.281; -0.783) -5.858 (-8.503; -3.228) -5.839 (-8.265; -3.131)
d 1AD pemetrexed vs. docetaxel -0.995 (-3.344; 1.381) -1.359 (-4.604; 1.885) -1.368 (-4.511; 1.814)
Pooled estimate for difference in β 2
d 2AB gefitinib vs. docetaxel    -0.007 (-0.016; 0.002) -0.006 (-0.015; 0.003)
d 2AC BSC vs. docetaxel    -0.053 (-0.082; -0.023) -0.052 (-0.079; -0.023)
d 2AD pemetrexed vs. docetaxel    -0.005 (-0.032; 0.023) -0.005 (-0.031; 0.022)
σ d0 Standard deviation of d 0 ; heterogeneity in difference of β 0 across comparisons      0.060 (0.002; 0.406)