Skip to main content

Table 4 Sensitivity analysis of the ‘proposed’ method depending on the underlying frailty distribution in terms of the averages of the relative bias (r.Bias) as well as the standard errors (SEM) and coverage probability (CP) when the type of the regression coefficients is ‘even’ and the LTF proportion is ‘moderate’

From: Additive-multiplicative hazards regression models for interval-censored semi-competing risks data with missing intermediate events

   N(0,0.01) U(−0.173,0.173) DE(0.007) G(100.5,0.01)
   (LTF(%)=34.2) (LTF(%)=34.4) (LTF(%)=34.5) (LTF(%)=34.5)
Parameter True value r.Bias (%) SEM (×105) CP (%) r.Bias (%) SEM (×105) CP (%) r.Bias (%) SEM (×105) CP (%) r.Bias (%) SEM (×105) CP (%)
α 01 0.01 18.8 15299 94.0 40.9 15169 94.2 36.7 15215 94.8 -28.7 15089 96.2
α 02 0.01 1.3 24633 93.2 -83.8 24639 97.0 89.0 24401 94.4 45.0 24623 96.6
α 03 0.01 -143.8 17210 94.4 58.9 17266 93.8 -110.5 16925 93.8 119.9 17198 93.6
α 12 0.01 181.1 25925 96.4 -27.9 26141 94.8 27.1 26318 93.2 159.3 25682 95.2
α 32 0.01 -16.5 33404 95.4 -80.1 33158 94.0 -22.5 33619 94.0 -82.4 33595 91.8
β 01 0.004 -4.3 91 94.8 -5.8 90 94.0 -3.3 91 94.0 -6.7 90 92.2
β 02 0.004 2.7 99 96.4 4.9 100 94.8 5.4 101 97.0 4.4 99 95.4
β 03 0.004 -2.2 109 93.2 -2.9 109 91.0 -2.8 110 94.4 -2.9 108 91.8
β 12 0.004 0.1 111 96.6 2.3 111 95.0 0.6 110 94.8 -0.3 111 95.6
β 32 0.004 8.0 174 97.0 0.8 169 96.0 2.9 171 96.8 5.2 170 96.4
σ 2 0.01 955.8 8431 89.2 974.2 8325 89.4 926.1 8284 89.6 928.7 8462 89.2