A Flexible Bayesian Parametric Proportional Hazard Model: Simulation and Applications to Right-Censored Healthcare Data

Table 12

Posterior properties summaries and the information criterion values for the considered GLL PH model and its competing models for the larynx cancer data.

Summaries	Posterior characteristics
Parametric competitive models	Parameter(s)	Posterior mean	Posterior SD	Pr (>|0|data)	HPD interval (95%)
GLL-PH model (DIC = 294.412)
	Alpha	1.539	0.215	1.000	(1.157, 2.005)
	(stage 2)	−0.182	0.454	0.352	(−1.108, 0.667)
	(stage 3)	0.376	0.337	0.870	(−0.289, 1.030)
	(stage 4)	1.222	0.411	0.998	(0.396, 2.010)
	(age)	0.187	0.144	0.906	(−0.091, 0.476)
	(diagyr)	−0.111	0.149	0.227	−0.730, 0.181)
	Eta	0.869	0.247	1.000	(0.197, 1.412)
	Kappa	0.336	0.077	1.000	(0.112, 0.507)
Weibull-PH model (DIC = 296.776)
	Alpha	0.908	0.105	1.000	(0.713, 1.118)
	(stage 2)	−0.380	0.446	0.198	(−1.275, 0.468)
	(stage 3)	0.174	0.318	0.711	(−0.483, 0.781)
	(stage 4)	1.095	0.393	0.997	(0.329, 1.857)
	(age)	0.176	0.141	0.899	(−0.092, 0.461)
	(diagyr)	−0.012	0.146	0.468	(−0.294, 0.274)
	Kappa	0.154	0.041	1.000	(0.081, 0.236)
Gompertz-PH model (WAIC = 297.560)
	Alpha	0.134	0.031	1.000	(0.076, 0.196)
	(stage 2)	−0.138	0.455	0.392	(−1.040, 0.737)
	(stage 3)	0.393	0.328	0.886	(−0.252, 1.041)
	(stage 4)	1.544	0.397	1.000	(0.776, 2.308)
	(age)	0.206	0.149	0.919	(−0.084, −0.501)
	(diagyr)	0.075	0.155	0.685	(−0.230, 0.374)
	Kappa	0.552	0.186	1.000	(0.227, 0.919)