BMC Medical Research Methodology

Table 3 Comparison of Longitudinal Effect on Survival (Distribution = Weibull, N = 100, Link = γ)

From: Revisiting methods for modeling longitudinal and survival data: Framingham Heart Study

Scenarios		TSA					MLA
Censoring	γ	Estimate	SE	CP	Bias	MSE	Estimate	SE	CP	Bias	MSE
10%	0.000	0.000	0.247	0.952	0.000	0.125	0.006	0.074	0.955	0.006	0.011
	0.500	0.512	0.251	0.950	0.012	0.128	0.609	0.103	0.848	0.109	0.033
	1.000	1.041	0.266	0.954	0.041	0.142	1.141	0.152	0.904	0.141	0.066
50%	0.000	0.009	0.319	0.950	0.009	0.208	0.006	0.096	0.945	0.006	0.020
	0.500	0.509	0.323	0.954	0.009	0.166	0.572	0.134	0.938	0.072	0.041
	1.000	1.049	0.315	0.945	0.049	0.211	1.190	0.170	0.804	0.190	0.100
90%	0.000	−0.005	0.644	0.948	−0.005	0.868	−0.041	0.188	0.909	−0.041	0.054
	0.500	0.556	0.666	0.947	0.056	0.970	0.588	0.251	0.878	0.088	0.112
	1.000	1.098	0.759	0.940	0.098	1.403	1.243	0.409	0.844	0.276	0.234
		BSJM					TDCM
Censoring	γ	Estimate	SE	CP	Bias	MSE	Estimate	SE	CP	Bias	MSE
10%	0.000	0.019	0.073	0.910	0.019	0.014	−0.001	0.188	0.938	−0.001	0.073
	0.500	0.592	0.099	0.920	0.092	0.025	0.336	0.191	0.845	−0.164	0.103
	1.000	1.127	0.145	0.900	0.127	0.053	0.666	0.200	0.595	−0.334	0.191
50%	0.000	−0.004	0.094	0.930	−0.004	0.021	0.013	0.244	0.949	0.013	0.124
	0.500	0.610	0.127	0.930	0.110	0.039	0.354	0.252	0.913	−0.146	0.154
	1.000	1.107	0.156	0.930	0.107	0.059	0.688	0.238	0.712	−0.312	0.214
90%	0.000	0.027	0.191	0.970	0.027	0.076	−0.002	0.494	0.943	−0.002	0.502
	0.500	0.778	0.242	0.933	0.278	0.161	0.373	0.510	0.919	−0.127	0.590
	1.000	1.009	0.269	0.990	0.009	0.116	0.758	0.577	0.907	−0.242	0.838

TSA Two Step Approach, MLA Maximum Likelihood Approach, BSJM Bayesian Semi-Parametric Joint Modeling, TDCM Time Dependent Covariate Modeling, CP Coverage Probability, SE Standard Error, MSE Mean Square Error

Back to article page

ISSN: 1471-2288

Contact us

General enquiries: journalsubmissions@springernature.com