Correcting for misclassification and selection effects in estimating net survival in clinical trials

Table 3 Performance in terms of parameters estimation (\( {\widehat{\beta}}_{age},{\widehat{\beta}}_{treatment},\kern0.5em \widehat{\boldsymbol{\alpha}}\Big) \) with rescaled B-spline (RBS) and Cox model in various scenarios

Method & Misc.	Scenario 1 (α = 1)			Scenario 2 (α = 0.5)			Scenario 3 (α = 2)			Scenario 4 (α = 4)
Method & Misc.	RBias	RMSE	ECR	RBias	RMSE	ECR	RBias	RMSE	ECR	RBias	RMSE	ECR
\( {\widehat{\beta}}_{age} \)(β_age = 0.05)
CSS methods
Cox & 0%	0.528	0.392	94.5	0.360	0.388	95.7	0.170	0.389	94.8	0.802	0.388	94.8
Cox & 20%	0.552	0.392	94.4	0.355	0.389	96.3	0.128	0.391	95.2	0.709	0.389	94.6
Cox & 30%	0.612	0.392	94.7	0.446	0.388	96.1	0.141	0.391	95.1	0.832	0.389	94.6
PNS methods
RBS	−0.489	0.444	96.4	−1.749	0.295	97.1	0.417	0.686	94.04	2.006	1.065	91.7
\( {\widehat{\beta}}_{treatment} \)(β_treatment = − 0.5)
CSS methods
Cox & 0%	0.869	0.392	94.8	−0.075	0.388	95.5	0.245	0.389	94.8	−0.311	0.388	93.9
Cox & 20%	1.027	0.392	95.2	0.042	0.389	95.3	0.710	0.391	94.6	−0.044	0.389	93.7
Cox & 30%	0.820	0.392	94.7	−0.271	0.388	95.3	0.773	0.391	95.8	0.095	0.389	94.3
PNS methods
RBS	1.920	0.234	95.6	3.209	0.155	97.1	0.774	0.391	95.8	−1.083	1.032	92.3
\( \widehat{\alpha} \)	0.060	0.444	92.6	15.754	0.305	90.8	−5.863	0.696	93.6	−7.419	1.105	94.9

Misc Misclassification rate, RMSE root mean square error, ECR empirical coverage rate, CSS cause-specific survival, PNS population-based net survival, RBS rescaled B-spline, RBias relative bias

ISSN: 1471-2288