Estimating and modelling cure in population-based cancer studies within the framework of flexible parametric survival models

Table 3 Parameter estimates from flexible parametric survival models

Parameter	Standard FPM splines calculated backwards	Restricted FPM incorporating a cure proportion	Restricted FPM including time-dependent effects
Intercept	0.385 (0.032)	0.350 (0.031)	0.279 (0.033)
Age group
<50	-	-	-
50-59	0.089 (0.034)	0.089 (0.034)	0.085 (0.035)
60-69	0.214 (0.030)	0.212 (0.031)	0.184 (0.031)
70-79	0.385 (0.030)	0.382 (0.030)	0.284 (0.031)
80+	0.788 (0.032)	0.785 (0.032)	0.532 (0.035)
LRT	p <0.0001	p <0.0001	p <0.0001
Calendar period
1953-1964	-	-	-
1965-1974	-0.339 (0.029)	-0.340 (0.029)	-0.264 (0.032)
1975-1984	-0.741 (0.028)	-0.743 (0.028)	-0.585 (0.030)
1985-1994	-0.922 (0.027)	-0.924 (0.027)	-0.769 (0.029)
1995-2003	-1.190 (0.027)	-1.194 (0.027)	-1.025 (0.030)
LRT	p <0.0001	p <0.0001	p <0.0001
LRT comparing to the previous model		p <0.0001	p <0.0001
AIC	92285.22	92323.95	91590.12
BIC	92398.63	92429.26	91825.04

Parameter estimates (standard errors) from flexible parametric survival models
FPM, flexible parametric survival model; LRT, likelihood ratio test; AIC, Akaike's information criterion; BIC, Bayesian information criterion.

ISSN: 1471-2288