Table 3 : Estimation of coefficient, 95% confidence interval (C.I.), and p-value from
the joint model with (i) random effects sub-model and Cox survival sub-model and
(ii) random effects sub-model and Weibull survival sub-model.


Model Covariate Estimate 95% CI p-value

(i) Longitudinal
random effect
sub-model
Intercept 23.79 22.79 24.81 < 0.0001
ART treatment $ -4.606 -5.83 -3.38 < 0.0001
HCV@ -3.03 -3.83 -2.23 <0.0001
Time -0.093 -0.13 -0.05 <0.0001
Time*ART treatment 0.097 0.05 0.14 0.0001

Cox
survival
sub-model
ART treatment -1.369 -2.19 -0.55 0.001
HCV 1.462 0.39 2.54 0.008
Association parameter -0.102 -0.16 -0.04 0.001

(ii) Longitudinal
random effect
sub-model
Intercept 21.42 20.31 22.54 < 0.0001
ART treatment -3.348 -4.58 -2.11 < 0.0001
HCV -1.333 -2.09 -0.58 0.0005
Time -0.115 -0.16 -0.07 <0.0001
Time*ART treatment 0.108 0.05 0.16 0.0001

Weibull
survival
sub-model
Intercept -6.288 -9.80 -2.77 0.0005
ART treatment -1.389 -2.25 -0.53 0.002
HCV 1.477 -0.74 3.70 0.19
Association parameter -0.125 -0.21 -0.04 0.003

*ART naïve is the reference group.
@no-HCV is the reference group.

Lim et al.Journal of Medical Statistics and Informatics  2013 1:1DOI : 10.7243/2053-7662-1-1