Longitudinal Joint Modelling of Ordinal and Overdispersed Count Outcomes: A Bridge Distribution for the Ordinal Random Intercept

Table 2

The results of longitudinal joint modeling of eGFR and the frequency of acute kidney transplant rejection.

Variable	Estimate	Standard error	value	95% CI		OR
Ordinal process (estimated glomerular filtration rate)
Age	0.047	0.010	<0.001	0.028	0.066	1.05
Sex (female)	0.884	0.278	0.002	0.335	1.434	2.42
Diabetes (no)	-0.195	0.288	0.499	-0.765	0.375	0.82
ATN (no)	0.346	0.696	0.620	-1.031	1.723	1.41
DGF (no)	2.140	1.231	0.084	-0.295	4.575	8.50
CAN (no)	-0.814	0.248	0.001	-1.305	-0.324	0.44
Time	-0.005	0.030	0.879	-0.064	0.055	0.99
Random effect standard deviation	0.322	0.041
Count process (acute kidney transplant rejection)
Sex (female)	-0.447	0.380	0.242	-1.200	0.306	0.640
ATG (no)	0.967	0.540	0.076	-0.102	2.035	2.630
Hypercalcemia (no)	0.964	0.536	0.075	-0.097	2.025	2.622
Proteinuria (no)	-0.641	0.393	0.105	-1.418	0.136	0.527
Hypertension (no)	-0.450	0.426	0.293	-1.294	0.393	0.638
Time	-0.046	0.054	0.389	-0.152	0.060	0.955
Random effect standard deviation	0.004	0.280
Negative binomial parameter	0.172	0.045	<0.001	0.083	0.261
Correlation between the random intercepts	0.765	0.101	<0.001	0.563	0.967

ATG: antithymocyte globulin; DGF: delayed graft function; ATN: acute tubular necrosis; CAN: chronic allograft necrosis; OR: odds ratio; CI: confidence interval.