A wide range of missing imputation approaches in longitudinal data: a simulation study and real data analysis

Table 6 Results of random effects expectation–maximization (REEM) tree algorithm for systolic blood pressure (SBP)

Imputation method	MSE	RMSE	MAD	Deviance
Complete cases	0.854	0.924	0.689	54241.55
Interpolation LOCF	0.849	0.922	0.666	174247.8
Interpolation global	0.854	0.924	0.639	181020.8
Interpolation local	0.859	0.927	0.624	185848
Interpolation bisector	0.854	0.924	0.636	181482.6
copyMean.LOCF	0.849	0.922	0.667	174344.9
copyMean.global	0.854	0.924	0.641	181116.8
copyMean.local	0.859	0.927	0.624	186016.9
copyMean.bisector	0.855	0.924	0.638	181544.8
LOCF	0.850	0.922	0.672	174940.5
NOCB	0.850	0.922	0.671	175030.9
Traj mean	0.847	0.920	0.636	171740.2
Traj median	0.847	0.920	0.639	172071.6
Traj hot deck	0.850	0.922	0.677	175267.3
Cross mean	0.866	0.930	0.680	175619.2
Cross median	0.866	0.931	0.678	175871.9
Cross hot deck	0.881	0.939	0.692	185584.4
FCS-LMM	0.874	0.935	0.693	181290.1
FCS-LMM-het	0.874	0.935	0.693	181290.1
FCS-GLMM	0.854	0.924	0.70	176799.5
FCS-LMM-LN	0.854	0.924	0.692	176975
FCS-LMM-LN-het	0.854	0.924	0.701	177478.9
JointAI	0.854	0.924	0.70	176904.3
hmi	0.854	0.924	0.70	176767
JM-SMC	0.854	0.924	0.70	176812
JM-SMC-het	0.854	0.924	0.70	176818.3
JM-MLMM	0.854	0.924	0.699	176732.2
JM-FJ	0.854	0.924	0.699	176758.1

ISSN: 1471-2288