Table 4.

Results of folate pseudo IPD

Model		Estimate	SE	95% CI	AIC REML	−2RES LogLik	#FE	#RE	AIC ML	−2LogLik	${\tau }_1^2$
Fixed study‐specific intercepts and
fixed treatment effect text	${\sigma }_{ϵ,ij}^2$	−3.24	0.15	(−3.5 to −2.9)	31 767.9	31 643.9	32	62	31 851.5	31 663.5	‐
	${\sigma }_{ϵ,i.}^2$	−3.29	0.13	(−3.5 to −3.0)	31 963.9	31 901.9	32	31	32 053.3	31 927.3	‐
	${\sigma }_{ϵ,.j}^2$	−2.98	0.36	(−3.7 to −2.2)	35 181.2	35 177.2	32	2	35 307.1	35 239.1	‐
	${\sigma }_{ϵ,..}^2$	−3.13	0.38	(−3.8 to −2.4)	35 374.5	35 372.5	32	1	35 504.3	35 438.2	‐
Fixed study‐specific
intercepts and random
treatment effects	${\sigma }_{ϵ,ij}^2$	−3.87	0.63	(−5.1 to −2.6)	31 636.1	31 510.1	32	63	31 748.5	31 558.5	9.81
	${\sigma }_{ϵ,i.}^2$	−3.91	0.63	(−5.2 to −2.6)	31 810.5	31 746.5	32	32	31 920.8	31 792.8	9.97
	${\sigma }_{ϵ,.j}^2$	−3.68	0.63	(−4.9 to −2.4)	35 164.0	35 158.0	32	3	35 308.2	35 238.2	6.41
	${\sigma }_{ϵ,..}^2$	−3.71	0.62	(−4.9 to −2.4)	35 358.6	35 354.6	32	2	35 501.4	35 433.4	6.31
Random intercept and
random treatment effect	${\sigma }_{ϵ,ij}^2$	−3.98	0.64	(−5.2 to −2.6)	31 841.3	31 711.3	2	65	31 848.3	31 714.3	9.97
	${\sigma }_{ϵ,i.}^2$	−4.03	0.64	(−5.3 to −2.7)	32 015.9	31 947.9	2	34	32 022.9	31 950.9	10.22
	${\sigma }_{ϵ,.j}^2$	−3.65	0.63	(−4.9 to −2.3)	35 369.8	35 359.8	2	5	35 376.8	35 362.8	6.71
	${\sigma }_{ϵ,..}^2$	−3.67	0.63	(−4.9 to −2.3)	35 564.7	35 556.7	2	4	35 571.8	35 559.8	6.74
Aggregate RE DL		−3.80	0.48	(−4.7 to −2.8)	‐	‐	‐	‐	‐	‐	5.17
Aggregate RE REML		−3.88	0.63	(−5.1 to −2.6)	‐	‐	‐	‐	‐	‐	9.88
Aggregate RE REML
Hartung‐Knapp		−3.88	0.63	(−5.1 to −2.6)	‐	‐	‐	‐	‐	‐	9.88

Abbreviations: AIC, Akaike information criterion (smaller is better); CI, confidence interval; #FE, number of fixed effect parameters; LogLik, log likelihood; ML, maximum likelihood; #RE, number of (co)variance parameters; RES, restricted; SE, standard error; ${\sigma }_{ϵ,ij}^2$, study‐ and arm‐specific variances; ${\sigma }_{ϵ,i}^2$, study‐specific variances; ${\sigma }_{ϵ,j}^2$, two variance parameters, one for control and one for treatment; ${\sigma }_{ϵ,..}^2$, one overall variance; ${\tau }_1^2$, between‐study variance.