Table 3.

Effect sizes in meta-analysis of studies on GAD in BD

		k	n	Prevalence	95% CI	Q	p Value	I2 (%)	95%CI
Point studies	FE model	12	2975	9.6%	8.5% to 10.7%
	RE model	12	2975	11.5%	6.6% to 17.4%	182.1	<0.001	94.0%	91.2% to 95.9%
	RE model without outliers	11	2467	12.9%	8.2% to 18.6%	84.1	<0.001	88.1%	80.7% to 92.7%
	RE in subgroup analysis 1	11	2785	11.6%	6.2% to 18.3%	183.4	<0.001	94.5%	92.0% to 96.3%
	RE in subgroup analysis 2	12	2212	11.1%	6.2% to 17.1%	170.1	<0.001	93.5%	90.5% to 95.6%
Subgroup analysis 1: without Simon et al12 and Taman and Ozpoyraz40 Subgroup analysis 2: with BD-I subsample in Simon et al12 and without BD-II subsample in Simon et al,12 BD-II subsample in Chang et al (2012) and Otto et al (2006)
Lifetime studies	FE model	18	4919	15.8%	14.8% to 16.9%
	RE model	18	4919	15.1%	9.7% to 21.5%	318.3	<0.001	94.7%	92.8% to 96.0%
	RE in subgroup analysis 1	8	3411	20.1%	12.7% to 28.7%	48.8	<0.001	85.7%	73.7% to 92.2%
	RE in subgroup analysis 2	13	1616	12.5%	6.4% to 20.1%	122.3	<0.001	90.2%	85.1% to 93.6%
	RE in subgroup analysis 3	11	1603	11.2%	5.0% to 19.4%	110.8	<0.001	91.0%	85.9% to 94.2%
k=number of included studies n=number of patients in the included studies Subgroup analysis 1: with BD-I samples only Subgroup analysis 2: with BD-II and mixed samples Subgroup analysis 3: with mixed samples only

BD, bipolar disorder; FE, fixed-effects model; GAD, generalised anxiety disorders;

RE, random-effects model with Bayesian estimator.