Table 6.

Confidence intervals for the I² estimates of MAs.

	Author (Year)	Number of trials	I2	Low interval (95% CI)	High interval (95% CI)	Statistical model	Assessed outcomes
Heart failure
	Clark 2007 [27]	5	0	0	79	Random effects	All-cause mortality
	Clarke 2011 [28]
		10	51	0	76	Fixed effect	All-cause mortality
		6	59	0	83	Fixed effect	All-cause hospitalization
		6	0	52	75	Fixed effect	CHF-related hospitalization
		4	82	0	93	Fixed effect	All-cause emergency visits
	Inglis 2010 [31]
		11	0	0	60	Fixed effect	All-cause mortality
		8	0	56	68	Fixed effect	All-cause mortality follow-up period >6 months
		8	78	70	89	Fixed effect	All-cause hospitalization
		6	85	0	93	Fixed effect	All-cause hospitalization follow-up period >6 months
		4	39	0	79	Fixed effect	CHF-related hospitalization
		4	39	0	79	Fixed effect	CHF-related hospitalization follow-up period >6 months
	Polisena 2010 [34]
		6	0	0	75	Random effects	All-cause mortality
		4	5	0	85	Random effects	All-cause hospitalization
Hypertension
	Omboni 2011 [38]
		11	65.8	15	82	Random effects	Systolic blood pressure changes
		11	56.6	44	78	Random effects	Diastolic blood pressure changes
		6	77.9	50	91	Random effects	Blood pressure control
		5	79.1	50	91	Random effects	Number of antihypertensive drugs
Diabetes
	Farmer 2005 [44]	9	0	20	65	Fixed effect	Glycemic control - Changes in HbA1c
	MAS 2009 [46]
		7	65	0	84	Random effects	Glycemic control - Changes in HbA1c (All studies)
		4	45	0	82	Random effects	Glycemic control - Changes in HbA1c (subgroup analysis)
	Montori 2004 [47]
		8	33.8	0	71	Random effects	Glycemic control - Changes in HbA1c
		7	0			Random effects	Glycemic control - Changes in HbA1c (post-hoc subgroup analysis)