Table 2.

Methodological features of included studies

Study		Inclusion and exclusion criteria	Outcome		Sequence generation	Allocation concealment	Blinding	Sample		Attrition, %	Funding
First author	Location		1ry	2ry				calc.	size
Fairall et al., 2012 [25]	ZA 2	✓	✓	✓	A	A	d	✓e	≥200	<20	G
Fairall et al., 2012 [25]	ZA 1	✓	✓	✓	A	A	d	✓e	≥200	≥20	G
Houweling et al., 2011 [27]	NL 3	✓	✓	✓	I	A	NP	✓	≥200	<20	G
Andryukhin et al., 2011 [18]	RU 1	✓			U	I	c	✓e	<200	≥20	None
Dierick-van Daele et al., 2009 [24]	NL 2	✓			A	A	NP		≥200	≥20	G
Chan et al., 2009 [22]	UK 6	✓			A	A	b	✓	<200	<20	nr
Hesselink et al., 2004 [26]	NL 1	a	✓	✓	U	U	b	✓	≥200	≥20	nr
Denver et al., 2003 [23]	UK 5	a	✓	✓	I	I	NP	✓e	<200	<20	nr
Kernick et al., 2000 [28]	UK 4	✓	✓		A	U	U	✓e	<200	≥20	Ind.
Kinnersley et al., 2000 [34]	UK 3	✓	✓	✓	A	A	NP	✓e	≥200	≥20	G
Shum et al., 2000 [32]	UK 2	✓			A	A	NP	✓e	≥200	≥20	G
Campbell et al., 1998 [19–21, 29–31, 33]	UK 1	✓			A	I	b	✓	≥200	≥20	G

Studies are listed by year (y) of publication, in decreasing order. A tick indicates the specific criteria fulfilled

Blinding: whether patients, care providers and/or outcome assessors were blinded; UK: United Kingdom; NL: the Netherlands; ZA: South Africa; RU: Russia. I: inadequate; A: adequate; U: unclear; NP: not performed; G: government; Ind.: industry; P: private; nr: not reported

^aInclusion criteria only

^bBlinding of outcome assessors

^cSingle blinding

^dData analysts partly blinded

^eIntention to treat strategies to deal with missing data