Table 2.

 Summary of quality assessment of included randomised controlled trials

Study	Was assignment of treatment really random?	Was allocation concealed and concealment method described?	Were groups similar at baseline?	Were eligibility criteria specified?	Who was blinded to treatment allocation?	Was intention to treat analysis used and were drop outs accounted for?
Batraw1	Yes; computer generated list	Yes; sealed envelopes	Yes*	Yes	Participants, therapists, and outcome assessors	Yes, yes
Bolligerw2	Yes; computer generated list	Yes; sealed envelopes	Yes*	Yes	Participants, therapists, and outcome assessors	Yes, yes
Hausteinw3	Yes; computer generated list	Yes; sealed envelopes	Yes	Yes	Participants, therapists, and outcome assessors	Yes, yes
Rennardw4	Likely, but method not described	Likely, but method not reported	Yes	Yes	Participants, therapists, and outcome assessors	Yes, yes
Wennikew5	Yes (stratified by Fagerström score); computer generated list	Yes; sealed code list	Yes*	Yes	Participants, therapists, and outcome assessors	Yes, yes
Wood-Bakerw6	Yes (stratified by Fagerström score); computer generated list	Yes; sealed envelopes	Yes	Yes	Participants, therapists, and outcome assessors	Yes, yes
Etterw7	Yes; computer generated list	Unclear	Yes	Yes	Participants and outcome assessors	Most outcomes†, yes

When extensive unpublished study reports were available, they were used for quality analysis.

*Except for small imbalance in sex distribution.

†Not intention to treat for product usage and for completeness of blinding of participants (determined at six months).