Six RCTs with 350 patients (181 and 169 in control and experimental groups, respectively).
Relapse rates. Analysis was carried out in 2 ways: the first was with data from the original studies as reported, whilst the second was an effectiveness analysis that included analysis on a basis of intention to treat. In this analysis, drop-outs were handled in what would be the worst case scenario, i.e. those in the treatment group were assumed to have had a relapse and those in the control group were assumed to be free from relapse. The results of the effectiveness analysis are listed here.
6 months: odds ratio (OR) 0.65 (95% confidence interval, CI: 0.21, 1.51, p<0.27).
9 months: OR 0.59 (95% CI: 0.35, 1.00, p<0.07).
Number-needed-to-treat: 2 to 5.
24 months: OR 0.80 (95% CI: 0.47, 1.42, p<0.44).
Expressed emotion status (3 studies): OR 0.38 (95% CI: 0.18, 0.87, p<0.06).
Drug compliance (3 studies): OR 0.35 (95% CI: 0.16, 0.73, p<0.44).
Hospital admission (2 studies): OR 0.43 (95% CI: 0.16, 1.12, p<0.04).