Abstract
BACKGROUND: There is evidence that psychological therapies including cognitive behaviour therapy (CBT) may be effective in reducing postnatal depression (PND) when offered to individuals. In clinical practice, this is also implemented in a group therapy format, which, although not recommended in guidelines, is seen as a cost-effective alternative. To consider the extent to which group methods can be seen as evidence-based, we systematically review and synthesise the evidence for the efficacy of group CBT compared to currently used packages of care for women with PND, and we discuss further factors which may contribute to clinician confidence in implementing an intervention. METHODS: Seventeen electronic databases were searched. All full papers were read by two reviewers and a third reviewer was consulted in the event of a disagreement on inclusion. Selected studies were quality assessed, using the Cochrane Risk of Bias Tool, were data extracted by two reviewers using a standardised data extraction form and statistically synthesised where appropriate using the fixed-effect inverse-variance method. RESULTS: Seven studies met the inclusion criteria. Meta-analyses showed group CBT to be effective in reducing depression compared to routine primary care, usual care or waiting list groups. A pooled effect size of d = 0.57 (95% CI 0.34 to 0.80, p < 0.001) was observed at 10-13 weeks post-randomisation, reducing to d = 0.28 (95% CI 0.03 to 0.53, p = 0.025) at 6 months. The non-randomised comparisons against waiting list controls at 10-13 weeks was associated with a larger effect size of d = 0.94 (95% CI 0.42 to 1.47, p < 0.001). However due to the limitations of the available data, such as ill-specified definitions of the CBT component of the group programmes, these results should be interpreted with caution. CONCLUSIONS: Although the evidence available is limited, group CBT was shown to be effective. We argue, therefore, that there is sufficient evidence to implement group CBT, conditional upon routinely collected outcomes being benchmarked against those obtained in trials of individual CBT, and with other important factors such as patient preference, clinical experience, and information from the local context taken into account when making the treatment decision.