Introduction and objectives Guidelines recommend that COPD patients admitted to hospital with an exacerbation should be assessed and considered for supported self-management interventions although it is not known how effective or cost-effective such an intervention would be when instigated during admission or shortly after discharge. We conducted a systematic review and evidence synthesis to answer this question.

Methods Key databases eg MEDLINE, EMBASE, CENTRAL, were searched up to May 2012 for studies of any design where patients admitted with an acute exacerbation of COPD were included in a supported self-management intervention (or important components) within 6 weeks of discharge. Citation lists were checked and authors of relevant conference abstracts since 2010 were contacted. There were no language restrictions. Data were extracted and risk of bias assessed independently by 2 reviewers.

Results Of over 16000 initial search hits, 14 papers have been provisionally included which report 8 randomised controlled trials (RCTs), 1 controlled clinical trial and 4 pre-post studies/arms. Study quality was variable and interventions heterogeneous. Of the RCTs, 4 described multi-component self-management packages, 1 was a cluster RCT providing support to both nursing home staff and patients, 1 was a home-based exercise trial and 2 were integrated care/case management packages with significant self-management components. RCT follow-up ranged from 3–12 months with a total of 1113 (range 33–464) patients enrolled. Results from n = 4 RCTs indicate a reduction in re-admissions of borderline significance (OR 0.65 (95% CI 0.42, 1.00)) but no significant effect on mortality (OR 1.22 (95% CI 0.79, 1.86)). Effect on overall quality of life was heterogeneous with large loss-to-follow-up. There were no cost-effectiveness studies.

Conclusions There is a paucity of good quality large RCTs of supported self-management delivered at discharge. Interventions are disparate and few studies report significant benefits in important outcomes. However, effect sizes for reduction in admissions are consistent with published evidence of self-management interventions delivered whilst patients are stable.