METHODS

Data sources:

previously published systematic reviews and meta-analyses; Medline, EMBASE/Excerpta Medica, CINAHL, and SciSearch (to March 2003); the Cochrane Library; specialised rehabilitation databases; health technology assessment web sites; clinical trial registries; bibliographies of selected articles; and contact with experts and agencies.

Study selection and assessment:

randomised controlled trials (RCTs) comparing exercise based cardiac rehabilitation (alone or combined with psychosocial or educational interventions) with usual care that did not include any form of structured exercise training or advice in patients with CHD and had ⩾6 month follow up. Study quality was assessed in terms of the method of randomisation, allocation concealment, blinding of outcome assessment, and loss to follow up.

Outcomes:

all cause mortality, cardiac mortality, non-fatal myocardial infarction (MI), revascularisation, change from baseline in modifiable cardiac risk …