Cardiac resynchronisation (biventricular pacing) for heart failure

Uses a structured format?Background:YesObjectives:YesSearch strategy:YesSelection criteria:YesData collection and analysis: YesMain results:Pooled results reportedReviewers’ conclusions:Reports results IntroductionDetailed introduction that includes description of health problem, current treatment options and a brief overview of CRT MethodsSearching:Number of electronic databases searched (search strategies provided) plus FDAwebsite and handsearching of major cardiac journal and annual conference abstracts.Unpublished data sought from authors. Searches up to June 2002Selection:Inclusion and exclusion criteria listed. Note scope of review beyond CRT-P and CRT-D (included ICDs as comparator). Included RCTs onlyValidity assessment:Reported internal validity quality by categoryData abstraction:Undertaken using standard forms by a two unblinded reviewers independentlyStudy characteristics:Tables of study characteristics – patients, intervention and comparatorQuantitative data synthesis: Binary outcomes summarised as odds ratios and pooled using random effect meta-analysis (sensitivity analysis for fixed effects). Statistical heterogeneity assessed ResultsTrial flow:Flow diagram for study inclusion includedStudy characteristics:Table of study characteristicsQuantitative data synthesis: Meta-analysis for all outcomes reported together with corresponding Forest plots.Sensitivity analyses undertaken to assess effect of CRT of ICD vs no trials trials andNYHA II–IV vs III–IV and trial selection DiscussionThe discussion summarises key findings, and discussed in the context of the potential limitations of the systematicreview/meta-analysis. Future research agenda are suggested Health Technology Assessment 2007; Vol. 11: No. 47