Abstract
BACKGROUND: Care bundles are used widely to prevent surgical-site infections (SSIs). Recent systematic reviews suggested larger effects from bundles with more interventions. These reviews were largely based on uncontrolled before-after studies and did not consider their biases. The aim of this meta-analysis was to determine the effectiveness of care bundles to prevent SSIs and explore characteristics of effective care bundles. METHODS: A systematic review, reanalysis, and meta-analysis of available evidence were undertaken. RCTs, controlled before-after studies, and uncontrolled before-after studies with sufficient data for reanalysis as interrupted time series studies (ITS) were eligible. Studies investigating the use of a care bundle, with at least one intraoperative intervention, compared with standard care were included. RESULTS: Four RCTs, 1 controlled before-after study, and 13 ITS were included. Pooled data from RCTs were heterogeneous. Meta-analysis of ITS resulted in a level change of -1.16 (95 per cent c.i.-1.78 to -0.53), indicating a reduction in SSI. The effect was larger when the care bundle comprised a higher proportion of evidence-based interventions. Meta-regression analyses did not show statistically significant associations between effect estimates and number of interventions, number of evidence-based interventions, or proportion of evidence-based interventions. CONCLUSION: Meta-analysis of ITS indicated that perioperative care bundles prevent SSI. This effect is inconsistent across RCTs. Larger bundles were not associated with a larger effect, but the effect may be larger if the care bundle contains a high proportion of evidence-based interventions. No strong evidence for characteristics of effective care bundles was identified.