Fifteen papers were included in the review; two studies included percutaneous coronary intervention (PCI) or coronary artery bypass grafting (CABG) data. One study provided seven of the 13 effect sizes for the CABG analysis and one study provided three effect sizes for the PCI analysis.
PCI (ten studies, n=1,322,342 patients in 1,746 hospitals): Mortality was lower in hospitals with a high volume of procedures compared with those with a low volume of procedures (OR 0.87, 95% CI 0.83 to 0.91). Moderate heterogeneity was found (I2=38%). Results were similar and heterogeneity was reduced by excluding one study, suggesting an increased mortality in high-volume hospitals (I2=20%). Results were not influenced by source of data used for case mix or completeness of case mix. Meta-regression showed no interaction with the cut-off point. There was no evidence of publication bias.
CABG (seven studies, n=1,470,990 patients in 2,040 hospitals): Mortality was lower in high-volume compared with low-volume hospitals (OR 0.85, 95% CI 0.79 to 0.92). Moderate heterogeneity was found (I2=47%). Results were similar and heterogeneity was reduced by excluding one study suggesting an increased mortality in high-volume hospitals (I2=33%). Sensitivity analyses did not change results. Other analyses showed no interaction with the cut-off point and no significant trend over time. The funnel plot was suggestive of publication bias, but Egger’s test was not significant.