Twelve observational studies (n=356,102 patients, range 807 to 233,247) were included in the review. Five studies were based on administrative data, six were based on clinical data collected in cancer registries, and one used both types of data.
Two studies assessed in hospital mortality after breast cancer surgery. Mortality was low (0.1 to 0.2%), but was significantly lower in high volume hospitals (OR 0.40, 95% CI 0.22 to 0.74). There was little heterogeneity (I2=12%).
Eight studies examined the effects of hospital volume on survival. Six studies reported results as hazard ratios and reported a significant beneficial effect in favour of high volume hospitals (HR 0.84, 95% CI 0.76 to 0.93), but there was substantial heterogeneity (I2=76%). There was no significant difference between high and low volume hospitals based on the two studies that reported results as relative risks, but there was substantial heterogeneity (I2=96%).
Seven studies assessed the effects of surgeon volume on survival. There was a significant beneficial effect in favour of high volume surgeons based on both hazard ratios (HR 0.82, 95% CI 0.72 to 0.93; four studies) and relative risks (HR 0.85, 95% CI 0.60 to 0.91; three studies). There was evidence of heterogeneity among studies reporting hazard ratios (I2=59%), but studies reporting relative risks were homogeneous.
Sensitivity analysis showed a larger effect size for studies that adjusted for differences in comorbidity for both the studies of hospital (p=0.003) and surgeon volume (p=0.007).
There was some evidence of publication bias for hospital volume (p<0.03) but not for surgeon volume (p=0.41).