Simultaneous relationships between procedure volume and mortality: do they bias studies of mortality at specialty hospitals?

Specialty hospitals have lower mortality rates for cardiac revascularization than general hospitals, but previous studies have found that this advantage disappears after adjusting for patient characteristics and hospital procedural volume. Questions have been raised about whether simultaneous relationships between volume and mortality might have biased these analyses. We use two-stage least squares with Hospital Quality Alliance scores and estimated market size as instruments for mortality and volume to control for possible simultaneity. After this adjustment, it is still the case that specialty hospitals do not have an advantage over general hospitals in mortality rates after cardiac revascularization. We find evidence of simultaneity in the relationship between volume and mortality.