Fig. 2

Mortality difference and individual hospital learning effect between hospital groups according to trends in procedures volume from 2010 to 2014. CABG coronary artery bypass grafting, AAA abdominal aortic aneurysm, PCI percutaneous coronary intervention. a Comparison of patient mortality across hospitals by comparing hospitals with increasing or decreasing volume with hospitals with unchanged volume. Hazard-ratios estimated from Cox model with adjustment regarding patient (age, gender, Elixhauser list of comorbidities, type and year of procedure, transfer, emergency admission, and median income) and hospital characteristics (hospital status, volume of procedures, specialization degree, and attraction rate). b Analyse to determine if mortality improved or deteriorated over time within hospital that increased or decreased its volume. The ratio of hazard ratio (RHR) compared the change in the mortality rate between two groups. A RHR greater than 1 suggests that the increase of mortality over time was greater in hospitals experiencing volume increase/decrease than in hospitals with unchanged volume