The Italian National Outcome Programme has assessed the performance of Italian hospitals regarding several clinical performance indicators, including 30-daymortality after admission for acute myocardial infarction. Risk adjustment was obtained using demographic and comorbidity data based on the hospital discharge databases in the index admission, as well as in those of the previous two years. Noticeably, the ICD-9-CM 410.7* classification coding for NSTEMI (Non-ST elevation myocardial infarction)myocardial infarction, i.e. the less severe form, was not used, due to known variability in its use. We found that hospital-specific adjusted relative risk of death versus the national mean, as computed by the programme, is negatively associated with the proportion of NSTEMI infarctions at each Tuscan and Florentine hospital, coherently with the hypothesis of a selection by the emergency network, which addresses STEMI (ST elevation myocardial infarction) patients to hospitals offering haemodynamic laboratory with reperfusive services. Individual level clinical data of 3,200 patients in the AMI-Florence study in the period April 2008-March 2010 found that ICD-9-CM410.7* is underused. The analysis based on hospital discharge diagnoses (410.7* vs. other 410* codes) cannot explain differences in mortality among Florentine hospitals, as opposed to the use of a classification of myocardial infarction type (STEMI vs. NSTEMI) based on clinical data collected in AMI-Florence.