Risk models to predict outcomes in cardiac surgery are increasingly being used to identify patients who might be better treated using percutaneous techniques than by surgical intervention. In this Review, Thalji and colleagues critically evaluate how the databases from which these models are derived might influence risk prediction in clinical practice, particularly for selecting patients to undergo transcatheter aortic valve implantation.
- Nassir M. Thalji
- Rakesh M. Suri
- Hartzell V. Schaff
