Abstract
The authors say public reporting of even the most well-defined end points—survival after treatment—can be misleading when used to compare outcomes across providers, and these data can suffer from patient-selection and treatment-intensity biases.