Hospitals, health insurers and patients often rely on patient death rates in hospitals to compare hospital quality. Now a new study questions the accuracy of that widely used approach and supports measuring patient deaths over a period of 30 days from admission even after they have left the hospital.