Discriminative Accuracy of Physician and Nurse Predictions for Survival and Functional Outcomes 6 Months After an ICU Admission

In JAMA, Michael Detsky and colleagues, including Michael Harhay and Scott Halpern, examine the accuracy of intensive care unit (ICU) physician and nurses’ predictions about patients’ survival and functional outcomes. Clinician predictions about a patient's chance of survival have been shown to influence decision-making for critically ill patients, yet little is known regarding their accuracy. The study was conducted in five ICUs in three hospitals in Philadelphia, and enrolled patients who spent at least three days in the ICU and required mechanical ventilation, vasopressors, or both. Attending physicians and bedside nurses were asked to make predictions about a patient’s chances of in-hospital mortality as well as six-month functional outcomes, including mortality, return home, ability to independently go to the bathroom and walk up 10 stairs, and cognition measures. They rated their confidence in the prediction on a scale of one to five. The study finds that the clinicians’ accuracy in predicting six-month outcomes of the  patients varied depending on the outcome being predicted. Physicians most accurately predicted six-month mortality, while nurses most accurately predicted in-hospital mortality. Both groups least accurately predicted cognition measures. Accuracy was in line with the self-reported confidence measures. The authors find that in comparison to a predictive model including objective clinical variables, a model that also included physician and nurse predictions had significantly higher discriminative accuracy for in-hospital mortality, six-month mortality, and return to original residence.