Clinicians' predictions about their patients’ recovery prospects can be a valuable part of shared-decision making with patients and families. However, clinicians often hesitate to provide these predictions or prognoses given their lack of certainty. Moreover, the accuracy of these predictions had yet to be vigorously studied. We assessed ICU physicians’ and nurses’ binary predictions of in-hospital mortality for more than 300 critically ill patients across five University of Pennsylvania Health System intensive care units, as well as their prognostic judgments of patients' abilities to return to their homes, walk up 10 stairs, communicate clearly, and toilet independently six months after ICU admission. Clinicians also rated their confidence in their predictions using a 5-point Likert scale.
Otto Haas Charitable Trust
Physicians' and nurses' predictions of 6-month outcomes for critically ill patients varied, depending on what outcome was being predicted, whether physicians and nurses were confident and whether they were concordant in their predictions. Overall, predictions were most accurately predicted for 6-month survival (positive likelihood ratio (LR), 5.91 [95% CI 3.74-9.32], negative LR 0.41 [95% CI 0.33-0.52]) and toileting (positive LR, 6.00 [95% CI 3.18-11.30], negative LR 0.51 [95% CI 0.35-0.75]) . When physicians and nurses were confident, their predictive accuracy improved, and when they were both confident and agreed, there predictive accuracy was even better (positive LR, 40.35 [95% CI, 5.73-284.28], negative LR, 0.18 [95% CI, 0.06- 0.50] ) and toileting (positive LR, 15.75 [95% CI, 4.04-58.94]; negative LR, 0.11 [95% CI, 0.02-0.68]). Further work is being done to better understand how clinicians arrive at prognostic judgments and to understand what clinician factors make them more accurate predictors.