The authors reply
We agree with Torbey et al (1) that withdrawal of life support without adequate time to assess prognosis can lead to a “self-fulfilling prophecy” and that high-quality care—whether a palliative care consultation is involved or not—requires adequate time before predictions about neurologic prognosis can be made. Important exceptions to this rule exist, for example, when a patient with chronic life-limiting illness has a baseline quality of life prior to the acute brain injury that is already below what that patient considers acceptable. We increasingly see this latter group of patients among elderly patients with advanced chronic illness or multimorbidity.
We also agree that ICU length of stay is only one outcome for family-centered care, but it was one that was ranked relatively highly by our guidelines panel and by the patients and family members that we recruited to review outcomes (6, 8). In the studies cited in our guidelines, ICU length of stay was used primarily as an outcome among patients who die in the ICU, as a marker of the prolongation of dying. However, it is also important to examine whether reducing ICU length of stay is associated with higher mortality, which would be an important unintended consequence of any type of palliative care or family-centered intervention.
Finally, we agree that the wording of our recommendation 4.1 regarding palliative care consultation could have included explicit acknowledgment of the importance of providing adequate time for prognostic assessment before implementation of withdrawal of life support (8). This is true for all critical illness, and especially important in severe acute brain injury.