Objective: Surrogate decision-makers (“surrogates”) and physicians of incapacitated patients have different views of prognosis and how it should be communicated, but this has not been investigated in neurocritically-ill patients. We examined communication preferences in surrogates and physician practices during the outcome prognostication for critically-ill traumatic brain injury (ciTBI) patients in neuroICUs.

Design: Qualitative study using in-person semi-structured interviews with surrogates of ciTBI patients and physicians with expertise in TBI.

Setting: Two neuroICUs at two level-1 trauma centers (surrogates); seven academic U.S. medical centers (physicians).

Subjects: Sixteen surrogates for 15 ciTBI patients and 20 attending physicians from neurocritical care, neurosurgery, trauma and palliative care.

Interventions: Not applicable.

Measurements and Main Results: We used qualitative content analysis and descriptive statistics of transcribed interviews to identify themes in surrogates and physicians. The majority of surrogates (82%) preferred numeric estimates describing the patient’s prognosis, as they felt it would limit prognostic uncertainty, which, in turn, surrogates perceived as frustrating. On the other hand, 75% of the physicians reported intentionally omitting numeric estimates during prognostication meetings due to low confidence in family members’ abilities to appropriately interpret probabilities, worry about creating false hope, and distrust in the accuracy and data quality of existing TBI outcome models. Physicians felt that TBI outcome models are for research only and should not be applied to individual patients. Surrogates valued compassion during prognostication discussions, and acceptance of their goals-of-care decision by clinicians. Physicians and surrogates agreed on avoiding false hope.

Conclusions: We identified fundamental differences in preferences for the communication of prognostic information between surrogates of ciTBI patients and physicians during goals-of-care discussions. A decision aid could potentially bridge this chasm by providing surrogates consistent and patient-centered information, however, with qualitative rather than quantitative estimates of ciTBI prognosis and an open disclosure of uncertainty.