The neurointensive care physician is frequently called upon to lead in the determination of neurological prognosis and in the decision-making that follows. This is a complex task that must integrate objective clinical information and test results together with published evidence and professional guidelines or recommendations when available. A number of scoring systems have been developed on the basis of multivariable models that typically integrate clinical severity indicators as well as features extracted from neuroimaging and neurophysiological testing and, in some cases, serum biomarkers. Performance of these models, evaluated with indices of discrimination and calibration, is often insufficient to enable prediction at the individual patient level—perhaps due to underpowered samples, lack of external validation, and failure to consider treatments as predictive features. The aim of effectively delivering prognostic information may fail because of the lack of consistency and coordination in communication between treatment teams and families and the unqualified use of population-based evidence to formulate point estimates of individual outcome. To help overcome this barrier, a model of “shared decision-making” has been proposed in which providers, patients, and surrogate decision-makers work collaboratively to make medical decisions that consider the best scientific evidence while integrating the patient’s values, goals, and preferences.