Since then, I have stopped giving patients specific predictions about their life expectancies. I recognize that patients need to know their prognoses to make treatment decisions and plan their affairs. However, I have found that relatively nonspecific prognoses are sufficient. I might say, “I cannot predict the future, but in my experience, patients with your illness typically live a matter of months, not years,” or “Many people in your condition will live for only a matter of weeks, but some live significantly longer. I do not know what your fate will be.” In these conversations, I discuss concrete treatment goals with patients. I do not hesitate to say when I think the goal should shift from cure to palliation. When things are grim, I suggest that it is time to visit with friends and family because “it is better to be safe than sorry.” I give enough prognostic information to help patients make decisions, but I avoid using numerical wording that suggests I have a prognostic crystal ball.