"Doc, how much longer you think I've got?"

One of the things I've been most struck by since I graduated fellowship has been the need for patients to hear their prognosis. Some beat around the bush, asking vague questions like, "So, doc, how'm I doing?" Some come out and say, "Level with me, doc, how much longer have I got to live?" It's a constant, even among people who have remarkably different life philosophies.
What's even more amazing is how poorly I can answer that question. I usually start by saying "I don't know, only God knows that." I really believe that. People usually think I'm hiding the truth from them when I say that, and it can undermine trust.
What amazes me is how poorly prepared people are when they start to confront their mortality.
I think I understand where this question comes from. It comes from a limited vocabulary for discussing mortality. People don't generally sit around thinking about the grim reaper, so when faced with a mortal challenge, they fall back on that constant language, TV and movies. They've seen these medical dramas, and recall fictional conversations between benevolent sandy-maned physicians and their helpless patients who die in the next reel.
Here's a thought I had today: if I tell you you'll be dead in six months, I'll be really embarrassed when you return on the seventh month. How does one recover from a wrong prediction of prognosis? We all have heard the old warhorse, "Dr. Blackcloud told me I only had two weeks to live, and that was eight years ago!"
So what are people really asking when they are asking "How much longer, doc?" I take encouragement from the article in the oncology newsletter that addressed the problem of patients saying they want to die. They concluded that these patients are really trying to bring up the subject of their own mortality and symptoms, and that's just the language some people use to do it. Parenthetically speaking, it's hard to operationalize that knowledge into effective communication.
Maybe it's just a manifestation of fear. People would think you were weird if you answered the question with, "Don't be afraid; I'm here to help!"
Maybe they want to hear some encouragement. It's hard to give hope without giving false hope. False hope creates unrealistic expectations, and can interfere with good decisionmaking. Statistically, the survival distribution has a definitive mean, but also has a very long tail. Someone could quantify that by comparing the mean to the median survival from time of diagnosis for different cancers--a long tail means a skewed mean vs. median. I try not to get too bogged down in statistics with patients--some patients don't handle that all that well.