Susan read yesterday's post, and informed me that in her conversations at the dinner in question there was discussion of one of our friends' sisters having a recent colonoscopy—in which the intestine was inadvertently punctured, with a nasty infection ensuing. (The victim, uh, patient, did live—I guess that's something.) Could it be that the odds of a screwed-up colonoscopy are higher than the odds of detecting a problem relatively early enough to justify the risk? I don't know the answer in this instance, but I do know that in any number of situations "Stay the f#^* away from the hospital" is the statistically correct choice.