Alex Tabarrok writes,

I consider a 1% chance of death to be very risky, perhaps worthwhile for some morbidly obese people but when 1 in every 100 patients doesn’t make it off the table that is not good odds.

What I find most interesting, however, is that I don’t think that any drug, even one with net benefits, could pass FDA trials with a mortality risk of 1%.

He is referring to bariatric surgery for obesity.

I think this is a very important point. We accept high risk when we go to a hospital at all. A paper from the Centers for Disease Control mentions (on p. 7)

Hospital-acquired infections from all causes are estimated to cause >90,000 deaths per year in the United States and are the sixth leading cause of death nationally.

The public wants to think in black and white terms. Something is “safe” (going to a hospital) or “unsafe” (taking Vioxx). The fact that in statistical terms the hospital visit is much more dangerous than the drug is something that is beyond people’s capability to grasp, because they cannot think in terms of the gray area of probability.

UPDATE: lots of good pushback in the comments.