Robin Hanson posts some carefully conducted studies suggesting that certain cancer screenings don’t seem to have any effect on overall mortality. It’s not that they don’t work: The evidence says they catch cancers. They just don’t save lives.

I think Hanson goes a bit far in the conclusions he draws, for reasons that some of his readers articulate in the comments. But “a bit far” isn’t the same as “wrong.” So if you’re anything like me, take a moment to think about how much you don’t want to believe that lung-cancer screenings or breast-cancer screenings may not actually work. And then think about trying to convince yourself they don’t work when your doctor is strongly urging you to get screened.

That’s why, if you want to control health-care costs, you somehow need to convince, incentivize or otherwise conscript doctors into doing it for you. Robin Hanson can write as many blog posts as he wants, but as long as doctors are telling scared and uncertain patients that they need to get screened, they’re getting screened. The moment they stop telling patients to get screened, screenings will plummet. In health care, doctors are really the relevant decision-makers. And right now, they don’t have the evidence to make good decisions nor the incentives to make cost-effective decisions.