An NPR story on the medical effects of prayer has elicited an interesting colloquy among bloggers of the Reality Based Community. Michael O'Hare (from whose post you can link back to the others) makes what for me is the key point:
There's an ocean of difference between a delicate, hard-to-observe, ephemeral effect that might lead to real scientific advance but will probably come to nothing, and treatments and practices that can be shown again and again to have big useful effects. It's cruel to describe the former in a news context so that it might be confused with the latter.
He also makes (in the same post) a nice point about the use and misuse of lies for life:
But doing this as policy is an intermediate Russian doll; every context outside it has to maintain the same lie; one can easily imagine this leading to a tangled mess of unaccountable mendacity "for our own good" that ensnares an army, for example, to lie about Pat Tillman. If it's OK for us to believe wrong things to maintain certain medical efficacy, is it OK for us to believe wrong things to maintain military effectiveness?
I'd just note that the argument applies beyond the narrow context of social policy. While "[t]he falseness of a judgment is for us not necessarily an objection to a judgment," chances are still high that upon any invitation to believe a falsehood, we should want to decline, because any method you want to cook up for incorporating "healthy" fictions into our lives threatens to degenerate into incontinent, debilitating forms of dishonesty.