The Placebo Solution

“I’m not suggesting that doctors start secretly prescribing sugar pills at their discretion,” writes Ezra Klein “or treating angina with a chest incision rather than an angioplasty.” But maybe we should take that option a bit more seriously. Here’s a chart Ezra borrowed from Peter Orszag:

One way of reading this chart is as evidence that “internal mammary artery ligation” doesn’t actually work very well at relieving chest pain. Indeed, I think that’s the conventional way of reading a placebo result. But it’s not really the right way. The ligation gets impressive results. Those results are just undermined by the fact that the placebo chest incision gets results that are almost as good. And since the incision is dramatically cheaper than the “real” surgery, you could do the incision and spend the money you saved vaccinating children and reap substantial public health benefits and long-term savings.

The only real problem here — and admittedly it’s a big one — is that it seems unethical to tell doctors to secretly give their patients a placebo. But does it really make sense to let a failure of language and imagination prevent us from making a switch that could be hugely beneficial? What if instead of something designed to be a placebo it turned out that there was some cheap to implement traditional Chinese medical practice that, like the incision, got results almost as good as the full-scale surgery at a fraction of the cost? Would it be so crazy to say that we ought to incorporate the traditional Chinese technique into the Western medical cannon? Would it really be so unscientific to let our medical practice by guided by the evidence about what works, rather than by our a priori notions about what kinds of treatment “really” work and which are just placebos? It seems to me that when we discover cost-effective, efficacious treatments we ought to be exploring ways to implement them.