The plight of the payer

Confirmation bias is the defining cognitive error of physicians.  Physicians want to help you. Therefore, by definition, whatever they do to you must be helping you. Plus, you are demanding, and if they actually do manage to say "well, actually, there is nothing we can really do about that", you get mad and search around for someone who will finally do something. Finally, they get paid more if they do something than if they don't. They will tell you that that has no effect on their decisions, but they are, of course, wrong.

Hence my essential healthcare cost equation:

Bias-to-action, income-requiring physicians + patients demanding that someone do something = excess spending on healthcare

(Please note:  I don't mean to imply that physicians are not actually performing useful services, or that patients are hypochondriacs.  I'm just pointing out a tendency)

And who has to do that excess spending? Health insurers. And by necessity...you, the premium payer.

So if insurers discover that some medical procedure really, genuinely doesn't do anything useful, can they stop paying physicians to do it?

Consider the dubious practice of vertebroplasty, where someone injects cement into a vertebra to alleviate the pain of vertebral compression fractures. Back pain is awful and debilitating, and sufferers will go to almost any lengths to find something that works. Doctors who provide vertebroplasty are sure that it works.

An interesting article in Health Affairs, "Can Coverage Be Rescinded When Negative Trial Results Threaten A Popular Procedure? The Ongoing Saga Of Vertebroplasty" (HT: The Incidental Economist) covers this issue in detail. (The article was ungated when I looked at it, but seems to be gated now.  Pressure from vertebroplasty providers? In any case, unfortunate: it's a good article.)

There is no good clinical trial that shows that it works.  Insurers started covering the procedure here and there, and as a result are stuck:

 ...after insurers had covered vertebroplasty for so long, however, they feared that any attempt to change their “yes” to a “no” would be publicly perceived as cutting off access to beneficial care.

Because insurance covered the procedure, that very coverage had created a landscape of providers, manufacturers, and patients with a vested interest in continuing the status quo.

Because the procedure is covered, it is hard to recruit for a clinical trial by offering access to a new procedure. Two studies were done comparing the results of a vertebroplasty against a sham surgery, both indicating that the procedure, as you might expect, does nothing. There is a lot of "need to act" and money (and ego) involved now, so those who do these procedures are pushing back hard:

Because vertebroplasty is a technique requiring manual skill, the clinicians who perform the procedure often believe that the results of any related study will be driven by the training and expertise of the study practitioners.

It would be nice if "does the procedure do anything?" was the standard of coverage, but this is the real world here:

The vertebroplasty trials illuminated a divide between the highest-quality scientific evidence available and the dominant opinion in the clinical community about the benefits of the procedure. Payers reported that when making policy, they must strive to take both into account.

There is no good solution to this conundrum. I'm sure it happens in single-payer systems too. At the edges of provable benefits, patients become big babies and doctors turn into self-deluding egomaniacs. That's the rock-solid foundation under every healthcare system.