THE RETURN OF EVIDENCE-BASED MEDICINE. It's easy to forget how much of American medicine is a guessing game, how your treatments are a composite result of your doctor's experiences, biases, treasured anecdotes, and personal reactions to his own training. Most folks think medicine operates off a rigidly defined set of standards: If you have symptom A, your doc orders tests B, C, and D. Not quite. According to a new study, doctors are ordering useless tests for asymptomatic patients at staggering rates. Of tests that aren't recommended for patients with a particular batch of complaints, we're spending between $12 million and $63 million. Worse yet, for tests with risks that outweigh the benefits for certain patients, doctors are ordering them against recommendations over 40 percent of the time, for a total cost reaching into the hundreds of millions. And that's not even getting into the ricochet tests and expenses that come from false positives found by unnecessary diagnostics.

The problem, basically, is that medicine often doesn't know it's doing. Technology is moving too fast, and we have too many treatments, for the profession to keep up. Add in that most doctors aren't spending their spare time reading journals, we're defunding medical research, we do very few comparative studies on new treatments, and one-third of journal studies are overturned anyway, and you've got a problem -- medicine becomes guesswork, and massively expensive guesswork at that.

A few examples: Remember when everyone used to get their tonsils out? Now we know that does kids little good. Back surgery? Apparently no more effective than drugs and physical therapy. Ulcers? They're bacteria, not stress. Angioplasties? We're hair trigger with these things, as we are with bypass surgeries, to the tune of hundreds of thousands unnecessary procedures a year. Prostate cancer? We don't actually know how to treat it, or at least how to choose between treatments for it. Masectomies? Turns out they're generally no more effective than the far less intrusive lumpectomies. Our culture, both medical and general, has a preference for the most aggressive treatments, but the data often doesn't bear that intuition out. This, by the way, is the world conservatives want patients navigating their own way through -- one that lacks enough digestible evidence for even the most highly trained, well-paid professionals to effectively traverse.

But we do need change, as we're wasting billions and billions on unnecessary, and even counterproductive, treatments. What can we do? A couple things: What's needed here is a long-time buzzword, "evidence-based medicine." We need proper data, filtered through standardizing committees, released in a comprehensible fashion. We need increased funding for NIH research, and more support for computer models that carry out virtual studies. We need laws regulating the nexus between Big Pharma, insurers, doctors, hospitals, and medical technology producers. We need a payment system that rewards outcomes rather than treatments. And we need a Congress that's not constantly distracted by BS issues like tort reform and insurance regulation. In 1994, Bill Clinton folded much of this into his health reform bill. But evidence-based medicine can be easily decoupled from insurance reform, and it should. Unfortunately, there are few points to be scored or donations to be had in such a technocratic, important exercise.

Update: Changed the line that said doctors didn't know they were doing. It was hastily written, and not what I meant to convey. The point isn't ignorance on their part, but a lack of digestible research, agreed-upon standards, and compensation practices that reward outcomes rather than treatments.

--Ezra Klein