The magical cure for all our health care woes is, we are told, pay-for-performance as dictated by comparative effectiveness research. That will ensure that only the very best is delivered by each physician to each patient. Right. Pity there is no empirical support for the notion.

In fact, most of the experience we've seen contradicts the premise. A recent article in The New York Times illustrates the problem. The article says, "It sounds like a simple idea for improving health care: draw up guidelines on how best to treat a particular illness and then pay doctors to follow them."

But alas, "setting guidelines that are good for every patient, it turns out, can get messy, with some experts warning that a big national plan of this sort poses risks. A recent case involving treatment for diabetes, one of the nation's most pervasive illnesses, illustrates the difficulties."

It seems that the National Committee for Quality Assurance "adopted [guidelines] in 2006 that called for aggressive control of blood sugar, or glucose." But, last year it abruptly pulled those guidelines "after a large federal study indicated that lowering glucose too quickly or too much in some patients could harm or even kill them."

The article says that many experts warned at the time, "that it was medically ill-advised for some patients," but the panel plowed ahead anyway possibly under pressure from drug companies.

But the bigger issue, according to the article, is "that many guidelines are based not on rigorous studies like clinical trials but on weaker types of medical evidence. And critics like Dr. Groopman have argued that the guideline-setting process is often influenced by industry or by medical ideologues looking to advance their personal agendas." Plus, "To many experts, the diabetes case shows how setting one guideline that works for all patients suffering from the same disease can be tricky."

The article concludes, "faulty guidelines can pose risks to patients, particularly when linked to doctors' pay, which is an idea under consideration in Washington."