When doctors persist in useless treatments: a few of the promises and pitfalls of evidence-based medicine
For many centuries medical practice was a black art. What physicians did was based upon theories of how the body worked that turned out to be fanciful at best, dangerous at worst. The late nineteenth century brought breakthroughs in the biological sciences, such as the identification of bacteria and new understandings of physiology, which increasingly placed medical practice on a scientific basis. That process has continued over the past 150 years, but parts of medicine remain a kind of black art; some of what we do is still based upon tradition, theories never completely validated, and sometimes just intuition and guesswork. I wish that weren’t the case, but there it is. Unfortunately, this can mean subjecting a patient to dubious, even dangerous therapies for which we have only sketchy evidence of efficacy. Once established, such practices can be hard to change because physicians, like everybody else, become attached to pet ways of doing things. The recent movement toward what is generally termed evidence-based medicine is an attempt to change this. Non-physicians are typically surprised, even shocked, to learn that much of what we do is not very evidence-based. The situation becomes even more interesting, if that is the right word, when we in fact have evidence that something doesn’t work but we keep doing it anyway.
A recent fascinating essay in The Atlantic, entitled “When evidence says no, but doctors say yes,” provides a good discussion of what can be at stake in this issue. In this example the financial subtext also becomes text because several of the therapies in question makes a lot of money for the doctor. So there may well be other motives here. The therapy of stenting for coronary artery disease, discussed in detail in the essay, is a good example of how common sense can be wrong.
When the coronary arteries, which supply blood flow to the heart muscle, get narrowed by disease the heart can be starved of oxygen and respond with pain. At worst the result can be a myocardial infarction, a heart attack. Forty years ago the only way we had to open up the blockages was to bypass them with a graft — major surgery. There are still times this is the best option. Subsequently cardiologists started doing other things to open up the vessels without surgery. This involved passing a thin device, a catheter, up through the vessel to the narrowing and stretching it open. Soon after came the notion of keeping the narrowed part open by placing a stent, a kind of wire expander, inside the offending vessel(s). That procedure makes intuitive sense to anybody with a rudimentary knowledge of plumbing; if the pipe is narrowed, open up the pipe and then prop it open. The procedure benefits some patients in some situations, but not most of them, especially not patients who are otherwise stable. It turns out to be better to use medicines to both eliminate the pain and reduce the chances of a heart attack. In the words of one expert, “Nobody that’s not having a heart attack needs a stent.” Importantly, putting a piece of expandable metal in the heart is not risk-free; it can cause serious complications, even death. So here we have a therapy that usually doesn’t help and can kill you. But many, even most invasive cardiologists keep doing it. The essay examines why this might be. It also gives several other examples besides coronary artery stents where research does not support continued use of a drug or a procedure.
There are a couple of issues in play — one specific, one general. The specific one is that here we have a procedure widely done apparently for reasons that solid research has shown are not valid. Why? The essay’s author concludes physicians are driven very much by social rather than scientific reasons. The highly regarded Dartmouth Atlas of Health Care has been studying for years how physician culture, not science, substantially determines what they do. The more general issue is that it may surprise you to read we have solid proof of benefit for a disturbingly small list of things doctors do. I think the main reason for that is we do many, many things in medicine and conducting a controlled trial on all of them, even most of them, is impossible. Evidence-based medicine is fine for those things for which we have evidence. But for those things for which we only have intuition and sometimes guesswork it is often best to remember the famous formulation of Loeb’s laws. Many times it is best to go by this dictum, when tempted to forge ahead into the mist: “Don’t just do something, stand there”!
At any rate, the Atlantic article is well worth a careful read.