During the run-up to health-care reform, a lot of people went looking for what I dubbed "the magic pot of money." It would be a large expense that could be cut out of the health-care system with very little political pain because it would not cause anyone to miss out on life-saving (or life-enhancing) treatment, and it also did not come attached to a large and powerful interest group that would fight like deranged weasels to save their pet benefit.
There were many candidates for the role of "magic pot of money": Medicare Advantage overpayments, preventive medicine, uncompensated care provided to the uninsured, electronic medical records, fighting obesity, ending unnecessary treatment and so forth. One by one, most of these turned out to be less promising than they had initially seemed. Medicare Advantage "overpayments" turned out to be providing extra benefits to those enrolled, and so far the administration has chickened out whenever the time has actually come to make deep cuts. Uncompensated hospital care turns out to be a small fraction of U.S. health-care costs (about $40 billion on total spending of more than $2.7 trillion), and it also didn't fall as far as people had expected in Massachusetts. Preventive medicine turned out to raise costs, not cut them (whatever its other benefits for patients). No one knows how to make people thin. And so we were left with overtreatment.
Everyone agrees that we treat people too much. Unnecessary back surgery is almost always Exhibit A: Most studies show that you're better off with physical therapy. I was very fond of castigating back surgeons about this (though not to their faces). And then one day, after a panel where I mentioned the epidemic of unnecessary back surgery, a nice middle-age fellow came up to me. Awkwardly, he turned out to be a back surgeon.
"Do you think we don't read the studies?" he said. "We know physical therapy is better for you than most back surgery. What we don't know is how to make sure that the patients go through with physical therapy."
I was reminded of this moment when I read Lisa Rosenbaum's excellent article in this week's New Yorker, in which she chronicles her attempt to follow "best practices" with regard to arterial stents, whose overuse has recently become Exhibit B in the overtreatment oeuvre:
Heart attacks happen when a small plaque in the artery suddenly ruptures and a blood clot rushes to heal it. In many instances, one moment the artery is mostly open and the next moment it is mostly closed. In those cases, using a stent to open an artery before the heart muscle dies is life saving. Because the benefits of stents for patients having heart attacks are so clear, hospitals equipped to perform the procedure are penalized if a stent is not placed within ninety minutes of the patient's arrival in the emergency room. This is partly why rates of death associated with cardiovascular disease have fallen by about half in the past forty years.
But for stable patients with chronic heart disease, when blockages build up over time, the benefit of opening the artery with a stent is far less certain. These long-standing blockages tend to have hardened exteriors, which can shield the plaque from exposure to the clotting factors in the blood passing by it. Chronic blockages aren't benign -- they give many patients chest pain (often called angina) and also portend a much higher risk of future heart attacks. But when it comes to treating chronic disease, medications, such as statins, aspirin, and beta blockers, are often as effective as stents.
However, this has not stopped clinicians from using stents for patients with chronic heart disease, leading to widespread concern about their overuse. The stories about cardiologists behaving badly validate the conviction, common among both policymakers and the public, that misaligned financial incentives drive doctors to do things that they shouldn't. This narrative has taken root for a number of reasons, one of which is that, in addition to villains, it has heros -- health-care reformers who will swoop in and put an end to unnecessary care, simultaneously saving the economy and improving the health of the population.
A patient of Rosenbaum's, a grocer, came into the emergency room with heart pain; he was diagnosed with a heart attack, and a stent was put in to open up a blocked artery. All the guidelines indicated conservative management: medication and lifestyle changes -- and a single stent, not several stents, as many patients walked out with. So that's what she prescribed. But her patient, who didn't speak much English, didn't show up for follow-up visits to have his condition carefully monitored. When she saw him again, she realized that he'd stopped moving because of his chest pain, causing him to waste. Eventually, he visited another doctor closer to his home, who put in another stent -- and clearly made the patient feel better.
Consider the claim that medications are as good as stents for treatment of stable coronary artery disease. This is the idea upon which most accusations about the overuse of stents are predicated. Several studies support this assertion, but the seminal one is the COURAGE trial, which randomly assigned patients with stable blockages to either medication or a stent. The study found that those treated with medications lived just as long as those with stents. COURAGE is a super-star trial, the best of its kind. So why can't we say, once and for all, that it's inappropriate to use stents for patients with stable coronary disease?
The answer is that it's because such a statement is a colossal oversimplification. The fundamental challenge of translating data into practice is what we call generalizability: Can we extrapolate the findings from a trial to real life? If you are a doctor who is trying to practice evidence-based care, the first thing you want to ask yourself is, Would my patient have been enrolled in the trial? Sun Kim would not have been eligible for the COURAGE trial, which excluded all patients with high-risk features -- or nine out of ten otherwise eligible patients.
But let's say that your patient is among the ten per cent who would have been enrolled in COURAGE. Can't you say now, with certainty, that your patient should not get a stent? Still no. Real life rarely resembles clinical trials, which are, by definition, rarefied environments -- well-oiled machines with incredible depth of resources and a staff to orchestrate patient care.
If, as in the case of COURAGE, you set out to show that medications are as effective as stents in treating chronic disease, you want to make sure that the patients in the trial are actually taking those medications. In reality, the rate of adherence to medications is about fifty per cent, but COURAGE not only provided medications for free -- it also hired nurse managers who saw patients regularly and adjusted dosages. Not only did these patients adhere to medications at a far higher rate than patients usually do -- this adherence also translated to excellent blood pressure and cholesterol control.
If it's hard to apply the findings from any one trial to the treatment of a particular patient, it's harder still to use data from many trials to create guidelines that can be applied to any patient. When a group of expert cardiologists were asked to do just that, they recognized that there are many factors to be considered in addition to medication -- including the acuity of the disease, the patient's degree of chest pain, the results of stress tests, and which of four main arteries are blocked. When you account for all of these factors, you come up with over four thousand clinical scenarios for which stenting may or may not be appropriate, many of which can't be mapped precisely to a clinical trial.
A study just published in the New England Journal of Medicine showed that, contra the guidelines she was following, preventive stenting does seem to save lives and improve patient outcomes.
Is this to say that there is no way to save money in health care without hurting patients? Of course not. But we should be wary of asserting that it will be easy to save money without hurting patients as long as we simply follow the best evidence. Studies are, of course, the foundation of science. But they also have limits. The questions they can answer are often much narrower than the questions we would like to answer -- and often other studies give different answers. There's a decent amount of evidence that hospitals which do more intensive interventions get better outcomes.
Magic pots of money are very hard to find. Science can hopefully point the way ... but probably not without a lot of false starts.
This column does not necessarily reflect the opinion of Bloomberg View's editorial board or Bloomberg LP, its owners and investors.
To contact the author on this story:
Megan McArdle at firstname.lastname@example.org