After I challenged Doug Bremner for claiming that all angioplasty (by which he meant, primarily, coronary stenting) was useless and should be eliminated from future health care budgets he updated his post thusly:
[...Even so they are still estimated to be about 1/3 which is too many and some cardiologist lately have gone to jail for performing PCI on people with little or no heart disease. So my initial statement that 25 billion dollars could be saved is not correct. It is more like, um, 8 billion.]
So the question Dr. Bremner and I now have on the table is not whether to cut out all stent procedures, but whether one third could be slashed from the budget. We're getting closer but we need to examine a little more rigorously what the clinical trials said about that third. They are the folks who have stable coronary disease. Neither the BARI 2D nor the COURAGE trial showed a reduction in major events with stenting as compared to medical therapy. But what else was learned from the two studies? First, both trials addressed the lowest risk stable angina patients. If the patients were too high risk they simply were not included for study. Second, many medically managed patients in both groups crossed over to revascularization---around 30% in courage and around 40% in BARI 2D. In other words many patients failed medical therapy. That's not to say it was inferior to PCI, because many PCI patients had to have repeat revascularizations. It is to say that many stable CAD patients initially managed medically will develop legitimate indications for unblocking their arteries sooner or later. In fact, the authors of the very paper Dr. Bremner was referring to (the BARI 2D trial, or at least I think it was—Dr. Bremner didn't provide the citation) said this in their discussion section:
It is important to note that all the patients who were assigned to receive medical therapy underwent careful clinical monitoring, and 42.1% had changes in the clinical course that called for later revascularization during 5 years of follow-up. In clinical practice, the initial treatment strategy for a patient with diabetes and coronary disease rarely remains constant over a 5-year period.
In the real world these stable patients get stented for better angina relief, which study after study shows is a benefit of revascularization even if major events are not reduced. So when Dr. Bremner calls for a denial of PCI for these patients he can't base it on evidence. He has to make a value judgment about whether the better quality of life these patients would experience is worth it. Will Obama's “comparative effectiveness” panels make similar value judgments?
Now if Dr. Bremner will concede that some stable angina patients have a legitimate need for PCI for symptom relief not afforded by optimal medical therapy then we can reduce this discussion down to the real question: how many of those revascularizations are truly unnecessary? How many cardiologists yield to the oculostenotic reflex and stent lesions just because they're there and because they can, in stable minimally symptomatic patients? There's little doubt such non-evidence based stenting takes place, but we don't know how much, do we? Without such numbers how can Dr. Bremner begin to estimate they monies potentially saved by eliminating non-evidence based PCI?
Maybe Dr. Bremner's problem is that he uses the same EBM pyramid as Marcia Angell. He seems to have gotten his information from this article in Business Week. Lets hope the folks in the Obama administration who evaluate comparative effectiveness spend a little more time reading the NEJM.