Meta Department: Comparative Effectiveness Research

The Washington Post published an interesting article a few days ago asking why cardiologists preferred an invasive treatment regimen over a low tech intervention? This case is more interesting than many of the comparative effectiveness types, as it turns out the treatments they’re comparing—whether, in cases of patients with stable angina (i.e. chest pain in response to stress), to give patients an angioplasty with stenting or to opt for medication with lifestyle changes—are basically equivalent in effectiveness. It turns out the latter treatment is cheaper, which means that, over time, the health care system should opt increasingly for the medication/lifestyle change one-two punch.

As it turns out, that hasn’t really happened—well, as far as we can tell. To me, the real secret of the article isn’t just that it’s difficult to get doctors to switch to treatments better situated on the cost/effectiveness curve, but that we don’t really seem to know what’s being done in the first place. Here’s the evidence from the article:

It may be that fewer people are getting PCI [the more aggressive option] for stable angina now. That was the case in a registry of 26,000 patients in northern New England, where that diagnosis was given in 21 percent of procedures in the year before [the study] was published but in only 16 percent in the years just after it. Whether that reflects a national trend also isn’t known.

One question that naturally occurs is—while the 5% decrease is appreciated, is it the right percentage? Should it be around, say, 10% in the average population? 5%? The other question is to wonder about the paucity of evidence. 26,000 patients would be quite a substantial study in most contexts, limiting it to northern New England is a problem given what we know about geographic differences in Medicare spending. If you use that as a rough proxy for overall medical culture, it suggests that some populations might be more receptive to more evidence-based approaches to medicine.

(And there’s some doubt whether the drop was due to the study, anyway:

The authors noted that the decline in overall PCI volume started before the [study] results were presented, which they attributed to analyses from European registries linking drug-eluting stents with fatal stent thrombosis that were reported in August 2006....That might have contributed to the response clinicians had to the [study] results six months later, they speculated.

)

It’s hard to know how effective your comparative effectiveness research studies are unless you do…comparative effectiveness research on the effects of comparative effectiveness research?