Gregory Conko writes the worst sort of logical fallacy today: about how we can't make decisions for the common good because someone really special might suffer. The FDA is decertifying Avastin as a treatment for breast cancer, largely because it has a few side effects:
The agency is making this recommendation after reviewing the results of four clinical studies of Avastin in women with breast cancer and determining that the data indicate that the drug does not prolong overall survival in breast cancer patients or provide a sufficient benefit in slowing disease progression to outweigh the significant risk to patients. These risks include severe high blood pressure; bleeding and hemorrhage; the development of perforations (or “holes”) in the body, including in the nose, stomach, and intestines; and heart attack or heart failure.
Of course, you already have holes in your nose, stomach, and intestines – how else are you supposed to get air or food or poo? Someone in the Obamunist government didn't think that one through at all, the lazy Trotskyite.
The hard part about these decisions is that there are always people who don't experience side effects and for whom the treatment is effective, much the same way you always have that one asshole who manages to not get food poisoning from the steam tray food at the reception. The argument that Conko lays out:
The weakness of the FDA's reasoning here is that averages ignore that individual patients respond differently to treatments. Particularly with life-threatening illnesses, where the downside of any treatment is relatively small, average or median survivability too often masks the fact that some patients respond very well.
When well-known scientist Stephen Jay Gould was diagnosed with a rare form of lung cancer in July 1982, he was told the diagnosis meant a median survival time of just eight months. His doctor gave up on him. But he lived another 20 years.
"Means and medians are the abstractions," he wrote in Discover magazine in 1985. "Therefore, I looked at the mesothelioma statistics quite differently—and not only because I am an optimist . . . but primarily because I know that variation itself is the reality."
For the sake of argument, we're going to pretend that Gould's doctor talking about about average survivability based on a diagnosis (which is what Gould's essay was about - .pdf link) is the same thing as average response to and risk from a treatment. Actually, wait, no. We're not. They're different things.
Gould's essay is a paean to the fact that cancer (and, by extension, illness) is an individual struggle, with statistics on your survival gleaned from the aggregated experiences of the masses. What it's not is an argument that we are all Randian supermen and women who should escape the oppressive dictation of the state on what is or isn't an effective treatment.
The purpose of the FDA isn't to provide individual recommendations on the efficacy of each treatment for each patient, because they aren't doctors or patients – they're a regulatory agency. Their threshold isn't Gould's, because the FDA doesn't just serve Stephen Jay Gould (or tomorrow's version) – it serves everyone in that distribution curve. By arguing that we should only serve the Goulds (or those on the rightward side of the distribution, as he would have put it), we tend to ignore those on the leftward side of the curve.
I understand that women with hemmorhaging stomach holes who are about to die of breast cancer aren't as compelling as charmingly literate science writers with tales of perseverance and determination, but someone should be looking out for them as well. It's nice that the FDA does.