A recurring link in my facebook feed today was to an article in the NY Times Magazine on rationing health care. It’s worth reading, but the this made me squirm a bit:
But even in emergency rooms, people without health insurance may receive less health care than those with insurance. Joseph Doyle, a professor of economics at the Sloan School of Management at M.I.T., studied the records of people in Wisconsin who were injured in severe automobile accidents and had no choice but to go to the hospital. He estimated that those who had no health insurance received 20 percent less care and had a death rate 37 percent higher than those with health insurance. This difference held up even when those without health insurance were compared with those without automobile insurance, and with those on Medicaid — groups with whom they share some characteristics that might affect treatment. The lack of insurance seems to be what caused the greater number of deaths.
Oh correlation/causation fallacy, I long for your demise. At least he said “seems.”
An Ergodic Walk 
Sure, that two facts are correlated doesn’t mean one is causing the other. But it still demands an explanation: can you think of a better one than what the author suggests? Can you think of any other evidence that contradicts what he presents?
Sure I’m willing to believe it’s a causal factor — for one, there is a causal mechanism. But there is a step that is elided in the description in the piece. A is correlated with B, there is a mechanism by which A has a causal relationship with B, therefore we infer that A is responsible for B.
The problem I have is that people make this elision all too easily in false situations and it requires some care to make this clear.
The thing is, there other factors which may also be relevant to the patient outcomes, and the lack of insurance seems like it *should* be the dominant one, but how was that validated?
You can read the paper yourself here:
Click to access doyle_insurance_restat.pdf
From my 10 minute skim, it seems that the Times summary is a clear case of this: http://www.phdcomics.com/comics/archive.php?comicid=1174
Erin-
The thing is, the piece wasn’t written by a journalist. It was written by a respected scholar who should really know better. I think the choice of “seems” reflects his awareness that he was pushing the boundaries of inference.
More generally, while the piece was interesting, I despair of it changing the rhetoric on the health care debate. Frank Luntz has identified “rationing” as one of the buzzwords that the GOP has to use when arguing against the Democratic reform bills. A lot of someones need to get up and say that we already have rationing – it’s just imposed by people who view actually paying for health care as an easily reducible business expense.
Singer isn’t a “respected scholar.” In fact, he’s very DISrespected among many people, hewing to a radical agenda in the name of “philosophy.” His animal rights opinions make me cringe as someone who is a moderate believer in animal rights and who believes that folks like Singer turn off other moderates who might otherwise want to get involved and self-educated about the issue. Such people live their lives oblivious to how everyday choices they make affect the welfare of animals, meaning that, if I’m correct, Singer causes far more animal suffering in the world than he prevents.
Singer’s article is also ludicrous as it accepts his opponents’ premises rather than challenging the details. Just because government pays for some things doesn’t mean that health care the government denies payment for is totally unavailable. A grandmother on Medicare can get cosmetic surgery if she wants. Just because it isn’t on Medicare doesn’t mean she can’t get it. Rationing in America will not be about rationing medical care – this isn’t North Korea – but about rationing costs.
Anyway, once I got disgusted enough with the first part of the article I skimmed the rest, but it doesn’t surprise me in the least that Singer would use this most common of fallacies. It’s very easy to see why someone who didn’t have insurance might just happen to be someone more likely to have their car accident be fatal, independent of any effect of the insurance itself.