Tyler Cowen suggests a number of ways to think about health care spending; more on that later. But here's a metric we might use to compare our various policy options. According to a study that even the New Republic's Jon Cohn admitted he thought was probably exaggerated, being uninsured killed 18,000 people a year this decade. Methicillin-resistant Staphylococcus aureus, on the other hand, apparently kills 19,000 a year.

MRSA is the result of inadequate hygiene in hospitals, and indiscriminate use of antibiotics. Doctors and other medical workers have gotten lazier about hygiene since the invention of antibiotics, in line with other sorts of risk-taking behavior (more than one wag has suggested that the best way to eliminate auto accidents would be to mount a spear on the steering wheel pointed straight at the driver's heart.) They also prescribe antibiotics even when they are not clearly indicated, "just in case" . . . or worse, when they know they won't do any good, but want to get an ignorant and demanding patient out of their office. Third place for blame must, of course, go to the patients who do not finish taking their drug courses, which allows partially resistant bugs to survive and eventually breed highly resistant bugs . . . and what with my awful memory, you can put me in the dock along with almost everyone else.

MRSA is just one of the infections that are thriving in this environment. What would be the cost of a war on infection in hospitals? One suspects it would be a lot less than insuring 44 million people.

This might be one of the items that Tyler suggests libertarians should think about. There's a clear public health cause here, as with vaccines; doctors who prescribe indiscriminately, or people who don't take all their pills, are in fact placing a substantial burden on the rest of the public.

Non libertarians can, of course, go along wishing that we would have national health care and a War on Infection. But it's worth asking yourself: in a world of scarce resources, where you could only have one, which would you choose? And by what principle?

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.