While working on some of my recent posts about the Oregon study, I came across this report from the CDC on changing causes of death over time. If you spend any time thinking about the history of health innovation in America, it's pretty fascinating.

The first thing you notice is that how we die hasn't changed all that much since the Great Depression: the leading causes of death today are cancer and heart disease, just as they were in 1935. But in other ways, everything has changed.

For decades, pneumonia was known to doctors as the "old man's friend" because it was viewed as a better way to die than cancer or heart disease, and among the elderly, it had a mortality rate approaching 100%. Babies, too, were apt to succumb. Thanks to antibiotics, this category has largely been vanquished. Kidney disease, formerly a major cause of death, has basically dropped out of the chart thanks to the invention of dialysis. Stroke fatalities have notably decreased, because we've gotten pretty good at controlling hypertension. Accident deaths are down too, for which you can thank both safer jobs and better trauma care. Cancer and heart attacks are expanding only because we've all got to die of something eventually; if you look at the annual mortality statistics, you'll see we're better at treating those, too.

This graph tells the happy tale:

Nearly 100% reduction in childhood mortality! Over 50% even among the elderly. Whenever you're tempted to argue that there's some halcyon past of simplicity and natural harmony that we should all aim to return to, come back and look at these graphs. We live in the best of times, full stop.

Except, possibly, for the future. I was on record, during the Obamacare debate, as opposing its passage on many grounds, but most notably because I feared that the US government would find itself too tempted to control costs by stamping out innovation. Those charts above represent true miracles in human affairs. I didn't want anything that might stop the miracle.

While we're reassessing the evidence on health insurance and mortality, it seems worth revisiting this assumption, too. Even before Obamacare takes effect, health care cost growth has started to slow. Some of this is the recession, and some of this is due to market shifts: higher deductibles are pushing consumers to shop around more, Medicare payment reform may be reducing hospital readmissions, and a number of blockbuster drugs like Lipitor are coming off patent with nothing in the pipeline to replace them.

I'm pretty sure that many of these changes are unambiguously good news: giving consumers incentives to shop smarter is great, and it's generally a good idea to keep patients out of the hospital. On the other hand, the prescription drug data is not good news. I mean, it's great that everyone can get their Lipitor cheaper. But it's worrisome that there's nothing else in the pipeline behind it.

Of course, it's tempting to blame this on the chilling effect of Obamacare, but I don't think that explains the changes in either health care costs, or drug discovery. If anything, we should blame both on the market.

I've never been particularly hawkish on health care costs, for two reasons. The first is that, per the graphs above, spending on health care is pretty great; "not dying" is one of the most important goods society can offer. So what if we end up spending 25% or 50% of national income on health care? We have a whole section of the supermarket devoted to stuff that makes your towels softer. We're really rich. What better to spend our money on than health?

The other reason I don't panic is the sure and certain knowledge that, in the immortal words of Herb Stein, "If something can't go on forever, it will stop." One way or another, we were never going to spend 100% of national income on health care. The market, or voter preferences, was going to fix our "cost problem". To me, this is only a problem insofar as the government is involved, because the government doesn't have any very good way to smoothly and slowly stop spending money on something. Instead, the inevitable correction tends to come with a very disruptive jolt.

Over the last five to ten years, employers and consumers have started to crack down on expenditures, and this is putting pressure on the margins for new inventions. Pharmaceutical benefit managers, for example, have started conducting their own clinical trials to try to narrow down the patient populations where new drugs are actually more effective than cheaper old ones. This is mostly all to the good. But of course, it has to reduce the return on innovation, and therefore, the number of new drugs and technologies that will be invented.

However, I don't know that we can blame the market all that much, either. We may simply be hitting the frontier of health improvements.

The 19th century saw dramatic declines in mortality thanks to a few major advances: the germ theory of disease; anaesthesia; the public health infrastructure to track and isolate sick people who were spreading disease; clean drinking water and sanitary waste treatment; better nutrition; and larger, better ventilated living and working spaces. (Don't underestimate the power of those last two; most of the decline in tuberculosis mortality came before we developed streptomycin to treat it.) The 20th century added vaccination, refrigeration, smoking cessation, and antibiotics. Most people are now dying because their bodies wear out, one way or another and stop repairing themselves. It's just not clear that we can push things much further. To be sure, it's not clear that we can't, either. I'm just saying, we can't bet on it.

What if we're hitting the edge of the possible improvement? We're not really ready for that world. We're used to the idea that every 10 or 20 years, someone comes along with a whizbang new idea that cures something which used to kill you. This is actually not normal at all; it just happens to have been true for all of our lives.