Stayin' Alive

No, I'm not running out of titles, that just seemed right for this one. A few days back I wrote that when I first got into this business, in the '80s, it was legitimate to argue that medical services (so-called "health care") made only a small contribution, at best, to population health and longevity; but that this is no longer true. We spend more and more on health care, but we are getting something for it. Still, is it worth it?

In the new NEJM, David Cutler, Allison Rosen and Sandeep Vijan try to answer this question. (Subscription only, here's the abstract.) They looked at gains in life expectancy from 1960 to 2000, compared with increases in medical spending. Now, life expectancy is a slightly dodgy construct. We talk about life expectancy at various ages, from birth on up, but how do we know how long an infant born today can expect to live? We can't possibly know, the calculation is made by assuming that the current death rates at various ages will apply to the infant as she or he grows up and ages. In fact those rates could change for better or for worse and the average baby born today could live a longer or shorter time than the life expectancy at birth. Whatever.

Next, they assumed that 50% of the gains in life expectancy since 1960 are due to medical intervention. That sounds as though they picked the number arbitrarily but actually they make a pretty good case that it's probably somewhere close to accurate.

So, first of all, you'll be happy to know that life expectancy at birth increased from 69.9 in 1060 to 76.87 in 2000. At least you'll be happy to know that if you're a newborn baby. Life expectancy at age 65 increased from 14.39 years to 17.86, about half as much. In other words, if today is your 65th birthday, and you're the average person, you can expect to live to be nearly 84 years old. Of course, you aren't that person. You'll probably do better if you are female, and worse if you are male, by the way. But you knew that.

Their analysis is a bit misleading in that they spend most of their time talking about the average total expenditure per person since 1960, which yields a figure of $19,900 spent per year of life gained per person. That doesn't sound like a bad deal at all.

But it's much more appropriate to look at trends over time, and over the life course. Here it gets a bit more worrisome. We spend much more on people 65 and older, obviously, and that differential has increased. The average cost per year of life gained for people 65 years old was $75,100 from 1960 to 1970, but $145,000 from 1990 to 2000. Of course, if their estimate of the percentage of gains in life expectancy that results from health care is off, the numbers go up or down accordingly.

Would you say it's worth it to spend $145,000 so you can die at 81 rather than 80? Or so somebody you love can? Maybe, but as with everything else, there's an opportunity cost. That same $145,000 could, literally, save the lives of hundreds of poor children who will die for lack of a $4 bed net to prevent malaria, or a village well to deliver clean water, or a condom so they don't become an orphan. But then again, we could get the money just as easily by ending the occupation of Iraq, on which we have already squandered $300 billion. So it's not a direct tradeoff. Medical care could also have additional benefits -- relieving pain, preventing or correcting disability -- and costs -- yup, causing pain, causing disability. Probably there is some additional net benefit there but I'm not convinced of it. Unnecessary or failed tests and procedures cause a lot of damage.

Then of course there is the issue, which I've about beaten to death here, that European countries and Canada get even better results from their health care systems, while spending half as much or less. They have seen the same percentage increases over time, but the gap remains. So even if we want to keep spending the big bucks to squeeze out another year in the rocking chair, we're still wasting half of it.

I'm not telling anybody what we ought to do, but those are some numbers on which to base your opinions.