Health spending & life expectancy

Does it matter how much a country spends on healthcare? My chart, taken from OECD data for 2005, raises this question. It plots total health spending (public and private) as a share of GDP against life expectancy, for 32 OECD nations.
What leaps out here is that the US is a huge outlier. If we ignore it, there seems to be a strong correlation (0.67, R-squared = 45.3%) between health spending and life expectancy, with each percentage point of GDP spent on health associated with 14 months longer life.
Taking life expectancy as the measure of a healthcare system suggests*, then, that the US’s system is hugely wasteful. The experience of other OECD nations suggests it could achieve its life expectancy, of 77.8 years, with health spending of 8% of GDP - a saving of 7.2% of GDP, or roughly $1 trillion a year.
By contrast, the UK, with health spending of 8.2% of GDP and life expectancy of 79.1, is in the middle of the pack. You’d not guess from our chart that there’s anything grievously wrong with the NHS as a system.
However, the strong correlation between health spending and life expectancy exists largely because there are a handful of poor countries (such as Turkey, Poland and Mexico) which spend little on health and have low life expectancy. If we exclude the seven countries in our non-US sample with health spending below 7.5% of GDP, the correlation between spending and life expectancy falls to a statistically insignificant 0.16.
This doesn’t necessarily mean that there ceases to be a causal link from health spending to life expectancy once we exceed a certain  level of spending. It might instead be that the link is confounded by another factor. For example, if a nation lives healthily, it would have long life expectancy and low health spending; this is true for Japan. 
This raises the question. Could it be that if we want a healthy citizenry we should think less about the level of health spending, or even how it is organized, and more about lifestyle factors that could promote better health?
* Of course, life expectancy isn’t the only metric. But is there an observable measure by which Americans are so much healthier than others as to justify the US’s exceptional spending?

Comments

[Jackart]

is there an observable measure by which Americans are so much healthier than others as to justify the US’s exceptional spending?

Yes. US health consumers subsidise drug development for the entire world.

[john b]

@Jackart: that isn't "an observable measure by which Americans are so much healthier than others".

Nor is it relevant even on its own merits. Total global healthcare R&D spending was US$53bn for 2004 [1] (it won't have order-of-magnitude changed since then). US healthcare spending is $2.4 trillion [2]. Even if you assume the US does literally pay every penny associated with global pharma R&D, that's only *2%* of total US healthcare spending.

[1] http://www.drugresearcher.com/Research-management/R-D-spending-per-NME-declines
[2] http://www.nchc.org/facts/cost.shtml

[Phil (Dex)]

I thought the BOM's blog post today about it had quite a nice table showing survival rates from typical cancers both in the US and in Europe. (http://burningourmoney.blogspot.com/2009/08/gee-thanks-guys.html)

but without having read the report that its based upon i couldn't tell you if that included poor countries like poland in the europe catergory, there by lowering the average for europe or some other bias like selective reporting for the states.

[Mike Woodhouse]

I wonder what would happen if you were able to split the US data by those who have adequate health insurance and the reported 45m who do not?

[john b]

The cancer survival charts are meaningless, though: the US system encourages generally-unnecessary medical testing, which means that cancers are found earlier than in Europe [*], which obviously means that patients are more likely to survive five years even in the complete absence of medical treatment (this is particularly obvious for prostate cancer, which nearly all men have by the time they die, but which kills very few of them). There's no clear evidence that, aside from that statistical artefact, US treatment is more effective.

[*] not finding a cancer until it's inoperably fatal is Very Very Bad. But finding it as early as they're found in the US is no better for survival rates, once the effect above is allowed for, than in Europe.

[john b]

"which nearly all *old* men have by the time they die", even.

[MattM]

John, I agree that 5 year survival rates are generally a very bad way to compare outcomes (due to the sampling bias you refer to). However, when you attempt to disaggregate the effects of this so called "lead time bias" as this paper does:

http://repository.upenn.edu/cgi/viewcontent.cgi?article=1012&context=psc_working_papers

by looking at the change in mortality rates over time for diseases for which there have been significant, recent advances in treatment, you see that the US does better than other countries

[charlieman]

CD: "For example, if a nation lives healthily, it would have long life expectancy and low health spending; this is true for Japan."

I'm not sure about the accuracy of the above statement.

1. Japan is a great country for smokers. Even outside Tokyo, the tobacconists have great product ranges.

2. On the positive side, red meat consumption is low and seafood consumption is high. A typical chicken ready meal from the supermarket will contain about one ounce of fowl.

3. A lot of food is consumed outside the home. Tempura with rice has the same nutritional value as fish and chips.

4. Any list of vegetables available in a Japanese supermarket would horrify anyone who likes their veg.

5. Japanese people are not noted for their indulgence in running around sports. However big cities have pavements and pedestrian routes that encourage walking or cycling.

6. Japanese bean curd products are enjoyable, unlike the filth that is sold in the UK. All the same, when Japanese people get the opportunity for a steak meal on company expenses, they use it.

So I am not convinced that modern Japanese lifestyle contributes to longevity. Those who are 80+ years today are survivors of austerity years, people who were able to live on a very humble diet.

If we are trying to work out the efficacy of health spending, shouldn't we be looking at mortality across the age spectrum?

Aside: I noticed today the oldest man in the UK died this week. So that was three "oldest person in the UK" passing away in a matter of months. I wish the new oldest person a few fruitful years.

[splinteredsunrise]

Well, despite the best efforts of the United Appeal for the Dead, death still remains our nation's biggest killer. And unfortunately the elderly are especially vulnerable.

[Tim Worstall]

My impression (and it's no more than that) is that the two things that all that money buys Americans are:

1) No waiting (if you are indeed insured, obviously).

2) Huge amounts of possibly wasteful near end of life treatment.

It may well be that a rational planner would say that, well, a little bit of waiting doesn't do that much harm, that squeezing such out of the system isn't worth the cost. That hip replacements for people with terminal cancer (Obama's mohter for example) are a waste.

It may also be that whatever the rational planner might think on the matter the people disagree.

This is essentially Johnny Munkhammar's argument: That Americans spend so much on health care because they've the most free market system around, the one that actually allows consumer preferences to be expressed.

Dunno how valid the argument is but it is at least interesting.

"Even if you assume the US does literally pay every penny associated with global pharma R&D, that's only *2%* of total US healthcare spending."

Er, no JohnB. For of course in order to pay for the R&D you've got to have the marketing arm, the sales people, the beancounters and all the rest. R&D is about 15% of pharma total revenues. But all of the revenue is required to fund the R&D. Which means that using your calculation then the cost of R&D is 13% of total US health care spending.

[john b]

No, because most of the sales and marketing infrastructure is not only unnecessary but actively harmful on a system-wide basis.

The right comparison would be looking at the fees that:

1) Pfizer pays to a small biotech to license a promising new compound
2) A CRO charges to do external trials on said compound up to Phase III and launch

That's the market price of drug R&D. The rest of a big pharma company is just rent-seeking. And the rent seeking is pretty much entirely how the big pharma companies' profits are derived (nearly all drug development is licensed in rather than in-house).

[Ironman]

"But is there an observable measure by which Americans are so much healthier than others as to justify the US's exceptional spending?"

Yes. Natural life expectancy:

http://politicalcalculations.blogspot.com/2007/09/natural-life-expectancy-in-united.html

But if you want to know why Americans spend so much more on health care than other nations, it has a lot to do with being wealthier than other nations (certainly on a per capita basis):

http://politicalcalculations.blogspot.com/2009/07/redefining-health-care-debate-part-2.html

Perhaps a better question to ask is why do so many nations with such relatively high GDP per capita figures work so hard to suppress spending for health care? Who benefits from that?

[Bruce]

They may not live longer in the US, but the place abounds with enhanced bodies, plastic faces and Osmond smiles.

[botogol]

your X-Axis is health spending *as a percent of GDP*, so that doesn't really tell you anything useful - a higher %tage doesn't mean more spending on health.

You need to recast this chart with the X_axis being *health spend per person*.

That would tell you something more useful.

[Roger]

So if the US habit of running tests because they can charge for them is unnecessary why does it correlate with better cancer survival rates?

It may be arguable that inefficient as it is from a purely economic standpoint US healthcare does 'save more lives' precisely because their isn't a single payer able to impose proper budgetary discipline.

Some US neoliberals like Mickey Kaus have also pertinently asked why does it matter if 15% of GDP is spent on healthcare?

If the portion of that 15% that comes out of taxes is in fact quite similar to the percentage in Europe then the difference is effectively just another form of productive economic activity that shouldn't be regarded any differently to widget-manufacturing or junk-bond trading.

In fact (again arguably) the taxes paid by US health workers and insurers probably pay for a significant amount of the tax subsidy.

(to a lesser extent this applies here as well - if a million NHS workers pay say an average of £9,000 a head in taxes that accounts for 10% of the NHS budget and about 1% of that percentage of GDP spent on healthcare)

As the more perceptive liberal supporters of reform (IIRC Jonathan Cohn in the New Republic for one) have pointed out, Obama's biggest mistake may have been framing the whole discussion in what are fundamentally negative and conservative terms of cost-control.

[MB]

First of all, I ask you to forgive my English: it is not my first language.

May I observe that the choice of life expectancy may not be best suited to judge the quality of the American medical system? I suggest looking at infant mortality, which gives really frightening results. Thus, the average infant mortality in the United States is higher than Cuba. The Black infant mortality is twice the national average at 1.4%, it is roughly three times the British one and the results are dangerously close to Russia. Hispanics, although they are less insured than others, do not fare so badly.

Finally, the United States is one of the few countries where infant mortality has recently increased.

A small chart, may be?

Here : http://en.wikipedia.org/wiki/List_of_countries_by_infant_mortality_rate

There, p. 85-86 : http://www.cdc.gov/nchs/data/hus/hus05.pdf

[Tom]

@Roger: "So if the US habit of running tests because they can charge for them is unnecessary why does it correlate with better cancer survival rates?"

The usual measure of cancer survival is defined relative to when the cancer is diagnosed. Earlier diagnosis will automatically increase the five-year survival rate, even if the treatment is no more successful.

@MB - I wonder if the infant mortality rate for the US is at all linked to their strong anti-abortion stance? I can imagine (that is to say, I have absolutely no figures to back this up) there might be a lot of babies carried to term in the US when it's known that they're unlikely to survive long outside the womb, when in many other countries they would be aborted at an early stage.

[MB]

@ Tom: Actually, I'm following Emmanuel Todd's thesis: infant mortality may be the best way to measure the hardness of life. I had no idea about abortion rates in different countries. It seems that the percent of known pregnancies ending in legal abortions in the USA is actually higher than that of most of the European countries.

Source : http://www.johnstonsarchive.net/policy/abortion/wrjp333pd.html

[buy fioricet]

i think it was depend on the person. life expectancy comes from the person itself, if he/she had habit and doesn't want to change then her/his life expectancy will ended soon.

[buy soma online]

I think we need to be determined to maintain our healthy lifestyle by a eating healthy!

[breitling watches]

This is just one idea, and perhaps displays no more than my limited imagination. If there are better ideas out there, that amount to more than "implement something called "market socialism" and then - alacazam! - full employment!" then I'd love to hear them. http://www.watchgy.com/ mostly bank deposits, fell by £143.2bn in Q1. And of course there’s no guarantee such buying will continue.
http://www.watchgy.com/tag-heuer-c-24.html
http://www.watchgy.com/rolex-submariner-c-8.html

[Invertir Dinero]

Great site about Health spending & life expectancy,this information really helped me , I really appreciate it,I will visit when ever i have found the stuff That i have been searching for in all the web for, keep up the great work!

[lots in Samara Costa Rica]

Hello guys very interesting this post I really like it and I want to read more blogs like this about health.

[Jo]

Great post. Thanks a lot!

[Lasix]

What the scatterplot really made me realize was the arbitrariness of the scaling of the parallel coordinate plot. In particular, the posted graph gives a sense of convergence, that spending is all over the map but all countries have pretty much the same life expectancy--look at the way the lines converge to a narrow zone as you follow the lines from the left to the right of the plot.

http://www.lasixmedication.info

Actually, though, once you remove the U.S., there's a strong correlation between spending and life expectancy, and this is super-clear from the scatterplot.

The only other consideration is novelty. The scatterplot is great, but it looks like lots of other graphs we've all seen. This is a plus--familiar graphical forms are easier to read--but also a minus, in that it probably looks "boring" to many readers. The parallel-coordinate plot isn't really the right choice for the goal of conveying information, but it's new and exciting, and that's maybe why one of the commenters at the National Geographic site hailed it as "a masterpiece of succinct communication." Recall our occasional discussions here on winners of visualization contests. The goal is not just to display information, it's also to grab the eye. Ultimately, I think the solution is to do both--in this case, to make a scatterplot in some pretty, eye-catching way.

P.S. I never know how much to trust these purchasing-power-adjusted numbers. Recall our discussion of Russia's GDP.

P.P.S. And here's the R code. Yes, I know it could be cleaner, but I just thought some of the debutants out there might find it helpful:

The first step was to download the data, open the Excel file and save the relevant data matrices as three .csv files. I don't know how to extract information from .xls files in R, so I went into Excel and did that part manually. After that, I did the following: