Sunday, August 16, 2009

US vs. Europe: Life Expectancy and Cancer Survival

Is the U.S. system inferior to those in other developed countries based on life expectancy and cancer survival rates? Not according to economists Robert L. Ohsfeldt (Texas A&M) and John E. Schneider (University of Iowa), who argued in their 2006 book, The Business of Health: The Role of Competition, Markets and Regulation (AEI Press), that the U.S. system actually compares very favorably to the health-care systems of other nations.

1. The top chart above (data here) shows both: a) unadjusted life expectancies for the U.S. and other OECD countries, and b) standardized life expectancies which are adjusted for the effects of premature death resulting from non-health-related fatal injuries. For unadjusted life expectancy, the U.S. ranks #14 out of 16 countries, but for the adjusted standardized life expectancy the U.S. ranks #1.

2. The bottom chart displays five-year age-adjusted cancer survival rates for the U.S. and selected European countries, showing that the U.S. has the best record for five-year survival rates for six different cancers. In some cases the differences are huge: 81.2% in the U.S. for prostate cancer vs. 41% in Denmark and 47.4% in Italy; 61.7% in the U.S. for colon cancer vs. 39.2% in Denmark; 12% in the U.S. for lung cancer vs. 5.6% in Denmark.

Also interesting is the fact that there is often a significant difference between white and black cancer survival rates in the U.S., e.g. prostate cancer - 82.7% for whites vs. 69.2% for blacks (see red circled data in bottom chart). But even in that case, the five-year survival rate for blacks (69.2%) is still higher than for all European countries except Switzerland.

See recent
related CD post on five-year cancer survival rates.

Originally posted at Carpe Diem.


At 8/16/2009 11:54 AM, Anonymous Anonymous said...

Most Whites can afford "Private" Health Insurance. Many (most?) Blacks cant't?

I would venture that many more blacks find themselves in the "uninsurability" trap. An example of this would be when you develop a health condition, lose your job, and then find yourself "uninsurable" for the affliction.

At 8/16/2009 11:59 AM, Anonymous Anonymous said...

the U.S. system actually compares very favorably to the health-care systems of other nations.

Sheesh, I would hope so. The US is a statistical outlier when it comes to health care spending as a % of GDP and even more so in terms of purchasing power parity per capita.

At 8/16/2009 3:31 PM, Blogger juandos said...

"I would venture that many more blacks find themselves in the "uninsurability" trap. An example of this would be when you develop a health condition, lose your job, and then find yourself "uninsurable" for the affliction"...

Oh dear! Now where did I lay the crying towel?

Got love that 'race card' action brought to us courtesy of the rufus...

Let me clue YOU in on what a major domestic based air carrier does...

Hernia? Need an operation? Get it done with workman's comp insurance...

Get your old job back? Not unless you can pass the 'new & improved' physical testing that allows the company doctor to give you a clean bill of health...

White, black, brown, it doesn't matter if you can't pass the test, you're done and so is your company purchased health insurance...

A few, a very, very few lucky ones might get a shot at SSI...

There is a wide range of maladies where this is the routine...

At 8/16/2009 8:22 PM, Blogger Ironman said...

Here's an old post from 2007 that has a dynamic table version of the natural life expectancy data - you can sort the data in the table by clicking the column headings.

MP: Survival rates for the black population of the U.S. for many chronic health conditions is below that for whites. As a result, U.S. life expectancy figures tend to be pulled lower compared to those of other nations that don't have such a significant portion of their populations made up by those of African descent.

Here's a big reason why that is. (Follow the links at the bottom of the post for the entire series.)

At 8/19/2009 10:12 AM, Anonymous Anonymous said...


Survival rates are not an important stat without the "age at onset" statistic. In the US we have younger onset of most diseases and then these younger patients survive longer. It is a bogus stat that the right wing enjoys using. Think about it. If you get diabetes at 80, you are likely to die. If you get it at 35 you will probably live many years and be listed as a survivor. Sorry, US health fans, we are a sick and dying country.

At 8/19/2009 2:16 PM, Anonymous Mike Jones said...

Thanks for the interesting article!

At 8/20/2009 3:24 PM, Blogger Unknown said...

According to "OECD Economic Surveys: United States 2008", p. 137:
"It has been claimed (Ohsfeld and Schneider, 2006) that adjusting for the higher death rate from accident or injury in the United States over 1980-99 than the OECD average would increase US life expectancy at birth from 18th of of 29 OECD countries to the highest. In fact, what the panel regression estimated by these authors shows is that predicted life expectancy at birth based on US GDP per capita and OECD average death rates from these causes is the highest in the OECD. The adjustment for the gap in injury death rates between the United States and OECD average alone only increases life expectancy at birth marginally, from 19th on average among 29 countries over 1980-99 to 17th. Hence, the high ranking of adjusted life expectancy mainly reflects high US GDP per capita, not the effects of unusually high death rates from accident and injury."

In other words, the figures in Table 1-5 are not U.S. life expectancy adjusted for fatal injuries, but rather a model that assumes that both the relationship of life expectancy to per capita GDP and injuries in the U.S. follow OECD trends.

That is - they are giving the U.S. credit for having the same basic life expectancy as other other high GDP OECD countries, when in fact it is markedly lower.

Check it out for yourself, their report is at:
See p. 20-21.

At 8/20/2009 8:54 PM, Anonymous Anonymous said...

At 8/20/2009 9:42 PM, Blogger Hoiles said...

Thanks for the link Carey. I dug up the equation used for these calculations:

LifeExpit = 50.78 + 3.020 • log(GDPPCit) – 0.077 • [mean(Trans)]– 0.137 • [mean(Falls)] – 0.133 • [mean(Homicide)]– 0.0326 • [mean(Suicide)] + year-effectsit

This model appears to "standardize" life-expectancy by:

- Giving credit to countries with higher GDP per capita (such as the US)
- assuming every country has equal homicide and accident rates

As Carey said, the estimate isn't really about adjusting for homicides and injuries. More so, it gives a "standardized" estimate of what life expectancy -should- be, based largely on GDP per capita. It predicts that Americans -should- have higher life expectancies based on higher GDP per capita, not necessarily that they -do- have higher life expectancies.

In short, saying that this is life-expectancy adjusted for higher homicide or accident rates appears to be a clear misinterpretation of results.

Please feel free to inform me if you feel I am making an error anything.

At 8/21/2009 2:14 AM, Anonymous Lasse said...

Considering how the statistics from the European countries will include almost every citizen, while the US will not cover a whole lot more than those who are actually insured, you can be sure that the US numbers are in reality a lot worse. This is perhaps slightly witnessed via the white/black difference in the second table.

At 8/21/2009 8:15 AM, Blogger Ironman said...

@Lasse: Evidence, please.

@Carey, Hoiles: The standard international comparisons of life expectancies are flawed in that they do not consider demographic differences between OECD high GDP nations. For example, if the US (~12.6%) had the UK's percentage (~3% in 2006) of black population, the US' life expectancy from birth would rise by roughly 0.5 years. The use of GDP per capita data provides an indirect method of "homogenizing" the respective national populations, which is required since many nations lack detailed statistics for identifying the ancestral origins of their minority populations, which would allow this factor to be accounted for directly.

At 8/21/2009 9:50 AM, Blogger Hoiles said...

Ironman, please provide any evidence that GDP per capita is related to demographics. US has one of the higher black populations in OECD countries and also one of the highest GDP. UK has a lower black population and a slightly lower GDP. Norway has an even lower black population and a slightly higher GDP.

If you wanted to "adjust" for a higher black population, using GDP is a very poor way of doing it.

At 8/21/2009 2:18 PM, Blogger Ironman said...

Hoiles said: "If you wanted to "adjust" for a higher black population, using GDP is a very poor way of doing it."

I don't disagree with you. As I noted, statistics (and studies, for that matter) related to the health of ethnic minorities with the larger population for other nations are often lacking - Canada is a good example here.

As to your other question, let's try a thought experiment.

Compare the income earned by identical twins, one who is healthy vs. one who has a chronic health condition. Statistically, you would expect that the healthier twin will have a higher income - all other things being equal, they're less likely to miss work and as a result, will earn more income, whether by number of days worked, or by being able to take advantage of additional opportunities for advancement that the unhealthy twin might not be present to see.

Now take that up to a national scale - say where you have a nation where most people are healthy compared to one that has endemic levels of disease. You would expect the GDP per capita of the nation without the endemic disease problem to be higher than that for the nation with the endemic disease, simply because more people within that nation are healthier and are thus more able to generate income within their nation.

By and large, that's exactly what we see in the real world.

Now, let's look at the U.S. Here we find that the black population has a relatively higher incidence of chronic health conditions than does the white population. This has two consequences: it lowers the overall life expectancy figures for the U.S. (as blacks are more likely to die at younger ages) and it lowers the U.S.'s GDP per capita, although by a lesser amount, as for many, it's possible to continue earning an income even with having to deal with a chronic health condition.

As a result, we see that GDP per capita automatically embeds the economic effects of the nation's underlying demographics where the relative health of various ethnic populations within the nation are concerned. It's not as desirable as a direct method, but when the data you need isn't available, it's a reasonable proxy.

At 8/21/2009 4:24 PM, Blogger Hoiles said...

thanks for you reply, ironman. However you did not really address my issue. Adjusting for demographic differences is fine. However, using another variable (GDP) that happens to correlate with it to boost the life expectancy estimates is something that any experience statistician would have issue with.

My original assertion (as well as that of the OECD economic survey) is that the "adjusted" life expectancy reflects more the GDP than anything else.

If anything, this table implies that the US should have the one of the highest true life expectancy rates due to their higher GDP (which they already DO have in spite these so-called demographic differences). Furthermore, the higher standard of living is more important than possible differences in homicide rate, accidents, etc. But instead, they have life expectancies in the middle of the pack.

At 8/22/2009 8:40 AM, Blogger sly uncle said...

Professor Perry, I'm just wondering if you plan to update this post with an acknowledgement that the study you cited is wrong and uses bad models(as described in comments). That seems like the responsible thing to do.


Matt D

At 8/23/2009 1:37 AM, Anonymous Anonymous said...

Apples and oranges - as alluded to above. If your comparing US survival rates you should compare them to PRIVATE european clinics. As the american data will be missing all those who died because they were uncovered/not insured.

Try you must young padawan

At 8/24/2009 1:43 PM, Blogger Unknown said...

are there more reliable statistics for judging health care available than life expectancy?

Like what percentage of people in different countries live up to the age of 50? of 70?

Unicef publishes infant mortality rates and also death rates for children up to age five that are compared across nations. Anyone doing an older age version of this?

At 3/22/2012 3:53 PM, Blogger Ye Olde Wielder O' Projects said...

Sorry to dredge up an old entry but it seemed to me that this discussion is missing a critically important consideration. That is simply that 5-year survival has improved to the current levels in the US for those over 30, but for young adults (under 30) it has not improved at all. See here for info: the cancer journal for clinicians, DOI: 10.3322/canjclin.57.4.242

In fact I suspect that it's the same for that demographic in all the countries. Why the discrepancy? Two reasons: (1) Early detection, and (2) lifestyle changes. Early detection efforts are almost entirely to credit for the improvement in cancer survival ... which is why you see survival rates so much better in those countries that have put early detection efforts at the forefront. It is also why we see no survival improvements for young adults: "early detection" efforts aren't early enough ... they start at about age 30. It's not the drugs - both age groups get the same drugs ... the same treatment.

In short, this indicates that the new drugs used to treat cancer are no better than those used 30+ years ago, and the above data does not make that apparent ... in fact it wrongly seems to suggest that it's the rich countries that can afford the most expensive latest cancer drugs that have the best survival - this, as I have proven, is an illusion (and one that if I were a pharmaceutical sales rep, I'd exploit).


Post a Comment

<< Home