6 ms·
Is life expectancy a flawed measure of health care quality?
- yummyfajitas 16y agoPoor scientists. Their conclusions are correct, but a significant portion of the reasoning which would allow them to get there is politically off limits. That's ok, I'm politically incorrect. I'll stick to the comparison between the US and Japan, since numbers are easy to come by[1]. On the face of it, it looks as if Japan has us beat: 82.6 years to 78.4 years. Now lets break things down by subgroup: People of Japanese descent: 84.5 years in the US, 82.6 years in Japan. This long lived group makes up <5% of the US (too lazy to look up exact #, but east Asians in total are about 5% of the US), and about 99.4% of Japan. http://minorityhealth.hhs.gov/templates/browse.aspx?lvl=2&lvlID=53 http://minorityhealth.hhs.gov/templates/browse.aspx?lvl=2... People of European descent: 78 in the US, unknown (but irrelevant) in Japan. This group makes up about 80% of the US, and < 0.6% of Japan. People of African descent: 72 in the US, unknown/irrelevant in Japan. This short lived group makes up about 12% of the US, and as far as I'm aware Japan only knows about this group via video games. http://www.cdc.gov/nchs/data/nvsr/nvsr53/nvsr53_15.pdf http://www.cdc.gov/nchs/data/nvsr/nvsr53/nvsr53_15.pdf So if the US had the same demographics as Japan, our life expectancy would be 2 years longer. [1] Japan is the only industrialized nation with the property that the (possibly great great) grandchildren of Japanese immigrants is classified by the census a "Japanese American".
- ugh 16y agoAh yeah, I know that argument ;-) Look, this is not to say that it is invalid, far from it. It’s just that I think you have to be a lot more careful. It could be that Japanese vs Japanese Americans is one of those Apples vs Oranges comparisons. Are Japanese Americans really like Japanese?
- yummyfajitas 16y agoGenetically yes, Japanese Americans are like Japanese [1]. Their diets are also closer to Japanese diets than standard American diets. I may only be comparing lemons to oranges, which I agree is only marginally better than apples to oranges. But that's ok - I'm certainly NOT trying to make the claim that US health care is better than Japanese health care as a result [2]. I'm simply pointing out that if you include slightly better controls, the effect completely reverses itself. This strongly suggests the original comparison is crappy. [1] Genes matter. A fact I've heard - Indians have narrower blood vessels than non-Indian Whites, leading them to be more vulnerable to hypertension. The UK NHS takes this into account when giving medical advice, though supposedly US doctors do not (that would be racist). [2] To draw conclusions like that, you need to take groups of statistically similar people with the same disease and measure health outcomes. No one does this, it's very hard.
- carbocation 16y agoPoint #1 is not quite true. US doctors certainly take ethnicity into account when it is a medically relevant indicator of genetic risk.
- borisk 16y ago>>People of Japanese descent: 84.5 years in the US, 82.6 years in Japan. What if someone was to add something like "People of Japanese descent: $85 000 average family income in the US, $50 000 in Japan." (I've made up the numbers from personal experience) Obviously richer ppl live longer.
- yummyfajitas 16y agoIf you assume Japanese Americans are average for Asian Americans, the actual number is $65,000 (see my link). In any case, I agree with you. Income is a major effect, although it's far from clear that this is due to income buying health care. In fact, most of the evidence suggests health care has nothing to do with it. For instance, during the RAND experiment, richer people were healthier than the poor regardless of which category they were randomly dropped into. (Observational studies since then draw less clear conclusions, since observational studies are crappy in comparison to randomized experiments.) Another fun fact. The shortest lived people in the US are the Lakota, who live 35 years less than Japanese Americans (not a typo, life expectancy = 50). This group gets free health care via IHS (I believe it is structured similarly to Britains NHS), so it seems that there are other factors at work. Huh, imagine that. Life expectancy is, in fact, not a great measure of health care, since it is affected by so many other factors.
- anamax 16y ago> This group gets free health care via IHS (I believe it is structured similarly to Britains NHS) IHS is pretty much a horrorshow. It's one of the best arguments against govt-run healthcare. It doesn't much matter if some govt somewhere can run healthcare well. The US govt runs healthcare badly. Fix that first.
- notahacker 16y agoThe trouble isn't political so much as the inherent complexity of the causal relationship between life expectancy and virtually every lifestyle/demographic factor you can think of. Not all lifestyle factors are isolated from the quality of healthcare either - there may be many significant cultural reasons why the US diet is more unhealthy than the Japanese one, but it's easy to argue that there is a [weak] causal link between quality of healthcare and healthier eating habits.* Probably the best measure of heathcare quality you're going to get is looking at survival rates from specific treatable diseases whose incidence is largely uncorrelated with climate/genetics/lifestyle. *i) [free] accessibility to specialist advice from medical practitioners should improve the number of people willing and able effectively managing their weight/cholesterol level etc (ii) healthcare spending can include preventative measures such as campaigns to encourage healthy behaviour (which depending on their efficacy may be be an optimal use of resources, particularly for a heathcare provider that views treatments as an overhead rather than a revenue source)
- qq66 16y agoUsing immigrants as a proxy for anything is very suspect since emigrants to the US are a self-selected subset of a country's population. Compare the average intelligence, ambition, and work ethic of an Indian-American immigrant, that immigrant's American-born child, and the average resident of India.
- yummyfajitas 16y agoI wasn't talking about immigrants. Most Japanese Americans are the grandchildren of manual laborers who immigrated to the US between the Chinese Exclusion Act and the Asian Exclusion Act. Why don't you go up to a Japanese American and ask, "hey, how long have you been in the US?" (Be sure to duck immediately after saying it.)
- carbocation 16y agoYummyfajitas' argument regarding life expectancy of different ethnic groups that have come to the US is thoughtful. I've always felt that the life expectancy stat only works in a statist, hyper-controlled system. Once you let people make personal decisions about their health, the utility of the stat declines. The scientists in this article try to distinguish health-related from non-health-related causes of death. But even our health-related causes of death are not exactly under the control of the medical establishment. People eat food as if they had an active death wish.
- matwood 16y agoI've always thought that life expectancy was a poor measure of health care quality. My reasoning is simple, the US is the most obese country in the world, and obesity is known to cause numerous health problems and shortens life expectancy. The health problems and shortening of life expectancy ends up having very little to do with poor health care quality (poor health education maybe). I wonder if an argument could be made that given the USs propensity towards unhealthy lifestyles that the US is in fact great at healthcare because the life expectancy is as high as it is? Finally, the best measures I've seen to date on healthcare quality are looking at individual outcomes. For example, if you get prostate cancer what are the odds of living in the US versus other countries. I can't find the exact article now, but various countries do better at various health issues with the US leading many of them. http://www.webmd.com/cancer/news/20080716/cancer-survival-rates-vary-by-country http://www.webmd.com/cancer/news/20080716/cancer-survival-ra... The highest survival rates were found in the U.S. for breast and prostate cancer, in Japan for colon and rectal cancers in men, and in France for colon and rectal cancers in women, Coleman's team reports.
- loewenskind 16y ago>I wonder if an argument could be made that given the USs propensity towards unhealthy lifestyles that the US is in fact great at healthcare because the life expectancy is as high as it is? This is no argument to be made about the US having great health care. I don't know how a place where a random illness can wipe out your and your family's finances can even be considered first world. To address your specific claim "we must be good because we make horrible life choices and don't die!" I would point at the UK and alcohol consumption. Their bad life choices are driving costs up to the point that they'll probably have to make some big changes to deal with them but so far the system is doing a reasonable job of keeping them from dying from it.
- matwood 16y agoThis is no argument to be made about the US having great health care. I don't know how a place where a random illness can wipe out your and your family's finances can even be considered first world. I didn't claim anything and simply wondered aloud. Also, the article was about life expectancy and quality of healthcare and spoke nothing about costs. Your statement does bring up another question. Would that same hypothetical family prefer to have their finances ruined or have their loved saved from some random illness? Of course the best solution is perfect care at zero cost, but I think we can agree that it is impossible. If the US is actually getting better health outcomes than other systems perhaps the cost is worth it? And to address your specific claim, I've always thought that the government should provide health related catastrophic fallback insurance similar to the way they do for floods. I don't think the government needs to be involved in the day to day healthcare or insurance matters, but I agree that getting cancer for example should not ruin your families finances.
- matrix 16y agoI take what this guy writes on statistics with a large grain of salt; despite the domain name the author is not an economist. He is an MD and it goes without saying that he's very smart and accomplished -- but not in the field of statistics. This is relevant to this particular article, because he may be missing subtle, but important factors that a lay-person would miss. For example, see yummyfajita's post in this discussion.
- DanielBMarkham 16y agoI'm sure I'll get clobbered for this, but I am wary of economists using statistics to make political arguments. Between "economist","statistics", and "politics", that's three yellow flags right at the outset. Unless, of course, they agree with me, in which case they are simply wise academics trying to share their wisdom with the unwashed masses. :) There is a logical problem here, and it's much deeper than life expectancy. It's a problem with this form of argument. The problem is that, if you use statistics to make political arguments, you're going to be arguing forever. Whatever this guy wants to do -- and god bless him I hope he gets his way -- next year the same guy will be using statistics to tell us to do something else. That's because you can slice statistics in an almost infinite amount of ways, and I can guarantee you that one of those ways will show some kind of imbalance or injustice that needs addressing. That's all fine and dandy if the entire country could turn on a dime. I wish it could but it can't. Not reality. Instead we live in a system where it takes decades to make even one change, and it's extremely rare to stop doing things that don't work -- we just do more new stuff. I hate to rag on the IRS, but take a look at the tax code. It's so bad that I'm willing to wager nobody alive knows how to file taxes exactly correct. This puts us in an infinite cycle of where we're always measuring things with statistics, finding some social cause worth pursuing, painfully making changes, then making more measurements. Each time through the loop the system gets more and more unwieldy and complex. It's a feedback loop that increases government dysfunction. So instead of arguing about one cause or another, I'd like to suggest a rule for such discussions: if you want to make a case for some kind of social policy change, then make it in terms of a trade-off from whatever the existing status quo is. I realize there are a lot of problems with this suggestion, but at least it starts putting things in terms of trade-offs instead of wish-lists. Just from looking around, it seems to me that the real world is a lot more about trade-offs and compromises than moral crusades and justice. But that's just me. The only thing that I am observing is that this pattern can be repeated indefinitely, and each time through the loop conditions will continue to deteriorate. It's reactive leadership by statistics, and that almost never works.
- nihilocrat 16y agoOn one hand you are suggesting "we can't know anything, let's not bother" but on the other hand you are saying "let's talk about things we should actually do". Ok! I moved to Canada (Quebec) from the US. The healthcare system is overloaded, there are weird things like a) it's extremely hard to get a family doctor, b) you have to go to a clinic early in the morning to get a number + time of day and then come back at that time to be seen, and c) you still need to pay your own way (or get insurance) for dental / vision. You have to wait weeks/months to have a surgery, but you have to do that in the US and then promptly file for personal bankruptcy. That's still a whole hell of a lot better than a random unexpected illness or injury sending an employed, insured person into bankruptcy. Seeing a doctor is now a possibility instead of a dangerous worst-case scenario that could ruin my life. No more hoping the pain/infection/whatever just goes away on its own. Thus, I still think moving towards a government program / single payer / whatever you want to call it system is the best idea. Now, I based this all on personal anecdotal experience and not statistics, so I wonder if anyone will actually agree with me. :)