"Life Expectancy" Doesn't Measure How Long You're Expected To Live | Hacker News Reader

"Life Expectancy" Doesn't Measure How Long You're Expected To Live

blog.edwardmarks.com

155 pointselmarks88 comments

This is a great post. Because I'm a nerd, I'm compelled to make an orthogonal point:

Life expectancy is not entirely medical and is not a particularly good way to compare countries. Consider that among the top causes of death in the US are "accidents" (predominantly motor vehicles) and suicide, both of which are anomalously high in the US.

Consider also that if you go to the top US cause of death (heart disease, clearly medical) and then compare the world rankings for heart disease deaths, you'll find many of the countries that beat the US in the life expectancy rankings do worse than the US in the heart disease mortality rankings.

You'll tend to see similar effects for other medical issues, particularly cancer, in which the US notoriously outperforms many of the European countries that outrank it on life expectancy.

Slight nitpick - accidents are not a top cause of death in the US, they are merely a top cause of early death.

But this has a disproportionate effect on low life expectancy, because early death has a disproportionate effect on life expectancy stats. 1 person dying at age 20 (instead of 80) has the same effect on life expectancy as 60 people dying at age 79 (instead of 80).

I have my stats memorized. Accidents are at the very least in the top ten cause of death every year. Every year, around 100K people die in accidents, of which 40K were car accidents. However, that is dwarfed by 600K people dying of heart disease and 570K people dying of cancer every year.

Not sure how we get to the word "dwarfed" in that analysis, since despite ranking below Europe on life expectancy, the US does significantly better on heart disease and cancer mortality.

Point being: something's going on here that confounds life expectancy comparisons.

I didn't disagree with your analysis. I was only pointing out that accidents(ranked 5th) are indeed one of the top cause of death in the US, which I didn't make it clear in my last comment.

I believe we also count infant mortality in life expectancy. Not sure.

Your statistics are simply wrong. Which is why they numbers don't seem to add up correctly. US has a mixed heath-care system with some people reviving enough treatment that it actually lowers there chances of survival and others getting little to no treatment.

PS: It's hard to get good data. Japan has twice the US suicide rate and a longer life expectancy. However, some of their numbers are off due to people not reporting deaths so they can continue to collect pension benefits.

Is there data to back up your argument? What you're saying seems to amount to this:

If you're wealthy, the US health care system is so good that not only does it outdo Europe's system at handling heart disease and cancer, but it does so much better than Europe's that the effect is evident from mortality statistics even though 75MM people in the US make less than $22,000/yr. At the same time, you're saying that regardless of the fact that US hospitals are required by law to treat indigent patients regardless of cost, enough people receive so little care that they bias the mortality statistics.

This seems like an extraordinary claim.

I am not an apologist for the US health care system, as a cursory look at HNSearch will show you. But I don't think the problem is that people don't get care in the US; it is, as policy wonks will tell you, better (from a health care perspective) to be homeless in 2012 than it was to be President in 1955. The problem is that the US health care system routinely bankrupts patients.

Your assuming at the lower end of the scale are accurately diagnosed.

However, for a simple idea of how bad the US healthcare system really is look at this: http://www.cddep.org/tools/methicillin_resistant_staphylococ...

In 2007 there were 8,324 deaths linked to MRSA in the UK. In 2005, there were between 4,429-8,850 (6,639 w/ 95% CI) deaths linked to MRSA in the US. There are more than 5 times as many people in the United States than in the UK.

I obviously do not have the specific stats at my fingertips to refute the anecdote you've supplied, but if there were a way for us to bet on which of our arguments the correct statistic is going to support, I'd bet on my argument.

Where do you get those numbers? MRSA is thought to have caused 1,652 deaths in 2006 in UK up from 51 in 1993.[104] http://en.wikipedia.org/wiki/Methicillin-resistant_Staphyloc...

Still, for those numbers to be useful the rate of infection, detection, treatment, death, AND linking deaths to MRSA must be considered before you can compare those healthcare systems as well as a near constant rare of infection.

Sorry, I got Clostridium confused with MRSA. But 1,652 normalized for population is still higher than the US MRSA fatality number!

I'm not sure what the second sentence means. If you're saying we don't know how to properly attribute deaths to MRSA, what does any MRSA statistic say about health care?

You have a long row to hoe with the overall argument you're making. It is, for instance, not hard to link deaths to heart disease, and heart disease is the leading cause of death in the US. The US has fewer heart disease deaths per 100,000 people than Austria, Sweden, Norway, Iceland, the UK, Finland, the Czech Republic, Ireland, Hungary, and Slovakia, and is closer to Germany and Denmark, the #14 and #15 followers to the US's #12, than it is to Austria's #11.

Yes, I am saying in many cases we don't know cause of death. In most of the US we do a crap job of discovering cause of death in many cases. When you dig into the numbers specific coroners often have different old guy died in sleep dumping grounds. 'Hart disease' often ends up as a grab bag for any number of quiet killers such as: http://en.wikipedia.org/wiki/Pulmonary_Embolism.

PS: Perhaps the most humorous being http://en.wikipedia.org/wiki/Fan_death, but the sadist and most blunt is probably: http://en.wikipedia.org/wiki/Sudden_infant_death_syndrome which is literally 'death without obvious cause'.

No, he's not assuming anything like that. He's observing that if the US is slightly worse on average (for the sake of argument), but you have what you claim are a large number of people being actively harmed by the system, then to make up for it you must have a large number of being receiving unbelievably good outcomes, or you can't end up with "slightly worse on average" in the end. It's a very simple mathematical point.

I'd observe that in general, the complaint with the US medical system is that it is too expensive, or that for what we pour into it it ought to be clearly the best everywhere across all measures instead of merely near the top. (Which is rather more accurate; it isn't "slightly below average", it's "slightly behind best", and there are rather a lot of individual measures in which it is the best.) It isn't that it's a terrible system in general. You have to game stats pretty hard to make it an actively bad system in terms of raw outcome.

Simpson's Paradox says you can have rich people country A outperform rich people in country B and poor people in country A outperform poor people in country B and have country B outperform country A when both groups are put together.

This is probably not the case, but since this is hacker news I'm being a dork about the math.

The problem is that the US health care system routinely bankrupts patients.

This is unclear as well. This stylized "fact" only became a talking point in the last election, and is drawn almost exclusively from a deeply flawed study by Elizabeth Warren.

http://www.theatlantic.com/business/archive/2009/06/elizabet...

http://www.theatlantic.com/business/archive/2010/12/will-hea...

Among other flaws, the study doesn't measure causation or even correlation.

Here is a list, I feel like it's fairly representative, of conditions that will generate an "automatic decline" for a regional health insurer. I chose it because it was the first on the Google SERP; I can be more rigorous if I need to be:

http://echealthinsurance.com/health-insurance-advisor/wp-con...

The fun stuff starts on age 18.

A huge number of Americans have conditions with a "D" in the right-hand column, which is "automatic decline coverage". Many of those people cannot get insurance on the private market at any cost.

Why do you have a hard time believing that our health care system needlessly bankrupts people?

Is it something along the lines of, "there is a savvy way to negotiate this system without having health insurance"? I'm prepared to concede that someone knows someone who's mom paid pennies on the dollar for care by playing hardball with providers. I just don't think that's a viable solution to the problem overall; providers will more often than not just send the bills to collections.

I submit that the system does needlessly bankrupt some people, but there are also a number of alternatives which allow a large percentage of the uninsured to avoid such a fate.

If I may submit two personal stories in place of rigorous data:

my incredibly poor sister recently injured herself and was facing nominal charges of about $40,000. She was able to get that written down to a few grand and then get donations from family, friends, and charities to cover her bills. She didn't have to "play hardball"; she just explained her situation and the hospital bent over backwards to help her. The hospital staff treated it as a routine occurrence.

My wife and I do not have health insurance at all. We use something called "Christian Healthcare Ministries" [0] which is not insurance, and (at our level of participation) doesn't cover routine visits or small problems, but covers us in case of conditions that might otherwise bankrupt us. So while I personally cannot get insurance on the open market, and routinely get counted in the "not insured" category, I still have coverage.

(Note that I'm not defending the US health care system overall, just introducing some information about alternatives that help people avoid bankruptcy.)

[0] https://www.chministries.org/

They might not call it heath insurance, but it still is de facto heath insurance. Also, nominal charges of 40k is small potatoes in the healthcare world, that can literally be a few days of moderate care in some hospitals. However, a lot of medical bankruptcys are around end of life care, where medicare forces people to 'go broke' before picking up the rest of the tab.

PS: I have awesome medical coverage, that limits out of pocket expenses to 2k per year with an unlimited sealing. However, many plans cap lifetime expenditures to 1million in coverage because people really do spend that much and far more.

> "it still is de facto heath insurance"

But without "automatic decline" conditions, which were critical to tptacek's point.

> "nominal charges of 40k is small potatoes"

Sure. But the cover story of CHM's latest newsletter [0] is a woman with a $300k bill that was reduced to $20k. Elsewhere in the newsletter is a request for donations to cover about $60k of a bill that has had $200k of reductions. There are, occasionally, bills that break the $1 million mark, though they usually come with reductions in the $500k+ range. Point being, very substantial reductions and charitable donations are common.

Again, I'm not specifically defending the status quo. I'm just saying, there are definitely options that allow some people with pre-existing conditions or catastrophic illness to avoid bankruptcy.

[0] https://www.chministries.org/downloads/newsletters/CHMSeptem...

Why do you have a hard time believing that our health care system needlessly bankrupts people?

I'd like to see good evidence that it's a significant problem. You haven't provided any. No one I've asked has provided any, beyond Elizabeth Warren's study and the claim that it's "obvious".

As for "savvy way", it's paying in cash, up front. Prices magically drop when payment is convenient. In my experience, and those of the many uninsured people I know, it's that simple. I'm going to speculate that you don't know very many uninsured people...

(Sadly, there is little data on this. If you have some, I'd love to see it.)

> Every year, around 100K people die in accidents, of which 40K were car accidents.

I don't know about overall accidents, and I don't have numbers memorized, but I know that vehicle accident deaths have dropped pretty significantly over the last couple of decades. Trends are relevant, too.

Well, then the upcoming self-driving vehicles will give us another bump up in the life expectancy figure.

Probably some part of the vehicle deaths will "leak" over to other categories, but there should be an overall reduction.

I haven't studied the data, but I would be surprised if motor vehicle deaths were "anomalously" high. My hypothesis would be that it is an entirely expected outcome given the comparative lack of public transportation in this country relative to other industrialized nations such as Japan and those in Europe. More driving miles (or, perhaps, more hours spent driving) per capita should lead to more motor vehicle accidents. And if the streets and highways are more crowded than in other countries, it might scale faster than linearly.

Though if what you meant was that motor vehicle accidents were not a good reflection of the quality of our medical care, I expect that would be a fair assessment.

It's easier to maintain a dense public transportation system when the system's goals are (for instance) to move 80MM people around in an area the size of Colorado. Colorado, by comparison, has a population of 5MM.

But yes, I was making the latter point.

Does life expectancy tell us something about the differences between life in the United States and Finland? Yes... but what if all it's telling us is "the United States is simultaneously less dense and comparably urban to Europe"?

I wouldn't expect, say, Wyoming to have a well developed public transportation system, or even Colorado as a whole. But once you get past a fairly small number of major cities, our public transportation systems for metropolitan areas are still underdeveloped.

Colorado's front range, for example, would almost certainly benefit from having a well-developed commuter rail system, with heavy passenger rail connecting the corridor from Colorado Springs to Fort Collins (with a spur to Boulder) and light rail taking people from the terminals to points spread further out.

Doing something like this would take a long time and cost a lot of money, but our choice (so far) not to spend that money in most metro areas has a direct effect on the quality of life in this country.

Fortunately, it isn't wholly hopeless in this country.

The Front Range is indeed making a large investment in commuter rail, though not yet all the way to Colorado Springs & Fort Collins: http://www.rtd-fastracks.com/systemmap.php

The Los Angeles area is likewise making significant investments in its rail system.

You picked an interesting comparison -- Finland is one of the rare European countries that is actually less dense than the US. The population density of the whole of US is ~34/km², while Finland clocks in only 16/km². Even when you account for the fact that the northernmost third of Finland is essentially unpopulated, Finnish population density is still well under the American one. (And I quess you guys have the Rockies, Alaska and a few deserts too...)

I picked Finland out of a hat, but the numbers I used were for Germany.

Do most people in Colorado commute from a corner of the state to the other? I would expect most people to live in or near the city where they work, just as in Europe. Having hundreds of miles of desert or farmland around a city does not make it any more difficult to build a metro system, for instance.

There are very real economic reasons to build out rather than up, when you have suitable land nearby. Especially in areas like Colorado where land is cheap and regulation is light, building nice big cheap (in terms of construction costs) single family homes sounds pretty good, vs. building expensive (per square foot of usable space) high-rise condos. Construction costs are lower, and you end up with a bunch more living space, a garage, etc, etc. And from a buyer's perspective, you have a lot more control over a single family home.

And yeah; you can have public transit in the 'burbs? but it's not going to be popular. Because you have so much more living space than a city dweller, you are likely to get more utility out of the hauling capacity of a car. And having that extra living space also dramatically lowers the cost of owning the car; You don't have to pay for parking, and you can do basic work in the garage. (you can save a giant wad of cash doing basic stuff like swapping rotors yourself.)

Personally, I think the economic realities of the suburbs mean that only the poorest of the poor are going to not own personal vehicles; because of this, public transit is built for and associated with the poor in all but the most built-up parts of America. "Domestic help" is not fashionable here, either, so for the middle class, there isn't a lot of upside to letting poor people into their neighbourhoods. (I'm wondering if this is going to change with the rise of in-home elder-care. There certainly are plenty of Americans willing to do that sort of work for wages the middle-class could pay.)

I would be surprised if motor vehicle deaths were "anomalously" high

Depends. Does getting shot while driving count? ;-)

Some other factors to think about:

Road quality in the United states is low compared to other OPEC countries

Seatbelts and motorbike helmets are still not mandatory in all US states

Car quality and safety in the United States is lower than other OPEC countries

EDIT: I meant OECD, not OPEC.

In only one state in the US are seatbelts not mandatory for adults (New Hampshire). Otherwise, the only differences are between states in which you can be pulled over for being observed not wearing a seatbelt (the "primary enforcement" states) and those in which you can only be issued a seatbelt ticket after being pulled over for some other reason.

I also think the road and car quality questions are up for debate.

Aren't there a lot of states where back seat passengers don't have to wear seat belts, and people can ride in the back of pick-ups etc.

Obviously road and car quality are up for debate, but once you spend some time driving around Europe in European cars, you'll see what I mean.

For every 1 passenger killed in the US, almost 3 drivers are killed. Given that the mean occupancy per car in the US is something between 1.3 and 1.7, it seems very unlikely that back seat restraints are a significant contributor to these statistics.

It's exactly that kind of thinking that explains why road deaths in the US are high (the original topic of this thread).

Whenever a stat says something interesting that could be improved about America, it seems Americans find a way of dismissing the data as "not really appropriate for reasons x,y,z", thus leaving the original problem unimproved.

I'm sorry, 'grecy, I've been dismissive of you in this thread, but I simply don't understand what you're trying to say here. It is evident from the statistics that lack of mandatory rear passenger seat belts are probably not a significant cause of death in the US.

If you're just trying to make a case that we should mandate rear passenger seatbelts, I'm not arguing with you. I'm talking about life expectancy stats.

I thought OPEC was the oil producers club, and that the US isn't a member. Did I miss something?

Sorry, confused my acronyms.

OECD

[WARNING: Egregiously incorrect comment follows]

In other words, you think car quality in the US is lower than that of countries like Tajikistan, Swaziland, Kazakhstan, and Ecuador. When you say "OECD", you might as well say "the entire world". I officially doubt your statistics.

No.

http://www.oecd.org/general/listofoecdmembercountries-ratifi...

You win this time, guy who can read a web page properly! :)

You'll tend to see similar effects for other medical issues, particularly cancer, in which the US notoriously outperforms many of the European countries that outrank it on life expectancy.

FWIW, I looked into this using the tables on http://www.worldlifeexpectancy.com/cause-of-death/all-cancer... - not just the all cancers table, but also the separate tables for different types of cancer. The implication of what you write - that the US has better medical healthcare - didn't seem to stand out. Rather, it looked like different European countries have markedly different death rates from different cancers. Things like diet, lifestyle, prevalence of smoking, etc. seem like a better explanation for the variance. France has especially low heart disease deaths, for example, but slightly higher cancer deaths than the US. Etc.

And of course we all die of something, so I would expect cancer deaths to be higher in a country with a higher life expectancy even if the medical success in treatment was higher. Third world countries generally do not have high deaths from cancer.

The United States ranks #56 (higher numbers are better), ahead of basically the entire EU, on colorectal cancer mortality.

The United States ranks #41, ahead of much of the EU but behind Germany and France, on ovarian cancer.

The United States ranks #61, again ahead of almost all of the EU, on breast cancer.

Here I might point out that to come in much higher than the US in these numbers, ie, to be Gabon, you have to have a lot of people dying before they can get cancer. Moving on:

The United States ranks #27 on leukemia, besting France, Poland, the Czech Republic, and Hungary.

The United States is #170 on stomach cancer, ahead of all of Europe.

The United States does worse than Europe on lung cancer, and is right in the middle of the pack on skin cancer. Other than that, the narrative is pretty clear.

But even by what you write, it's not convincing, though, is it? Behind some major countries on a number of common cancers, but in the lead on others. I mean, it's clear that the US is not lagging. But it's not a convincing lead either. And you need to look at the variances. From looking at the per 100K deaths, many of the US leads are in the region of 10%, but the variances between different European countries is much larger than that.

But the US is definitely in the lead on breast cancer. No doubt about that.

Correlate the cancers in which the US is ahead of 2 out of 3 of France, Germany, and the UK, to the cancers that kill the most people (incidence and mortality). Pull out lung cancer, for which there is a clear causal explanation.

Now compare that result to life expectancy rankings. That's the point I'm making.

Do Europeans really smoke less than Americans? I was under the impression that the US was unusually anti-smoking.

Cigarette advertising is still legal in the USA right? And a packet of cigarettes costs less than an hours work at minimum wage because the taxes aren't that high, right?

No, and no. Also, it's invariably illegal to smoke indoors or within 10 feet of a building entrance. And there is constant anti-smoking propaganda. And people on the street will make negative comments about you smoking.

The other thing to consider is how the causes of death interact, and who it is that's contributing to mortality/survival rates. For example, perhaps people in France have higher cancer mortality rates because they are not dying from heart disease at a younger age. Perhaps the US has better survival rates for cancer because the incidence rates are higher among lower age groups, and people's general health is better at those ages, and so on.

You may mean homicide rather than suicide: the US doesn't seem especially high in this ranking of suicide deaths -- http://en.wikipedia.org/wiki/List_of_countries_by_suicide_ra... -- below many other rich or large countries, and not far from culturally/developmentally similar countries (like Canada/NZ/UK). The US does stick out among its usual developmental peers on the homicide rates -- http://en.wikipedia.org/wiki/List_of_countries_by_intentiona... -- doubling or tripling our English-speaking or western-European peers.

Usually the way I hear this statistical point made is that the US leads the world in life expectancy when accidental injuries and violent crime are omitted. (Usually the point is deployed to caution people against making quick conclusions about US diet or medical care quality from national life expectancies, as for example here: http://www.forbes.com/sites/aroy/2011/11/23/the-myth-of-amer...)

That wikipedia link is interesting - almost in every single country, male suicides are much higher than female suicides, in some cases 4 to 5 times more.

Actually, female attempt rates are higher than male. There are a variety of factors that come into play but essentially females choose methods of suicide that are less effective so male rates ultimately end up higher.

I have heard it described as an attempt to get help versus actual deadly intent.

I'd hazard a guess it's due to two things, each not necessarily independent. First, males in many countries are prone to societal "honor" and may kill themselves in times of great shame (poor Indian farmers loosing everything to shady Monsanto salesmen and then drinking pesticide). And second, women are often mothers, and while a suicidal man tends to not "think of the children" women (being mothers and all) do consider the fates of their children, and likely forestall killing themselves out of compassion/concern for their offspring.

Given the shit the world generally dumps on women, I'd say the female gender is simply built of sterner emotional stuff than their male counterparts.

Plain and simple, across and countries and all historical periods, males act more dangerously. They put themselves in physically dangerous situations more often than women, and die in mid-life at a much higher rate. Male children are more likely to live until adolescence, but after their 20s their portion of the population falls back in line with that of females.

This sounds a lot like the idea former N.O.W. boardmember Warren Farrell wrote in Why Men Earn More:

that bias-based unequal pay for women is largely a myth, and that women are most often paid less than men not because they are discriminated against, but because they have made lifestyle choices that affect their ability to earn.

http://www.amazon.com/Why-Men-Earn-More-Startling/dp/0814472...

I buy that. However, we're talking about suicide, not dangerous behavior in general.

> Given the shit the world generally dumps on women,

that they actually get sympathy for, instead of being told to 'suck it up' and 'be a man' when they attempt to find help.

that they actually get sympathy for, instead of being told to 'suck it up' and 'be a man' when they attempt to find help.

This makes sense. I suppose women ask for help more, and sooner than men (and probably get it too). Does anyone know, what is the percentage of men vs women, who actually go to psychiatrists? That might give some idea

Surely the rate of death from both accidental injuries and violent crime is affected by access to emergency medical care, though? You can't disentangle the two that easily.

That's right. If trauma medicine was still at the levels of the 70's the US murder rate would be about five times higher than it is.

That quiet little phrase may be the most intriguing thing I read all year.

Looking at things like http://en.wikipedia.org/wiki/Crime_in_the_United_States, it implies that the US murder rate would have doubled, rather than halved.

I'd suggest it is worth a separate post.

>" Consider that among the top causes of death in the US are "accidents" (predominantly motor vehicles)"

If I have a higher-than-average chance of dying in an automobile accident in a specific country, that's good information to have. In fact, I want to know all the particularly common ways I could die somewhere, be it via car or mortar attack. I think that's a great way to compare countries and life expectancy. Maybe I miss your point?

I'm not saying life expectancy tells you nothing, just that it doesn't tell you what a lot of people seem to think it tells you. If you got sick and had to pick a country to be treated in solely based on your health outcome, you'd be making a mistake to choose based on life expectancy. In the same vein: if you were picking countries based on quality of public transit nationwide, sure, life expectancy could be a counterintuitive and convoluted way to arrive at the right conclusion. But a look at an Amtrak schedule would tell you pretty much the same thing without the tea leaf reading. Everyone knows that Europe and Japan have denser public transit than the US.

Maybe I am getting it wrong but does it matter how you die? Generally speaking (averaged out,) if you live in US and do what the average person does, you will get in accidents and deal with the pressures that push some to suicides (around 10 to 22 per 100,000 http://www.bloomberg.com/news/2011-04-14/suicide-rates-rise-... .) Even if you are careful, the recently licensed teenager, to stereotype, can ram his car into yours.

Yes, it matters a lot how you die. For instance, there are 2,900 hippopotamus-related fatalities worldwide every year. Zero of those occur in the US. If you compare the US to Subsaharan Africa by hippo fatality, all you learn is that Subsaharan Africa has hippos.

But they have less cars, to take an example. Dead is dead was my point and in whatever country you live, you will deal with the cars, lions, tundra or whatever.

Sure, and it does seem to be true: life expectancy in part tells you "this country has more cars than hippos". The car/hippo ratio is a very different concept than "quality and outcome of health care systems".

"Life expectancy is not entirely medical"

If by not entirely medical you mean overwhelmingly not medical, then yes. Of the 30 years that life expectancy increased in the 20th century, 25 of those years were do to non-medical factors according to the CDC.

Furthermore, if you actually look at the top 'medical' causes of death, they can overwhelmingly be prevented by non-medical means. To quote from Overdose In America, "Simply eating fish once a week reduces the risk of stroke by 22 percent. Controlling high blood pressure reduces the risk of stroke by 35 to 45 percent. And even moderate exercise for less than two hours a week reduces the risk of stroke in an elderly population by about 60 percent."

Furthermore, just 1.2% of Americans meet all 7 cardiovascular health metrics from 2005 to 2010, compared to 2% from 1988 to 1994. These metrics are not smoking, being physically active, having blood pressure under control, maintaining healthy blood glucose levels, maintaining healthy blood cholesterol levels, mainting a healthy body weight, and following a healthy and balanced diet.

Other than the cohort and period life expectancy, what is imp. to most people should be something else:

* How many more years do you expect to live if you are already X years old in 2012?

This is obviously different than just calculating from the cohort life expectancy of 2012-X since you you've already passed the vulnerable infancy stage. Also, say you are 90 today, so your cohort exp, for 1922 births must be something like, say 80. Your expectation of living more is certainly not -10 years!

A very interesting post indeed. Another common mistake many people seem to make is when comparing historical life expectancies. Most of our progress in prolonging human life is due to minimizing the infant mortality rate. This means that people who survived infancy and childhood could make it to an old age even in historical times.

This is true and a huge annoyance to me when people say "When social security was first put in place, most people didn't live to collect it!"

This is of course false. Back in the early 1900's when life expectancy was only around 60, if you survived your childhood, you'd likely live to 70 or 80.

Back in the early 1900's when life expectancy was only around 60, if you survived your childhood, you'd likely live to 70 or 80.

This didn't seem right so I checked. A 20 year old in 1900 could expect to live to 62. In 1939 (4 years after SS was first put in place) a 20 year old would live to 67.

Interpretted as a median age of death, you are right that more than 50% of 20 year-olds collected at least 1 check, but a significant fraction died before retirement age, and they generally didn't say in retirement for decades.

http://www.infoplease.com/ipa/A0005140.html I used white males, and women lived longer but non-whites lived shorter.

Thanks for looking up actual numbers!

I guess my point also has to do with what "averages" mean. In 1939, a 20 year old might _on average_ live to 67, but that means a number would die at 47, but a similar amount would live to 87.

It's also inappropriate to compare "infant mortality" across countries. A 490-gram baby that is born with trouble breathing and dies in 2 hours will count as a stillbirth in some countries and an infant mortality in others.

Also life expectancy changes as you age: http://www.ssa.gov/oact/STATS/table4c6.html

By the time you hit 60, you've gone from ~75 to ~80.

I don't like using averages because they can be misleading. A triplet of the 25th, 50th and 75th percentiles is more descriptive. Of course, the most complete picture is looking at the whole distribution.

I'd like to see a plot of the distribution of lifetime as a function of birth year. I.e. a 3D plot with current age on the X axis, age of death on the Z axis, and probability on the Y axis. (Where, as in Minecraft, X is east/west, Z is south/north.)

Cut off the portion representing people who have died already (Z < now-X), renormalize each cohort so its Y values add up to 1, and you'll get the distribution of lifetimes for people who are alive today. (I.e. a conditional distribution, conditioned on being alive today.) Which at least one other poster was asking about.

Just last week there was another Hacker News post that prompted me to look up some of the articles by demographer James W. Vaupel, one of the leading scholars on the demography of aging and how to adjust for time trends in life expectancy. His article

"Biodemography of human ageing" Nature. 2010 Mar 25;464(7288):536-42. doi:10.1038/nature08984

http://www.demographic-challenge.com/files/downloads/2eb51e2...

covers most of the essential issues. His striking finding is "Humans are living longer than ever before. In fact, newborn children in high-income countries can expect to live to more than 100 years. Starting in the mid-1800s, human longevity has increased dramatically and life expectancy is increasing by an average of six hours a day."

http://www.prb.org/Journalists/Webcasts/2010/humanlongevity....

Meanwhile, a person at any given age can look up period life tables for what the blog post author of the post submitted here correctly describes as a MINIMUM life expectancy at that person's current age.

http://www.ssa.gov/oact/STATS/table4c6.html

http://www.infoplease.com/ipa/A0005140.html

(The links shown are for United States data, but data like these are available for most developed countries.)

It is also possible to find life expectancy formulas adjusted for personal health status and lifestyle.

http://www.msrs.state.mn.us/info/Age_Cal.htmls

(Minnesota data)

http://www.bupa.com.au/health-and-wellness/tools-and-apps/to...

(Australia data)

The link posted last week

http://www.scientificamerican.com/article.cfm?id=longevity-w...

gives good data on trends in causes of death from 1960 to the present in the United States and in OECD countries as a whole. It showed that life expectancy at birth, at age 40, at age 60, at age 65, and at age 80 have all increased during the years shown on the chart.

The "rectangularisation of the life curve", i.e. a cohort facing low mortality (thanks to better social practices and healthcare) until a ripe old age when it starts dramatically dying off, implies a maximum human lifespan that we have not significantly manipulated in the history of human biology.

Thanks for making me realize I've been using life expectancy tables incorrectly for most of my life.

One comment regarding the yearly progress of the cohort rate. I would think this would fluctuate with significant medical advances (e.g. the first few years that bypass surgery/artificial hearts started to be used). I can't see how medical progression is perfectly linear at 1.0%/year.

I think you're overestimating the number of people who benefit from even a major advance.

In the UK (population 70 million), 28,000 heart bypass operations happen per year. The number of people who will have one is in the low single digits as a percentage of the population.

Even for those people who do have one there is no guarantee that it will have a significant impact on their lives. Yes for some it will but for others it's only reduced one possible cause of death and a lot of the behaviours that lead to heart problems also lead to, say cancer, or diabetes, or liver failure, or something else that will cut your life short.

Then compare that to the thousands of small improvements.

I can see it might not be constant over an extended period, but I don't think there are many, if any, things that would cause a significant spike.

> I can see it might not be constant over an extended period, but I don't think there are many, if any, things that would cause a significant spike.

While it's hard to predict, I personally wouldn't bet against a game changing spike in the coming decades. It's only very recently that medicine has started riding Moore's law. Robotics and miniaturization are speeding up basic research by orders of magnitude.

I don't think it's unreasonable to say that biology has achieved more in the past two decades than in all preceding history. And there's no sign that the exponential is running out.

Granted, there may be fundamental limits that we don't appreciate yet. That's the whole draw of science, we simply don't know.

Interestingly I think the one thing that would cause a spike happened outside the last two decades - the discovery / development of antibiotics.

You may be right but there will still be bottlenecks in the process around patient trials and the more "manual" stages of any treatment.

I think the next big improvement will likely not register on life expectancy because it will be in response to a specific problem. The next generation of anti-biotics for instance will have a massive impact but will likely only cancel out the negative impact of drug-resistant strains of bacteria.

It's not perfectly linear in the short term, though what tends to happen is a longer-term cyclical effect rather than spikes.

This makes sense - medical advances won't be widely adopted immediately, then there will be a gradual acceleration of adoption (and improvement rate) once its benefits are proven and costs come down, and then no further contribution to improvements once wide adoption is the norm.

What you also tend to see is that advances affect improvements by year of birth more prominently than by year of discovery. In the UK we have a 'golden cohort' for example, which you could Google for more info.

Spikes do tend to occur during and after big causes of death (spanish flu, world wars) as the improvement rate drops sharply and then recovers again.

Site seems to be down, Google cache: http://webcache.googleusercontent.com/search?q=cache%3Ablog....

When looking at these numbers, and thinking how the official retirement age in most countries has been raised to 67 or 69, does anyone else feel like old Boxer from Animal Farm [1] ?

They keep dangling that retirement carrot in front of us, but by the time we get to enjoy it, we're dead.

[1]en.wikipedia.org/wiki/Boxer_(Animal_Farm)

Reply on news.ycombinator.com