7 ms·

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

155 points by elmarks 14y ago

xm1994 14y ago
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.
- Tyrannosaurs 14y ago
  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.
  - ef4 14y ago
    > 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.
    - Tyrannosaurs 14y ago
      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.
- manifold 14y ago
  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.
tptacek 14y ago
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.
- yummyfajitas 14y ago
  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).
  - kiba 14y ago
    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.
    - tptacek 14y ago
      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.
      - 14y ago
baltcode 14y ago
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!
Groxx 14y ago
Site seems to be down, Google cache: http://webcache.googleusercontent.com/search?q=cache%3Ablog.edwardmarks.com%2Fpost%2F31337304505%2Flife-expectancy-doesnt-measure-how-long-youre http://webcache.googleusercontent.com/search?q=cache%3Ablog....
yesbabyyes 14y ago
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.
- refurb 14y ago
  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.
  - danielweber 14y ago
    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 http://www.infoplease.com/ipa/A0005140.html I used white males, and women lived longer but non-whites lived shorter.
    - refurb 14y ago
      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.
- danielweber 14y ago
  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.
eranki 14y ago
Also life expectancy changes as you age: http://www.ssa.gov/oact/STATS/table4c6.html http://www.ssa.gov/oact/STATS/table4c6.html By the time you hit 60, you've gone from ~75 to ~80.
tokenadult 14y ago
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/2eb51e2860ef54d218ce5ce19abe6a59/dc_biodemography_of_human_ageing_nature_2010_vaupel.pdf 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.aspx 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.ssa.gov/oact/STATS/table4c6.html http://www.infoplease.com/ipa/A0005140.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 http://www.msrs.state.mn.us/info/Age_Cal.htmls (Minnesota data) http://www.bupa.com.au/health-and-wellness/tools-and-apps/tools-and-calculators/quick-health-age-check 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-why-we-die-global-life-expectancy 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.
grecy 14y ago
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)
csense 14y ago
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.
JumpCrisscross 14y ago
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.

155 points by elmarks 14y ago

xm1994 14y ago

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.

Tyrannosaurs 14y ago
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.
- ef4 14y ago
  > 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.
  - Tyrannosaurs 14y ago
    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.
manifold 14y ago
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.

tptacek 14y ago

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.

yummyfajitas 14y ago
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).
- kiba 14y ago
  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.
  - tptacek 14y ago
    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.
    - 14y ago

baltcode 14y ago

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!

Groxx 14y ago

Site seems to be down, Google cache: http://webcache.googleusercontent.com/search?q=cache%3Ablog.edwardmarks.com%2Fpost%2F31337304505%2Flife-expectancy-doesnt-measure-how-long-youre http://webcache.googleusercontent.com/search?q=cache%3Ablog....

yesbabyyes 14y ago

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.

refurb 14y ago
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.
- danielweber 14y ago
  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 http://www.infoplease.com/ipa/A0005140.html I used white males, and women lived longer but non-whites lived shorter.
  - refurb 14y ago
    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.
danielweber 14y ago
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.

eranki 14y ago

Also life expectancy changes as you age: http://www.ssa.gov/oact/STATS/table4c6.html http://www.ssa.gov/oact/STATS/table4c6.html By the time you hit 60, you've gone from ~75 to ~80.

tokenadult 14y ago

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/2eb51e2860ef54d218ce5ce19abe6a59/dc_biodemography_of_human_ageing_nature_2010_vaupel.pdf 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.aspx 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.ssa.gov/oact/STATS/table4c6.html http://www.infoplease.com/ipa/A0005140.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 http://www.msrs.state.mn.us/info/Age_Cal.htmls (Minnesota data) http://www.bupa.com.au/health-and-wellness/tools-and-apps/tools-and-calculators/quick-health-age-check 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-why-we-die-global-life-expectancy 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.

grecy 14y ago

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)

csense 14y ago

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.

JumpCrisscross 14y ago

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.