Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 


Forgot your password?
Close
typodupeerror
×
Science

Researchers Are Creepily Close To Predicting When You're Going To Die (arstechnica.com) 77

Posted by BeauHD from the future-predictions dept.
An anonymous reader quotes a report from Ars Technica: If death is in the cards, it may also be in your blood. Measurements of 14 metabolic substances in blood were pretty good at predicting whether people were likely to die in the next five to 10 years. The data was published this week in Nature Communications. A team of researchers led by data scientists in the Netherlands came up with the fateful 14 based on data from 44,168 people, aged 18 to 109. The data included death records and measurements of 226 different substances in blood. Of the 44,168 people, 5,512 died during follow-up periods of nearly 17 years.

The researchers then put their death panel to the test. They used the 14 blood measurements to try to predict deaths in a cohort of 7,603 Finnish people who were surveyed in 1997. Of those Finns, 1,213 died during follow-up. Together, the 14 blood measurements were about 83% accurate at predicting the deaths that occurred within both five years and 10 years. The accuracy dropped to about 72% when predicting deaths for people over 60 years old, though.
This discussion has been archived. No new comments can be posted.

Researchers Are Creepily Close To Predicting When You're Going To Die More

Researchers Are Creepily Close To Predicting When You're Going To Die

Comments Filter:

  • There is a lot I can do with this information (Score:4, Insightful)

    by OrangeTide ( 124937 ) on Wednesday August 21, 2019 @10:37PM (#59111326) Homepage Journal

    With this I can purchase just enough whiskey to last me the rest of my life.

    • But whiskey is the main metabolic substance that they're testing for.

    • A death clock? What's next? Some sort of Fing-Longer? https://futurama.fandom.com/wi... [fandom.com]
    • Come on moderators, this is insightful. Even if it isn't whisky, enough pot to stay stoned till the end. Or what ever someone wants to be happy on the way out.

  • as I die with my boots on!

  • If you know you are going to die in 10 years, you don't want to waste the time you have working until the day you die. I would much rather enjoy my final years aging gracefully as I swear like a sailor at children on Fortnite. ;)

  • I want to know where I will die (Score:4, Insightful)

    by chrism238 ( 657741 ) on Wednesday August 21, 2019 @10:56PM (#59111354)
    ...then I'll just avoid going there.

    • What if the answer is "in bed"?

    • Re: (Score:2)

      by JeffTL ( 667728 )
      Stavromula Beta

    • Re: (Score:2)

      by dissy ( 172727 )

      I want to know where I will die ...then I'll just avoid going there.

      Oh, well that's super easy!
      The answer is: where ever you happen to be on March 20th 2023 at 10:25 am UTC

      PS, I had to borrow some blood for other purposes, sorry about that.
      PPS, but we had some of it tested for you, cheers!

    • "The Appointment in Samarra"
      (as retold by W. Somerset Maugham [1933])
      The speaker is Death

      There was a merchant in Bagdad who sent his servant to market to buy provisions and in a little while the servant came back, white and trembling, and said, Master, just now when I was in the marketplace I was jostled by a woman in the crowd and when I turned I saw it was Death that jostled me. She looked at me and made a threatening gesture, now, lend me your horse, and I will ride away from this city and avoid my fat

  • Recently, giant volcano in news, big asteroid, political turmoil, so How Soon?

    • Recently, giant volcano in news, big asteroid, political turmoil, so How Soon?

      So when I look at most politicians these days . . . and I mean all, internationally . . . I can only answer not soon enough!

      • Maybe if we fill the giant volcano with politicians it will cool down. Actually, this sounds like something that should be done proactively.

  • In TFA there is a single mention of using the data to "targeted prevention of mortality" but I would have thought that should be the focus on the next steps.

    It sounds like there hasn't been a correlation to the markers to the causes of death (which I would think would be pretty easy) and for things like cardio-pulmonary, cancer, diabetes and other diseases causing death then there should be approaches taken to avoid the expected deaths, resulting in longer lives and less expensive (to health care systems) disease treatments.

    The Dutch scientists have some useful data, let's see it being used to prevent early deaths.

    • Re:Rather than Predicting how about Preventing Dea (Score:4, Insightful)

      by phantomfive ( 622387 ) on Wednesday August 21, 2019 @11:57PM (#59111432) Journal
      Yeah, the headline is misleading. They aren't saying, "We know the exact day when you will die," they are saying, "we measure that you are unhealthy so you will die soon."

      And it's not creepy, it's normal.

    • The publication's discussion goes in detail about the markers and their meaning.

      A bunch of those are associated with treatable diseases such as Hypercholesterolemia, Diabetes Type 2, etc.

      (other are more generic, such as inflammation markers).

      You can see in a couple of years this publication being used as an argument to justify introduction of new public health policies for the ongoing efforts to better detect and treat the diseases.
      (at least here around in our Evil Eurocommunist countries. In the US, your f

  • Obligatory Woody Allen Quote: (Score:5, Funny)

    by mykepredko ( 40154 ) on Wednesday August 21, 2019 @11:01PM (#59111364) Homepage

    "I'm not afraid of dying, I just don't want to be there when it happens."

    • The Woody Allen quote I recall is "I don't want to be immortalized through my work. I want to be immortalized by not dying."

      • There are quite a few:
        - My relationship with death remains the same, I'm strongly against it.
        - It is impossible to experience one's death objectively and still carry a tune.
        - There are worse things in life than death. Have you ever spent an evening with an insurance salesman?
        - Eternal nothingness is fine if you happen to be dressed for it.
        - I don't believe in after life, but I am bringing a change of underwear.
        - I don't want to achieve immortality through my work. I want to achieve it through not dying.
        - Li

  • bullshit - worse than majority classifier (Score:4, Insightful)

    by someoneOtherThanMe ( 1387847 ) on Thursday August 22, 2019 @12:49AM (#59111474)

    Out of 7603 people, 1213 died. So if I just predict "you won't die", I'll be 84 % accurate. And similarly, I'll be less accurate for people over 60.

    • Out of 7603 people, 1213 died. So if I just predict "you won't die", I'll be 84 % accurate. And similarly, I'll be less accurate for people over 60.

      If you predict you won't die, that would make you dead wrong, not merely "less accurate". We humans are still batting 1000 when it comes to death, so let's not factor out the obvious.

    • Out of 7603 people, 1213 died. So if I just predict "you won't die", I'll be 84 % accurate. And similarly, I'll be less accurate for people over 60.

      Well I've never died once, so based on research I've done on myself in over 40 years of study, I must be immortal.

    • If you predict "you won't die" you will be 0% accurate

    • This is the most interesting graph from the study that in a way addresses your concern:
      https://www.nature.com/article... [nature.com]

      If I understand it correctly, the graph is a function of where you put the threshold for predicting death or not (which actually makes both axes into dependent variables).

    • Re: (Score:1)

      by atpage ( 941136 )
      I think the summary should not have used the term "accurate". They actually did get better than majority classification results. 0.83 was the AUC.

  • Found the cause (Score:5, Funny)

    by dromgodis ( 4533247 ) on Thursday August 22, 2019 @01:28AM (#59111508)

    Of those Finns, 1,213 died during follow-up.

    I would really suggest not going to those follow-ups.

  • To quote Ricky Gervais again (Score:3)

    by nospam007 ( 722110 ) * on Thursday August 22, 2019 @01:58AM (#59111524)

    "Being dead is like being stupid. It's only painful for the people around you."

  • ... without a bloodtest, just by looking at and watching people.

    20% accuracy in predicting a 5-10 year window of death? Seriously?

    Watch an old lady go and get a new hip, see her a year later with grey hair, an onset dementia and a general vibe of out-of-this-world and you know she's up soon on the gaunt mans list.

    I bet dollars to doughnuts that any caretaker of old people has a higher accuracy that this testing thing some desknerd came up with.

    My 2 cents.

    • ... without a bloodtest, just by looking at and watching people.

      20% accuracy in predicting a 5-10 year window of death? Seriously?

      Watch an old lady go and get a new hip, see her a year later with grey hair, an onset dementia and a general vibe of out-of-this-world and you know she's up soon on the gaunt mans list.

      I bet dollars to doughnuts that any caretaker of old people has a higher accuracy that this testing thing some desknerd came up with.

      My 2 cents.

      Nurses and Nurses aids, health care people in general have that sort of accuracy.

  • Sick and twisted insurance tool (Score:3)

    by geekmux ( 1040042 ) on Thursday August 22, 2019 @03:35AM (#59111622)

    This will be abused by insurance companies to deny coverage for certain claims. Why would they bother "wasting" their money covering a critical surgery when the new magical blood test gimmick shows that you'll likely be dead in five years? (The obvious answer is they won't.

    And given the accuracy isn't exactly in that 99% range, this is stupid, mean and downright dangerous to tell people this information. Consider your prediction is part of that 16% who were wrong in their analysis. Can you imagine the mental and psychological havoc that can easily manifest into real physical problems once a person is told their fate is sealed within the next decade? How many people might be sent off the mental deep end after being told they don't have much to live for because death is knocking on their door? Seems there's a lot of shit those sitting inside labs all day staring at Petri dishes don't think about.

    • Re: (Score:3)

      by Whibla ( 210729 )

      I agree that over-screening, especially without any explanation or contextualisation by medical professionals, is rarely a net positive. However...

      Consider your prediction is part of that 16% who were wrong in their analysis. Can you imagine the mental and psychological havoc that can easily manifest into real physical problems once a person is told their fate is sealed within the next decade?

      I did scan the full study but it was not immediately clear to me the difference between the false positive and the false negative %ages. Of course it's possible they were both 17%, but I'd guess that to be fairly unlikely. I'd also guess that different people would have varying opinions as to which is the worse situation, not being flagged as 'at risk' when you a

  • Test the blood and predict if 109 year old is going to die within next 10 years.
  • Two hunters are out in the woods when one of them collapses. He doesn't seem to be breathing and his eyes are glazed. The other guy whips out his phone and calls 911. He gasps, "My friend is dead! What can I do?" The operator says, "Calm down. I can help. First, let's make sure he's dead." There is a silence; then a gun shot is heard. Back on the phone, the guy says, "OK, now what?"
  • If you look at the abstract of the paper they improved from 77% prediction accuracy which is based on general data (e.g. age, sex, location, weight) to 79% prediction accuracy involving blood panel. No causal mechanism found for addition of 2%, pure statistics. Made a lot of headlines though.

  • So.... they guess a date and they consider it correct if you die in some 20 year range (+10 years -10 years of date picked). I wonder who wasted money on this pointless research.

    • no, there is no minus. they didn't include dead people. It was those that would die in next 5 to 10 years. that is very useful, not just for prediction but to tell people to change their lifestyle as they will be in mortal danger if they don't shape up and do the right things.

  • "The accuracy DROPPED to about 72%"
    does not look bad to me at all when forecasting life expectancy

  • I bet I could be 80% accurate if I had a +/- 20 year range. I'd just guess 75 for everyone and be pretty damn close.

  • The factors are fats, sugar, and amino acids. All are adjustable by diet, supplementation, and in the worst case moderately priced medication.

    I think that if you're knowledgeable enough in statistics (I'm not), you can infer optimum levels of these markers and modify your behavior accordingly. I wish the authors had presented the data in a way that made recommended levels obvious.

    Is there somebody here on slashdot that can figure out those levels?

  • https://www.baen.com/Chapters/... [baen.com]

    Watch out for those insurance company death squads!

    • Re: (Score:2)

      by shanen ( 462549 )

      Didn't Asimov do a version of this story, too? And there were probably versions before either or them.

      At a higher level, I see the topic in terms of dying boots-on or boots-off. These days boots-on deaths are increasingly rare.

Slashdot Top Deals

Business is a good game -- lots of competition and minimum of rules. You keep score with money. -- Nolan Bushnell, founder of Atari

Close