4 ms·

"If you smoke, your chances of getting cancer are only 3 in 10!" is not actually very comforting.

by rafram 3y ago

max_ 3y ago
I don't mean it's comforting I prefer to have it at 0/10. I am just trying to say that things that cause cancer are not as deterministic as we think.
- panabee 3y ago
  this is correct. it's clear that smoking elevates cancer risk, but why doesn't it cause cancer in all smokers? to develop a cure, we must better understand the causal mechanisms. this starts with acknowledging what we know and don't know about a devilishly complex disease that is arguably better conceptualized as a broad category rather than one monolith -- similar to how the flu, cold, and covid could be grouped under one mega classification, but are better identified as distinct conditions.
  - s1artibartfast 3y ago
    I think we understand the causal mechanism pretty well. it is just laypeople struggle with binary thinking vs probability. The reason why only 20% of smokers get cancer is similar to why a person doesn't get cancer after 1 cigarette. This is only counterintuitive if your default thinking is that smoking=cancer. In reality, there are a lot of variable chemical and biological processes involved, but ultimately it ultimately boils down to a cumulative risk, not guarantee.
    - pokerface_86 3y ago
      [dead]
jejeyyy77 3y ago
20% of lung cancer cases in the US are people who have never smoked.
- hackeraccount 3y ago
  Poor Stephen Jay Gould. :-(
d1sxeyes 3y ago
Also worth remembering that cancer isn’t the only bad thing smoking does to you. 7/10 smokers (2/3) die of a smoking related illness.
- jjgreen 3y ago
  Any idea of what percentage of non-smokers die of a smoking related illness?
  - d1sxeyes 3y ago
    Short answer: no. Longer answer: Logically, the answer should be zero, but it's actually not that simple to categorise either 'non-smokers' or 'a smoking-related illness'. Taking lung cancer as an example, approximately 10-20% of lung cancers occur in folks who are non-smokers (defined as 'never' or 'fewer than 100 cigarettes in their lifetime'). Researchers estimate that of these, around 20% are caused by second hand smoking, meaning that around 8-16% of lung cancers are non 'smoking related' in non-smokers. Not all lung cancers result in death. So if we take the same approach, and consider the 80% of lung cancers which occur in smokers, presumably similarly 10-20% of those would have developed lung cancer regardless of whether they smoked or not, so potentially we would need to say that less than 65% of lung cancers are caused by smoking. Is it then fair to call lung cancer a 'smoking-related illness', if more than a third of people who develop lung cancer did not do so because they themselves smoked[0] (the number changes to 70% if we consder the folks who hypothetically develop lung cancer due to second-hand smoking as 'smoking-related'). There are also a lot a problems with the methodology here. How do we know whether someone really smoked 100 cigarettes or fewer? Is there any statistically significant difference between people who smoked 100 and 1 000? 100 000? Does it matter if someone smoked all of those 10 000 in the last year (almost 30 a day) or whether they smoked those 10 000 over the past 50 years (less than a pack a month)? All of these things seem to my mind to be problematic and make it impossible to answer your question. I suppose that means that I was overconfident in my assertion in the parent comment that '7/10 smokers die of smoking-related illnesses'. It was not unresearched (although uncited) and I got this number from this article: https://www.washingtonpost.com/news/wonk/wp/2015/02/26/the-terrifying-rate-at-which-smokers-die-from-smoking/ https://www.washingtonpost.com/news/wonk/wp/2015/02/26/the-t..., which in turn reported on this study: https://bmcmedicine.biomedcentral.com/articles/10.1186/s12916-015-0281-z https://bmcmedicine.biomedcentral.com/articles/10.1186/s1291.... However, perhaps I wasn't critical enough in my reading, and instead simply parroted the stat. [0]: big caveat here is that my ramblings don't consider the possibility that a 50 year-old who would otherwise have developed lung cancer at 80 develops lung cancer as a result of smoking, and lung cancers caused by smoking stop lung cancers developing from other causes.

"If you smoke, your chances of getting cancer are only 3 in 10!" is not actually very comforting.

by rafram 3y ago

"If you smoke, your chances of getting cancer are only 3 in 10!" is not actually very comforting.

max_ 3y ago

I don't mean it's comforting I prefer to have it at 0/10. I am just trying to say that things that cause cancer are not as deterministic as we think.

panabee 3y ago
this is correct. it's clear that smoking elevates cancer risk, but why doesn't it cause cancer in all smokers? to develop a cure, we must better understand the causal mechanisms. this starts with acknowledging what we know and don't know about a devilishly complex disease that is arguably better conceptualized as a broad category rather than one monolith -- similar to how the flu, cold, and covid could be grouped under one mega classification, but are better identified as distinct conditions.
- s1artibartfast 3y ago
  I think we understand the causal mechanism pretty well. it is just laypeople struggle with binary thinking vs probability. The reason why only 20% of smokers get cancer is similar to why a person doesn't get cancer after 1 cigarette. This is only counterintuitive if your default thinking is that smoking=cancer. In reality, there are a lot of variable chemical and biological processes involved, but ultimately it ultimately boils down to a cumulative risk, not guarantee.
  - pokerface_86 3y ago
    [dead]

jejeyyy77 3y ago

20% of lung cancer cases in the US are people who have never smoked.

hackeraccount 3y ago
Poor Stephen Jay Gould. :-(

d1sxeyes 3y ago

Also worth remembering that cancer isn’t the only bad thing smoking does to you. 7/10 smokers (2/3) die of a smoking related illness.

jjgreen 3y ago
Any idea of what percentage of non-smokers die of a smoking related illness?
- d1sxeyes 3y ago
  Short answer: no. Longer answer: Logically, the answer should be zero, but it's actually not that simple to categorise either 'non-smokers' or 'a smoking-related illness'. Taking lung cancer as an example, approximately 10-20% of lung cancers occur in folks who are non-smokers (defined as 'never' or 'fewer than 100 cigarettes in their lifetime'). Researchers estimate that of these, around 20% are caused by second hand smoking, meaning that around 8-16% of lung cancers are non 'smoking related' in non-smokers. Not all lung cancers result in death. So if we take the same approach, and consider the 80% of lung cancers which occur in smokers, presumably similarly 10-20% of those would have developed lung cancer regardless of whether they smoked or not, so potentially we would need to say that less than 65% of lung cancers are caused by smoking. Is it then fair to call lung cancer a 'smoking-related illness', if more than a third of people who develop lung cancer did not do so because they themselves smoked[0] (the number changes to 70% if we consder the folks who hypothetically develop lung cancer due to second-hand smoking as 'smoking-related'). There are also a lot a problems with the methodology here. How do we know whether someone really smoked 100 cigarettes or fewer? Is there any statistically significant difference between people who smoked 100 and 1 000? 100 000? Does it matter if someone smoked all of those 10 000 in the last year (almost 30 a day) or whether they smoked those 10 000 over the past 50 years (less than a pack a month)? All of these things seem to my mind to be problematic and make it impossible to answer your question. I suppose that means that I was overconfident in my assertion in the parent comment that '7/10 smokers die of smoking-related illnesses'. It was not unresearched (although uncited) and I got this number from this article: https://www.washingtonpost.com/news/wonk/wp/2015/02/26/the-terrifying-rate-at-which-smokers-die-from-smoking/ https://www.washingtonpost.com/news/wonk/wp/2015/02/26/the-t..., which in turn reported on this study: https://bmcmedicine.biomedcentral.com/articles/10.1186/s12916-015-0281-z https://bmcmedicine.biomedcentral.com/articles/10.1186/s1291.... However, perhaps I wasn't critical enough in my reading, and instead simply parroted the stat. [0]: big caveat here is that my ramblings don't consider the possibility that a 50 year-old who would otherwise have developed lung cancer at 80 develops lung cancer as a result of smoking, and lung cancers caused by smoking stop lung cancers developing from other causes.