3 ms·

>They start by assuming a causal relationship that we simply don't know actually exists They assume this causal relationship based on decades of randomised con

>They start by assuming a causal relationship that we simply don't know actually exists They assume this causal relationship based on decades of randomised controlled trials (the only ethical way to get this information in humans). Their estimate for the risk of cancer based on alcohol consumption comes from previous work which would have carefully summarised the data from these trials and conservatively estimated the risk. As for the counter that alcohol may reduce cancer risk. Sure, that is true. You can also demonstrate that sunscreen causes cancer. But does the increased cancer risk from sunscreen outweigh the risk of getting too much sun exposure? Probably not (keep wearing sunscreen!). It is easy to imagine a similar equation for alcohol.

maxerickson 5y ago
There are people that drink what the study says to drink, every day? What else would 'controlled' mean in this context?
- ImaCake 5y ago
  >There are people that drink what the study says to drink, every day? Not sure, honestly. They seem to pull this data from a WHO database on health statistics [0]. In stricter terms, you are asking if their estimate of what people drink reflects what they really drink in real life. That is, is the estimate biased? I would expect it to be biased in this case. I am not sure how much it really matters unless you are in charge of large sums of money for public health policy. The best evidence we have suggests alcohol does increase the risk of cancer. Faulty population level alcohol consumption data is not being used for that observation, so it won't matter. Although it would impact the quality of this particular paper. "Controlled" doesn't appear in the original paper so not sure what you mean there. 0. https://www.who.int/data/gho/data/themes/global-information-system-on-alcohol-and-health https://www.who.int/data/gho/data/themes/global-information-...
  - maxerickson 5y ago
    You plainly stated that evidence about alcohol use comes from randomized controlled trials. It doesn't, it all comes from observational studies.
TaupeRanger 5y ago
It's really not so simple, even for sunscreen: https://www.outsideonline.com/health/wellness/sunscreen-sun-exposure-skin-cancer-science/ https://www.outsideonline.com/health/wellness/sunscreen-sun-...

maxerickson 5y ago

There are people that drink what the study says to drink, every day? What else would 'controlled' mean in this context?

ImaCake 5y ago
>There are people that drink what the study says to drink, every day? Not sure, honestly. They seem to pull this data from a WHO database on health statistics [0]. In stricter terms, you are asking if their estimate of what people drink reflects what they really drink in real life. That is, is the estimate biased? I would expect it to be biased in this case. I am not sure how much it really matters unless you are in charge of large sums of money for public health policy. The best evidence we have suggests alcohol does increase the risk of cancer. Faulty population level alcohol consumption data is not being used for that observation, so it won't matter. Although it would impact the quality of this particular paper. "Controlled" doesn't appear in the original paper so not sure what you mean there. 0. https://www.who.int/data/gho/data/themes/global-information-system-on-alcohol-and-health https://www.who.int/data/gho/data/themes/global-information-...
- maxerickson 5y ago
  You plainly stated that evidence about alcohol use comes from randomized controlled trials. It doesn't, it all comes from observational studies.

TaupeRanger 5y ago

It's really not so simple, even for sunscreen: https://www.outsideonline.com/health/wellness/sunscreen-sun-exposure-skin-cancer-science/ https://www.outsideonline.com/health/wellness/sunscreen-sun-...