3 ms·

That's not a given, it's the first two predicates of an illustrative example of how dietary advice is more complicated than "which diet is best?" If all three o

by strken 4y ago

That's not a given, it's the first two predicates of an illustrative example of how dietary advice is more complicated than "which diet is best?" If all three of those predicates hold true, then your advice has killed them, therefore all-cause mortality rate is important. It doesn't matter what the exact reason for higher all-cause mortality rates on low carb diets is unless you can give people advice on how to avoid them. What's important is the outcome of giving the advice, not a theoretical ivory-tower diet that most of the population won't or can't eat.

Semaphor 4y ago
That’s like saying you should never give any advice unless you exhaustively list all possible ways for someone to fuck up the implementation. Telling someone the pharmacy is on the other side of the street will raise mortality unless you also tell them that they should check for cars before crossing. And not run into cars. Or walls close to the pharmacy.
- strken 4y ago
  No. You should give advice that causes good outcomes. You don't need to do anything else. We know that low carb diets have higher all-cause mortality. Because of this, telling people to eat low carb will (probably) cause a bad outcome. If we want a good outcome, we need to work out more specific advice. To be clear, advice doesn't always have to be very specific. This particular advice needs to be improved, because it looks like it would kill more of them than current dietary advice. We can tell it would kill them because the meta-analysis several comments above shows an increase in all cause mortality. This is not the same as telling people where the pharmacy is, because A) we have no studies at all about going to the pharmacy, B) people need to get important medicine from the pharmacy and going to it could plausibly lower mortality, and C) telling someone where to find a building is not lifelong public health advice and the risk for one person crossing one road is miniscule anyway.
  - Semaphor 4y ago
    So the part I objected to was not actually relevant for your argument.

by strken 4y ago

Semaphor 4y ago

That’s like saying you should never give any advice unless you exhaustively list all possible ways for someone to fuck up the implementation. Telling someone the pharmacy is on the other side of the street will raise mortality unless you also tell them that they should check for cars before crossing. And not run into cars. Or walls close to the pharmacy.

strken 4y ago
No. You should give advice that causes good outcomes. You don't need to do anything else. We know that low carb diets have higher all-cause mortality. Because of this, telling people to eat low carb will (probably) cause a bad outcome. If we want a good outcome, we need to work out more specific advice. To be clear, advice doesn't always have to be very specific. This particular advice needs to be improved, because it looks like it would kill more of them than current dietary advice. We can tell it would kill them because the meta-analysis several comments above shows an increase in all cause mortality. This is not the same as telling people where the pharmacy is, because A) we have no studies at all about going to the pharmacy, B) people need to get important medicine from the pharmacy and going to it could plausibly lower mortality, and C) telling someone where to find a building is not lifelong public health advice and the risk for one person crossing one road is miniscule anyway.
- Semaphor 4y ago
  So the part I objected to was not actually relevant for your argument.