Hacker News Reader

Comment by kelseyfrog | Hacker News Reader

> Has this actually been confirmed in humans[...]?

There's no way such a study would pass an ethics review board. It is structurally unknowable knowledge.

It is kind of funny it's "ethical" to give millions of children a potential dangerous chemical but if you track it then it's unethical.

Wow, I never thought about it from that perspective.

Maybe studies about normal intake volumes of GRAS[0] food and additives should have some more leeway than a naive ERB might otherwise consider were it any other chemical. If they're already legal and in the food system (in some cases, in MASSIVE volumes), there should be very few barriers to gathering good data on it about humans.

Maybe there is some ethical concern about testing a hypothesis that includes harm to humans, but at the same time, many people are already eating this stuff. I'm not sure how to reconcile those -- I do know that my feeling of "Come on! Look at reality!" correlates well with making some logical or ethical fallacies, but in this case, I'm not sure haha.

[0] https://www.fda.gov/food/food-ingredients-packaging/generall...

> Maybe there is some ethical concern about testing a hypothesis that includes harm to humans, but at the same time, many people are already eating this stuff.

People are already eating this stuff, so it’s obviously not “drop dead dangerous”. In order to test it, we’d either need to change dosage/frequency or try some longitudinal study by tracking what people eat more exhaustively.

Increasing dosage/frequency can make something “drop dead dangerous”, which is where the ethical qualms arise. Doing a long term longitudinal study is just really hard to draw statistically significant outcomes considering the variety we eat and the massive amount of bad things in the world.

I agree with the central point that testing legal things (for public safety) is a thing we may need to do.

Kind of funny? Academia gets strict ethical regulations (as it should be) and global corporations get a pass.

It shows how brazenly ethical concerns are subverted by profit motives.

Such a controlled trial would not pass ethics review. But for the human side, hopefully we can run statistical analyses to show correlation at least?

I think a mechanism of action in vitro plus a biologically confirmed trial in rats plus correlation in statistical analyses is fairly convincing?

It depends on the audience. Nothing short of RCTs is admissible on HN.

With enough accurate data (and time, as these toxins aren't really fast acting) you could get to causality with the right kind of statistical analysis. Typically, counterfactual analysis and some DAGs.

This reminds me of the old tootsie pop commercial

https://www.youtube.com/watch?v=5ZtbCOpx8Sk

How many skittles does it take to....

Reply on news.ycombinator.com