Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

aelephant comments on Living Forever is Hard, or, The Gompertz Curve - Less Wrong

46 Post author: gwern 17 May 2011 09:08PM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (86)

You are viewing a single comment's thread. Show more comments above.

Comment author: aelephant 23 March 2012 05:15:09AM *  1 point [-]

Actually I did google Mike Darwin, but that didn't give me any reason to believe all of the claims he made. He's the founder of some cryonics company that I've never heard of. How does that qualify him as an expert on aspirin? The Wikipedia page about him is also riddled with warnings, which make me skeptical of the other content I might find there.

I also googled aspirin and Reye's syndrome. Apparently you did too, but didn't read what popped up very carefully. Aspirin has been ASSOCIATED with Reye's syndrome in EPIDEMIOLOGICAL studies. Epidemiological studies are not capable of proving causation.

I already know NSAIDs can increase the risk of having a GI bleed. You aren't teaching me anything new. What is at question is whether the risks of taking aspirin outweigh the benefits.

In light of your comments and attitude, I plan to disregard your opinion as well from here on out.

Edit: I thought the Vioxx vs. NSAID death comparison sounded funny too, so I took a look at some of the numbers.

It is thought that Vioxx over 5 years caused between 88,000 to 140,000 serious cases of heart disease. Per year, that works out to 17,600 to 28,000.

Of those who developed heart disease, it is estimated that 30-40% died. That gives us 5,280 to 7,040 on the low end and 8,400 to 11,200 on the high end.

What about NSAIDs?

It is estimated there are about 60,000,000 NSAID users annually. 1-2% of people who take NSAIDs develop GI events, like a hemorrhage. That's 600,000 to 1,200,000 people.

Estimates of deaths caused by NSAIDs are between 3,200 to 16,500 per year.

So we have, looking at Vioxx vs. NSAIDs:

5,280 to 11,200 deaths per year vs. 3,200 to 16,500 deaths per year

The low estimate puts NSAIDs ahead, the high estimate puts them below. They seem roughly comparable to me. When you take into account how widespread NSAID use is compared to Vioxx, I think you have a strong argument that NSAIDs are much safer than Vioxx.

At its peak, Vioxx was taken by about 20,000,000 Americans.

Comment author: JoshuaZ 23 March 2012 05:19:30AM 2 points [-]

I also googled aspirin and Reye's syndrome. Apparently you did too, but didn't read what popped up very carefully. Aspirin has been ASSOCIATED with Reye's syndrome in EPIDEMIOLOGICAL studies. Epidemiological studies are not capable of proving causation.

No, but you asked if there was evidence. Correlation doesn't imply causation but it is Bayesian evidence for causation.

Comment author: aelephant 23 March 2012 05:37:47AM 0 points [-]

I asked if it was proven that aspirin causes Reye's syndrome. The correct answer is, "No, it is not."

Comment author: JoshuaZ 23 March 2012 05:43:05AM 2 points [-]

Yes, I see you did use the word "proven". That does lesson the force of my comment but not by very much. I'm curious actually what you mean by proven outside a mathematical context. In this context, the correlation is extremely strong, and there's no obvious alternate hypothesis. I'm not completely sure what you consider in this context to be an acceptable amount of evidence, but the medical consensus seems clear, and the links Gwern gave show that the correlation exists even when one tries to control for other variables.

Comment author: gwern 23 March 2012 03:23:40PM 2 points [-]

I think he wants randomized controlled trials. Which of course no ethics board would ever approve because avoiding aspirin in kids is not that hard or expensive, and the cost of confirming it would be too high.

("You want to test whether slamming your face into the wall causes pain, and doesn't just correlate with it? Why on earth?")

Comment author: aelephant 25 March 2012 02:54:22AM 0 points [-]

We have a plausible mechanism (confirmed by massive amounts of science) by which slamming your face into the wall causes pain; for the link between Aspirin and Reye's syndrome there is no such plausible mechanism.

Comment author: aelephant 23 March 2012 06:04:25AM *  0 points [-]

Well, at least in the drug world it is more convincing if you have a rational mechanism by which a drug could cause some effect.

A plausible alternative hypothesis is that children who get sick sometimes are developing Reye's syndrome and take Aspirin. The problem with epidemiological studies is that it is IMPOSSIBLE to control for other variables. This is why randomized controlled trials are essential and give you a kind of information epidemiological studies never can.

My point is not that Aspirin doesn't cause Reye's syndrome, but that it is impossible to say one way or the other. Maybe it does, maybe it doesn't. The data is unavailable and likely never will be. So if you are an honest person who values the truth, you can't say that it does.