Comment by carbocation | Hacker News Reader

Comment by carbocation | Hacker News Reader

When interpreting the observational literature (including my own work) I find it helpful to think like a threat actor: how can this association be broken?

Note that this is an article making huge claims about Alzheimer's but is published in Molecular Psychiatry. Given that signal, I'm just going to look at the abstract and not read the full text.

Here, the study population is patients with rheumatoid arthritis. This tells you up front that we don't have a typical population, but a highly selected one.

The use of HCQ is observational, not randomly assigned.

So, how could HCQ look protective even if it had no effect on Alzheimer's biology?

1. Maybe more rheumatoid arthritis patients should be on HCQ. It's a disease-modifying drug, after all. So it may be the case that people with better care, or more wealth, are more likely to be on HCQ. And that could be associated with other lifestyle factors that are never fully accounted for in an observational analysis.

2. Or it could be the case that people with dementia are less likely to have their HCQ prescriptions renewed, so the arrow of causation could even be reversed.

Additionally, one could imagine that HCQ might actually protect against Alzheimer's disease in people with rheumatoid arthritis—but not in a general population. That would be very big news, but would require further study.

And sure, it could be the case that HCQ reduces Alzheimer's risk in everyone. But you'd need a lot more evidence to buy that.

I think the study is somewhat more robust than you're letting on, since it compares HCQ against an alternative rheumatoid arthritis drug - Methotrexate. It's possible that the choice of prescribed drug itself is correlated with some other confounding factor, but it's certainly less likely to be so compared to HCQ vs nothing.

Also, the study itself is clear about the amount of evidence needed to validate whether HCQ actually reduces Alzheimer's risk: "This hypothesis merits testing through adequately powered clinical trials in at-risk individuals during preclinical stages of disease progression."

In that case, couldn't it be that Methotrexate makes the risk worse rather than Hydroxychloroquine reducing the risk?

Methotrexate is standard of care, considered “strongly recommended” over hydroxychloroquine (source: https://onlinelibrary.wiley.com/doi/epdf/10.1002/acr.24596 ) for rheumatoid arthritis. If I read the paper right, hydroxychloroquine may be favored in “low disease” states.

So I would certainly expect some sort of difference in patient population. It would suggest those initiated on hydroxychloroquine would normally have less severe rheumatoid arthritis (or have some other difference that their doctor would choose to not use methotrexate by default), although I don’t know if that’s the case in reality.

Not having prescribed either since I was a PCP years ago, I would just say that I agree with this line of inquiry.

If both drugs were equally good and were pretty much used at random, then they could be good instruments. But if there is confounding by indication, then it’s harder to get value out of an analysis that compares them.

It doesn’t mean the conclusions are wrong (and nor does my comment above), it just tempers my interpretation.

You could try and reduce the possible effect by controlling for disease severity. If it reduced the effect a large amount it'd strengthen the case that the effect is actually due to RA disease severity. Actually I'd be a bit surprised if the authors didn't do that for an observational study.

Agreed. I didn't see it described in their methods, although it seemed that the main focus of this paper (based on length dedicated to the material) was the biological/model data rather than the EHR data.

This summary right here is what I would like to accompany any medical study that pop up online. Thanks for helping us stay skeptical (doesn’t mean that I’m not hopeful and that the claims is true)

It might be that among the causes of dementia, inflammation plays a decent role, which is also the cause of rheumatoid arthritis

Yes, I think the hypothesis is conceptually interesting. But rather than thinking about how it could be true, I try to think about how the data could be misleading. I do this to try to avoid making broader claims in my own work than I should.

Reply on news.ycombinator.com