Of course. It would be bizarre if there weren't a relationship between Lewy Body dementia, Alzheimer's, or vascular dementia (which in old people, means you've gone into heart failure) and repetitive negative thoughts. For one, you know you've got an incurable disease that will inevitably destroy your mind, and you've become one of the rare class of people for which assisted suicide has almost no controversy, it's something you're putting down payments on. For two, you can't finish thoughts.

My father was just diagnosed with Parkinson's a few months ago, and he already has trouble following any conversation, and knows it. If that didn't lead to depression, that's what would be notable. And any insight that he reaches that gives him comfort might be gone an hour later.

It just seems like a silly study.

Wait down payments? Is that metaphorical? How much does that cost?

They don't claim it's a cause. In fact, they explicitly state more research is needed to determine the relationship.

The first thing that jumps out at me is the concept of perseveration (repeated fixed obsessions) that happens in dementia syndromes. It would be interesting to consider whether this is a chicken-or-egg scenario, whether individuals tended to ruminate in earlier life.

No one’s saying anything about it being a cause though … association is not cause

Several comments already seem to assume that already though

I believe there have been other studies showing people with a history of depression develop dementia at higher rates. There are some that have shown the structural/signal changes that happen after longterm depression as well. These are things that occur years or decades before the dementia.

This is the problem of correlation being reported in media, people read it as "causation found"

When really it's "We've found an interesting association, and we are going to explore it more to see if there's an causation that we can influence"

I think they need to use a different word than associated. That's what's causing the confusion?

> Personally I think that reporting of correlations should be dropped altogether…

It really should be a shared responsibility to report and understand the meaning of statistically significant correlation. Unfortunately, few journalists seem to have much interest in understanding it. And given that their readership likely has about average (i.e poor) numeracy and iffy understanding of probability, it’s a bad combination. The widespread misunderstanding about the iterative way that science converges on truth also contributes to this problem.

That said, I would rather know about interesting findings such as this if for no other reason than to start digging for the original paper.

I mostly agree. Although for many years the only evidence for harm from smoking was correlational.

As far as I know, it still is. I'm happy to be shown otherwise.

A lot of modern medicine is based on just correlations and an more or less educated guess. I do not think ignoring them makes sense, it is just that the reporting needs to be more clear and less sensationalist.

That is really because we don't actually understand the human body, and how it reacts to various stimuli.

We used to think stress caused ulcers, based on a correlation. We now know the actual cause is a bacterium.

Yes, sometimes we are right, sometimes we are wrong.

Which is more damaging to society?

Having things that are wrong, but we don't know any better, or having things that are right, but we don't have the skills to prove it?

We often laugh how people in times before thought "crazy" things about health maintenance, but we're no better.

True, but it still might be a good proxy to estimate how well a client is doing.

Gwern on correlation and causation: https://gwern.net/correlation