> The researchers used innovative statistical methods to predict average trajectories of depressive symptoms for approximately 15,000 participants ages 20 to 89, divided into three life stages: older, midlife and young adulthood. They then applied these predicted trajectories and found that in a group of approximately 6,000 older participants, the odds of cognitive impairment were 73 percent higher for those estimated to have elevated depressive symptoms in early adulthood, and 43 percent higher for those estimated to have elevated depressive symptoms in later life.
Can somebody explain how this could show causation, instead of correlation?
Setting aside the "innovative statistical methods" that "predicted" depressive symptoms (machine learning?), the actual conclusion seems to be: "depression in early adulthood is correlated with dementia later in life".
(Actually "predicted depression in early adulthood", which I don't trust one bit if it's actually coming from a machine learning model.)
The results were adjusted for a lot of things, but not for other existing mental illness, for example. I can't imagine that there wouldn't be confounding factors that can cause both depression and dementia, especially knowing that a lot of mental illnesses are already known to be linked (e.g. autism and ADHD, bipolar disorder and depression).
The title seems completely dishonest to me. Maybe someone who knows statistics better can explain what I missed.
One thing for sure is that life becomes non linear quick, people who have just the right foundation in their 20s will grow solid beneficial experiences (SO, parenting), social status (these days most companies are parenting-friendly).
If you're out the system runs a bit against you, as is time, but now you're fighting harder, alone.
below a certain threshold your life goes to misery real fast.. above a certain amount you have other problems too (wealth can destroy your social life, or emotional balance, badly managed it may turn into overwhelming debt)
just enough wealth gives you mental safety, allows for patience and errors thus grows, lets you live in positive contexts where your social network helps reinforcing your wealth even more.
ps: I've experienced low level jobs and contemplated homelessness, you really don't want to live near these levels at all. It's a black hole of issues... you'd be surprised how fast your mind goes insane (I literally punched my car screaming "damn snobs" while late to deliver newspaper in the morning.. I wanted to restore some good old french tradition quite strongly if you know what I mean)
The French tradition of using the guillotine once on rich people followed by using it punitively on poor people that got caught for the next 200 years straight
This seems right, dementia is correlated sleep quality and consistency which is correlated with the personality trait neuroticism (govens depression, anxiety, stability and negative emotion largely). This simply confirms the existing evidence that poor sleep hygiene over a long period of time harms the brain, possibly along with lack of fitness, both of which can be blamed on personality
Not sure why you assumed this. ML isn't the only "innovative statistical method". I haven't read the paper though.
> "depression in early adulthood is correlated with dementia later in life"
That is how all inferences are made. It is very _very_ rare to be able to demonstrate strict causation in medical sciences. This is why the title says "may".
But the path from "we found a correlation" to "we have good evidence of causation" is a very long path where 99% of hypotheses drop off. And this one is especially suspect. It's possible that early life happiness is protective against dementia but mid and late life less so, but our model of biology says that's less likely because almost all similar findings that we have found to be casual recency increases the effect. For instances what you eat and how you exercise this year impacts your current health more than what you ate and how you exercised 2/5/10/15/20 years ago.
I think "may protect" is misleading because while it might be the case that if you increase someone's happiness in early adulthood it will decrease their risk of dementia, it's certainly not likely. For instance if you found some depression RCT from 50 years ago in 20 somethings, what are the chances the control group and the intervention group are going to have significantly different dementia outcomes?
So while it's technically true that it "may protect against dementia", it was also technically true before someone did this study. And honestly studies like this have so many degrees of freedom they are almost always unintentionally p-hacked.
>> Not sure why you assumed this. ML isn't the only "innovative statistical method". I haven't read the paper though.
I have to assume things since I don't seem to have access to the actual paper. "Innovative" methods that predict things made me think of machine learning, but I'd be happy to learn more if I could read the paper.
> It is very _very_ rare to be able to demonstrate strict causation in medical sciences. This is why the title says "may".
The title is a lie. It very clearly suggests that the study gives some evidence of a causation, while the method and results don't support that interpretation at all.
If whoever wrote this wanted to honestly inform people about the result, they could have written "people who are depressed early in adulthood have higher chance of dementia later in life", which makes it clear for someone paying attention that this doesn't exclude a confounding factor.
I agree with your first post but to call the title a lie is not right. It seems to me that “happiness in early adulthood may protect against dementia” is logically consistent with your proposed rewrite, “people who are depressed early in adulthood have higher chance of dementia”
On reading the title alone, I assumed a university marketer was trying to reframe in an interesting way what had been found in the study (something akin to your proposed summary)
> It seems to me that “happiness in early adulthood may protect against dementia” is logically consistent with your proposed rewrite, “people who are depressed early in adulthood have higher chance of dementia”
Saying "the moon is white" is logically consistent with saying "the moon may be made of white cheese". Logical consistence isn't an excuse for a major university to make people believe that research supports something that isn't actually supported by their work.
> On reading the title alone, I assumed a university marketer was trying to reframe in an interesting way what had been found in the study (something akin to your proposed summary)
Yes that's definitely the reason. I couldn't find the article on the journal's website, but I imagine that the paper was honest about this being a correlation. We can debate whether the title (and some of the article) fits the definition of a lie, but it's 100% clickbait that will make some people believe something that has no basis in research, and they are exploiting that for PR reasons.
And I actually wouldn't be surprised if the causality they suggest is true... But that kind of exaggeration makes it impossible to follow what we actually have evidence for, which is the opposite of what this kind of popular science articles should be doing.
To me it's not the same statement: it's one thing to say that people who might be prone to be depressed as early adults might have higher risk of dementia later. To say that happiness at that period is protective, suggests that if you cheered them up at the right time, they'd no longer be at higher risk. Is it something about their brain structure or approach to adversity, or is it specific, moment to moment lived experience? If it's the latter, then a dementia prevention strategy could be giving out free amusement park tickets to 20-somethings. I'm being facetious, but the "happiness" title really does shift the focus. So I didn't think it was equivalent.
As a general statement, these university press releases on research findings are some of the lowest-quality posts on HN.
Brenowitz, W. D., Zeki Al Hazzouri, A., Vittinghoff, E., Golden, S. H., Fitzpatrick, A. L., & Yaffe, K. (2021). Depressive Symptoms Imputed Across the Life Course Are Associated with Cognitive Impairment and Cognitive Decline. Journal of Alzheimer's disease : JAD, 83(3), 1379–1389. https://doi.org/10.3233/JAD-210588
Methods: "Using a pooled study of 4 prospective cohorts (ages 20-89), we imputed adult life course depressive symptoms trajectories based on Center for Epidemiologic Studies Depression Scale-10 (CESD-10) and calculated time-weighted averages for early adulthood (ages 20-49), mid-life (ages 50-69), and late-life (ages 70-89) for 6,122 older participants. Adjusted pooled logistic and mixed-effects models estimated associations of imputed depressive symptoms with two cognitive outcomes: cognitive impairment defined by established criteria and a composite cognitive score."
Yes, same feeling here. I consider my life like a tragedy in the litterature sense of the term: I knew what was going to happen, I fought and attempted to get people onboard, had psys and all, and at 38 it had exactly the trajectory I had expected.
(Not that I’m not enterprising, I’ve moved abroad, created a company, but love and affection went missing because I have long been depressed and didn’t have any clue in how to make it work when young, and now I can’t stand people socially anymore, and Covid lockdowns were the nail in the coffin - I see the dementia coming like an elephant, or at least the “crazy old man” part of it).
I obviously don't know your life or circumstances, but try to ensure it's not a self-fulfilling prophecy where you envision a tragic life for yourself and then steer yourself in that direction.
> didn’t have any clue in how to make it work when young, and now I can’t stand people socially anymore
1. You're still very young. I'm a similar age, and I'm fairly certain when we're in our 50s and 60s we'll reminisce about our younger years in our 30s and 40s.
2. There's a lot of people in the world. It's a bit unfair to group everyone together and say you "can't stand people socially anymore". You likely only need a few close people in your life for it to feel rewarding, and I'm quite certain they're out there.
When I think I “can’t stand people” (which happened a lot to me from my teens to adulthood), it’s always because I’m in a bad place mentally, which can be identified and solved, or because I’m hanging out with the wrong people, which is easy to solve by meeting new ones.
It’s not always easy or fast, but it’s something you can fix.
I hope this paper accounts for the well established evidence of dementia relating to poor sleep hygiene across the lifespan, and the also well established relationship between poor sleep and neuroticism - the personality trait responsible that broadly accounts for negative emotionality, as this would explain that this paper simply observes the causation of dementia from poor sleep and other behaviours across the lifespan, which is originally attributed to personality developement (neuroticism is very plastic and is established before the age of 6).
I'm not sure the methodology here is sound, but if true, then Gen Z is doomed because statistically it's one of the most depressed generations ever. We have a serious and growing problem with mental health in the West, and it seems strongly tied to sociocultural outlooks and economic realities. I know it seems too easy to say, but the greed of Boomers has really screwed future generations in so many facets it's hard to account for, seems this is just one more way.
I think you are right. Its really up to our generation to let go of a lot of things we were told are guaranteed in our life. We really do have the technologies and labor force to fix the problems in the world in a matter of years, we just need some good leadership and discipline, but, of course, the human condition...
Can somebody explain how this could show causation, instead of correlation?
Setting aside the "innovative statistical methods" that "predicted" depressive symptoms (machine learning?), the actual conclusion seems to be: "depression in early adulthood is correlated with dementia later in life".
(Actually "predicted depression in early adulthood", which I don't trust one bit if it's actually coming from a machine learning model.)
The results were adjusted for a lot of things, but not for other existing mental illness, for example. I can't imagine that there wouldn't be confounding factors that can cause both depression and dementia, especially knowing that a lot of mental illnesses are already known to be linked (e.g. autism and ADHD, bipolar disorder and depression).
The title seems completely dishonest to me. Maybe someone who knows statistics better can explain what I missed.