Reminds me of how Psychiatrists have one if the highest suicide rates of all professions. Seems unusual; like an accountant who's insolvent, or a mechanics who's car always breaks down.
Nah, I'm a computer guy who's computers are never put together fully, as I don't want to mess with it when I get home. Happens all the time. Well, that's my desktop. My servers get TLC.
I wrote to a journalist called Dan Diamond who had written an article covering how FTXs 'collapse' would set back health philanthropy. He would not directly answer my reasonable questions except to say "I'm a health journalist ".
Like, if I write about how it's a sad a murderer is going to jail because they worked in a charity shop on Thursdays, that seems like distraction to the point of complicity.
It really sniffs of agenda. Crypto folk are too familiar with paying for articles. But this was in Washington Post.
They did adjust for confounders available in the Framington heart study.
Observational studies are less good at showing causation, sure.
Whilst I can't see the original article, the news summary insinuated a dose related effect. A sample size of >2000 is also decent.
Naturally in an ideal world you'd want a 3 arm, 20 year long, randomised placebo controlled double blinded study. But who would ever 1. Pay for it, 2. Wait that long.
The problem with large sample sizes is that random deviations show up as significant difference unless you have a responsible statistician in the team.
Maddening paywalled article;
https://doi.org/10.1038/s41586-022-05258-z
Honestly why bother doing the science for public good on public money if the information can't be disseminated freely.
I do commend the use of science and data analysis that the author is advocating.
There is a caveat. Most science, even when high calibre and delivering clear results, doesnt reliably predict wwhat individuals should do. It only speaks to an average across the population. Individuals have unique responses, which can vary by degree and also by direction. The gold standard is always careful experimentation on a case by case basis. This phenomena is more apparent in diet science where dissecting causation and elaborating mechanism is particularly difficult. Many studies are narrow in scope, with short duration, and applied to limited demographic types... limiting generalisability. Larger, longer studies in more natural contexts produce results that often conflict, presumably because of factors we don't fully understand.
So my two cents would be to remember that results of studies are not always true for individuals, annoying as that is.
If you're serious about change you'll need to experiment.
>Individuals have unique responses, which can vary by degree and also by direction.
I think the tension is that that are errors in both directions here. Sure - some people will have worse results than optimal because they followed advice meant for the "average" person when they have unique body chemistry. But on the contrary many people who think they are unique will shun advice for the "average" person and have worse results.
My hunch is that more people make the second mistake than the first.
Is the claim that Asian women suffer lower rates of osteoporosis despite lower diary because they consume an alkaline diet something which has been studied at a scale and depth to support a causal link?
Better think the serotonin/ monoamine hypothesis of depression is akin to saying a server load balancing failure is because of "electricity imbalance". Whilst electrons are undeniably involved, the explanation is not useful, and distracts from better truths.
We have such a scarcity of safe and effective tools to probe and influence the human brain. Herein lies the hope that the diseases we invent might be treated neatly by the tools we posses. Yet, given the origin of disease and the tools are sparsely interlinked, it is too often just a fantasy.
Yet to "do nothing" is usually not acceptable and so we persist, and iteration or invention become the next hope.
