Mixed bag: dietary and supplement interventions on patient outcomes

[jim mcnamara]

TL;DR

Conclusions:
[quote]
Reduced salt intake, omega-3 LC-PUFA use, and folate supplementation could reduce risk for some cardiovascular outcomes in adults. Combined calcium plus vitamin D might increase risk for stroke.
[/quote]

This is a meta-analysis. It looks at data sets from other studies.

https://annals.org/aim/article-abst...diovascular-outcomes-umbrella-review-evidence
For cardiovascular diseases and stroke:
this paper shows that interventions to improve patient survival and outcomes through dietary changes and supplements displays modest gains and some losses. Data uncertainty was an important issue.

The takeaway is probably: dietary change won't hurt you, and may help. No guarantees. Some supplement regimes may hurt you.
Sounds like life in general. Not everything comes out great.

 

[jrmichler]

I find it interesting that many people focus on some factor so minor that large meta analyses can barely tease out any effect at all, while the overall healthy lifestyle (diet, not smoking, exercise, weight, and moderate alcohol) is known to add ten years to one's life.

This source says that the healthy lifestyle adds 14 years life for women and 12 years for men: https://www.ahajournals.org/doi/full/10.1161/CIRCULATIONAHA.117.032047.

Other studies have shown that by the healthy lifestyle:
Strokes are 90% preventable.
Heart disease is 80% preventable.
Cancer is overall about 1/3 to (maybe) 1/2 preventable.
Dementia is at least 1/3 preventable.

But the news coverage is about some minor factor that maybe showed up in a small study...

 

[jim mcnamara]

@jrmichler
Medical/Nutrition studies are among the worst studies because humans are terrible test subjects. Traditionally when you want to isolate an effect you change just one parameter in your study. Each human has his/own set of parameters. You cannot do the one variable kind of study. So to get around that mishagosh - If you have huge monetary resources - you design and monitor a random controlled trial or a large longitudinal study like the https://www.framinghamheartstudy.org/

It is so bad out there the NIH has a staff dedicated to RCT's so you as a researcher can winnow out the chaff.
Example:
https://prevention.nih.gov/educatio...-randomized-trials-public-health-and-medicine
Otherwise you work with what you can afford. Piecemeal small studies cobbled together into a meta-analysis.
So the quality of the data takes a hit. And the study is restricted to what is in somebody else's data set and methods. Note the data uncertainty factor. But that is not how the story gets out on news feeds, just as you noted.

Reporting, ugh...
News media get paid for clicks - not quality science reporting. You have a reporter with a degree in journalism writing on newly discovered asteroids. Good luck. The result looks like Armageddon is just around the corner.
Example: https://www.express.co.uk/news/scie...close-earth-approach-science-orbit-atira-near

Note the headline. Pure yellow journalism. Or clickbait as it is called nowadays.

 

[Laroxe]

I find it interesting that many people focus on some factor so minor that large meta analyses can barely tease out any effect at all, while the overall healthy lifestyle (diet, not smoking, exercise, weight, and moderate alcohol) is known to add ten years to one's life.

This source says that the healthy lifestyle adds 14 years life for women and 12 years for men: https://www.ahajournals.org/doi/full/10.1161/CIRCULATIONAHA.117.032047.

Other studies have shown that by the healthy lifestyle:
Strokes are 90% preventable.
Heart disease is 80% preventable.
Cancer is overall about 1/3 to (maybe) 1/2 preventable.
Dementia is at least 1/3 preventable.

But the news coverage is about some minor factor that maybe showed up in a small study...

I would be careful about claims like this, basically what is suggested that when you compare groups who have a healthy lifestyle with groups with all these risk factors there is an average difference of 10 years. In reality there are few people that would fit into either group particularly well so any advantage would likely be much smaller and there are other significant mediators of longevity, genetics and environmental influence spring to mind.
I also think there are a number of issues in these estimates which need to be highlighted, the main one being that they are estimates and usually generated by particular interests groups. Its difficult to make reliable claims when the evidence doesn't support the models of disease causation or when the underpinning pathology simply isn't well understood. Whether various dementia's can be prevented beyond avoiding the premature loss of neurons remains a cause of considerable debate. We should also remember the predicted cardiovascular apocalypse linked to obesity, in fact cardiovascular disease rates have fallen and this can only be explained in part by better care.
Its been suggested that in medical epidemiology you really need to find relative risk associations of 1 : 2 or 3 to consider something to be worth even looking at, smoking has a RR of around 1:20 and there is no doubt about its potential for harm. However for most risk estimates the figures rarely get as high as 1:2, they are simply not very credible.