Afternoon.
Here is the link to the study that is being discussed. The bulk of it is behind a paywall.
https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(18)31809-9/fulltext
I'm guessing the reason many are having a pop at this is these types of study are relatively common, are routinely reactionary (Fat bad! Carbs good!) and are actually pretty insubstantial when you look at the details.
In terms of 'confirmation bias' what struck me is this from the abstract; they say they wanted "to establish an evidence base for quantitative recommendations for intakes of dietary fibre." That sounds very much to me like they 'know' we should be eating fibre and wanted to find out just how much and lo and behold they discovered that we should be eating fibre at a rate that 90% of the population doesn't. If you 'know' what you want to come out of your study then you are not doing a study you are engaged in marketing.
Their fibre recommendation is very precise which is peculiar to say the least from observational studies; there is a 4 gram leeway; between 25 g and 29 g is the best. How much worse is 24 g than 25 g? Is 35 g really no good as well?
Another thing is the oblique use of numbers versus improvements. Look at this sentence (from the study):
"Observational data suggest a 15–30% decrease in all-cause and cardiovascular related mortality, and incidence of coronary heart disease, stroke incidence and mortality, type 2 diabetes, and colorectal cancer when comparing the highest dietary fibre consumers with the lowest consumers Clinical trials show significantly lower body weight, systolic blood pressure, and total cholesterol when comparing higher with lower intakes of dietary fibre"
15-30% meaning what? Is that absolute or relative? (i.e. the difference between a 4% and a 3% chance of all-cause mortality is 33% relative reduction or the less glamorous 1% absolute). Which is this? If they don't say I'm guessing it's relative. Or are they saying that in the high fibre group 30% of the population avoided all-cause mortality? That they are now immortal? That would be good; but probably not what they mean, so why not detail the average life extension in those lucky 30% - could it be that it's insubstantial? Eat 200 kg of kale a day and live 3 days longer on average? Not so good.
How do those numbers cover those different outcomes is it 15-30% for each or in aggregate; i.e. are they all added up to get to that 15-30% figure or is for each bad outcome?
All in all this is a study looking at observational dietary studies (i.e. asking people often over a very long period what they eat which is famously problematic - what did you eat last year for instance?) and seeing how that correlates with their health.
There's nothing causative about it at all. What is the biological mechanism whereby fibre prevents coronary heart disease and type 2 diabetes? They don't know and/or aren't telling, probably because there is none.
Presumably, they discounted other health impacting habits like smoking? But they don't say that so it's possible that people who eat yams and don't smoke also tend not to do other health impairing things either (drink, drugs, no exercise and so on) and therefore live longer. To put that down to their yam consumption is stretching it a bit isn't it?
As mentioned above though these data specifically excluded people with chronic disease, i.e. diabetes, and so can't be held to be applicable to us as none of the 4,000 or so people who were the subject of these trials were diabetics.
I would say the risk from running high blood sugars which has been proven by clinical trials ( always keep this in mind; it's the basis for all our treatment on the NHS -
https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/GetPdf.cgi?id=phd000390 ) and well understood mechanisms of arterial and neurological damage far outweighs the risk from fibre 'deficiency' that might (or might not) arise by avoiding carbohydrates.
In short the confirmation bias on display here is in that study not in the pretty measured response on this forum.
Dillinger