Author Topic: Any science-based article summarizing a healthy diet?  (Read 4317 times)

0 Members and 1 Guest are viewing this topic.

Offline bachfiend

  • Not Any Kind of Moderator
  • Frequent Poster
  • ******
  • Posts: 2105
Re: Any science-based article summarizing a healthy diet?
« Reply #75 on: September 28, 2018, 08:37:47 PM »
CarbShark,

There are no random controlled trials on diet and nutrition.  At best, they’re semi-random (the participants have to know which manipulation they’re being requested to follow, and they have to be motivated to actually follow the manipulation) and semi-controlled (for the same reasons).

In true RCTs, the participants don’t know which treatment they’re receiving, nor do the investigators.  In trials of statins, for example, half of a test group receive the drug being investigated and the other half receive a placebo or another accepted statin, and the two groups are then treated identically.  And at the end of the trial, the results are collated to see if there’s a difference in the results for the two groups.

You are conflating two disparate concepts: randomization and blinding.  A trial is randomized when subjects are assigned to treatments via a random process.  Trials may be either randomized or not.  Trials, whether randomized or not, may be single blind, double blind, triple blind, or open label.  In a single-blind trial, subjects do not know what treatment they are receiving; in a double-blind trial, neither do the investigators; in a triple-blind trial, neither do the data analysts.  In an open-label trial, everybody knows everything.

There have been, and there continue to be, lots of dietary randomized controlled trials, some even double blind.  What there have been very few of are long-term diet trials, that is, on a time scale (years to decades) needed to ascertain the effects of diet on hard chronic disease endpoints.  The problem, with such trials, as you have explained, is compliance.  Subjects will not stick to an assigned diet for a decade.  That's why effects of diet on chronic disease have to be ascertained by using a combination of basic science, short-term metabolic trials, and long-term observational studies.

No, I’m not conflating blinded random controlled trials with random controlled trials.  And how do you have a double blind dietary trial?  Dietary trials are at best semi-random (the participants can make the decision to drop out or not comply if the diet is too onerous) and semi-controlled (the investigators have limited means of ensuring compliance).

Really, long term observational studies are the only way of determining the health effects of diet, when the end points are chronic disease or death.  Basic science and short term metabolic studies are interesting, but hardly conclusive.  Steve Novella has warned against using basic science to extrapolate to results in complex systems, because science simplifies systems, and there’s always added factors you haven’t considered.  And metabolic trials only demonstrate proxies of risk, not the risk itself.  If your aim is to reduce the incidence of chronic disease or increase life expectancy by some dietary manipulation, then you need to actually show that it happens.

Anyway.  I’m wasting too much time on this thread.  My view regarding diet is expressed in my third comment on page 2 when I recommended ‘Dietary Guidelines for Americans 2015-20’ as a good reference for the science of diet and health.  If I’m recommending ‘the Dietary Guidelines for Americans 2015-20’ I’m also recommending the guidelines.  If more people followed the guidelines, then we’d be better off.  Willett wrote in his book that the ‘Guidelines’ are finally getting the science much closer to right by removing vested interests such as food producers from the editing process.

People such as CarbShark want the guidelines to be chucked out and replaced by recommendations for their extreme low carbohydrate/high fat ketogenic diets.  I know my diet is extreme, at the other end of the spectrum, but I’m not claiming that it’s healthier.  Nor am I recommending it to other people or demanding that it ought to be included in the ‘Guidelines.’

I’m just pointing out that I’m getting all the benefits claimed for low carbohydrate/high fat ketogenic diets (no feelings of hunger during the day, low BMI and body fat and perfect blood chemistry) on a high carbohydrate/low fat diet.
Gebt ihr ihr ihr Buch zurück?

Online jt512

  • Frequent Poster
  • ******
  • Posts: 2623
    • jt512
Re: Any science-based article summarizing a healthy diet?
« Reply #76 on: September 28, 2018, 09:36:01 PM »
CarbShark,

There are no random controlled trials on diet and nutrition.  At best, they’re semi-random (the participants have to know which manipulation they’re being requested to follow, and they have to be motivated to actually follow the manipulation) and semi-controlled (for the same reasons).

In true RCTs, the participants don’t know which treatment they’re receiving, nor do the investigators.  In trials of statins, for example, half of a test group receive the drug being investigated and the other half receive a placebo or another accepted statin, and the two groups are then treated identically.  And at the end of the trial, the results are collated to see if there’s a difference in the results for the two groups.

You are conflating two disparate concepts: randomization and blinding.  A trial is randomized when subjects are assigned to treatments via a random process.  Trials may be either randomized or not.  Trials, whether randomized or not, may be single blind, double blind, triple blind, or open label.  In a single-blind trial, subjects do not know what treatment they are receiving; in a double-blind trial, neither do the investigators; in a triple-blind trial, neither do the data analysts.  In an open-label trial, everybody knows everything.

There have been, and there continue to be, lots of dietary randomized controlled trials, some even double blind.  What there have been very few of are long-term diet trials, that is, on a time scale (years to decades) needed to ascertain the effects of diet on hard chronic disease endpoints.  The problem, with such trials, as you have explained, is compliance.  Subjects will not stick to an assigned diet for a decade.  That's why effects of diet on chronic disease have to be ascertained by using a combination of basic science, short-term metabolic trials, and long-term observational studies.

At best, long term observational studies will only show correlations, and in part due to the weakness of the methodology (FFQ; few samples for each subject; limited follow up) they are not suited for a complex field like nutrition, and certainly not suited for proving causation between specific foods and chronic diseases.


At best, observational studies demonstrate causation.  That's what all that controlling for confounders is all about.

Quote
When they show correlations, there must be followup with further studies (RCTs, etc.) to show any causation. That's not just true for nutrition, but anywhere long-term observational studies are used.


Why do you keep saying that when it has been explained to you over and over again that long-term RCTs are not feasible in nutrition.  For the umpteenth time, the effects of nutrition on chronic diseases can ONLY be ascertained by using a combination of basic science, short-term metabolic studies, and long-term observational studies.  Those are THE tools.
Ich verstehe nur Bahnhof.

Online jt512

  • Frequent Poster
  • ******
  • Posts: 2623
    • jt512
Re: Any science-based article summarizing a healthy diet?
« Reply #77 on: September 28, 2018, 10:02:20 PM »
CarbShark,

There are no random controlled trials on diet and nutrition.  At best, they’re semi-random (the participants have to know which manipulation they’re being requested to follow, and they have to be motivated to actually follow the manipulation) and semi-controlled (for the same reasons).

In true RCTs, the participants don’t know which treatment they’re receiving, nor do the investigators.  In trials of statins, for example, half of a test group receive the drug being investigated and the other half receive a placebo or another accepted statin, and the two groups are then treated identically.  And at the end of the trial, the results are collated to see if there’s a difference in the results for the two groups.

You are conflating two disparate concepts: randomization and blinding.  A trial is randomized when subjects are assigned to treatments via a random process.  Trials may be either randomized or not.  Trials, whether randomized or not, may be single blind, double blind, triple blind, or open label.  In a single-blind trial, subjects do not know what treatment they are receiving; in a double-blind trial, neither do the investigators; in a triple-blind trial, neither do the data analysts.  In an open-label trial, everybody knows everything.

There have been, and there continue to be, lots of dietary randomized controlled trials, some even double blind.  What there have been very few of are long-term diet trials, that is, on a time scale (years to decades) needed to ascertain the effects of diet on hard chronic disease endpoints.  The problem, with such trials, as you have explained, is compliance.  Subjects will not stick to an assigned diet for a decade.  That's why effects of diet on chronic disease have to be ascertained by using a combination of basic science, short-term metabolic trials, and long-term observational studies.

No, I’m not conflating blinded random controlled trials with random controlled trials.

Did you actually read what you wrote?  I did, and you conflated blinding and randomization.  They are independent concepts: trials can employ either, neither, or both.

Quote
And how do you have a double blind dietary trial?

The same way you have double blind drug trials.  Except instead of drugs, you have meals.  Meals in the two arms are formulated to be indistinguishable, and investigators are not privy to which meals are being fed to which subjects.

Quote
Dietary trials are at best semi-random (the participants can make the decision to drop out or not comply if the diet is too onerous) and semi-controlled (the investigators have limited means of ensuring compliance).

Subjects in any study doop out if the intervention is too onerous as well as a myriad of other reasons.  And in a strictly controlled diet trial, investigators have at least as much control over compliance as in a drug trial.  We can feed every subject every meal in a metabolic ward and encourage them to eat everything.  We can even keep them in a metabolic ward for the duration of the study to insure they eat nothing else.  Clearly, such studies cannot go on for years to decades, but they can and do go on for weeks.

Quote
Really, long term observational studies are the only way of determining the health effects of diet, when the end points are chronic disease or death.  Basic science and short term metabolic studies are interesting, but hardly conclusive.

But none of these types of studies alone is usually conclusive.  We need basic science and metabolic studies to back up results from observational studies.  We look for consistency of results among all these studies.  If the studies consistently back each other up, then that is a good indication that we've discovered a causal relationship; if observational studies disagree with more basic science, then something somewhere is wrong, and there is more work to do.

Quote
Steve Novella has warned against using basic science to extrapolate to results in complex systems, because science simplifies systems, and there’s always added factors you haven’t considered.

You are again misapplying what Novella is talking about.  Basic science and metabolic studies are integral to discovering cause-and-effect relationships between diet and disease. 

Quote
And metabolic trials only demonstrate proxies of risk, not the risk itself.  If your aim is to reduce the incidence of chronic disease or increase life expectancy by some dietary manipulation, then you need to actually show that it happens.

But "showing that it happens" using observational studies is rarely sufficient due to the difficulty in interpreting observational studies.  To "show that it happens" convincingly, basic science, metabolic trials, and observational studies have to paint a coherent etiologic picture.
Ich verstehe nur Bahnhof.

Offline bachfiend

  • Not Any Kind of Moderator
  • Frequent Poster
  • ******
  • Posts: 2105
Re: Any science-based article summarizing a healthy diet?
« Reply #78 on: September 28, 2018, 10:16:51 PM »
jt,

And I insist that the ONLY way the effects of nutrition on chronic diseases can be ascertained is by long-term observational studies.  If long-term observational studies show NO effect on the incidence of chronic disease, mortality or life expectancy, then that’s the end of the question.  Period (unless newer larger better studies are carried out).

If long-term observational studies do ascertain an effect, then you can look at determining causation, which could be explained by basic science or short-term metabolic effects.  Obviously though, basic science has to inform you of what dietary manipulations are worth trying.  Getting the participants to face north when eating would hardly be a plausible dietary manipulation.  Getting the participants to chew 20 times before swallowing would be.

Basic science and short-term metabolic studies leave out the psychology of the person.

I’m very much a behaviourist.  The most important factors associated with chronic disease such as heart disease, type 2 diabetes and dementia are being overweight or obese, being sedentary or consuming excessive amounts of simple sugars.  Most people who are overweight or obese are so because of bad habits indulged in for years or decades.  Losing weight, and maintaining a healthy body weight, requires losing the bad habits and acquiring good habits.

Basic science and short-term metabolic studies are incapable of predicting psychology.  And can’t predict unpredictable or unintended consequences.  There’s some evidence that high fat diets increase the incidence of some cancers such as prostatic.  It would be hardly a benefit for a person to die earlier of cancer, unpredictable with basic science or short-term metabolic studies, but with absolutely ‘clean’ coronary arteries.

Which is the reason for studies such as that one in the Lancet, which was looking at all-cause mortality.

And I’m not confused and conflating random controlled trials with blinded random controlled trials.  I wrote that they are at best semi-random and semi-controlled trials.  The investigator might randomly allocate participants to various dietary arms, but the participants can self-select and drop out or not participate if the proscribed diet doesn’t suit, removing some of the randomness.  For example, you might decide to do a study of the health benefits of fish in the diet, and randomly allocate me to the ‘fish’ arm.  I’d refuse, or take part (and just not consume any fish to ruin your results).
« Last Edit: September 28, 2018, 10:47:43 PM by bachfiend »
Gebt ihr ihr ihr Buch zurück?

Online jt512

  • Frequent Poster
  • ******
  • Posts: 2623
    • jt512
Re: Any science-based article summarizing a healthy diet?
« Reply #79 on: September 28, 2018, 11:13:02 PM »
jt,

And I insist that the ONLY way the effects of nutrition on chronic diseases can be ascertained is by long-term observational studies.  If long-term observational studies show NO effect on the incidence of chronic disease, mortality or life expectancy, then that’s the end of the question.  Period (unless newer larger better studies are carried out).

It's not that black and white.  Observational studies can fail to show a true exposure-disease relation for numerous methodological reasons, and if the basic science is felt to be strong enough, it will outweigh the epidemiology.  A case in point is that of saturated fat and heart disease.  Observational studies do not consistently support this connection, yet because the underlying basic science is as strong as it is, most health authorities believe, and recommend, that saturated fat intake should be reduced.

Quote
If long-term observational studies do ascertain an effect, then you can look at determining causation, which could be explained by basic science or short-term metabolic effects.


Are you aware that almost always it's the other way around: that observational studies are undertaken to test whether an exposure-disease relationship suggested by basic science is true?

Quote
And I’m not confused and conflating random controlled trials with blinded random controlled trials.

Anyone can read your post and see for themselves that you are.

Quote
The investigator might randomly allocate participants to various dietary arms, but the participants can self-select and drop out or not participate if the proscribed diet doesn’t suit, removing some of the randomness.  For example, you might decide to do a study of the health benefits of fish in the diet, and randomly allocate me to the ‘fish’ arm.  I’d refuse, or take part (and just not consume any fish to ruin your results).

The investigator might randomly allocate participants to various drug arms, but the participants can self-select and drop out or not participate if the prescribed drug doesn’t suit, removing some of the randomness.  For example, you might decide to do a study of the health benefits of statins, and randomly allocate me to the statin arm.  I’d refuse, or take part (and just not consume any statins to ruin your results).


(BTW, if you didn't eat the fish, we'd know.)
« Last Edit: September 28, 2018, 11:16:28 PM by jt512 »
Ich verstehe nur Bahnhof.

Offline bachfiend

  • Not Any Kind of Moderator
  • Frequent Poster
  • ******
  • Posts: 2105
Re: Any science-based article summarizing a healthy diet?
« Reply #80 on: September 29, 2018, 12:25:37 AM »
jt,

How would you know that I didn’t eat the fish?

Anyway, you’ve got it around the wrong way.  First comes the observational studies, then comes the basic science.  People observe that lung cancer seems to be more common in smokers, they do larger observational studies, determine that it is, then do the basic science to see if there’s a mechanism which could explain it.  People see that there’s a group in the population that is using a special diet, and swears by it, and wonder if the diet is effective.  So they do an observational study to see if there’s any effect.  And if there is, then they wonder - why?

There’s an almost infinite number of possible diets and dietary manipulations which are possible and could have some effect on consideration of basic science.  You can’t test all of them.  You can only test the ones that are being already used.

And I never claimed that random controlled trials need to be blinded.  I never wrote that random controlled trials are actually at best semi-random semi-controlled trials because they’re non-blinded.

And if reducing saturated fat in the diet can’t be demonstrated to have any effect in reducing mortality in good long-term observational studies, then regardless of whether the basic science indicates that it should, then it’s not worth doing it.  And the medical authorities are wrong.  Thomas Huxley’s comment that a single ugly fact has destroyed many beautiful theories applies here.
Gebt ihr ihr ihr Buch zurück?

Online jt512

  • Frequent Poster
  • ******
  • Posts: 2623
    • jt512
Re: Any science-based article summarizing a healthy diet?
« Reply #81 on: September 29, 2018, 12:57:29 AM »
jt,

How would you know that I didn’t eat the fish?

Biomarkers.

Quote
Anyway, you’ve got it around the wrong way.  First comes the observational studies, then comes the basic science.  People observe that lung cancer seems to be more common in smokers, they do larger observational studies, determine that it is, then do the basic science to see if there’s a mechanism which could explain it.

Cigarette smoking was determined to cause lung cancer without any long-term observational studies, so your example is not applicable to the question.  Long-term epidemiological studies of an hypothesized exposure-disease relationship cannot, by definition, be conducted until there is actually such a hypothesis.  The hypothesis has to come from somewhere other than long-term observational studies. 

Quote
And I never claimed that random controlled trials need to be blinded.

Reread what you wrote.  You claimed that by definition they are.

Quote
I never wrote that random controlled trials are actually at best semi-random semi-controlled trials because they’re non-blinded.

I'm certainly glad to hear that!

Quote
And if reducing saturated fat in the diet can’t be demonstrated to have any effect in reducing mortality in good long-term observational studies, then regardless of whether the basic science indicates that it should, then it’s not worth doing it.  And the medical authorities are wrong.

There's a difference between "can't" and "haven't been," and there are limits to how "good" good nutritional epidemiology is.  The best long-term prospective cohort studies can fail to find a true diet-disease relation for many methodological reasons.  When basic science conflicts with epidemiology, then one or the other has to be wrong.  But there is no obvious reason to presume it is the epidemiology.  Rationally, you must consider the strengths and limitations of each.  Strong evidence from metabolic studies should increase your probability that a hypothesized diet-disease relationship is true, negative results from observational studies should decrease it, but to determine how much, you really need to weigh the strengths and limitations of the studies.  You're very naive about nutritional epidemiologic methodology if you think that strong results from metabolic studies should be ignored in light of null results from nutritional epidemiology.
« Last Edit: September 29, 2018, 01:16:16 AM by jt512 »
Ich verstehe nur Bahnhof.

Offline bachfiend

  • Not Any Kind of Moderator
  • Frequent Poster
  • ******
  • Posts: 2105
Re: Any science-based article summarizing a healthy diet?
« Reply #82 on: September 29, 2018, 03:52:41 AM »
kt,

I never wrote that the harm of cigarette was first examined  with long-term observational studies.  I wrote that first it was observed that lung cancer appeared to be more common in cigarette smokers.  First comes the observation, then more observation, and then the studies to determine the mechanism.

And I wrote that if good long-time observational studies (and they have to be good) fail to demonstrate a benefit of reducing dietary saturated fat intake intake, then it’s not worth restricting saturated fat intake.  It’s an indication you need to do larger, better, bigger studies.
Gebt ihr ihr ihr Buch zurück?

Online jt512

  • Frequent Poster
  • ******
  • Posts: 2623
    • jt512
Re: Any science-based article summarizing a healthy diet?
« Reply #83 on: September 29, 2018, 10:54:52 AM »
kt,

I never wrote that the harm of cigarette was first examined  with long-term observational studies.  I wrote that first it was observed that lung cancer appeared to be more common in cigarette smokers.  First comes the observation, then more observation, and then the studies to determine the mechanism.


We were not talking about casual observation.  We were talking about long-term observational studies.  They are big and expensive, and are thus normally undertaken to follow-up on hypotheses that are already supported by basic science and shorter, cheaper metabolic studies.

Quote
And I wrote that if good long-time observational studies (and they have to be good) fail to demonstrate a benefit of reducing dietary saturated fat intake intake, then it’s not worth restricting saturated fat intake.


Then either you and CarbShark are wrong, or just about everybody else is.  Maybe you should reconsider your understanding of nutritional epidemiology.

Ich verstehe nur Bahnhof.

Offline bachfiend

  • Not Any Kind of Moderator
  • Frequent Poster
  • ******
  • Posts: 2105
Re: Any science-based article summarizing a healthy diet?
« Reply #84 on: September 29, 2018, 03:17:20 PM »
jt,

Unless you, or more relevantly medical authorities, can convincingly demonstrate that there’s a benefit in restricting dietary intake of saturated fat on health and life expectancy, people like CarbShark will be able to claim that the nutritional guidelines published by medical authorities are wrong, and should be replaced with guidelines more in keeping with their ideology including their low carbohydrate/high fat ketogenic diets.

And you can’t base dietary recommendations just on basic science and short-term metabolic studies, even if both are completely correct.  If long-term observational demonstrates consistently fail to demonstrate a benefit, then it means that either 1.  There’s some basic science you’ve failed to consider, which has negated the basic science you’ve used to develop your hypothesis, or 2. Reducing saturated fat has had some other non-metabolic effect increasing mortality, or there’s some other reason increasing mortality despite some causes being reduced, or 3. You need better, bigger studies to demonstrate that there’s a benefit in reducing dietary saturated fat intake.

I’m not doubting that there’s a benefit in restricting dietary saturated fat intake.  I’m on a low fat diet after all.  And it’s not true that nutritional advice was initially based on basic science and short-term metabolic studies.  It all started with the Framingham study in 1948, which was a purely observational study when very little were known about risk factors, and atherosclerosis, hypertension and hypercholesterolaemia were assumed to be normal concomitant of ageing.

There’s a similar study going on just down the coast from me in Busselton.

First comes the observations, then the basic science.  To give one of your pet obsessions, no one developed the hypothesis that the ratio of omega-3/omega-6 fatty acids was important in health based on the basic science of the biochemistry of essential fatty acids and their conversion into various inflammatory autacoids, some beneficial, some damaging.  And developed the hypothesis that manipulating the ratios in the diet might influence health.  And examined the ratio in various foods and discovered that oily fish seemed to have the right ratio.  And then did the long-term observation on dietary fish.

No.  First came the observation that fish eaters seemed to be healthier.  Followed by larger observational studies.  Then followed by the basic studies.

You’ve got a distorted view of science, perhaps produced by your reading of the science journals, which presents the work in a misleadingly way.  Most really significant scientific studies are made as a result  of a surprising observation, demanding new hypotheses.  Such as Kepler’s Supernova in 1604, which challenged the ruling picture of a central unmoving Earth and a perfect unchanging cosmos, with the movement of the planets influencing the life of humans on Earth (and astrology was a science).
Gebt ihr ihr ihr Buch zurück?

Offline bachfiend

  • Not Any Kind of Moderator
  • Frequent Poster
  • ******
  • Posts: 2105
Re: Any science-based article summarizing a healthy diet?
« Reply #85 on: September 29, 2018, 03:49:06 PM »
I should think about starting a new thread.  There was a very interesting article in this morning’s ‘Sunday Age’ reporting that the Melbourne (Victoria, Australia) zoo has been forced to remove fruit from the diets of many of their animals because they’re becoming obese.  The fruit have become too sweet and sugary after thousands of years of selective breeding to suit human tastes, and the captive animals share the drive to gain a calorie rich nutrient which otherwise was limited or difficult to find naturally.

There’s no returning to the diet eaten by humans before the agricultural revolution, even if there was a single diet - which there wasn’t.  Almost everything we eat has been genetically modified, both selectively and inadvertently, including the oceanic fish we eat, which have become smaller and maturing reproductively earlier in response to overfishing and the benefit (to the fish) of individual fish being able to reproduce earlier having an advantage over fish reproducing later (because the latter get caught before they get a chance to reproduce).
Gebt ihr ihr ihr Buch zurück?

Offline PANTS!

  • One leg at a time.
  • Poster of Extraordinary Magnitude
  • **********
  • Posts: 12004
  • What seals? I auditioned for this job.
Re: Any science-based article summarizing a healthy diet?
« Reply #86 on: September 29, 2018, 04:07:11 PM »
Eat a variety of food, not too much, mostly plants.

Would that mean that various meat-substitutes, around here mostly soy-based, are a good thing? Or do they bring the problems of meat even though they are not meat? They are meant to substitute meat after all.

For example, just recently I ate potato wegdes, oumph, and broccoli for dinner. Was that an improvement compared to if I had eaten meat instead of oumph?

I suggest you ask CarbShark and bachfiend for their answers to that question.

Boy you just gave him both barrels didn't you.
Now where I come from
We don't let society tell us how it's supposed to be
-Uptown, Prince 👉

The world is on its elbows and knees
It's forgotten the message and worships the creeds

Online jt512

  • Frequent Poster
  • ******
  • Posts: 2623
    • jt512
Re: Any science-based article summarizing a healthy diet?
« Reply #87 on: September 29, 2018, 04:12:13 PM »

Unless you, or more relevantly medical authorities, can convincingly demonstrate that there’s a benefit in restricting dietary intake of saturated fat on health and life expectancy, people like CarbShark will be able to claim that the nutritional guidelines published by medical authorities are wrong, and should be replaced with guidelines more in keeping with their ideology including their low carbohydrate/high fat ketogenic diets.


So?

Quote
And you can’t base dietary recommendations just on basic science and short-term metabolic studies, even if both are completely correct.


You can't?  Given the inconsistent findings on the relation between saturated fat and heart disease form long-term observational studies, that is precisely what public health officials around the world are doing.  You'd better let them know they can't!


Quote
If long-term observational demonstrates consistently fail to demonstrate a benefit, then it means that either 1.  There’s some basic science you’ve failed to consider, which has negated the basic science you’ve used to develop your hypothesis, or 2. Reducing saturated fat has had some other non-metabolic effect increasing mortality, or there’s some other reason increasing mortality despite some causes being reduced, or 3. You need better, bigger studies to demonstrate that there’s a benefit in reducing dietary saturated fat intake.


Right.  So what?  When did you forget how Bayesian inference works?

Quote
I’m not doubting that there’s a benefit in restricting dietary saturated fat intake.


Wait, what?  If you're not doubting the benefit, but the epidemiology doesn't clearly show one, then you must be basing your opinion on basic science?  Unless you're just being irrational.


Quote
And it’s not true that nutritional advice was initially based on basic science and short-term metabolic studies.  It all started with the Framingham study in 1948, which was a purely observational study when very little were known about risk factors, and atherosclerosis, hypertension and hypercholesterolaemia were assumed to be normal concomitant of ageing.


That's a valid example, but it's not 1948 any more.  Modern long-term nutritional epidemiological studies are primarily used to test associations suggested by cheaper, more-preliminary research.  In fact, the items selected for inclusion in modern FFQs are not intended to represent a person's complete diet, but rather are specifically chosen because of their likelihood of being risk factors for disease or (confounders that need to be controlled for) based on prior research and current hypotheses.  Moreover, it is important for statistical validity that the hypothesis tested in long-term epidemiological studies be determined a priori, rather than being the result of exploratory digging through the data.

Quote
First comes the observations, then the basic science.  To give one of your pet obsessions, no one developed the hypothesis that the ratio of omega-3/omega-6 fatty acids was important in health based on the basic science of the biochemistry of essential fatty acids and their conversion into various inflammatory autacoids, some beneficial, some damaging.


Citation needed.

Quote
You’ve got a distorted view of science, perhaps produced by your reading of the science journals, which presents the work in a misleadingly way.


Quoted for posterity.
Ich verstehe nur Bahnhof.

Online CarbShark

  • Poster of Extraordinary Magnitude
  • **********
  • Posts: 12012
Re: Any science-based article summarizing a healthy diet?
« Reply #88 on: September 29, 2018, 05:00:59 PM »
Hey, Moderators, this very interesting conversation about a healthy diet has been derailed by an interesting discussion about the science of epidemiological studies vs. RCTs.

Maybe someone could move these posts into a new thread?
and Donald Trump is President of the United States.

I'm not a doctor, I'm just someone who has done a ton of research into diet and nutrition.

Online jt512

  • Frequent Poster
  • ******
  • Posts: 2623
    • jt512
Re: Any science-based article summarizing a healthy diet?
« Reply #89 on: September 29, 2018, 05:01:31 PM »
To see whether prospective cohort studies are hypothesis generating, as you suggest, or intended to confirm diet-disease relationships suggested by cheaper, shorter studies, I searched Pumbed for "Willett WC" and took the following excerpts from first (most recent) three reports from prospective cohort studies.  Here are the results.


1. Liu X, Li Y, Tobias DK, Wang DD, Manson JE, Willett WC, Hu FB. Changes in
Types of Dietary Fats Influence Long-term Weight Change in US Women and Men. J
Nutr. 2018 Sep 22. doi: 10.1093/jn/nxy183. [Epub ahead of print] PubMed PMID:
30247611.

Previous weight-loss trials have examined the effectiveness of varying amounts of macronutrients on weight loss and maintenance over short periods of time. However, these trials cannot be generalized to long-term weight gain prevention in the general population. Therefore, the present study aimed to investigate the associations between changes in intakes of varying types of dietary fat and long-term weight gain in women and men from 3 independent prospective cohort studies.


2. Tahir MJ, Michels KB, Willett WC, Forman MR. Age at Introduction of Solid Food
and Obesity Throughout the Life Course. Obesity (Silver Spring). 2018 Sep 11.
doi: 10.1002/oby.22277. [Epub ahead of print] PubMed PMID: 30204942.

Studies evaluating the association between age at SF introduction and obesity have inconsistent results. Research has shown higher odds of increased weight for age in infancy (8) and childhood overweight and obesity (9, 10) with earlier SF introduction, whereas other studies have demonstrated null associations in childhood and adolescence (11, 12). Results from one study indicated the association between early SF introduction and childhood obesity existed only among formula‐fed infants (13). Other research has suggested that delayed introduction of SF beyond 6 months may in fact increase the risk of childhood obesity (14, 15). Several of these studies have been limited by their relatively small sample sizes and an inability to adjust for multiple confounding variables or track results into adolescence or adulthood. Therefore, our aim was to evaluate the association between age at introduction of SF and obesity at different stages of the life course using a large ambidirectional cohort study with exposure and outcome data from two independent sources and adjustment for potential confounders.


3. Seidelmann SB, Claggett B, Cheng S, Henglin M, Shah A, Steffen LM, Folsom AR,
Rimm EB, Willett WC, Solomon SD. Dietary carbohydrate intake and mortality: a
prospective cohort study and meta-analysis. Lancet Public Health. 2018
Sep;3(9):e419-e428. doi: 10.1016/S2468-2667(18)30135-X. Epub 2018 Aug 17. PubMed
PMID: 30122560.

This paper, which we've discussed before, actually contains a section entitled, "Evidence Before This Study."  It begins, "Although many randomised controlled trials of low carbohydrate diets suggest beneficial short-term weight loss and improvements in cardiometabolic risk, mortality risk has typically not been investigated in light of the practical challenges posed by studies involving very long durations of follow-up.


Each paper was intended to clarify or resolve a question left open by short-term RCTs and/or more-limited observational studies.

Nowadays you can't even get funding for a major cohort study unless you can justify it on the basis that prior research.
Ich verstehe nur Bahnhof.

 

personate-rain