Crisis 2. More on the identical twin vegan study.

Habeas Corpus Datorum. Let us see the body of the data.

The landmark triumph of Zoë Harcombe and Malcolm Kendrick, along with the recent Noakes case brought out the distressing state of medical research and just how much it takes to introduce accountability and reasonable behavior.  Few of us have the determination or strength to follow theses examples.

The lack of standards and outright hostility is difficult to face. I have personally have been discouraged as anybody else. A reasonable first start of simply trying to get communication — sensibly the first requirement of scientific endeavor — is not reinforcing behavior.

Here’s an attempt to interact with authors of questionable publications. I try to keep it non-confrontational.  Cooperation is possible. After all, it is, in fact what I do if one of my working colleagues publishing something that raises questions. I simply ask. I don’t challenge them (unless we are close). I expect an answer and I recognize that I might actually learn something.

I recently came across a podcast with Christopher Gardner on a site called Zoe — not to be confused with our hero, umlaut-Zoë. Now, I knew Gardner from a couple of previous meetings. While we are not exactly on the same wavelength — he is a vegetarian or maybe a vegan — but he is not particularly doctrinaire. Also, he actually does experiments, most famously, the A-Z study. In any case, I had some differences of opinion on the podcast so I wrote suggesting we have a dialog or debate on the Zoe site. I was actually surprised it didn’t work out. In the course of this I also read a paper on a twin study on vegan diets discussed in my recent Substack. Anyway, here’s the exchange. The emails are in standard text. My comments (afterwards) in italics.

==================================================================

Date: Tuesday, March 5, 2024 at 11:13 AM
To: Christopher Gardner

Hi Chris,

Greetings. I saw you on a few of the ZOE series, particularly the ones on protein and carnivore.  I think you did a good job but, as you know, we have some disagreements.  I would suggest you and I go on ZOE and discuss some of the issues. I don’t think we are far enough apart to consider it a debate — I am probably the least doctrinaire of low carb/keto researchers — but I think we could bring out some relevant subject matter.  Would you be open to writing to Jonathan Wolf of ZOE with me to see if he would have us both as guests on the podcast?

Regards, RDF

==================================================================

March 5, 2024

Richard,

Sure. Happy to agree to disagree, or even to be moved to agree with something you challenge me on.

I recently had a “debate” with Stu Phillips about protein on a podcast, and after a few preliminary e-mail exchanges we agreed in advance it wouldn’t be much of a debate…..we agreed on most topics, and had different perspectives, but not strong disagreement, on others.

I think it was GREAT to share that kind of respectful exchange with listeners rather than having them hear what sounds like polarized and irreconcilable differences…

Christopher

==================================================================

March 11, 2024

Jonathan,

Hope you and Zoe and the podcasts are doing well.

Richard Feinman from SUNY has reached out to me to suggest a possible Zoe podcast…….Protein and Carnivore Diets

See below.

As he suggests….we aren’t far enough apart to have a debate, but we disagree on some of the issues, perhaps enough to make for an interesting session.

I’m open to this possibility.

Whadya think?

Warmly, Christopher

==================================================================

Hi Chris,

I haven’t heard from Jonathan and I was listening to something on Ultimatehealthpodcast and I wrote to Jesse Chappus about our doing the round table on his podcast and he seems interested. Want to try this?

Regards,

RDF

==================================================================

From: Jesse Chappus <jessechappus@gmail.com>
Date: Tuesday, May 14, 2024 at 6:07 PM

Hi Chris and Richard!

Let’s do this - it will be fun and the audience will love it :)

Talk soon. Jesse

==================================================================

Jesse and Richard,

Looping in *** who is our ** Director.

They are helping me to manage Podcast scheduling.

Thanx,

Christopher

==================================================================

Nothing moved forward. Meanwhile I checked Gardner’s paper on an identical twin study. It had obvious errors — in particular, a twin study is supposed to allow comparison of individual. The paper, however, reported group statistics. I discussed this in my Substack “The Crisis. 1. The Identical Twin Vegan Study.” Some time had passed so I wrote to Gardner.

==================================================================

From: Richard Feinman <feinman@icloud.com>
Date: Monday, July 8, 2024 at 7:06 PM
Subject: twins paper, etc.

Chris,

My original intention as I indicated was to get some discussion on keto and carnivore — still possible — but I came on this and I think it is potentilly useful. I give you a hard time on my Substack post but I think there's some possible benefit. Most of the objections I made were real: SD, not SEM is the correct statistic.  More to the point. If you have a twin study you compare the twins. In fact, I think most of the time group statistics is useless.  You can't talk about individualization and also show the group average. As in Lake Wobegon, nobody is average. So, if you use the matrix method — this is a simple method that we had published that gives a good intuitive view of results of an experiment — you only need 3 columns really -- and just fill in the data (or very second best, if you send me the data). The prediction is that the twin on diet V will have lower LDL than the twin on O for most of the pairs. I don’t think you could make a global prediction on this data but I would change my mind on outcomes within the study. We could branch out to the original question. Now is the moment for the breakthrough cooperation.  The reason is that people are already making up their minds. And the continual progression of data mining to no point is, as they say on YouTube, effing unproductive. This is a good idea. Right? (My emphasis).,

Regards, RDF

==================================================================

Richard,

Cc’ing our statistician, Dr. ****, with this response.

I couldn’t make much sense out of your communication.

Keto and carnivore?

SD vs. SEM? Those were not used for statistics, they were used for data presentation.

That’s what’s wrong. It is universally agreed that SEM (the standard error of the mean), a statistical measure that does not show the actual variation in the data points — will always make your data look better than it is. It is precisely using it for data presentation that is misleading. Presentation that shows the variability usually employs the standard deviation (SD).

For statistical testing, the model is described in the paper, and includes a term for the twin pairs. “Primary analysis included all available data. A linear mixed model was used and included fixed effects for diet and time (baseline as reference) and an interaction effect for diet x time and a random effect for twin pair to account for the correlation between identical twins (ie, random intercept allowed intercept to vary for each twin pair).”

I don’t know what this means but it can’t change the conclusion. We didn’t get to see the data, or even a statistical measure reflecting the variation.

???????

Christopher

==================================================================

July 9, 2024.

Richard,

Thanx for the interest. I have too many papers, too many studies, and too many vacation days coming up for this.

Sounds interesting, but my schedule simply won’t fit anything new into my calendar or brain.

Christopher.

==========================================================

To: Christopher Gardner 

No sweat.  As per the offer in the MS, you could you have someone send me the data.

Note: (Supplement 3 of their paper states

“Who can access the data: Researchers whose proposed use of the data has been approved.

Types of analyses: For any research purpose.”)

I don’t need everything. I can send a template to be filled in. I will, of course, not publish anything without first running it by you. Thanks.

Regards,

RDF

==========================================================

July 10, 2024

Richard,

Appreciate that you don’t need everything.

Good, since that is >1,000 variables (if you include all the diet variables, all the inflammatory markers, all the…..).

What DO you want?

How about….

LDL-C, at all time points, with randomization assignment, and identification of twin pairs. Is that enough for now?

Christopher

==========================================================

July 10

That would be poifict, as we say in Brooklyn. Look forward to it.

Discuss sometime, if you like, but it is presentation of data that is a major concern. I think traditional statistics is inherently misleading and different approach needed. I would, of course, be a more convincing critic if I really understood traditional statistics but the principle, as in my Substack, is that science comes first.

Anyway, thanks.

RDF

==========================================================

Some time elapsed and no message..

September 20 (sic).

To: Christopher Gardner <cgardner@stanford.edu>

1. We are all busy. I myself do not spend time binge-watching old Law and Orders.

2. Your paper offers to provide the original data.

Ok. Maybe I was getting a little impatient.

3. There is nothing more important in your paper than the comparison of the twin data.

Spreadsheet template below to be filled in.

==========================================================

The data were never sent to me. Very suspicious since I personally didn’t think they would show much but maybe the conclusion really doesn’t hold up.

==========================================================