Lightnews — Scholar-powered news

Health Nerd

@gidmk.bsky.social

30K followers 400 following 2.3K posts

Epidemiologist. Research Fellow. Doctor of Spreadsheets. Writer (Slate, TIME, Guardian, etc). PhD, MPH. Host of senscipod Email [email protected] he/him. Find my writing on Substack and Medium.

Posts Media Videos Starter Packs

Pinned

Health Nerd @gidmk.bsky.social · 8d

Cocoa supplements have been in the news as an anti-aging tool.

My new piece breaks down why this is a very dubious finding:

gidmk.substack.com/p/chocolate-...

Chocolate Can't Stop You From Aging

The dubious claims about cocoa supplements and anti-aging potential

gidmk.substack.com

2 2 29

Health Nerd @gidmk.bsky.social · 7d

Astonishingly, since posting this comment I've already identified more errors in the paper. This statement does not match the very first row of the table directly underneath it, for example:

1 6 41

Health Nerd @gidmk.bsky.social · 7d

I have just left a lengthy Pubpeer comment on the autism/Tylenol study being cited by the Trump administration. There appear to be a number of fairly straightforward errors in the paper.

pubpeer.com/publications...

PubPeer - Evaluation of the evidence on acetaminophen use and neurodev...

There are comments on PubPeer for publication: Evaluation of the evidence on acetaminophen use and neurodevelopmental disorders using the Navigation Guide methodology (2025)

pubpeer.com

5 49 150

Health Nerd @gidmk.bsky.social · 8d

Cocoa supplements have been in the news as an anti-aging tool.

My new piece breaks down why this is a very dubious finding:

gidmk.substack.com/p/chocolate-...

Chocolate Can't Stop You From Aging

The dubious claims about cocoa supplements and anti-aging potential

gidmk.substack.com

2 2 29

Health Nerd @gidmk.bsky.social · 8d

9/n We may not be able to say for certain what happened here, but we can be quite sure that the study reported by the authors in 2024 definitely did not happen as it was described.

It's plausible that it never happened at all.

Health Nerd @gidmk.bsky.social · 8d

8/n I have not seen the dataset myself, but having read this report I would be surprised if these are the only issues. Indeed, that's what the conclusion says.

1 1 11

Health Nerd @gidmk.bsky.social · 8d

7/n

7. There are signs in the dataset that the main outcome measure of BMI may have been intentionally manipulated to find a significant result. Specifically, BMI was incorrectly calculated for the high-dose and placebo groups in a way that made the effect appear larger.

1 1 11

Health Nerd @gidmk.bsky.social · 8d

6/n

4. The outcomes of the study had the same issue except worse.

5. The baseline data of the groups was highly unbalanced, meaning this dataset could not have been randomized.

6. The study drastically UNDERestimated the treatment effect(!).

1 2 12

Health Nerd @gidmk.bsky.social · 8d

5/n

3. The baseline values of the study - i.e. Table 1 - could not be replicated. In 42/84 of cases, the mean and SD of the dataset sent through by the authors was more than 10% out from what they stated in their manuscript.

1 2 10

Health Nerd @gidmk.bsky.social · 8d

4/n Some wonderful highlights of the statistical report:

1. The study didn't adhere to CONSORT guidelines.
2. The statistics section of the methodology was wrong.

1 2 10

Health Nerd @gidmk.bsky.social · 8d

3/n The really interesting thing here is that the BMJ commissioned a full statistical report into the data that the authors sent through for their study.

It is DAMNING and a great read. Sadly, it's just the supplementary to the retraction and not the main body: nutrition.bmj.com/content/earl...

Retraction: Apple cider vinegar for weight management in lebanese adolescents and young adults with overweight and obesity: a randomised, double-blind, placebo-controlled study

This paper1 is retracted by the journal, with the agreement of the authors.

nutrition.bmj.com

1 2 13

Health Nerd @gidmk.bsky.social · 8d

2/n The errors noticed by @jamesheathers.bsky.social and others in the original study were numerous. I'm not going to go into the issues except to say you should read James' blog about the whole debacle: jamesclaims.substack.com/p/research-i...

Research Integrity Is A Clown Car Which Continually Spills Forth A Truly Surprising Quantity Of Sad, Honking, Incompetent Clowns

The anatomy of an utter failure of academic governance

jamesclaims.substack.com

1 2 17

Health Nerd @gidmk.bsky.social · 8d

This is a fascinating story.

A Lebanese team published a study showing massive benefits for apple cider vinegar in 2024. People immediately noticed it was problematic, and complained.

BMJ asked the authors for the data, and they sent it through. The study has just been retracted and WOW 1/n

3 15 53

Health Nerd @gidmk.bsky.social · 9d

If you're in Sydney and know people/have patients with metabolic disease (diabetes, cardiovascular disease, high blood pressure, etc) - let them know about the PANDA trial! Being run by my former PhD supervisor:

www.powerlab.site/research/pan...

PowerLab - Phase 1 & 2 PANDA Trial

PANDA Trial Physical Activity in Nature for Cardiometabolic Diseases in People Aged 45y+

www.powerlab.site

5 20

Health Nerd @gidmk.bsky.social · 10d

Hahahaha yes. My daughter is 2, this tracks.

Health Nerd @gidmk.bsky.social · 11d

All of the studies have inadequate measures of exposure. In many ways, the Swedish one is the most robust (clinician report+prescribing vs mostly self-report) but regardless there's no scientific reason to believe that such biases flow in the direction the authors want them to.

1 4

Health Nerd @gidmk.bsky.social · 11d

Yes it's bizarre and arbitrary. There's no clear reason given for any of their decisions, which makes it feel like it's basically their opinions masquerading as science.

Health Nerd @gidmk.bsky.social · 11d

They put more faith in the 1,000-person Boston children's cohort, which only measured acetaminophen exposure in fetal cord blood and therefore not at all DURING pregnancy than in an analysis of every child born in Sweden which linked data to national drug records.

✨️science✨️

2 7

Health Nerd @gidmk.bsky.social · 11d

It's worth noting that their causal interpretation is largely based on arbitrarily downgrading the largest and most robust study that found no association btwn Tylenol and autism in favour of much weaker studies that did find an association.

2 8

Reposted by Health Nerd

Prof Gavin Yamey @gavinyamey.bsky.social · 12d

VERY good piece, as always, by the brilliant @gidmk.bsky.social

"The data thus far, especially the highest-quality studies out there, does not show that acetaminophen causes autism. It doesn’t even show a consistent association."

gidmk.substack.com/p/tylenol-us...

Tylenol Use In Pregnancy Probably Doesn't Give Babies Autism

Why acetaminophen/paracetamol probably isn't causing the autism epidemic.

gidmk.substack.com

6 21

Health Nerd @gidmk.bsky.social · 11d

But it's the public health experts who damage trust in science.

IT WAS ALWAYS A LIE.

Aaron Rupar @atrupar.com · 12d

Highly normal. Nothing to see here, folks.

Pregnant Women, DON’T USE TYLENOL UNLESS ABSOLUTELY NECESSARY, DON’T GIVE TYLENOL TO YOUR YOUNG CHILD FOR VIRTUALLY ANY REASON, BREAK UP THE MMR SHOT INTO THREE TOTALLY SEPARATE SHOTS (NOT MIXED!), TAKE CHICKEN P SHOT SEPARATELY, TAKE HEPATITAS B SHOT AT 12 YEARS OLD, OR OLDER, AND, IMPORTANTLY, TAKE VACCINE IN 5 SEPARATE MEDICAL VISITS! President DJT

4 4 47

Health Nerd @gidmk.bsky.social · 12d

New piece out - do soft drinks give you depression?

Probably not. The evidence is SUPER weak. gidmk.substack.com/p/soft-drink...

Soft Drinks Probably Aren't Making You Depressed

Why soda/pop/soft drinks probably aren't causing depression.

gidmk.substack.com

2 2 28

Health Nerd @gidmk.bsky.social · 12d

I would guess that the interpretation here also hinges on how the authors have defined the terms "right-wing", "left-wing", and "terrorist attack".

Health Nerd @gidmk.bsky.social · 12d

10/n Regardless, all of this fuss and noise about soft drinks and depression comes from a cross-sectional study that only assessed soft drink consumption at one time point and did not follow-up patients at all.

Which is absurd.

1 3 32

Health Nerd @gidmk.bsky.social · 12d

9/n If any causal inference people want to correct me, please do. I cannot see anything that would correct for the issues in study design here, but maybe I'm wrong. It just seems like a very complex way to generate a DAG and then examine the ensuing associations to me.

1 16

Health Nerd @gidmk.bsky.social · 12d

8/n They wax lyrical about how amazing this model is, but from what I can tell it does nothing to address the underlying weakness of the data itself. This is just a very fancy and convoluted way of estimating associations in a cross-sectional cohort study, not a miracle of causal inference.

1 15