Lightnews — Scholar-powered news

John Pickering

@kiwiskinz.bsky.social

1.5K followers 290 following 330 posts

Research Professor, but prefer Scrymaster. Embedded scientist and statistician in emergency medicine. See #NerdNite talk https://youtu.be/Gd180NiWSCg. Don't believe in statistical significance. Amateur astronomer and will post photos of the heavens.

youtu.be

Posts Media Videos Starter Packs

Pinned

John Pickering @kiwiskinz.bsky.social · Dec 13

Where 2 Emergency Medicine docs are gathered there are 3+ opinions. Follow them all here where multiple hashtags are included #EMedSky, #EMSky, #EmergencyDepartment, #EmergencyRoom, #EmergencyCare
bsky.app/profile/did:...

1 2 13

John Pickering @kiwiskinz.bsky.social · 2d

I agree with this ... I teach never to make dynamite plots simply because they give the wrong impression of data from the bottom of the plot to the mean/median and not much above it.

John Pickering @kiwiskinz.bsky.social · 4d

2. Get Nobel in Phys. ❌ 😞

John Pickering @kiwiskinz.bsky.social · 5d

I asked OpenEvidence & Asta medical literature trained #AI s "What does a troponin of 4 mean?"
Neither said the qn makes no sense or it depends on assay
Only Asta said it depends on units
Asta's references were very dodgy
OpenEvidence's better, but out of date
Caveat Emptor
#AI #EMedSky #CardioSky

John Pickering @kiwiskinz.bsky.social · 5d

If you've a great memory... then that works... but not for me (the search function isn't great). I have to put into groups (or tag) - groups for topics, and (temporary) groups for a projects/papers.

Reposted by John Pickering

Winfried Ruigrok @winfriedruigrok.bsky.social · 5d

If you sometimes feel uncertain about what #statistics do or do not tell us, this #podcast from early September is well worth listening to. (And while you are at it, most other episodes of the #InGoodCompany series are also well worth your time.) shows.acast.com/622618c7057f...

Sir David Spiegelhalter: Statistics, Communicating in Crises, and Living with Uncertainty | In Good Company with Nicolai Tangen

shows.acast.com

1 1

John Pickering @kiwiskinz.bsky.social · 5d

This week's plan:
1. Get Nobel in Med. ❌ 😞
2. Get Nobel in Phys.
3. Get Nobel in Chem.
4. Get Nobel in Lit. ... not eligble 'cos can't writ. 😜
5. Get Noble in Peace ... 🤣 ... I want to bang all their heads together!!!!!

1 1

Reposted by John Pickering

American College of Cardiology @accintouch.bsky.social · 21d

“We can’t think about one anymore without the other,” says Martha Gulati, MD, MS, FACC on the link between treating the kidney and protecting the heart.

"We should aim to reduce the progression of chronic kidney diseases in order to reduce the possibility of heart disease." #ACCLATAM #ACCIntl

4 5

Reposted by John Pickering

EMA Journal @emajournal.bsky.social · 8d

The potential use of clinical biomarkers as a rule-out test for intracranial pathology in head-injured patients

doi.org/10.1111/1742...

#OpenAccess

@wiley.com
#MedSky #EMSky #EMedSky #emimcc #ACEM

1 1

John Pickering @kiwiskinz.bsky.social · 6d

For #Stunday the Sculptor Filament Galaxy captured this week from Christchurch #NZ At 11M light years it’s really close! About 70% diameter of the moon in the sky.

Reposted by John Pickering

F. Perry Wilson, MD @fperrywilson.bsky.social · 10d

More detail / nuance in my full writeup about this article on @medscape if you want to dive deeper: www.medscape.com/viewarticle/...

/🧵

Is the Rise in Cancer in Young People... Fake?

A study looking at the eight fastest-rising cancers in young adults concludes that the real problem is overdiagnosis. But not everyone agrees.

www.medscape.com

1 1

Reposted by John Pickering

F. Perry Wilson, MD @fperrywilson.bsky.social · 10d

The facts everyone agrees on: Early-onset cancer rates (before age 50) ARE rising. Some cancers like thyroid, kidney, endometrial, and colorectal are up >1% per year since 1992.
The debate? Whether these new cases actually matter. 2/

1 1 1

John Pickering @kiwiskinz.bsky.social · 13d

I don't understand "too small" ? Do you mean it is a small separate study with only one or two results - then a research letter may be appropriate. Or is it simply a very small effect size? The size of the effect should not influence whether something is published or not (nor the p-value). Or...?

John Pickering @kiwiskinz.bsky.social · 14d

Comet 2025 R2 Swan from New Zealand soon after sunset. #stunday #astronomy #NZ

1 7

John Pickering @kiwiskinz.bsky.social · 20d

Greeted the sun from Christchurch, New Zealand this morning. #NZ #eclipse #stunday #astronomy

6 29

John Pickering @kiwiskinz.bsky.social · 24d

Just saying
#StatsSky #rstats #Statistics

Graphs demonstrating when a machine learning prediction model did not outperform the reliable statistical method, logistic regression.

1 5

John Pickering @kiwiskinz.bsky.social · 25d

Because it's been a while - for those working in Emergency Medicine there is a feed for you with all your fav tags:
bsky.app/profile/kiwi... #ED #ER #EMSky

John Pickering @kiwiskinz.bsky.social · 25d

To improve equity "opportunity" risk stratification for cardiovascular outcomes with a mammogram is intriguing. Whilst not convincingly better than the #NZ PREDICT score, this looks good work & certainly worth more investigation. Calibration is impressive (hidden in supplement :( ) #CardioSky

Eric Topol @erictopol.bsky.social · 25d

Current read out for a mammogram leaves a lot on the table.
A new study shows how risk of cardiovascular disease can be predicted as well as traditional cardiovascular risk scores (aka opportunistic A.I.)
heart.bmj.com/content/earl...

John Pickering @kiwiskinz.bsky.social · 25d

Darn..may have to redo our research which showed the newer (now older) scanners can be used instead of lumber punctures for up to 24h post headache onset to detect brain bleeds (SAH). ps - photo is of Phil Adamson #Cardiologist - one to watch in the academic space.

John Pickering @kiwiskinz.bsky.social · 25d

The Return of the Godwit

John Pickering @kiwiskinz.bsky.social · Sep 12

Took #rstats #ggplot 4.0.0 for a spin with @universityofotago.bsky.social colours to test my latest effort to fake biomarker data
Thanks @thomasp85.com - brilliant.
See www.tidyverse.org/blog/2025/09...

2 10

John Pickering @kiwiskinz.bsky.social · Sep 12

... does maths using absolute zero... actually 1.035 times warmer... but then that's all it takes - move to Ōtautahi everyone.
ps @Paul... will this mean more stunning photos of snow?

John Pickering @kiwiskinz.bsky.social · Sep 12

I'd thought - good journal - interesting cohort - this maybe a good #machinelearning paper
Sadly - trained with ~50 features yet only 23 events!
Validated with only 16 events!
No calibration!
Please don't do this
#cardiosky #statsky #ACS
www.ahajournals.org/doi/10.1161/...

Using Machine Learning to Risk Stratify Emergency Department Patients With Chest Pain but No Acute Myocardial Infarction: A Multicenter Retrospective Analysis | Journal of the American Heart Associati...

www.ahajournals.org

John Pickering @kiwiskinz.bsky.social · Sep 11

Trying to fix this ... or maybe just sowing more confusion. #statistics #regression

John Pickering @kiwiskinz.bsky.social · Sep 11

I just read this nice pre-print on interpretation of complex prediction models using the #rstats and #python marginaleffects package by @dingdingpeng.the100.ci & @vincentab.bsky.social . Very accessible & intuitive. osf.io/preprints/ps...
#statistics

OSF

osf.io

1 2 10

John Pickering @kiwiskinz.bsky.social · Sep 11

Fair rant...sounds like dodgy simultaneous submission-to-multiple-jnls behaviour?

I once gave a critical review with suggestions - paper was rejected. A few weeks later from another jnl I was asked to review the same paper - the authors had made no changes! So, I sent the same review. Grrr