Samuel Weisenthal
banner
samweisenthal.bsky.social
Samuel Weisenthal
@samweisenthal.bsky.social
stats ∩ medicine ∩ humanity

Phd statistics, ms data science, family medicine resident in community setting.

For the sick and those who care for the them.

https://samuelweisenthal.github.io
Pinned
Constant sens/spec assumption (see Dawid 1976, Moons and Harrell 2003, Guggenmoos 2000) + pre-test probability scores (eg Wells’) = Information Mismatch

Preprint: arxiv.org/pdf/2503.15382
arxiv.org
decision analysis encourages one to consider all outcomes (and their probabilities), including those that involve side effects

statmedhumanity.org/2025/08/25/d...
Decision analysis and the pharmaceutical industry
Summary: decision analysis encourages one to consider all outcomes (and their probabilities), including those that involve side effects. I previously described decision analysis, a well-known frame…
statmedhumanity.org
October 12, 2025 at 11:34 PM
“Pride cometh before the fall”
- regression to the mean
September 28, 2025 at 11:21 PM
One of the more difficult/ tedious parts of being a resident in family medicine is keeping track of vaccines for each patients.
September 21, 2025 at 4:58 PM
Reposted by Samuel Weisenthal
People are constantly asking me — what can I do? Between now and the 2026 midterm elections, here are five practical steps you can take to make a difference.
August 21, 2025 at 10:33 PM
Reposted by Samuel Weisenthal
The GOP governors of South Carolina, Louisiana, Mississippi, and Tennessee have all sent their states' National Guard troops to assist Trump's occupation of DC.

All four of those states have cities with crime rates higher than DC.

This was never about "fighting crime."
August 20, 2025 at 11:01 PM
If you’re going to use an example to help explain a concept, it may not be to your advantage to make the example more complicated than the concept itself
August 3, 2025 at 4:13 PM
4 hours debugging code saves you 10 minutes planning it
July 26, 2025 at 8:38 PM
If a trial is powered for blood pressure as a primary outcome, take care when drawing conclusions about noisier secondary outcomes, such as time-to-dialysis.
July 14, 2025 at 2:10 AM
Reposted by Samuel Weisenthal
Jewish leaders denounced President Trump on Friday for using an antisemitic term at a rally in Iowa the previous evening.
Jewish Leaders Denounce Trump’s Use of Centuries-Old Trope
President Trump said he did not realize that the term “Shylock” was viewed as antisemitic.
trib.al
July 5, 2025 at 11:55 PM
Hello everyone! I just finished writing a piece on large language models and medical decision making.

I argue that one who sets out to build a chatbot that solves medical problems will ultimately find themselves staring into the cold, hard eyes of evidence-based medicine.
July 4, 2025 at 12:27 AM
Reposted by Samuel Weisenthal
When bullshiters tell you they will soon be able to make "digital twins" of patients, remember articles like this about how hard it is to model a single cell.
Ewen Callaway reports on biology's rush for AI-enabled 'virtual cells'.

So far it "has a lot of hype, but not a lot of concrete results or a clear path to success".

"I don’t think there is a clear definition of a virtual cell", says CZI's Jonah Cool.
@nature.com
www.nature.com/articles/d41...
Can AI build a virtual cell? Scientists race to model life’s smallest unit
Several groups hope to develop artificial-intelligence models that can predict how cells behave.
www.nature.com
June 28, 2025 at 5:49 PM
If you’re going to start a statistics phd, write out a Taylor expansion 50 times
June 26, 2025 at 11:43 PM
Like this idea that research question should precede regression. Especially in the medical field, patient should always come before model
June 26, 2025 at 11:41 PM
Recommend!
Vital Statistics is my newsletter where I give you the tools to think clearly about data in an age of lies and distortions.

Statistics not as math, but as a pathway to seeing the world more clearly.

Launching in 5 weeks. Sign up now: kareemcarr.substack.com
June 24, 2025 at 1:35 AM
Large language model extrapolates—“hallucination.”

Human extrapolates—“creativity.”
June 24, 2025 at 1:34 AM
Reposted by Samuel Weisenthal
Launching the U.S. into what could be yet another unpopular and costly forever war, while simultaneously pushing for an unpopular budget bill at home that will take food and healthcare away from millions of people.

The "peace president."
June 22, 2025 at 4:01 PM
Reposted by Samuel Weisenthal
Samuel J. Weisenthal, Matthew Eckard, Askhan Ertefaie, Marissa Sobolewski, Sally W. Thurston
Optimality-based reward learning with applications to toxicology
https://arxiv.org/abs/2404.04406
April 9, 2024 at 8:09 PM
Wednesdays should be holidays
June 5, 2025 at 1:45 AM
Smartphones mine dopamine, dopamine floods lead to insensitivity, dopamine insensitivity leads to inattention (adhd) - all generations harmed
Are #smartphones and #socialmedia harming a generation?

This is a hotly debated and often polarizing debate. So we surveyed over 120 experts on the topic to see where there was genuine consensus (or not), like experts have previous done for climate change.

See our paper: osf.io/preprints/ps...
May 19, 2025 at 10:41 PM
Imagine if the early opioid trials had better measured addiction
In clinical trials it is not only potential benefits that should be measured - it is crucially important that any potential #harms are also investigated 1/7
#MethodologyMonday #118
May 19, 2025 at 10:18 PM
When deciding whether to take a med, we think about benefit but also side effects.

We need more risk models that predict whether someone will experience a side effect when taking a med.
May 18, 2025 at 1:39 PM