Benjamin P. Geisler
ben-geisler.com
Benjamin P. Geisler
@ben-geisler.com

Stipendiat in Oslo, 🇳🇴 | prior hospitalist | Stuck between 3 continents/still fresh off the boat | RT≠opinion≠others'

Economics 45%
Medicine 32%

Moltbook: 1.5M AI agents posting on a social network where humans just watch. Karpathy calls it "sci-fi takeoff-adjacent." Real tech capability or elaborate performance art? Probably both. Worth watching either way. 🦞🤖

Update: Now live on Kindle! 📱

Seeing "50 Studies Every Hospitalist Should Know" on Amazon feels surreal after months of work with brilliant contributors worldwide.

Paperback soon | www.amazon.com/Studies-Ever...

Just published: our roadmap for making innovative therapies affordable & accessible across Europe. Proud to be part of the ASCERTAIN team! #HEOR #HealthEconomics
www.sciencedirect.co...

New framework for evaluating medical LLMs published in Nature Medicine!

My take: Interesting standardization effort, but may be too general for real-world use & models evaluated are already outdated...
Holistic evaluation of large language models for medical tasks with MedHELM
Nature Medicine - MedHELM, an extensible evaluation framework including a new taxonomy for classifying medical tasks and a benchmark of many datasets across these categories, enables the evaluation...
www.nature.com

Fascinating essay uses ecological metaphors for academia. Having worked across 3 continents, Australia was most neoliberal (high metrics, high pay). But academia needs what metrics can't measure: time to think, boredom, permission to fail.
Jean Fan | Personal site
Personal website of Prof. Jean Fan
jean.fan

Claude reviews a prior authorization request!

Is this good (speedy, less costly) or bad (AI arms race on payor and provider sides)?
Speed up prior authorization reviews with Claude
Claude helps utilization management teams complete initial prior auth reviews in minutes—work that typically takes hours of manual cross-referencing across policies, databases, and documentation, which can slow patients’ access to care. Watch how Claude validates provider credentials in the NPI Reg
youtu.be

Claude write a clinical trial protocol with its new healthcare/life sciences connectors and skills.

Is this good (speeds up trial planning) or bad (deskills P.I.s, more "cookie cutter" trials, not enough critical review in some cases)?
Draft clinical trial protocols with Claude
Claude helps clinical development teams draft Phase 2 protocols in under an hour—work that typically takes days of cross-referencing FDA guidelines, competitor trials, and internal templates. Watch how Claude digests preclinical data, takes relevant FDA and NIH requirements into account, searches Cl
youtu.be

New 13-year study of 4K Norwegian pregnancies with RA: medication use during pregnancy up 31%→50% (2008-18), driven by TNF inhibitors. But gaps remain in preconception care.

w/ Eun-Young Choi, Marleen van Gelder, 오인선In-Sun Oh, and Hedvig Marie Nordeng
www.sciencedirect.co...

SMDM 2026 social event announced: Oslo City Hall in the Nobel Peace Prize Reception Hall. Bread and Games subsequently.

smdm.org/civicrm/eve...

This book was started by the wonderful Hugo Torres and Priyank Jain and the series by Michael Hochman.

Great post in Bob Wachter's new substack:

Will A.I. kill health care jobs, too (and not just coders', Uber drivers', or those in admin, manufacturing, retail ...)?

Bob's conclusion: yes, but it may not be the clinical jobs!

CDI? Compliance? More so!

substack.com/inbox/p...

New in Lancet Oncol: Which occupations have the highest cancer deaths?

For ♂s it was the rather broad construction/extraction category, for ♀s the transportation one.

Statisticians, chiropractors, & funeral directors had the highest breast cancer risk

http://bit.ly/4aZLP9c

Hospital medicine is evolving fast!

50 Studies Every Hospitalist Should Know (OUP), co-edited w/ Jeff & Kathy, distills landmark trials for everyday inpatient decisions -- who was studied, effect sizes, and key limits beyond guidelines.

Pre-order paperback (ebook coming): www.amazon.com/Studi...

🇬🇧🇺🇸 NICE raising cost-effectiveness threshold 25% (£20-30k→£25-35k/QALY) as part of UK-US pharma deal with 0% tariffs.​

Concern: Will funding for new drugs displace mental health services in fixed NHS budgets?
Landmark UK-US pharmaceuticals deal to safeguard medicines access and drive vital investment for UK patients and businesses
Milestone deal secures UK’s medicines access for patients, with a wide range of groundbreaking new treatments to reach NHS front-line quicker.
www.gov.uk

🔥 Don't miss THE debate of #SMDM2026 in Oslo: Tomas Philipson vs. Karl Claxton debate drug pricing & value of innovation. Two heavyweights with ?opposing views. Moderated by Mark Sculpher. Hoping for fireworks and that this is going to be epic.

smdm.org/meeting/48t...

"'It's going much too fast': the inside story of the race to create the ultimate AI

In Silicon Valley, rival companies are spending trillions of dollars to reach a goal that could change humanity – or potentially destroy it"

Guardian Interactive story:
www.theguardian.com/...

Exciting news! Prof. Sander van der Linden - leading expert on misinformation & decision-making - announced as keynote for @smdm.bsky.social 2026 in Oslo! June 28-July 1. Save the date! 🎉🇳🇴
smdm.org/meeting/page...
Keynote
smdm.org

I was an immigrant physician (even though I did not have an H-1B visa).

The threat that the H-1B visa fees pose to U.S. health care is real. It will first hit rural teaching hospitals. They will have to decrease available beds/appointment slots.

www.npr.org/sections...

Not sure what the CDC is referring to here?

www.cochranelibrary....

.. and a non-inferiority trial in shock all-comers (although this has already crept into practice -though that's dependent on center and attending- so this trial is prolly less of a surprise)

Whoa: two potentially paradigm-shifting new shock trials - one in early septic shock..
ANDROMEDA-SHOCK-2 is up next…. this looks at personalised phenotype-based, capillary refill time (CRT) targeted resuscitation in early septic shock, comparing to standard resuscitation. It is large (86 centres, 19 countries).

Read it free from JAMA: tinyurl.com/mjcj8s5w
#Lives2025 #criticalcare
Hemodynamic Resuscitation Targeting Capillary Refill Time in Early Septic Shock
This randomized clinical trial examines whether a personalized hemodynamic resuscitation protocol targeting capillary refill time was more effective than usual care in patients with early septic shock...
tinyurl.com
ANDROMEDA-SHOCK-2 is up next…. this looks at personalised phenotype-based, capillary refill time (CRT) targeted resuscitation in early septic shock, comparing to standard resuscitation. It is large (86 centres, 19 countries).

Read it free from JAMA: tinyurl.com/mjcj8s5w
#Lives2025 #criticalcare
Hemodynamic Resuscitation Targeting Capillary Refill Time in Early Septic Shock
This randomized clinical trial examines whether a personalized hemodynamic resuscitation protocol targeting capillary refill time was more effective than usual care in patients with early septic shock...
tinyurl.com
EVERDAC is hot off press @esicm.bsky.social #Lives2025 in @nejm.org open access currently:
www.nejm.org/doi/full/10....
Reports that in (non-blinded trial) patients with shock, management *without* early a-line insertion was noninferior to early catheter insertion…. #Lives2025 #criticalcare
Deferring Arterial Catheterization in Critically Ill Patients with Shock | NEJM
In patients with shock, whether noninvasive blood-pressure monitoring is an effective alternative to the recommended use of an arterial catheter is uncertain. In this multicenter, open-label, nonin...
www.nejm.org

Leo Celi's latest in Nature Medicine: "Teaching machines to doubt."

As we rush to deploy A.I. in healthcare, teaching algorithms to recognize their own limitations isn't just good science, it's a patient safety thing.
Worth your time.
Teaching machines to doubt
Nature Medicine - We must build safeguards against AI’s most dangerous feature: its ability to stop us thinking critically.
www.nature.com

Today I learned that AOL still has an email service (from the John Bolton indictment)

Professor Sue Goldie is a masterful educator and storyteller.

She decided to share her diagnosis on her own terms and brings up things here about Parkinson's (and about triathlons as a PD treatment) that no one ever does

www.nytimes.com/interactive/...
Sue Goldie Has Parkinson’s Disease
An acclaimed researcher is an expert at explaining complicated problems. Now she has to confront the most vexing question: What is happening to her?
www.nytimes.com

A stress test with a slight chance of MI or death... 😕

Do patients truly understand the risks we communicate?

European Hospital features discussion on medical risk communication across different cultures & systems.
“Very rarely a chance of heart attack or death” – Wait, what?
Hospitalists frequently discuss the risks associated with tests, treatments, and/or surgical procedures with their patients. But is everyone in the clear on what a “slight risk of complications” actually means? A session on the meaning of risk to patients and how to effectively communicate risk was discussed at SHM Converge 2025, the annual meeting of the Society of Hospitalist Medicine held in Las Vegas in April.
healthcare-in-europe.com

Pure gold from Adam Rodman on overconfidence of reasoning models. Does this mirror incompetent + overconfident residents?

x.com/AdamRodmanMD/s... (yes, X, but worth your while)