Benjamin Schnapp, MD MEd (@schnappadap) header
Benjamin Schnapp, MD MEd (@schnappadap) profile picture

Associate Professor at University of Wisconsin Dept. of Emergency Medicine. Education research, Flight Physician. Tweets are my own opinions, not my employer's.

Madison, Wisconsin
Joined January 2009
Photo by Benjamin Schnapp, MD MEd (@schnappadap)Photo by Benjamin Schnapp, MD MEd (@schnappadap)Photo by Benjamin Schnapp, MD MEd (@schnappadap)Photo by Benjamin Schnapp, MD MEd (@schnappadap)Photo by Benjamin Schnapp, MD MEd (@schnappadap)Photo by Benjamin Schnapp, MD MEd (@schnappadap)Photo by Benjamin Schnapp, MD MEd (@schnappadap)Photo by Benjamin Schnapp, MD MEd (@schnappadap)Photo by Benjamin Schnapp, MD MEd (@schnappadap)Photo by Benjamin Schnapp, MD MEd (@schnappadap)
Benjamin Schnapp, MD MEd retweeted
‼️ Stop Hitting Kids in the Head ‼️ And no, I don’t mean stop kids from playing sports. I mean stop putting young athletes in situations where repeated hits to the head are treated as part of the game - and therefore are at risk of getting CTE. Sports give kids so much: teammates, confidence, discipline, fitness, competition, and joy. I want kids to have all of that. But we can give them those benefits without making head impacts part of the experience. Flag football instead of tackle football. Soccer without heading. Hockey without checking. These aren’t anti-sports choices. They’re ways to reduce exposure to repetitive head impacts while keeping kids in the game. Research continues to show that long-term exposure to repetitive head impacts cause CTE, reducing that exposure is a vital step in CTE prevention. So when I say Stop Hitting Kids in the Head, I’m not saying take sports away from kids. I’m saying change the rules so they can play without getting CTE.
8
40
7
148
26,111
Benjamin Schnapp, MD MEd retweeted
This week’s featured @JournalofGME articles are, "Engaging Choice Architecture: Interventions in the Emergency Medicine Standardized Letter (eSLOE 2.0) to Adjust Rater Behaviors" and "Visualizing Qualitative Data: Challenges, Strategies, and Opportunities." Read all the open access articles in the latest issue of #JGME: ow.ly/OL1b50ZBExm #MedEd #MedX #Research
2
2
624
Benjamin Schnapp, MD MEd retweeted
Replying to @GomezDLeonMD
yes everyone should be asking this question the fact that people fixate in WBC and not neutrophil to lymphocyte ratio indicates a misunderstanding of evidence based medicine if you want to demean the NLR that's fine, but the only logically coherent position in that case is to pay no attention to WBC whatsoever emcrit.org/pulmcrit/nlr/
6
15
3
119
26,759
Benjamin Schnapp, MD MEd retweeted
🚨 BREAKING: @acgme just proposed the biggest resident work-hour overhaul since 2017 And if you're a resident or fellow, you need to read the fine print 👇
19
19
5
128
114,911
Benjamin Schnapp, MD MEd retweeted
Nueva revisión de hiperkalemia aguda (BMJ 2026). Lo que cambió, y que probablemente sigas haciendo: 🔴 Dextrosa al 10%, no al 50%. La glucosada hipertónica puede EMPEORAR la hiperkalemia. 🔴 Insulina 5 U (o 0.1 U/kg, máximo 10 U). Misma caída de potasio que con 10 U, con casi la mitad de hipoglucemias (OR 0.55). 🔴 1 g de gluconato de calcio no sirve. 97% necesitó tres dosis de 1 g y una sola dosis no funcionó en NINGUNO. Dosis: 1-3 g. 🔴 El calcio no "estabiliza la membrana". Restaura la conducción por canales de calcio. Traducción práctica: sirve cuando hay QRS ancho, no en la onda T picuda aislada. 🔴 Bicarbonato: no se recomienda de rutina. 🔴 La onda T picuda aparece en menos del 25% de los pacientes con K de 5.5-7.0. Un ECG normal NO descarta hiperkalemia. 🔴 Reevalúa el potasio a los 60 min y otra vez a las 2-4 horas. El rebote es frecuente El tema ya está re-escrito y actualizado en Memodi y Accio. Completo en el canal (t.me/memodiapparticulos).
12
798
6
2,523
116,769
Benjamin Schnapp, MD MEd retweeted
Big Signaling is the cure for Application Fever.
7
6
146
19,143
Benjamin Schnapp, MD MEd retweeted
Valproic acid is great for refractory agitated delirium. 🏆 Especially useful in patients with explosive/aggressive behavior in the context of a personality disorder (often exacerbated by delirium). Candidates for VPA: ✅ Not pregnant ✅ No severe liver dz or history of hyperammonemia ✅ No major drug-drug interactions with VPA ✅ Didn't respond to antipsychotics +/- alpha-2 agonists (dexmedetomidine). This isn't a hard requirement, but generally VPA isn't usually a front-line agent. Advantages of VPA: 👍 Cardiovascular stability 👍 Low risk of airway compromise (only mildly sedating) 👍 Antiseizure activity (e.g., can be used for sz prophylaxis in an agitated patient) 👍 Can be given IV or PO (1:1 conversion with immediate-release formulations) 👍 EM/ICU folks should be knowledgable and comfortable within it already Disadvantages of VPA: 👎 Dosing requires a little more thought than most agents. For patients on this for more than a few days you want to check a VPA trough and adjust it using the Fraser equation (to account for albumin). 👎 Can cause a lot of side-effects (but most of these are due to *chronic* VPA, rather than just a few days of therapy) Dosing 🎯 My preference is to use the same dosing regimen as for status epilepticus (40 mg/kg load max 3 grams, then 15 mg/kg/day in divided doses). You can up-titrate as needed while following VPA troughs if ineffective (max dose of 45 mg/kg/day). 🎯 Many studies have reported starting lower and up-titrating, but this delays its efficacy for 2-3 days. 🎯 Using more frequent doses (e.g., q6hr rather than q8hr) may avoid toxic peak levels while maintaining adequate trough levels. This isn't an option I use a ton, but it's a terrific tool to have in your toolbox for agitated delirium that isn't responding to usual front-line treatments. ⚠️ VPA should be weaned off prior to hospital discharge (unless the patient is seen by psychiatry and they are intentionally recommending VPA as chronic outpatient therapy for bipolar disorder etc - which would be uncommon). (more discussion on VPA pharmacology in the IBCC chapter on status epilepticus) Recent article on this in NeuroCritical care here:
6
42
168
19,644
Benjamin Schnapp, MD MEd retweeted
i have to do this course to keep my job as a resident physician
6
2
108
15,277
Benjamin Schnapp, MD MEd retweeted
The obsession with normalizing lactate might be the single worse "new" practice in the hospital Leads to a lot of silliness Are there other contenders?
20
6
64
19,605
Benjamin Schnapp, MD MEd retweeted
in ICU, the white coat is just a big walking fomite ideally, every piece of external clothing should be industrially laundered daily by the hospital (e.g. scrubs, green hospital-issued warm-up jackets) part of primum non nocere is not walking around covered in Cdiff spores
While the white coat remains a powerful symbol in medical education, its daily utility is shifting as clinicians prioritize comfort and approachability. Recent studies suggest that patient trust is built through demeanor and communication rather than attire, leading many physicians to favor fleece jackets or scrubs. Do you still view the white coat as an essential component of professional identity? mdsc.pe/3QARQBb
9
21
1
114
44,726
Benjamin Schnapp, MD MEd retweeted
You open ChatGPT. You type the question. A clean, structured answer comes back in three seconds. You read it, it makes sense, you move on. You feel like you learned something. Forty-five days later, a professor walks in and hands you a test you weren't expecting. You don't remember most of it. André Barcaui at the Federal University of Rio de Janeiro ran the experiment to find out if the feeling was accurate. 120 undergraduate business students, ages 18 to 24. All told to spend two weeks researching AI concepts, ethics, societal impacts, technical foundations, and prepare a 10-minute presentation. Sixty used ChatGPT freely. Sixty used textbooks, library databases, articles, and standard web search. Then, 45 days later, with no warning, a retention test. The ChatGPT group scored 57.5%. The traditional group scored 68.5%. Cohen's d was 0.68, a medium-to-large effect. In most grading systems, that's the difference between passing and failing. This is called cognitive offloading. When your brain delegates thinking to an external tool, it reduces the mental effort required during encoding. Effort is what makes memories durable. Struggling to find, synthesize, and connect information is not an inefficiency in the learning process. It is the learning process. ChatGPT removes the struggle and takes the encoding with it. Barcaui calls what the AI group experienced "borrowed competence." The answer was structured, the vocabulary was right, the reasoning felt sound. It just wasn't theirs. And 45 days later, it was gone. The AI group's forgetting curve was steeper and didn't stabilize the way the traditional group's did. The memories weren't just smaller. They were more fragile from the start. You didn't learn it. You borrowed it.
Students who used AI to study remembered less than those who did not.
113
1,156
144
4,709
783,667
Benjamin Schnapp, MD MEd retweeted
It’s hard to watch a young patient with severe addiction and mental illness, deemed to have capacity, choose to leave against medical advice. Often they are back within hours, sicker. Sometimes they do not come back at all, and we often find out why. Autonomy matters. But when illness is shaping the decision, it raises an uncomfortable question about where we draw the line.
1
4
1
19
6,086
Benjamin Schnapp, MD MEd retweeted
This recent article explores engagement and perceptions of feasibility and utility of EPA assessment implementation for participants in a pilot for #vascularsurgery #EPAs. Read more: ow.ly/LFkH50YPlNN
2
6
389
Benjamin Schnapp, MD MEd retweeted
It's always bothered me that medical research papers are static. A PDF, showing you what the authors thought were the important findings. I think we're ready for a paradigm shift and I did a little proof of concept...
10
12
84
25,217
Benjamin Schnapp, MD MEd retweeted
Four million physician hours are spent on mandatory modules each year. They represent not just clinical time lost but “evenings reclaimed and cognitive load reduced.” Redundant training and low-value modules should be a thing of the past. jamanetwork.com/journals/jam…
17
100
13
413
52,747
Benjamin Schnapp, MD MEd retweeted
Hey look! @nejm moved my writing to a new spot. Here's the first post, a rant about a particularly annoying requirement for those of us in work in hospitals (link below) 1/3
43
99
27
779
116,384
Benjamin Schnapp, MD MEd retweeted
Ninety nine times out of 100 you can replace the phrase “let’s try push-dose epi” with “we should’ve ordered levo ten minutes ago.”
4
8
61
8,662
Benjamin Schnapp, MD MEd retweeted
💊 Not a very good news for Medical LLMs. A new Mass General Brigham study shows leading LLMs often try to please the user in medical chats, and to do that, can output wrong advice. Paper shows that default models will confidently echo bad medical assumptions, and that a small behavior nudge plus a 300-example fine-tune can push rejection of harmful requests to 99% to 100% without hurting core knowledge. They built 50 trick questions that treat a brand drug and its generic as different, then asked 5 models to answer. GPT-4, GPT-4o, and GPT-4o-mini agreed with the wrong premise 100% of the time, Llama3-8B agreed 94%, and Llama3-70B rejected fewer than 50%. This behavior is sycophancy, the model goes along with a bad assumption even when it knows the 2 names are the same drug. Adding a refusal cue and asking the model to recall the brand to generic link first raised rejections to 94% for GPT-4 and GPT-4o, 92% for Llama3-70B, and 62% for GPT-4o-mini. Small supervised fine-tuning on 300 examples then generalized the skill, giving GPT-4o-mini 100% and Llama3-8B 99% rejection on new cancer drug tests. The models also explained their rejections correctly in 79% and 70% of those cases, and scores on 10 standard medical benchmarks stayed about the same. The recipe is simple, allow refusal, cue factual recall before answering, and fine-tune on illogical request pairs so the model spots and blocks false premises. This work isolates a real failure mode and shows a low-cost way to harden medical chat systems fast. Health systems should adopt the rejection hint + factual recall + small fine-tune pattern and monitor for regressions as base models change. --- nature. com/articles/s41746-025-02008-z
18
83
18
330
63,645
Benjamin Schnapp, MD MEd retweeted
My teaching was feeling a bit stale, so I picked up this book - quick read, would recommend. 10 core points summarized, with examples of where we fall for the illusions in MedEd
2
27
5
91
21,065
Benjamin Schnapp, MD MEd retweeted
The Dream Team 💛 “I chose UW–Madison because the collaborative, family vibes were obvious from the outset,” says a current resident. “I’m a big fan of a friendly work environment and was drawn to UW because everyone is so happy and welcoming!” #EMBound #WhyUWEM
1
2
2
344