Physician, educator, historian, author, podcaster, researcher @BIDMC_IM @HarvardMed @HarvardDBMI, host of @BedsideRounds, AE @NEJM_AI, studies 🤖+🧠. 🖖🚲
Boston, MA
- Tweets 16,362
- Following 1,502
- Followers 18,126
- Likes 15,145
Pinned Tweet
In case you were wondering why there hasn't been a new @BedsideRounds in a while ...
magazine.hms.harvard.edu/art…
4
6
58
17,606
Everyone is the hospital is working really hard right now, but I especially wanted to thank our residents.
You all are being tested like very few generations of trainees have been historically.
I really appreciate all that you're doing for your patients.
9
195
1,683
Why do we use godawful blue-background-with-bright-yellow-text for medical school lectures?
A 🧵on magic lanterns, darkrooms, path dependence, and “things we do for no reason”
👇
46
394
1,359
Preprint out today that tests o1-preview's medical reasoning experiments against a baseline of 100s of clinicians.
In this case the title says it all:
Superhuman performance of a large language model on the reasoning tasks of a physician
Link: arxiv.org/abs/2412.10849
A 🧵⬇️
32
217
1,114
349,464
I love that in this painting of an attending physician's rounds, you can see how much has changed (and how much hasn't), including the medical student who appears to be using his phone during the encounter (sorry time travelers, it's a notebook)
14
205
921
The entire idea of a "physical exam" is a 20th century invention. Seriously, in Rene Laennec could see us auscultating patients' lungs in whom we had just obtained a chest CT, he would lose it.
52
67
668
Huge update to our preprint today on the superhuman performance of reasoning models in medical diagnosis!
TL;DR – they don't just surpass humans in meaningful benchmarks, but in actual medical care from unstructured clinical data:
A 🧵⬇️:
Preprint out today that tests o1-preview's medical reasoning experiments against a baseline of 100s of clinicians.
In this case the title says it all:
Superhuman performance of a large language model on the reasoning tasks of a physician
Link: arxiv.org/abs/2412.10849
A 🧵⬇️
17
103
552
159,133
It's time for me to channel my inner @tony_breu -- which means it's Tweetorial time!
So let's talk about azotemia (elevated blood urea nitrogen) after an upper gastrointestinal bleed!
14
219
475
So many arguments about what's wrong with medicine today are predicated on imagined (and inaccurate) histories. Let's take some examples from my colleagues who imagine a "golden" age of the exam:
14
100
444
American progress notes are > 4 times longer than our colleagues' in other countries (source: pubmed.ncbi.nlm.nih.gov/2980…)
34
86
398
They're called discharge summaries for a reason, not "discharge copy-pasted-crappy-progress-note-that-lists-every-little-thing-that-happened-during-a-hospitalization."
9
38
379
70,148
Why are medical podcasts like @thecurbsiders, @BehindTheKnife, @emcrit, and @AFPpodcast so popular for learning? And who is making them? And can they be trusted?
We listened to (and coded) the top 100 podcasts on the Apple podcasts US medicine chart to find out!
A 🧵⬇️
14
103
359
It’s time for another #histmed Tweetorial -- this time I'm going to talk about the pesky definition of a fever, and where the 98.6 F average body temp came from!
Full disclosure: will use C AND F for temp, but no K or R.
9
121
323
The quality improvement movement hasn't improved healthcare quality, especially when taking into account the massive amount of money and manpower we spent collecting and analyzing quality metrics. A new paradigm is necessary.
19
25
316
67,277
Supposition: there are so many notes in the medical record of a hospitalized patient that if a physician ACTUALLY read each one in its entirety, there'd be no time for anything else.
I think much of our note writing is performative -- not driven towards actual patient care.
29
24
293
Can I still complain about that stupid article in JAMA that chastised physicians for using the phrase "chest x-ray" instead of radiograph? (because, of course, x-rays are actually invisible waves of energy we cannot see).
22
13
288
Our first RCT on using an LLM on diagnostic reasoning is out!
And the results are 🔥🌶️... adding ChatGPT did NOT improve diagnostic accuracy or reasoning, and the AI alone outperformed ALL the humans.
What does this mean? A 🧵⬇️
A small randomized trial of generative #AI for diagnosis again (as seen in a few previous studies) shows higher performance for #AI than physicians + AI. May indicate that physicians need to be trained on how to incorporate AI.
jamanetwork.com/journals/jam…
14
86
294
195,777
The other day, I walked in to one of my severely demented patients with her hair beautifully done in a complex braid. One of our aides had washed and braided it for her. Such simple, little acts show so much humanity.
One of the greatest honors in our profession is witnessing the beautiful acts of kindness nurses do for patients every day:
- sweet conversations about children and families
- relieving shame of the body and its functions
- respect for and relief of pain
6
30
280
Almost exactly a year ago, I had a modestly controversial tweet about routine daily physical exams -- and about how we should probably spend more time actually talking to our patients daily rather than pretending to examine then.
The reason to do a physical exam (or pieces of a physical exam) is because it helps you make decisions about caring for the patient.
If you want to connect with them, maybe spend that time talking to them, instead of pretending to listen to their heart and lungs?
10
54
276
There is a lot of buzz about our new paper in Nature Medicine on the effects of LLMs (GPT-4) on physician management reasoning! I had TONS of fun working on this -- but what it MEANS requires some unpacking.
A 🧵⬇️
A randomized trial of GPT-4 vs 92 physicians with or without this #AI LLM for performance on patient care tasks.
AI improved physician performance, on par with AI alone (based on 5 clinical vignettes) nature.com/articles/s41591-0… @NatureMedicine @AdamRodmanMD @jonc101x
8
91
278
91,000
This is officially my favorite Netter plate: "hormonal pathways in the rage reaction"
Anyone have a better Netter for a Netter-off?
16
51
239
Would anyone be up for publishing a controversial piece on "physical exam kabuki"?
Maybe"Things We Do For No Reason: the daily physical exam"?
The argument, "it builds connection with our patients" is the ultimate post hoc fallacy that I would love to push back against.
46
33
248
I have just finished the first draft of (what I think is) literally the best grand rounds I have ever prepared -- an overview of the links between our understanding of clinical reasoning and artificial intelligence, stretching back over a century but contextualizing today.
14
17
253
48,699
Our new study in @JAMAInternalMed looking at the reasoning abilities of GPT-4 compared with human physicians just came out.
Big picture: AI displays (much) better reasoning than humans, makes diagnoses similarly, but hallucinates considerably more.
A 🧵to put in context ⬇️
How good is #AI at clinical reasoning? An early, simulated assessment
jamanetwork.com/journals/jam…
“An LLM was better than physicians in processing
medical data and clinical reasoning using recognizable frameworks as measured by R-IDEA”
8
74
241
113,261
For anyone bemoaning the "good old days", an experiment in which 10 healthy medical students were made to drink blood to figure out the volume necessary to produce melena. TL;DR 50-80 mL @mondhiry
(jamanetwork.com/journals/jam… -- 1939!)
15
38
225
Can GPT-4 solve really hard medical cases and come up with a good list of differential diagnoses?
@zahirkanjee @byrondcrowe and my study is out in @JAMA_current , and the short answer is, “Yes.”
But what does this all mean? 🧵⬇️
In this study, a generative artificial intelligence (AI) model provided the correct diagnosis in its differential in 64% of challenging cases and as its top diagnosis in 39%. ja.ma/3X5M6i9
17
71
224
196,964
Thirty years that changed medicine -- all the infectious disease identified from 1877 (anthrax) to 1906 (pertussis).
8
101
226
Useless medical word of the day: upper and lower extremity. Are the words arm and leg too much? Or are we just allowing that we might eventually take care of patients with tentacles?
20
28
207
For all those who wanted a recording of my grand rounds on diagnosis, here it is! player.vimeo.com/video/53181…
2
52
209
Hey everybody. I'm putting @BedsideRounds on hiatus for the time being.
As a hospitalist (an inpatient internal medicine doctor) in Massachusetts, I expect the next several weeks will be some of the busiest of my life.
Good luck to everyone. We're all in this together.
18
8
190
During the 1889 flu ("Asiatic flu"), a panicked public turned to quinine and antipyrine (an early anti-pyretic) to save them, leading to reported shortages among druggists.
The historical parallel to chloroquine and its analogues in COVID-19 is striking.
7
61
190
This is fantastic. In this study, nurses' concern about a patient has a positive LR of 18 for ICU transfer, and *40* for the highest level of concern ("worry factor" as they term it in the paper). Area under the curve 0.92!
TL;DR -- when your nursing colleagues worry, LISTEN.
5
37
181
Okay, I think it's time to be a party pooper here. It's very hard to prove a negative -- but unless someone has a primary source that proves otherwise, I'm pretty sure this is a myth (that lasix is named because it "lasts six hours") A brief 🧵
14
32
179
We write so much useless research that no one reads that we now have invented a new type of useless research that no one reads -- the "scoping review".
17
27
176
77,637
I hereby propose Rodman's Law:
If there are two medical interventions without clear evidence supporting one over the other, the default should be the one that makes the patient less miserable.
9
27
149
Incredibly honored to be recognized as @bidhospitalists teacher of the year by my amazing residents @BIDMC_IM! I'm glad you guys appreciated my pressured ranting and late night deep dive emails (instead of just thinking I'm a crazy person) 😀
23
4
157
Honest question -- we call a troponin a "cardiac enzyme" because creatine kinase is an enzyme, and the troponin has replaced that test, right? And we just kept calling them enzymes despite the fact that a troponin is not an enzyme?
14
13
159
The reason to do a physical exam (or pieces of a physical exam) is because it helps you make decisions about caring for the patient.
If you want to connect with them, maybe spend that time talking to them, instead of pretending to listen to their heart and lungs?
13
12
145
The fingernails are the windows to the soul
(Actually, I suspect the clinical utility of most nail findings is nil, but most of my House MD exam moments have been through the nails)
3
32
152
Why you should always Google yourself: because otherwise, I wouldn't have found this gem of a review (in a list of the author's favorite podcasts):
"Bedside Rounds
A doctor with an annoying voice talks about the history of medicine. Not much more to it than that."
Um, thanks?
25
133
19th century physicians often had personal mottos.
Laennec's was "the whole medical art lies in observation"
Osler's was "equanimity"
Kind of sad that we've dropped this tradition. What would yours be?
86
33
135
Adam's daily 🔥take:
A SOAP note does not have discrete sections of subjectivity and objectivity but rather a continuous spectrum throughout.
5
6
117
My (preaching to the choir) 🔥 take: digital educational skills -- whether teaching on #MedTwitter, podcasting, or making videos -- are essential #meded skills for the 21st century. And we can teach these to future educators.
A Tweetorial🧵:
6
35
119
When I was a medical student at Tulane, I took an advanced internal medicine course taught by Chad Miller (and Ben Rothwell).
We read Rouche and Holmes. We went to the art museum to look at paintings and sat in the French quarter and made careful observations of people passing.
5
7
121
39,854
#spoileralert: a lot of things. And if you’re interested in that, I produce an entire podcast of medical history and how modern medicine came to be called @BedsideRounds. You can even get CME credit if you're a member of @ACPinternists!
5
12
123
Well, this was not expected! @BedsideRounds is now written up in the @bmj_latest as one of "five podcasts every medical student should listen to!" Thanks to Anna-Lucia Koerling (the author, who I'm not sure is on Twitter) -- I'm glad you enjoy, and incredibly honored!
7
6
116
"When there is no reason to believe there is any difference in outcomes between two interventions, choose the one that sucks less for the patient" - Rodman's Law
4
13
111
If the justification for a medical practice is, "that's what we've always done," it's probably a good idea to take a long, skeptical look.
3
33
118
TL;DR - 35mm 🎞️used🔵for visibility; early 🖥️used 🔵/🟡 (among others) for visibility, and PPT and other slide presenters adopted these standards by default. Great example of path dependence!
I just saved you 20+ Tweets 😄
4
6
118
Since my thread on the historicity of the exam can gained some traction, here's a reading list if you're interested in gaining perspective on the nature of clinical reasoning -- rather than "just so" stories about imagined halcyon pasts (the era of "the Giants")
3
20
118
This past Tuesday, I was honored to help plan and participate in what I think is the first ever clinicopathologic conference (CPC) with an artificial intelligence (Dr. GPT-4) discussant @BIDMC_IM.
It was… eerie... to say the least.
These were my experiences: 🧵⬇️
📢Who do you think reigns supreme for today's Clinicopathology Conference (#CPC) ?
Dr. @stephcabral_ vs Dr. Kevin He vs Dr. @AdamRodmanMD vs ChatGPT
8
38
114
123,423
This is that happens when you ask me to lecture about anything I want 😂
Coming Tuesday @BIDMC_IM! Putting all the reading I've done over the past two years to explore the McKeown thesis into some clinical context. Hope to see all my residents there!
7
6
112
What other things do we commonly do to patients today that have a similar evidence basis to “blue powerpoints with yellow text?
Why do we have a medical culture that largely accepts tradition?
What else do we only do because of path dependence?
3
12
109
Another wonderful piece by @VPrasadMDMPH. For those applicants who had high school or college jobs bagging groceries, working as a waiter, &c, PUT IT ON ERAS!
I know you don't believe me now, but food service has far more relevance to practicing medicine than pipetting in a lab
7
14
105
One of my favorite books on the history of diagnostic reasoning (fortunately it's still in print!) is Medical Thinking by Leslie King Snow. I haven't read it in its entirety in years, but the opening is a vivid reminder of medicine just a few generations ago.
7
14
109
Have you ever wondered where SOAP notes come from?
Do you wonder why we spend SO MUCH time at the computer, instead of with our patients?
Do you worry that we way we document warps our diagnostic thinking?
Then see you at BIDMC Gen Med Grand Rounds tomorrow!
12
4
106
How should we feel about the incoming deluge of AI-generated text in our electronic health records?
(if you don’t want to read, the answer is, “not good”).
🧵 from my piece with @arjunmanrai and @LiamGMcCoy ⬇️
Concerns about the potential adverse impact of LLMs on electronic health records
@NEJM today, by @LiamGMcCoy @arjunmanrai
@AdamRodmanMD
nejm.org/doi/full/10.1056/NE…
10
35
101
20,948
If you couldn't make it. I'll be live tweeting the @BIDMC_Education conference on AI, clinical reasoning, and the future of medical education! @iMedEducation
1
16
99
38,605
"Hi, my name is Adam Rodman and I'll be your doctor. You can call me Adam if you'd like. How would you like me to call you?" - every patient I meet for the first time
Asking someone how they'd like to be referred to is a mark of respect, not the other way around.
6
8
94
About a month ago, I commissioned @sukritibanthiya to draw this AMAZING image to celebrate six years of @BedsideRounds (and to finally provide some swag) -- and it turned out better than I could have ever imagined!
And now I've set up a @TeePublic store: tee.pub/lic/AU3rz_lX3yY
13
16
99
I am planning on creating a "mini-curriculum" (probably five talks) on medical history and epistemology targeted towards medical students.
Any ideas for topics? I already have:
➡️Hx of nosology (what is a disease?
➡️Hx of the physical exam
35
5
103
The reason that we do ANYTHING in medicine needs to be based on reasoning based on the medical conditions and realities of today -- and not imagined histories of the past.
6
12
99
This rotation, for a day I made my entire team drink only thickened liquids for a day (myself included). I ended up using six packets of Thickit, including for the Coke I drank in the AM (wasn't going to thicken my coffee ...)
10
19
97
Early chest x-rays were very poor at showing the lung parenchyma. These images are from Francis Williams' radiography department at Boston City Hospital. The patient would have to lie in this position for almost 45 minutes, producing the image seen below.
3
22
95
So excited to speak on what I think is one of the most exciting innovations in medical education in the last decade @BIDMC_Academy with @tony_breu and @ShreyaTrivediMD!
One of the things I find so wonderful is how much the format has evolved over the past year.
A brief 🧵👇
4
19
96
Another editorial cartoon from the Boston Daily Globe in 1889 showing quinine defeating "La Grippe" (a common name for influenza) while the standard of care (uh, whiskey) looks sullenly on.
3
23
91
Many of us already listen to amazing medical podcasts like @TheCurbsiders, @CPSolvers, @NephJC_Podcast, @medrants #AnnalsonCall in our own time
-- but will medical podcasting be integral to medical education?
9
26
84
Are there better colors to use in a modern powerpoint? @JenniferSpicer4 pointed me to this modern research on PPT presentations -- it's actually remarkably similar to research in the 1970s showing black on white is the most readable! (and blue among the least)
2
14
93
Do large language models have a probabilistic understanding of disease states? And what does this mean for the future of diagnosis and clinical reasoning?
I explore this with Thomas Buckley, @arjunmanrai, and @dr_dmorgan in our new paper in @JAMANetworkOpen.
A brief 🧵⬇️
8
27
91
40,338
If you couldn't attend my grand rounds at @WUDeptMedicine, it has been uploaded to YouTube! (piped.video/watch?v=z4aZVQhY…)
Again -- thank you so much for having me (and to @MohitHarshMD for a way-too-gracious introduction).
7
15
88
I am giving my most expansive grand rounds for the first time ever tomorrow at @BIDMC_Medicine -- Medicine Without Disease: AI, Diagnosis, and the Future of Medicine.
I haven't given a historical talk in almost a year...
5
11
93
9,462
Putting the finishing touches on a new grand rounds! (I'm trying to put together a new one each year which I refine throughout the year). A big picture explanation on the cognitive nature of diagnosis that I've been working on for the past two years.
4
6
88
Very exciting today -- for the first time, Bedside Rounds took the #2 spot in the Apple Podcasts medicine chart for the US! (and #22 in the science category, which I actually think is more exciting since it contains my idols @Radiolab). Thanks to all my listeners!!!!
8
5
88
I might be driving myself down a crazy rabbit hole. I keep reading descriptions of the Russian Flu of 1889 -- the description is SO similar to COVID-19.
We've previously assumed that the flu epidemics prior to the 20th century have been influenza -- but how can we be sure?
4
15
84
The "ancient" and "sacred" exam that we imagine Hippocrates, Avicenna, Vesalius, Laennec, and Osler all performed lovingly on their patients before us moderns got distracted by alienating technology NEVER EXISTED. It's a 20th century nostalgia, not based on historiography
1
11
84
So I don't know the actual answer (the idea of "panels" that are reimbursed at certain rates seems to be a post-WW2 phenomenon), but the tests (or their precursors) on the Chem7/BMP/&c were developed by Otto Folin c1909; phos and mag levels were developed until the 1920s
3
80
Do you make Tweetorials, infographics, medical podcasts, or streaming videos for medical education?
Or have you ever wanted to try?
You're a digital educator -- and we have exciting news! This year's @iMedEducation #DigitalEducation2022 conference has a teaching competition!
3
23
83
If you ever doubt how much path-dependence there is in medicine, the appropriate doses for quinine were based on cinchona decoctions in the Schedula Romana published in *1651*, and weren't changed until the 1970s.
3
12
71
All of our bacterial identification in 2019 is based on the application of textile dyes from the 1860s, in a method developed in the 1880s. That's kinda crazy, right?
This is wool dyed by crystal violet, the dye used in the gram stain.
2
14
74
We don't let long haul truckers do it.
We don't let pilots do it.
Remind me how we work for an "evidence-based" field again?
I worked an unexpected overnight last night
This 🧵 is an exhausted, muddy-headed takedown of the historic but continued saddling of medical trainees with ridiculous work hours
There will be no data and I will appeal to emotion a lot
Here we go!
#MedTwitter
1/
2
12
79
I'm giving grand rounds again! This time I'm going to give a talk that's a long-time coming (almost four years in the making!) -- a historical framework for how we know things it medicine! It'll be interesting, I promise.
Please come, and please come say hi!
6
7
80
Oh man -- I'm probably going to make a lot of people mad :)
But I think the folk belief exists because there's some truth to it (see, eg, this fascinating RCT: pubmed.ncbi.nlm.nih.gov/1628…)
10
7
79
Where did the godawful yellow-text-on-blue-background default powerpoint template that was all the rage in Med Ed for the past decade come from? I remember as a med student at Tulane being specifically told to do my powerpoints like that...
16
3
73
I am SO excited to give @BIDMC_IM Gen Med grand rounds tomorrow!
➡️What does it mean to make a diagnosis?
➡️Why do collect data from our patients the way that we do?
➡️Why do we call it "data"?
➡️Why are we so miserable with the EMR?
➡️Will computers replace us?
10
1
79
17,734
Very exciting news! @ShreyaTrivediMD and I co-founded @iMedEducation @BIDMC_Medicine earlier this year, dedicated to the advocacy and study of all types of digital education, and we are hosting a (virtual) national conference on January 22nd.
We want you to come!
A brief 🧵⬇️
5
16
71
The idea that APSO notes -- that is, SOAP notes with the assessment and plan at the top -- are unironically suggested as a way to "fix" medical documentation shows how devoid of real ideas we really are.
13
4
74
27,543
And if you're "a certain age" (as I am) you'll remember this was everywhere! My first programming language was QBASIC when I was 7 years old -- with a blue background and yellow/white/grey text!
1
4
78
When I'm working on a solo service, I tell my discharging patients that they have to ask me three questions before I'll leave their room.
Always get some interesting questions 😀
4
3
75
Hey guys! @tony_breu and I are looking professorly for our live podcast on the history of syphilis! Room 114 -- come and join us! #IM2019 #Curb19
4
2
77
The "ee" to "ah" change of egophony in lung consolidation was described by Shibley in his Chinese patients. He would have then count 1-2-3 (yi, er, san) to test for pectoriliquy, and noted that yi (ee) changed to "ah" over the pneumonia.
3
25
76
I am going to force my poor residents to play a Dungeons and Dragons roleplaying session to learn about myths in hospital medicine (maybe I can call it Wards and Wizards??)
Any ideas for classes for the game for their characters?
@cjchiu @BrighamSK
15
4
75
Well, now that angry tweet is a point-counterpoint-rebuttal series in @JHospMedicine!
The first piece is by me and @ShaneWarnockMD, and I cut right to the point: Routine daily physical exams in hospitalized patients are a waste of time.
🧵⬇️
or
journalofhospitalmedicine.co…
8
18
79
Should we mourn CMS ditching the review of systems documentation requirements? Or celebrate? Or something else?
@Gurpreet2015 and I I think is a HUGE opportunity for medical educators to better teach history taking -- and clinical reasoning. A 🧵⬇️
4
20
74
60,583
And that is how you end up with this guide to a national scientific conference in 2006 and the advice that @jbcarmody references to only use this at national meetings
4
4
75
Money quote from the Oath of Maimonides:
"May I never see in the patient anything but a fellow creature in pain"
1
21
77
So thankful to @BIDMC_Medicine for inviting me for super esoteric topics for gen med grand rounds! (and to @_JosephLi for an over-gracious introduction!)
5
4
72
This chart is the one that really freaks me out.
These are "no right answer" management cases designed by 25 different specialists to be as difficult as possible.
Humans appropriately struggled. But o1 – you don't need statistics to see how well it performed.
6
12
79
13,854
Today @AvrahamCooperMD and my essay on AI/LLMs and medical education comes out in @NEJM , about the challenges our field faces and what we should do about it.
I wanted to give some context for this essay, and talk about some of the things we can be doing NOW.
A brief 🧵⬇️
Medical educators have a crucial role in shaping the inevitable integration of AI into the healthcare.
@AdamRodmanMD and I wrote about it in NEJM ⬇️
nejm.org/doi/full/10.1056/NE…
1
25
74
57,331
