What I believe in Medicine
Dr. Bryan Vartabedian has a great post on 75 things he believes about medicine. Do read it. There are many gems. Bryan and I started our careers at about the same time.
Here are 45 things I believe:
I tell patients that I will help them. Even when you can’t cure, you can care. And that helps. (I learned this from Jerry Lacy, one of the best doctors I ever knew.)
When in doubt, take a good history, preferably sitting down with family in the room. Always sit down.
Never be overconfident. Go slow; be present. Mistakes happen most when you are too confident.
Re mistakes: they will happen, and you will regret them most when the procedure could have been avoided with a conservative approach. Think about this regret when recommending procedures.
Re, Re, mistakes: discuss them with colleagues. I like to tell stories of the dumbest things I ever did in medicine. There are many.
Resist the urge to do more when less is enough.
Read the seminal trials. Read current trials. Don’t rely on guideline summaries. If you read the post-MI beta-blocker seminal trials, you’d have zero surprise about their lack of effect in current trials.
Trials require expiration dates. See also the ICD in patients with non-ischemic cardiomyopathy.
I will bet you two espressos that quality improvement initiatives reduce quality.
Cite trials in the medical record. For instance, “we are stopping aspirin because the patient is on oral anticoagulation… AQUATIC trial.”
Your note does not have to be long, but it should create a memorable picture of the patient.
Notes are a great place to practice writing. Use short sentences. Avoid throw away phrases, such as, “it was pleasure seeing this wonderful patient.” Say something interesting, please. AI does not help me. I write my own notes.
Use ultrasound when puncturing major blood vessels. POCUS may be one of the most elegant innovations in medicine.
Repeat #1: say I will help you to patients.
Be truthful with patients, but be optimistic. Use the placebo effect. Avoid the nocebo effect. Words can harm or heaI. I often warn families after I implant a pacemaker that the patient will be hard to contain. Say this in front of the patient.
Never order a test just to know something. Tests are scary and you should always have a prior; the test modifies that prior.
Learn the Gerd Gigerenzer breast cancer problem that most doctors get wrong.
“The heart will not stop.” Dr Bill Dillon when called about asymptomatic bradycardia. Caveat that we have added: “The heart will not stop, unless a heart doctor gets overzealous.”
Learn the Miguel Indurain heart rate story. The 5-time winner of the Tour famously had a resting heart rate in the 20s.
Polymorphic VT can have two origins—with totally different treatments. A) is pause-dependent with QT prolongation. This is torsades de pointes, and it’s treatment is to increase HR and remove the QT prolonging agent. B) is malignant VT usually due to ischemia. Treatment is antiarrhythmic drugs and investigating the cause. Knowing the difference allows you to save lives.
Speaking of drugs, pharmacology is underrated. This should be taught later in medical training, so that it sticks better.
Don’t use enoxaparin or heparin acutely for AF. There are no trials. Oral anticoagulation provides its benefit over years not hours. Just give an oral dose of anticoagulation if you feel compelled.
For AF and rapid rates, don’t use IV diltiazem. Use IV beta-blockers, digoxin and oral meds. Also put people in a quite reassuring environment. Adrenaline drives tachycardia.
Don’t rely on the echocardiogram ejection fraction when the patient is in new AF with a heart rate of 130 bpm.
Never ever shock a person with less than maximum energy. 1J hurts as much as 200J. You look really dumb if you don’t use maximal energy.
Phone-a-friend is your most important tool in medicine. 2x better than any LLM.
Think about palliative care before patients have hours to live. Staci says palliative care people are pals for people with severe disease. Know that palliative care is not the same as hospice.
Not all cardiac devices need to be replaced. Tell pts this before the device reaches battery replacement. Generator change is the classic low probability/high consequence problem. E.g. device infection (low Pb) and subsequent need to extract 20 year olds leads (high consequence).
I hate adenosine for SVT. The “heart stopper” works, but it’s harsh, ugly and unnecessary. Calm the patient down, naturally lower adrenaline levels, and then vagal maneuvers or IV beta-blockers are more likely to work in a calm patient.
When in doubt, go see the patient.
Moving radiologists out of the hospital has been an awful development. I learned so much from them—even the one in our hospital who had no chairs in her office (so that people would not overstay their welcome.)
I hate telemedicine. I am allergic to Zoom.
Doctors who take night call are underpaid, especially those who go in to fix heart attacks. They have my respect.
Your most important day in Medicine is career day. Choose wisely. I am so glad I picked EP. Interventional fields seem sexy when you are young, but fixing MIs at 0400 is not so easy in middle age.
Speaking of EP: three of the most common procedures I do in 2025 (AF ablation, CRT, and conduction system pacing) did not exist when I trained. Not only did they not exist, they were not even thought of. Be prepared to learn lots.
Spironolactone is one my secret weapons for patients with any kind of heart failure.
Doctors who write and think about policy should travel. I cannot believe how little I knew about how healthcare could be delivered until I went to other countries. My favorite system is the Danish system.
Two of the most important questions to ask about a trial result: was the effect size clinically meaningful and was the effect statistically robust. For the former, look at absolute risk reduction; for the latter, look at the 95% confidence intervals.
Another overlooked evidence pearl: how many people were screened to get in a trial? Famously, to get into PARADIGM HF (sacubitril/valsartan) you had to tolerate both enalapril and sacubitril/valsartan for 2 and 4 weeks.
Nearly every treatment recommendation in modern medicine is preference sensitive. Never say someone “needs” anything. People need food and water. Almost everything else is sensitive to patient preferences.
Shared decision making is way overrated. Too often, it is used as a menu. Never present treatment options as a menu. Other times, SDM is used to guide patients to lower value or less beneficial procedures.
Stories are what make the hospital such a great place to work. Nearly a day goes by that I don’t hear a great story.
AI is an amazing tool that will only get better. But it won’t replace the need for humanity in medicine. That said, you’d better figure out a way to create value over and above a LLM.
Doctors overestimate how much control they have over outcomes. Dr John Lloyd when he read my essay on Medical Reversals.
No matter how bad the administrative overreach gets, Medicine remains an excellent career. Because we help people. The job has meaning. I love it.
Feel free to add some of yours. Happy Thanksgiving
JMM
It would be nice to have a similar list, but from the perspective of the patient.
Humility is essential