zlacker

[parent] [thread] 3 comments
1. OJFord+(OP)[view] [source] 2023-06-06 16:52:00
The basics definitely taught & required of all junior docs (UK) - but it's nuanced, not just a binary 'read an ECG correct or incorrect' - a specialist might spot something someone else doesn't; two specialists might disagree on whether it shows something or not. (And an interested (inherently non-specialist) junior might notice something a bored & rushed specialist doesn't.)
replies(2): >>antonj+D5 >>halduj+07
2. antonj+D5[view] [source] 2023-06-06 17:14:57
>>OJFord+(OP)
Seems like the perfect application for some kind of first pass (in the ambulance,even) automated ML diagnostic, given the data is 2D, well characterized, and mostly repeating.

Edit: aha--https://www.nature.com/articles/s41467-020-15432-4

replies(1): >>halduj+A7
3. halduj+07[view] [source] 2023-06-06 17:22:00
>>OJFord+(OP)
Agree in general, even if one knew this at one point or know the textbook appearances patients don't read textbooks and there's nuance to everything. Unless you're actively exposed to/interpreting ECGs in clinical practice you lose the skill.

When I was an intern I was 100x better than I am at this point in my career.

The basic atrial fibrillation and STEMI is something any doctor can interpret with confidence (I used myself as an example as I'm probably the least competent because I haven't looked at an ECG in 10 years). I doubt non-cardiac surgeons are much better on average as they don't really look at these themselves that often other than for basic things.

An ER or general internal medicine physician is expected to be competent in more advanced but common stuff like bundle branch blocks, left ventricular hypertrophy, non ST elevated MI.

Weird arrhythmias or conduction abnormalities is really only for cardiologists, and even then typically a subspecialist electrophysiologist.

◧◩
4. halduj+A7[view] [source] [discussion] 2023-06-06 17:24:09
>>antonj+D5
We've had computer-reads/AI on ECGs since I was a medical student. Really good at detecting normal, bad at everything else.

US/Can healthcare systems still pay a cardiologist a couple of bucks to "finalize" the interpretation whenever they get around to it. It's a bit ironic, someone could have an MI on Friday, get treated and discharged and we're still paying someone on the Monday to read 40 ECGs (note these would have been acutely interpreted by the cardiologist treating the patient, most places have rules against self-referrals so you can't formally interpret anything you order yourself).

[go to top]