⚡ A new cardiac technology enters the real world. How do we track its safety?
Pulsed field ablation (PFA) has rapidly expanded for atrial fibrillation treatment.
This study used natural language processing (NLP) to analyze nearly 7,000 FDA post-market reports, exploring:
🔹 Adverse event patterns
🔹 Device malfunction profiles
🔹 Real-world safety surveillance
Officially the funniest copy editor note comment I’ve seen on a manuscript. Not sure how to address it when @VivekReddyMD is an expert on nearly all ablation technologies
“We noticed that a scholar, Vivek Y. Reddy, is cited 9 times, accounting for approximately 42% of the total citations. To ensure a balanced and appropriate citation distribution, we kindly ask you to review these citations and reduce the number of references from this author to no more than 5 citations, retaining only those that are most relevant and necessary for the manuscript.”
New peer-reviewed evidence for AI-ECG in aortic stenosis. Big thanks to Dr. @MattSegar and the Texas Heart Institute team for leading this work.
academic.oup.com/ehjdh/advan…
AK-AVS is investigational and not FDA-cleared.
Development and Validation of a Novel Machine Learning-Based Algorithm to Predict Incident Atrial Fibrillation – A Multicohort Analysis
heartrhythmjournal.com/artic…
POLY-HF trial: Polypill containing metoprolol, spironolactone, and empagliflozin improved EF, QoL, and adherence, and reduced HF hospitalizations at 6 months in HFrEF patients compared to usual care. #AHA25 View the slides here: clinicaltrialresults.org/wp-…
Interesting case with PMCardio - Queen of Hearts.
On call today and paged to ED for chest pain. ECG looked completely “benign”. Clinical story suggested an ACS event. We decided to take him to the Cath lab anyways, but used the app to see if I missed anything.
QoH noted a STEMI equivalent. Patient ended up having a LCx occlusion. Ended up buying a subscription to the app for that alone.
@PowerfulMedical@MartinHerman_@RobertHermanMD
I experienced my first AI scribe hallucination today.
What I said: "Heart rate 71 bpm"
What the AI documented: "Electrocardiogram: Normal sinus rhythm"
No ECG was performed. The AI scribe extrapolated from basic vital sign data and generated documentation for a procedure that never occurred.
Remember, AI scribes are powerful tools, but they're not infallible. They can (and will) "hallucinate" by creating plausible-sounding but completely fictional content.
We cannot delegate our professional responsibility to validate documentation. The technology is impressive, but physician oversight remains non-negotiable.
Has anyone else experienced this?
AccurKardia launches clinical pilot for AK+ Guard. Three arm study, 150+ patients across care settings. AI hyperkalemia risk from Lead I ECGs, works with FDA cleared wearables. Details: businesswire.com/news/home/2…
(Investigational, not FDA cleared.)