Predicting Bleeding Risk in AF Patients Using AI (Smart Computer Tools).
Predicting Bleeding Risk in AF Patients Using AI (Smart Computer Tools).
Source: American Heart Journal
Published: April 2025
1. Researchers studied 24,468 patients with atrial fibrillation (AF) who were taking DOACs.
2. They tested different smart computer systems to see how well they could predict major bleeding compared to traditional risk calculators like HAS-BLED, ORBIT, and ATRIA.
3. One of the computer systems (called random forest) predicted bleeding better, scoring 0.76, while HAS-BLED scored only 0.57.
4. The study lasted up to 5 years. Bleeding happened in 2.3% of patients after one year, 5.8% after five years.
5. The computer tools also found new risk signs like body weight, cholesterol levels, and even type of insurance.
6. These tools can help doctors better understand which patients are more likely to bleed, and make more personalized treatment plans(Alternatives like left atrial appendage closure devices lower stroke risk with fewer nonprocedural bleeds).
https://www.ajconline.org/article/S0002-9149(25)00115-8/fulltext