Arrhythmias & electrophysiology
Using Atrial Fibrillation Burden Trends and Machine Learning to Predict Near-Term Risk of Cardiovascular Hospitalization
Authors' conclusion
Trends in atrial fibrillation were strongly associated with near-term CVH, especially above-average burden coupled with low patient activity.
Study context
- Population
- Patients with >1 days of ICM-detected AF, Optum Clinformatics database linked with Medtronic CareLink ICM database (2007–2019)1,998 (76%) had cardiovascular hospitalization (CVH)
- Participants
- 2,616 patients
- Exposure
- AF burden trend groups vs sinus rhythm
- Outcome
- Cardiovascular hospitalization (any hospital, ED or ambulatory surgical center encounter with a CV code), predicted 5 days before
Methodological detailsmachine learning, linked claims and ICM database
- Machine learning
- Linked claims and ICM database
Main finding
As reported in the abstract
Relative risk of CVH
Each group vs ARelative risk (95% CI)
B
4.49(3.74–5.40)
C
8.41(7.00–10.11)
D
11.15(9.10–13.65)
- Increase in CVH detection over prespecified burden thresholds (≥1 hour, ≥5%)
- 21%
Observational data: associations, not effects.
Further findings
Prediction of CVH (AUC)
- AUC using hourly burden amounts
- 0.55
- AUC using burden trends and decreasing patient activity
- 0.66
- Difference in AUC
- P < 0.001
- Increase in predictive power
- 20%
Original publication
Using Atrial Fibrillation Burden Trends and Machine Learning to Predict Near-Term Risk of Cardiovascular Hospitalization. Circ Arrhythm Electrophysiol. Published 24 October 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →