Arrhythmias & electrophysiology
Identification of recurrent atrial fibrillation using natural language processing applied to electronic health records
Authors' conclusion
When compared with a code-based approach alone, this study's high-performing automated NLP method identified significantly more patients with recurrent AF.
Study context
- Population
- Adults with new-onset atrial fibrillation (AF) initiating rhythm control therapy (ablation, cardioversion or antiarrhythmic medication)
- Participants
- 22,970 adults
- Comparison
- Natural language processing (NLP) algorithm vs code-based algorithm for identifying AF recurrence
- Outcome
- Identification of recurrent AF episodes
- Follow-up
- 12 months after initiating rhythm control therapy
Methodological details2 US integrated healthcare delivery systems
- 2 US integrated healthcare delivery systems
Performance
As reported in the abstract
AF recurrence within 12 months after starting rhythm control therapy, by site (site 1, site 2)
| Test or model | Ablation, site 1 | Ablation, site 2 | Cardioversion, site 1 | Cardioversion, site 2 | Antiarrhythmic medication, site 1 | Antiarrhythmic medication, site 2 |
|---|---|---|---|---|---|---|
| NLP | 60.7% | 69.9% | 64.5% | 73.7% | 49.6% | 55.5% |
NLP
Ablation, site 160.7%
Ablation, site 269.9%
Cardioversion, site 164.5%
Cardioversion, site 273.7%
Antiarrhythmic medication, site 149.6%
Antiarrhythmic medication, site 255.5%
Further findings
NLP algorithm performance vs physician-adjudicated reference standard
- F-scores, sensitivity and specificity all above 0.90 at both sites
Parts of this Brief are withheld: the automatic checks could not confirm them against the abstract.
Original publication
Identification of recurrent atrial fibrillation using natural language processing applied to electronic health records. Eur Heart J Qual Care Clin Outcomes. Published 30 March 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →