Valvular & structural heart diseaseCohort study
From Theory to Practice
- retrospective analysis
Authors' conclusion
GPT-based data extraction showed high accuracy across a broad range of variables, particularly continuous repeated measurements and rare intraoperative outcomes. Performance was lower for some infrequent postoperative events, reflecting sparse true positives. These findings support the feasibility of integrating LLM-assisted extraction into observational research workflows, with further validation needed in larger or multicenter cohorts.
Study context
- Population
- Patients undergoing transcatheter aortic valve replacement at Ca' Foncello Hospital (Treviso, Italy) between January and June 2024; medical and nursing records
- Participants
- 108 cases
- Exposure
- GPT vs Manual abstraction
Further findings
Accuracy
- valve brand
- 0.657
- gender, Barthel index, procedure timings, and several intraoperative complications
- 1.00
Sensitivity
- rare but clinically important events, including intraoperative neurological complications
- 1.00
Specificity
- most variables
- exceeded 0.90
Vital parameters (Bland-Altman)
- minimal bias and narrow limits of agreement
Original publication
From Theory to Practice. JACC Adv. Published 11 June 2026. doi:10.1016/j.jacadv.2026.102871
Summarised from the published abstract; every number is checked against it