Heart failure & cardiomyopathiesCohort study
Artificial Intelligence Methods for Improved Detection of Undiagnosed Heart Failure with Preserved Ejection Fraction
- Retrospective cohort study
- Natural language processing (NLP) pipeline applied to electronic health record (EHR)
- validated in a second, independent centre
Authors' conclusion
This study demonstrates that patients with undiagnosed HFpEF are an at-risk group with high mortality.
Study context
- Population
- Patients with a clinical diagnosis of heart failure (HF) identified from the EHR, 2010–20223,727 with left ventricular ejection fraction (LVEF) ≥50%
- Participants
- 8,606 patients
- Exposure
- Clinician-assigned heart failure with preserved ejection fraction (HFpEF) vs European Society of Cardiology (ESC) criteria
Main finding
As reported in the abstract
Diagnostic status among patients with HF and LVEF ≥50%
Observational data: associations, not effects.
Further findings
Hospitalisation, 5-year mortality, comorbidity, acute CV events
- Confirmed HFpEF hospitalized more frequently; ESC criteria group had higher 5-year mortality, despite being less comorbid with fewer acute CV events
Original publication
Artificial Intelligence Methods for Improved Detection of Undiagnosed Heart Failure with Preserved Ejection Fraction. Eur J Heart Fail. Published 11 January 2024. doi:10.1002/ejhf.3115
Summarised from the published abstract; every number is checked against it