Heart failure & cardiomyopathies
Deep Learning for Predicting Rehospitalization in Acute Heart Failure: Model Foundation and External Validation
Authors' conclusion
Our deep learning-based model using real-world data could provide valid predictions of HF rehospitalization in 1 year follow-up.
Study context
- Population
- Patients admitted for acute HF (AHF) to a tertiary hospital, January 2014–January 2019
- Outcome
- Prediction of HF rehospitalization (AUC)
- Follow-up
- 30, 90 and 365 days after AHF discharge
Methodological detailssingle-centre, deep learning (recurrent layers with gated recurrent units), real-world data
- Single-centre
- Deep learning (recurrent layers with gated recurrent units)
- Real-world data
Performance
As reported in the abstract
Discrimination for HF readmission (AUC)
AUC, readmission within 30 days
0.63
AUC, readmission within 90 days
0.74
AUC, readmission within 365 days
0.76
Further findings
Prognostic features
- Features significantly associated with HF rehospitalization
- 22
- Time-independent features
- 6
- Time-dependent features
- 16
- Features during follow-up (FU) contribute relatively more than features from other time points (Shapley value)
Original publication
Deep Learning for Predicting Rehospitalization in Acute Heart Failure: Model Foundation and External Validation. ESC Heart Fail. Published 9 July 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →