Heart failure & cardiomyopathies
Comparison of machine learning and conventional statistical modeling for predicting readmission following acute heart failure hospitalization
Authors' conclusion
Fine-Gray models had similar discrimination but superior calibration to the RSF-CR model, highlighting the importance of reporting calibration metrics for ML-based prediction models.
Study context
- Population
- Patients with HF (> 18 years) discharged alive from a hospital or ED in Ontario, Canada, 2004–20073,602 in validation sample
- Participants
- 10,919 patients
- Exposure
- Machine learning (random survival forests for competing risks (RSF-CR)) vs conventional statistical model (Fine-Gray) for predicting readmission
- Outcome
- Discrimination and calibration for CV and noncardiovascular readmission
- Follow-up
- 30 days
Methodological detailsprediction model development and validation, 2:1 random training/validation split
- Prediction model development and validation
- 2:1 random training/validation split
Main finding
As reported in the abstract
Discrimination (c-statistic) in validation set
| Group | C-statistic, CV readmission | C-statistic, noncardiovascular readmission |
|---|---|---|
| RSF-CR | 0.620 | 0.632 |
| Fine-Gray | 0.621 | 0.641 |
C-statistic, CV readmission
RSF-CR0.620
Fine-Gray0.621
C-statistic, noncardiovascular readmission
RSF-CR0.632
Fine-Gray0.641
Observational data: associations, not effects.
Original publication
Comparison of machine learning and conventional statistical modeling for predicting readmission following acute heart failure hospitalization. Am Heart J. Published 31 July 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →