Machine Learning Predicts Cardiovascular Events in Patients With Diabetes: The Silesia Diabetes-Heart Project
Problem
Patients with diabetes are at high risk of cardiovascular events, and predicting which patients will suffer events is difficult.
Why it matters
Cardiovascular events are a leading cause of death and hospitalization in diabetes patients.
What the AI did
A machine learning pipeline (neighborhood component analysis plus hybrid sampling/boosting, logistic regression, or clustering) was developed to predict new cardiovascular events in hospitalized diabetes patients.
Results
1735 patients analyzed; 150 (8.65%) had new CV events in follow-up; model performance metrics not stated in abstract
Impact
The evidence supports model development on real prospective data, but predictive accuracy is not stated.
Record
| Venue | Current Problems in Cardiology |
|---|---|
| Published | 2023 |
| Work type | article |
| Authors | Katarzyna Nabrdalik, Hanna Kwiendacz, Karolina Drożdż, Krzysztof Irlik, Mirela Hendel, Agata M. Wijata, Jakub Nalepa, Elon Correa et al. |
| Citations | 32 (OpenAlex, at import time) |
| References | 28 |
| Model or method | NCA with hybrid sampling/boosting; multiple logistic regression; hierarchical clustering |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1016/j.cpcardiol.2023.101694 |
| Open access | hybrid |
| Keywords | machine learning, diabetes, cardiovascular events, prediction, prospective study |
Importance rating
significant — thousands to millions affected
CV event prediction in hospitalized diabetes patients; diabetes CV complications are serious, but abstract only describes a 1735-patient single-center study with modest AUCs, no broader scale stated.