Article Impact Level: HIGH Data Quality: STRONG Summary of Clinical Cardiology, 46(10), 1234–1243. https://doi.org/10.1002/clc.24104 Dr. Yicheng Wang et al.
Points
- This research paper developed and validated a machine learning (ML) based prediction model for assessing the risk of heart failure (HF) in patients with prediabetes or diabetes.
- The study utilized data from the National Health and Nutrition Examination Survey (NHANES) from 2007 to 2018, including 3527 subjects aged 40 years and older with a prior diagnosis of prediabetes or diabetes.
- Independent risk variables associated with HF were identified through univariate and multivariable logistic regression analyses, including age, poverty-to-income ratio, myocardial infarction condition, coronary heart disease condition, chest pain condition, and glucose-lowering medication use.
- Five ML models, including random forest (RF), were built using different ML algorithms and validated on a separate validation set.
- The RF model demonstrated the best prediction performance, with an area under the curve (AUC) of 0.978, indicating its accuracy in predicting the risk of HF in middle-aged and elderly patients with prediabetes or diabetes.
Summary
This research paper aimed to develop and validate a machine learning (ML) based prediction model for assessing the risk of heart failure (HF) in patients with prediabetes or diabetes. The study utilized data from the National Health and Nutrition Examination Survey (NHANES) from 2007 to 2018, including 3527 subjects aged 40 years and older with a prior diagnosis of prediabetes or diabetes.
The researchers conducted univariate and multivariable logistic regression analyses to identify independent risk variables associated with HF. The subjects were randomly divided into training and validation sets in a 7:3 ratio. Five ML models, including random forest (RF), were built using different ML algorithms on the training set and then validated on the validation set. The predictive performance of the ML models was evaluated using receiver operating characteristic (ROC) curves, calibration curves, decision curve analysis, and the Bootstrap resampling method.
The results of the multivariate logistic regression analysis revealed several independent predictors of HF, including age, poverty-to-income ratio, myocardial infarction condition, coronary heart disease condition, chest pain condition, and glucose-lowering medication use. The RF model demonstrated the best prediction performance among the five ML models, with an area under the curve (AUC) of 0.978.
In conclusion, ML models can accurately predict the risk of HF in middle-aged and elderly patients with prediabetes or diabetes. The RF model, in particular, showed the highest predictive performance and can assist healthcare professionals in making clinical decisions regarding managing and treating these patients.
Link to the article: https://onlinelibrary.wiley.com/doi/10.1002/clc.24104
References Wang, Y., Hou, R., Ni, B., Jiang, Y., & Zhang, Y. (2023). Development and validation of a prediction model based on machine learning algorithms for predicting the risk of heart failure in middle‐aged and older US people with prediabetes or diabetes. Clinical Cardiology, 46(10), 1234–1243. https://doi.org/10.1002/clc.24104
