Key result
Simple logistic regression outperforms complex machine learning in predicting 14-day hospitalization risk in chronic HF.
Why the study?
Complex machine learning models have seen low clinical adaptation, prompting investigation into whether interpretable models using biweekly vital signs can predict acute hospitalizations in CHF patients.
Do interpretable machine learning models accurately predict 14-day nonelective hospitalization risk in chronic heart failure patients using biweekly vital signs?
Observational (n=122)
Yes
Do interpretable machine learning models accurately predict 14-day nonelective hospitalization risk in chronic heart failure patients using biweekly vital signs?
Absolute Event Rate: 0.657% vs 0.644%
Simple interpretable machine learning methods outperformed complex black-box models in predicting 14-day hospitalization risk in CHF patients, though overall predictive performance remained modest.
Modest accuracy cautions against clinical adoption of biweekly vital-sign models; leaves open whether larger cohorts can improve prediction in chronic heart failure.
Purpose We wished to investigate whether the risk of acute hospitalizations of chronic heart failure (CHF) patients, could be predicted from biweekly measurements of pulse, blood pressure and weight. We emphasized machine learning models with a high degree of interpretability, due to low adaptation of complex machine learning models in clinical practice. Methods Using 11,575 measurements of pulse, blood pressure and weight belonging to 122 patients, we trained three types of machine learning algorithms, logistic regression, Random Forest and “RuleFit” to predict nonelective hospitalization within the next 14 days. We used a fivefold cross validation framework to estimate performance metrics, including f-measure, “Receiver Operating Characteristic—Area Under the Curve” (ROC-AUC), sensitivity and specificity. Results A simple interpretable machine learning algorithm, logistic regression with least absolute shrinkage and selection operator (lasso), performed the best. The regression based on simple features performed with a ROC-AUC of 0.622 (sensitivity = 0.185, specificity = 0.93), while the regression based on a more complex feature set performed with a ROC-AUC of 0.657 (sensitivity = 0.212, specificity = 0.921). Conclusion In our study simple interpretable methods, outperformed more complex black box machine learning methods in predicting hospitalization of heart failure patients. This suggests that interpretable methods are appropriate in this context. However, the strength of results are slightly limited by the overall modest performance of the models and the small sample size. Clinical Trial Registration The original trial was registered at ClinicalTrials.gov, with the identification number NCT02860013 at the 9th of august, 2016.
No takes yet. Share an insight, caveat, or question.
Xylander et al. (2025) conducted an observational in Chronic heart failure (n=122). Logistic regression with lasso vs. Random Forest and RuleFit models was evaluated on Prediction of nonelective hospitalization within 14 days (ROC-AUC). A simple logistic regression model predicted 14-day hospitalization risk in chronic heart failure patients with a ROC-AUC of 0.657, outperforming more complex machine learning methods.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: