<![CDATA[The aim of this study was to identify clinical and demographic characteristics associated with heart failure and develop an interpretable risk model using binary logistic regression on hospital patient data. Early detection of heart failure is expected to support timely intervention and clinical decision-making based on routine measurements. This study analyzed 130 anonymized patient data with heart failure status as a binary outcome. The initial logistic regression model included all candidate predictors and was then simplified to improve stability and calibration. Results are presented as odds ratios with 95% CIs. Performance evaluation included ROC–AUC, classification metrics, the Hosmer–Lemeshow test, calibration plots, and 5-fold cross-validation. The final model was significant (LR p = 1.0×10⁻⁵; McFadden R² = 0.222) with an accuracy of 81.54%, sensitivity of 89.41%, specificity of 66.67%, AUC of 0.811, and a Brier score of 0.164. Cross-validation showed an average AUC of 0.774 and an accuracy of 0.762. Significant predictors included BMI, serum creatinine, serum potassium, and total cholesterol, with acceptable calibration (p = 0.0767). This model has potential use as an interpretive screening tool, although external validation is still needed.]]>
Copyrights © 2026
