Construction and evaluation of a 180-day readmission prediction model for chronic heart failure patients based on...

The high readmission rate of patients with chronic heart failure (HF) can cause waste of medical resources and economic losses. Establishing an effective HF readmission model can effectively alleviate medical pressure and improve the quality of treatment. In this study, we conducted a comprehensive analysis of clinical and laboratory data from 248 patients with chronic HF who received treatment at our medical center between January 2021 to January 2022. We also measured soluble CD40 ligand (sCD40L) levels to determine their association with readmission due to HF during follow-up. To analyze the data, we employed various statistical methods including one-way ANOVA, correlation analysis, univariate COX regression, and Least Absolute Shrinkage and Selection Operator COX regression. Using these techniques, we organized the data and constructed a predictive model that was both trained and validated. We developed a nomogram to assess the likelihood of readmission within 180 days for patients with chronic HF. Our findings revealed that monocytes, creatinine, sCD40L, and hypertension history were all independent risk factors for 180-day HF readmissions. Additionally, our model's AUC was 0.731 in the training dataset and 0.704 in the validation dataset. This study provides new insights for predicting readmission within 180 days for patients with chronic HF. And sCD40L is an important predictive indicator for readmission of HF patients within 180 days, and clinical doctors can develop appropriate treatment plans based on sCD40L.