Identification of Important Features to Predict 30 -Day Mortality in Heart Failure Patients with Cardiogenic Shock

Introduction: Cardiovascular diseases remain a leading global cause of mortality, with ischemic heart disease projected to account for 23.3 million deaths by 2030. Heart failure and cardiogenic shock account for a significant proportion of these deaths and require timely treatment as medical emergencies. This study aims to predict mortality within one month in patients experiencing cardiogenic shock secondary to heart failure using a concise set of predictive features. Method: An analytical cross -sectional study was conducted at Babol Razi Hospital, involving 201 adult patients (≥18 years) treated for cardiogenic shock in 2020. Data from 34 clinical variables, including age, history of cardiac surgery, pH levels, lactate concentration, diabetes status, and blood pressure, were meticulously analyzed. Mortality outcomes within one month were assessed via structured telephone follow -up. Logistic regression and Gradient Boosting Machine (GBM) algorithms were used for predictive modeling . Results: The average age of patients was 69.44 ±15.71 years. Among them, 47.7% died. The study identified age, lactate levels, diabetes, and initial confusion as significant predictors of mortality risk. Each additional year of age was associated with a 7% higher probability of mortality. Diabetic patients faced more than double the mortality risk compared to non -diabetics. Confusion at presentation increased the mortality risk fourfold, while elevated lactate levels raised it by 1.5 times . Conclusion: Logistic regression and GBM algorithms demonstrated effectiveness in predicting one -month mortality among cardiogenic shock patients with heart failure based on selected features. This approach holds promise for improving referral processes and reducing costs in healthcare settings .