Clinical and Echocardiographic Predictors of Moderate Heart Failure in Children With Congenital Heart Disease: A Prediction Modelling Study
Description
The dataset contains anonymized clinical and echocardiographic data of pediatric patients included in the study. It was collected retrospectively from medical records at a tertiary referral hospital. The dataset was used to analyze factors associated with heart failure severity in children with congenital heart disease and pulmonary hypertension. The data include demographic variables (age, sex), anthropometric measurements (body weight, nutritional status), clinical parameters (vital signs, Ross score), echocardiographic measurements (pulmonary systolic arterial pressure, tricuspid regurgitation gradient, mean pulmonary arterial pressure, ejection fraction, chamber dilatation), and diagnostic categories of congenital heart disease. Outcome variables and model-related variables used for statistical and machine learning analyses are also included. All patient identifiers were removed prior to analysis to ensure confidentiality. The dataset is provided in a structured tabular format and can be reused for secondary analyses, validation studies, or methodological research related to pediatric cardiology and predictive modeling.
Files
Steps to reproduce
1. Prepare the de-identified dataset and apply the predefined inclusion and exclusion criteria. 2. Define the primary outcome according to the main study definition. 3. Fit the primary predictive model using pulmonary systolic arterial pressure as the predictor. 4. Assess model discrimination by calculating the ROC-AUC and its 95% confidence interval. 5. Perform sensitivity analyses by replacing the primary predictor with tricuspid regurgitation gradient and mean pulmonary arterial pressure. 6. Conduct an additional sensitivity analysis using an alternative echocardiographic-based outcome definition.