Blood Transfusion Database for Pediatric Congenital Heart Disease Surgery

Description

This dataset contains perioperative data from 3,342 pediatric patients undergoing congenital heart disease surgery. Variables include demographics (age, gender, weight), preoperative laboratory values (hemoglobin, hematocrit, platelet count, coagulation parameters, liver function tests), surgical complexity (Aristotle Score), cardiopulmonary bypass use, and intraoperative blood transfusion volumes (RBC, plasma, platelets). The transfusion rates were 17.7% for RBC, 31.3% for plasma, and 2.0% for platelets. Missing values are denoted by "/". This data supports the development of machine learning models for predicting blood transfusion requirements in pediatric cardiac surgery. Missing data primarily occurred in coagulation parameters (PT, APTT) and some laboratory values that were not routinely measured in all patients. This study was approved by the Institutional Review Board. The requirement for informed consent was waived due to the retrospective nature of the study. All data were de-identified to protect patient privacy. This dataset is associated with the manuscript: "Machine Learning-Based Prediction of Blood Transfusion in Pediatric Congenital Heart Surgery: A Novel Framework for Unified Comparison"

Steps to reproduce

Data were retrospectively extracted from electronic medical records of pediatric patients undergoing congenital heart disease surgery. Preoperative demographics, laboratory values (complete blood count, coagulation parameters, liver function tests), surgical complexity scores (Aristotle Score), and intraoperative blood transfusion volumes were collected. All patient identifiers were removed for de-identification. Missing values are denoted by "/". Data were compiled into Excel format with 3,342 patient records and 21 variables.

Categories

Funders

• National Key R&D Program of China

  Grant ID: 2023YFC2706400

Licence