ABSTRACT
Objective: This study aimed to describe transfusion reactions of pediatric patients from a National Center for Children's Health in China and to examine reaction incidents, reaction types by blood transfusion, and the associated blood products resulting in transfusion reactions. Methods: We compared transfusion reaction rates, among platelets, plasma, and red blood cells (RBCs) using a retrospective analysis of pediatric patients treated with blood transfusion based on data from the National Center for Children's Health (Beijing, China) by a hemovigilance reporting system from January 2015 to December 2019. Results: Over the past 5 years, 165 reactions were reported, and the overall incidence was 1.35 (95% CI: 1.14-1.55; 165/122,652); for each separate year, the incidences were 1.25 (95% CI: 0.76-1.74; 25/20,035; 2015), 1.09 (95% CI: 0.65-1.52; 24/22,084; 2016), 1.66 (95% CI: 1.14-2.18; 39/23,483; 2017), 1.36 (95% CI: 0.92-1.81; 36/26,440; 2018) and 1.34 (95% CI: 0.93-1.75; 41/30,610; 2019). Transfusion reaction incidents by person included 0.37 (95% CI: 0.21-0.53; 21/56,815) RBCs, 2.98 (95% CI: 2.33-3.64; 79/26,496) platelets and 1.65 (95% CI: 1.25-2.05; 65/39,341) frozen plasma. According to the analysis by blood products, the incidence of transfusion was 0.34 (95% CI: 0.20-0.48; 23/66,958) for RBCs, 3.21 (95% CI: 2.50-3.92; 78/24,318.5) for platelets, and 0.94 (95% CI: 0.71-1.17; 64/67,912) for frozen plasma. Transfusion reactions were most commonly associated with platelets, followed by plasma and RBC transfusions. The types of blood transfusion reactions were mainly allergic reactions (86.67%) and febrile non-hemolytic transfusion reactions (FNHTRs, 4.24%). The disease types of pediatric patients with transfusion reactions were concentrated among those with blood system diseases. A total of 80.61% of children with transfusion reactions had a previous blood transfusion history. Conclusions: Transfusion reactions are still relatively common in pediatric patients, and additional studies are necessary to address the differences in reaction rates, especially allergic and FNHTRs. Robust hemovigilance systems do include a special section dedicated to children will further the understanding of these reactions and trends, and prospective randomized clinical controlled trials may need to be conducted to perform preventive and corrective measures.
ABSTRACT
BACKGROUND: Postoperative blood coagulation assessment of children with congenital heart disease (CHD) has been developed using a conventional statistical approach. In this study, the machine learning (ML) was used to predict postoperative blood coagulation function of children with CHD, and assess an array of ML models. METHODS: This was a retrospective and data mining study. Based on the samples of 1,690 children with CHD, and screening data based on demographic characteristics, conventional coagulation tests (CCTs) and complete blood count (CBC), with a precise data selection process, and the support of data mining and ML algorithms including Decision tree, Naive Bayes, Support Vector Machine (SVM), Adaptive Boost (AdaBoost) and Random Forest model, and explored the best prediction models of postoperative blood coagulation function for children with CHD by models performance measured in the area under the receiver operating characteristic (ROC) curve (AUC), calibration or Lift curves, and further verified the reliability of the models with statistical tests. RESULTS: In primary objective prediction, as decision tree, Naive Bayes, SVM, the AUC of our prediction algorithm was 0.81, 0.82, 0.82, respectively. The accuracy rate of the overall forecast has reached more than 75%. Subsequently, we furtherly build improved models. Among them, the true positive rate of the AdaBoost, Random Forest and SVM prediction models reached more than 80% in the ROC curve. These overall accuracy rate indicated a good classification model. Combined calibration curves and Lift curves, the better fit is the SVM model, which predicted postoperative abnormal coagulation, Lift =2.2, postoperative normal coagulation, Lift =1.8. The statistical results furtherly proved the reliability of ML models. The age, sex, mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration (MCHC), white blood cell count (WBC) and platelet count (PLT) were the key features for predicting the postoperative blood coagulation state of children with CHD. CONCLUSIONS: ML technology and data mining algorithms may be used for outcome prediction in children with CHD for postoperative blood coagulation state based on the bulk of clinical data, especially CBC indictors from the real world.
