Development and Validation of a Transfusion Risk Score for Patients Receiving Maintenance Hemodialysis

Background: In dialysis patients with anemia, avoiding red blood cell transfusions is preferable. We sought to develop and validate a novel transfusion prediction risk score for patients receiving maintenance hemodialysis. Methods: This retrospective cohort study used United States Renal Data System data to create a model development cohort (point prevalent hemodialysis patients on November 1, 2012) and a validation cohort (point prevalent hemodialysis patients on August 1, 2013). We characterized comorbidity, inflammatory conditions, hospitalizations, anemia and anemia management, iron parameters, intravenous iron use, and vitamin D use during a 6-month baseline period to predict subsequent 3-month transfusion risk. We used logistic least absolute shrinkage and selection operator regression. In an exploratory analysis, model results were used to calculate a score to predict 6- and 12-month hospitalization and mortality. Results: Variables most predictive of transfusion were prior transfusion, hemoglobin, ferritin, and number of hospital days in the baseline period. The resulting C-statistic in the validation cohort was 0.74, indicating relatively good predictive power. The score was associated with a significantly increased risk of subsequent mortality (hazard ratios 1.0, 1.22, 1.26, 1.54, 1.71 grouped from lowest to highest score), but not with hospitalization. Conclusions: We developed a transfusion prediction risk score with good performance characteristics that was associated with mortality. This score could be further developed into a clinically useful application allowing clinicians to identify hemodialysis patients most likely to benefit from a timely, proactive anemia treatment approach with the goal of avoiding red blood cell transfusions and attendant risks of adverse clinical outcomes.

Prediction Model of Perioperative Blood Transfusion for Cardiovascular Surgery Patients Based on Machine Learning: Retrospective Study Using Electronic Medical Records (Preprint)

BACKGROUND Blood transfusion was related to postoperative adverse events and increased medical costs in patients underwent cardiovascular surgery. Predicting transfusion risk or major bleeding risk will help reduce transfusion. Machine learning (ML) methods show good performance at predicting risk, but transfusion risk prediction based on ML models among Chinese population were unavailable. OBJECTIVE To establish and validate prediction models using ML methods for perioperative transfusion risk of patients undergoing cardiovascular surgery in the Chinese population. METHODS Analysis was performed using electronic medical records from patients underwent cardiovascular surgery in Fuwai hospital between January 1, 2016 and June 30, 2019. Based on the 66402 unique patients, a retrospective cohort (N=61892) and a prospective cohort (N=4510) were formed for model derivation and validation. Four ML algorithms including eXtreme Gradient Boosting (XGBoost), random forest, naive Bayes, logistic regression with least absolute shrinkage and selection operator were adopted using 10-folds cross-validation to build prediction models of perioperative blood, red blood cell, plasma and platelet transfusion. According to the model evaluation in the validation cohort, the optimal perioperative blood transfusion prediction model was selected to compare with the Association of Cardiothoracic Anaesthetists perioperative risk of blood transfusion score (the ACTA-PORT score) established in previous research. RESULTS Among ML models, the XGBoost(area under the receiver-operating characteristic curve[AUC]:0.823; 95% confidence interval[CI]: 0.810 to 0.836) outperformed other models for perioperative blood transfusion and showed better prediction ability than ACTA-PORT score (AUC:0.690; 95% CI: 0.673 to 0.707; P<.001) in the validation cohort. While ML prediction models for perioperative red blood cell transfusion, plasma transfusion and platelet transfusion, achieving good model performance as AUC levels were 0.836(95% CI: 0.823 to 0.849), 0.766(95% CI: 0.745 to 0.787) and 0.948(95% CI: 0.937 to 0.959) respectively. CONCLUSIONS The study retrospectively developed and prospectively validated discriminative perioperative transfusion prediction models, which may promote the early warning and intervention against perioperative transfusion, and benefit patient blood management.

Identification of Preoperative Clinical Factors Associated With Perioperative Blood Transfusions

Predicting patients' surgical transfusion needs preoperatively enables more efficient blood resource management. Identifying the significance of variables to use for transfusion predictions may be accomplished more reliably using machine learning, specifically artificial neural networks (ANN). A logistic regression model and two ANN programs are used to identify the contribution of nine variables selected following a literature review. The first ANN uses a sum of the weights method to identify variable contribution and the second ANN uses a leave one out strategy to identify variable contribution. All models indicated that hematocrit was the most significant variable for predicting perioperative blood transfusions. The weighted averages method indicated wRVU's and ASA score were the next most significant contributors. The leave one out method identified sex and INR as contributing to transfusion prediction. The importance of the variables other than hematocrit varied between techniques and may be dependent on the modeling method used.