Dataset Information

Development and validation of a machine learning method to predict intraoperative red blood cell transfusions in cardiothoracic surgery.

ABSTRACT: Accurately predicting red blood cell (RBC) transfusion requirements in cardiothoracic (CT) surgery could improve blood inventory management and be used as a surrogate marker for assessing hemorrhage risk preoperatively. We developed a machine learning (ML) method to predict intraoperative RBC transfusions in CT surgery. A detailed database containing time-stamped clinical variables for all CT surgeries from 5/2014-6/2019 at a single center (n = 2410) was used for model development. After random forest feature selection, surviving features were inputs for ML algorithms using five-fold cross-validation. The dataset was updated with 437 additional cases from 8/2019-8/2020 for validation. We developed and validated a hybrid ML method given the skewed nature of the dataset. Our Gaussian Process (GP) regression ML algorithm accurately predicted RBC transfusion amounts of 0 and 1-3 units (root mean square error, RMSE 0.117 and 1.705, respectively) and our GP classification ML algorithm accurately predicted 4 + RBC units transfused (area under the curve, AUC = 0.826). The final prediction is the regression result if classification predicted < 4 units transfused, or the classification result if 4 + units were predicted. We developed and validated an ML method to accurately predict intraoperative RBC transfusions in CT surgery using local data.

SUBMITTER: Wang Z

PROVIDER: S-EPMC8789772 | biostudies-literature | 2022 Jan

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Development and validation of a machine learning method to predict intraoperative red blood cell transfusions in cardiothoracic surgery.

Wang Zheng Z Zhe Shandian S Zimmerman Joshua J Morrisey Candice C Tonna Joseph E JE Sharma Vikas V Metcalf Ryan A RA

Scientific reports 20220125 1

Accurately predicting red blood cell (RBC) transfusion requirements in cardiothoracic (CT) surgery could improve blood inventory management and be used as a surrogate marker for assessing hemorrhage risk preoperatively. We developed a machine learning (ML) method to predict intraoperative RBC transfusions in CT surgery. A detailed database containing time-stamped clinical variables for all CT surgeries from 5/2014-6/2019 at a single center (n = 2410) was used for model development. After random ...[more]

PMID: 35079127

Similar Datasets

Project description:IntroductionMajor perioperative blood loss often necessitates allogeneic blood transfusions, leading to adverse outcomes and straining healthcare resources. Intraoperative cell salvage (ICS) offers benefits like reduced adverse reactions, optimised blood resource allocation, and cost-effectiveness. Yet, ICS adoption remains limited. Our aim is to formalise and expand ICS as a routine service to promote wider adoption and improve patient outcomes.MethodsThe model for improvement was adopted for an ICS quality improvement initiative across two Plan-Do-Study-Act cycles. Interventions followed the '4E' framework: education, empowerment, enforcement and embedding reminders. The primary outcome was the number of allogeneic red blood cell (RBC) units transfused per surgery with predicted blood loss >500 mL; the secondary outcome was the percentage of indicated surgeries using ICS. Process measure was percentage of anaesthetic unit nurses trained in ICS. Balancing measures included the number of ICS activations with no blood return and complications during and after ICS.ResultsA total of 152 surgeries used ICS postintervention (January 2023 to April 2024), marking an increase from preintervention levels. The median number of monthly mean allogeneic RBC units transfused per patient demonstrated a significant reduction from 3.10 (preintervention) to 2.44 (postintervention). The percentage of indicated surgeries utilising ICS increased from 1.4% to 12.3%. Cost savings were observed, with potential annual savings close to SGD 500 000. No complications related to the use of ICS were recorded.ConclusionThe formalisation and expansion of the ICS service within our institution resulted in notable improvements, including increased ICS adoption rates, reduced reliance on allogeneic blood transfusions and potential cost savings. Future quality improvement efforts should concentrate on further promoting ICS adoption, particularly for surgeries with clinical indications.

Project description:Clinical improvement after red blood cell (RBC) transfusions in preterm infants remains debated. This study aims to investigate the effect of RBC transfusion on the occurrence of desaturations and hypoxia, and other cardiorespiratory outcomes in preterm infants. In this longitudinal observational study, prospectively stored cardiorespiratory parameters of preterm infants who received at least one RBC transfusion between July 2016 and June 2017 were retrospectively analyzed. Sixty infants with 112 RBC transfusions, median GA of 26.7 weeks, were included. The number of desaturations and area < 80% SpO2 limit, as a measure of the hypoxic burden, were calculated in 24 h before and after RBC transfusion. A mixed effects model was used to account for repeated measurements. Overall, the mean (SE) number of desaturations per hour decreased from 3.28 (0.55) to 2.25 (0.38; p < 0.001), and area < 80% SpO2 limit decreased from 0.14 (0.04) to 0.08 (0.02) %/s (p = 0.02). These outcomes were stratified for the number of desaturations in 24 h prior to RBC transfusion. The largest effect was observed in the group with the highest mean number of desaturations (≥ 6) prior to RBC transfusion, with a decrease from 7.50 (0.66) to 4.26 (0.38) (p < 0.001) in the number of desaturations and 0.46 (0.13) to 0.20 (0.06) in the area < 80% SpO2. Perfusion index increased significantly after RBC transfusion (p < 0.001). No other significant effects of RBC transfusion on cardiorespiratory data were observed.Conclusions: RBC transfusions in preterm newborns could help decrease the incidence of desaturations and the area < 80% SpO2 as a measure of the hypoxic burden. The higher the number of desaturations prior to the RBC transfusion, the larger the effect observed. What is Known: •Red blood cell transfusions potentially prevent hypoxia in anemic preterm infants by increasing the circulatory hemoglobin concentration and improving tissue oxygenation. •There is not a predefined hemoglobin concentration cut-off for the occurrence of symptomatic anemia in preterm infants. What is New: •Oxygen desaturations and hypoxia in anemic preterm infants can be improved by RBC transfusions, especially if more desaturations have occurred before transfusion. •Cardiorespiratory monitor data may help identify infants who will benefit most from red blood cell transfusions.

Dataset Information

Development and validation of a machine learning method to predict intraoperative red blood cell transfusions in cardiothoracic surgery.

Publications

Development and validation of a machine learning method to predict intraoperative red blood cell transfusions in cardiothoracic surgery.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets