Prediction of massive transfusion in trauma patients in the surgical intensive care units (THAI-SICU study)

Development of blood demand prediction model using artificial intelligence based on national public big data

Objective

Modern healthcare systems face challenges related to the stable and sufficient blood supply of blood due to shortages. This study aimed to predict the monthly blood transfusion requirements in medical institutions using an artificial intelligence model based on national open big data related to transfusion.

Methods

Data regarding blood types and components in Korea from January 2010 to December 2021 were obtained from the Health Insurance Review and Assessment Service and Statistics Korea. The data were collected from a single medical institution. Using the obtained information, predictive models were developed, including eXtreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LGBM), and category boosting (CatBoost). An ensemble model was created using these three models.

Results

The prediction performance of XGBoost, LGBM, and CatBoost demonstrated a mean absolute error ranging from 14.6657 for AB+ red blood cells (RBCs) to 84.0433 for A+ platelet concentrate (PC) and a root mean squared error ranging from 18.5374 for AB+ RBCs to 118.6245 for B+ PC. The error range was further improved by creating ensemble models, wherein the department requesting blood was the most influential parameter affecting transfusion prediction performance for different blood products and types. Except for the department, the features that affected the prediction performance varied for each product and blood type, including the number of RBC antibody screens, crossmatch, nationwide blood donations, and surgeries.

Conclusion

Based on blood-related open big data, the developed blood-demand prediction algorithm can efficiently provide medical facilities with an appropriate volume of blood ahead of time.

Accuracy of Trauma on Scene Triage Screening Tool (Shock Index, Reverse Shock Index Glasgow Coma Scale and National Early Warning Score) to Predict the Severity of Emergency Department Triage: A Retrospective Cross-Sectional Study

Introduction

Prehospital trauma care includes on-scene assessments, essential treatment, and facilitating transfer to an appropriate trauma center to deliver optimal care for trauma patients. While the Simple Triage and Rapid Treatment (START), Revised Triage Sieve (rTS), and National Early Warning Score (NEWS) tools are user-friendly in a prehospital setting, there is currently no standardized on-scene triage protocol in Thailand Emergency Medical Service (EMS). Therefore, this study aims to evaluate the precision of these tools (SI, rSIG, and NEWS) in predicting the severity of trauma patients who are transferred to the emergency department (ED).

Methods

This study was a retrospective cross-sectional and diagnostic research conducted on trauma patients transferred by EMS to the ED of Ramathibodi Hospital, a university-affiliated super tertiary care hospital in Bangkok, Thailand, from January 2015 to September 2022. We compared the on-scene triage tool (SI, rSIG, and NEWS) and ED triage tool (Emergency Severity Index) parameters, massive transfusion protocol (MTP), and intensive care unit (ICU) admission with the area under ROC (univariable analysis) and diagnostic odds ratio (multivariable logistic regression analysis). The optimal cut-off threshold for the best parameter was determined by selecting the value that produced the highest area under the ROC curve.

Results

A total of 218 patients were traumatic patients transported by EMS to the ED, out of which 161 were classified as ESI levels 1-2, while the remaining 57 patients were categorized as levels 3-5 on the ESI triage scale. We found that NEWS was a more accurate triage tool to discriminate the severity of trauma patients than rSIG and SI. The area under the ROC was 0.74 (95% CI 0.70-0.79) (OR 18.98, 95% CI 1.06-337.25), 0.65 (95% CI 0.59-0.70) (OR 1.74, 95% CI 0.17-18.09) and 0.58 (95% CI 0.52-0.65) (OR 0.28, 95% CI 0.04-1.62), respectively (P-value <0.001). The cut point of NEWS to discriminate ESI levels 1-2 and levels 3-5 was >6 points.

Conclusion

NEWS is the best on-scene triage screening tool to predict the severity at the emergency department, massive transfusion protocol (MTP), and intensive care unit (ICU) admission compared with other triage tools SI and rSIG.

Development and internal validation of a nomogram to predict massive blood transfusions in neurosurgical operations

Objectives

A massive blood transfusion (MBT) is an unexpected event that may impact mortality. Neurosurgical operations are a major operation involving the vital structures and risk to bleeding. The aims of the present research were (1) to develop a nomogram to predict MBT and (2) to estimate the association between MBT and mortality in neurosurgical operations.

Material and Method

We conducted a retrospective cohort study including 3660 patients who had undergone neurosurgical operations. Univariate and multivariate logistic regression analyses were used to test the association between clinical factors, pre-operative hematological laboratories, and MBT. A nomogram was developed based on the independent predictors.

Results

The predictive model comprised five predictors as follows: Age group, traumatic brain injury, craniectomy operation, pre-operative hematocrit, and pre-operative international normalized ratio and the good calibration were observed in the predictive model. The concordance statistic index was 0.703. Therefore, the optimism-corrected c-index values of cross-validation and bootstrapping were 0.703 and 0.703, respectively.

Conclusion

MBT is an unexpectedly fatal event that should be considered for appropriate preparation blood components. Further, this nomogram can be implemented for allocation in limited-resource situations in the future.

A Machine Learning Approach for the Prediction of Traumatic Brain Injury Induced Coagulopathy

Background: Traumatic brain injury-induced coagulopathy (TBI-IC), is a disease with poor prognosis and increased mortality rate.

Objectives: Our study aimed to identify predictors as well as develop machine learning (ML) models to predict the risk of coagulopathy in this population.

Methods: ML models were developed and validated based on two public databases named Medical Information Mart for Intensive Care (MIMIC)-IV and the eICU Collaborative Research Database (eICU-CRD). Candidate predictors, including demographics, family history, comorbidities, vital signs, laboratory findings, injury type, therapy strategy and scoring system were included. Models were compared on area under the curve (AUC), accuracy, sensitivity, specificity, positive and negative predictive values, and decision curve analysis (DCA) curve.

Results: Of 999 patients in MIMIC-IV included in the final cohort, a total of 493 (49.35%) patients developed coagulopathy following TBI. Recursive feature elimination (RFE) selected 15 variables, including international normalized ratio (INR), prothrombin time (PT), sepsis related organ failure assessment (SOFA), activated partial thromboplastin time (APTT), platelet (PLT), hematocrit (HCT), red blood cell (RBC), hemoglobin (HGB), blood urea nitrogen (BUN), red blood cell volume distribution width (RDW), creatinine (CRE), congestive heart failure, myocardial infarction, sodium, and blood transfusion. The external validation in eICU-CRD demonstrated that adapting boosting (Ada) model had the highest AUC of 0.924 (95% CI: 0.902–0.943). Furthermore, in the DCA curve, the Ada model and the extreme Gradient Boosting (XGB) model had relatively higher net benefits (ie, the correct classification of coagulopathy considering a trade-off between false- negatives and false-positives)—over other models across a range of threshold probability values.

Conclusions: The ML models, as indicated by our study, can be used to predict the incidence of TBI-IC in the intensive care unit (ICU).

Artificial intelligence-based prediction of transfusion in the intensive care unit in patients with gastrointestinal bleeding

Objective

Gastrointestinal (GI) bleeding commonly requires intensive care unit (ICU) in cases of potentialhaemodynamiccompromise or likely urgent intervention. However, manypatientsadmitted to the ICU stop bleeding and do not require further intervention, including blood transfusion. The present work proposes an artificial intelligence (AI) solution for the prediction of rebleeding in patients with GI bleeding admitted to ICU.

Methods

A machine learning algorithm was trained and tested using two publicly available ICU databases, the Medical Information Mart for Intensive Care V.1.4 database and eICU Collaborative Research Database using freedom from transfusion as a proxy for patients who potentially did not require ICU-level care. Multiple initial observation time frames were explored using readily available data including labs, demographics and clinical parameters for a total of 20 covariates.

Results

The optimal model used a 5-hour observation period to achieve an area under the curve of the receiving operating curve (ROC-AUC) of greater than 0.80. The model was robust when tested against both ICU databases with a similar ROC-AUC for all.

Conclusions

The potential disruptive impact of AI in healthcare innovation is acknowledge, but awareness of AI-related risk on healthcare applications and current limitations should be considered before implementation and deployment. The proposed algorithm is not meant to replace but to inform clinical decision making. Prospective clinical trial validation as a triage tool is warranted.