Explainable Machine Learning Based a Packed Red Blood Cell Transfusion Prediction and Evaluation for Major Internal Medical Condition
0
조회
0
다운로드
1
인용
0
참고문헌
키워드
초록
Efficient use of limited blood products is becoming very important in terms of socioeconomic status and patientrecovery. To predict the appropriateness of patient-specific transfusions for the intensive care unit (ICU)patients who require real-time monitoring, we evaluated a model to predict the possibility of transfusiondynamically by using the Medical Information Mart for Intensive Care III (MIMIC-III), an ICU admissionrecord at Harvard Medical School. In this study, we developed an explainable machine learning to predict thepossibility of red blood cell transfusion for major medical diseases in the ICU. Target disease groups thatreceived packed red blood cell transfusions at high frequency were selected and 16,222 patients were finallyextracted. The prediction model achieved an area under the ROC curve of 0.9070 and an F1-score of 0.8166(LightGBM). To explain the performance of the machine learning model, feature importance analysis and apartial dependence plot were used. The results of our study can be used as basic data for recommendationsrelated to the adequacy of blood transfusions and are expected to ultimately contribute to the recovery ofpatients and prevention of excessive consumption of blood products.
Abstract
Efficient use of limited blood products is becoming very important in terms of socioeconomic status and patientrecovery. To predict the appropriateness of patient-specific transfusions for the intensive care unit (ICU)patients who require real-time monitoring, we evaluated a model to predict the possibility of transfusiondynamically by using the Medical Information Mart for Intensive Care III (MIMIC-III), an ICU admissionrecord at Harvard Medical School. In this study, we developed an explainable machine learning to predict thepossibility of red blood cell transfusion for major medical diseases in the ICU. Target disease groups thatreceived packed red blood cell transfusions at high frequency were selected and 16,222 patients were finallyextracted. The prediction model achieved an area under the ROC curve of 0.9070 and an F1-score of 0.8166(LightGBM). To explain the performance of the machine learning model, feature importance analysis and apartial dependence plot were used. The results of our study can be used as basic data for recommendationsrelated to the adequacy of blood transfusions and are expected to ultimately contribute to the recovery ofpatients and prevention of excessive consumption of blood products.
참고문헌 (16)
- 노주혜, "Blood Supply and Demand in Korea: What is in Store for the Future?", Yonsei Medical Journal, 61(5), pp.400~405 (2020), 연세대학교의과대학[DOI]
- 윤수진, "주요 수술의 최대혈액신청량 및 Adjacent Diagnosis Related Groups 환자분류체계에 따른 질환의 적혈구제제 평균수혈량의 설정", Laboratory Medicine Online, 10(3), pp.235~241 (2020), 대한진단검사의학회[DOI]
- O. Blatchford, "A risk score to predict need for treatment for uppergastrointestinal haemorrhage", The Lancet, 356(9238), pp.1318~1321 (2000)
- D. L. Shung, "Validation of a machine learning model that outperforms clinical risk scoring systems for upper gastrointestinal bleeding", Gastroenterology, 158(1), pp.160~167 (2020)
- C. Jo, "Transfusion after total knee arthroplasty can be predicted using the machine learning algorithm", Knee Surgery, Sports Traumatology, Arthroscopy, 28(6), pp.1757~1764 (2020)
- Shun Liu, "Machine learning models to predict red blood cell transfusion in patients undergoing mitral valve surgery", Annals of Translational Medicine, 9(7), pp.530~530 (2021), AME Publishing Company[DOI]
- Steven Walczak, "Prediction of perioperative transfusions using an artificial neural network", PLOS ONE, 15(2), pp.e0229450~ (2020), Public Library of Science (PLoS)[DOI]
- Riccardo Levi, "Artificial intelligence-based prediction of transfusion in the intensive care unit in patients with gastrointestinal bleeding", BMJ Health & Care Informatics, 28(1), pp.e100245~ (2021), BMJ[DOI]
- Dennis Shung, "Neural network predicts need for red blood cell transfusion for patients with acute gastrointestinal bleeding admitted to the intensive care unit", Scientific Reports, 11(1) (2021), Springer Science and Business Media LLC[DOI]
- S. E. Volpato, "Red blood cells transfusion in intensive care unit", Revista Brasileira de Terapia Intensiva, 21(4), pp.391~397 (2009)
- R. Baradarian, "Early intensive resuscitation of patients with upper gastrointestinal bleeding decreases mortality", American Journal of Gastroenterology, 99(4), pp.619~622 (2004)
- K. Oakland, "Acute lower GI bleeding in the UK : patient characteristics, interventions and outcomes in the first nationwide audit", Gut, 67(4), pp.654~662 (2018)
- A. L. Goldberger, "PhysioBank, PhysioToolkit, and PhysioNet : components of a new research resource for complex physiologic signals", Circulation, 101(23), pp.215~220 (2000)
- Alistair E.W. Johnson, "MIMIC-III, a freely accessible critical care database", Scientific Data, 3(1) (2016), Springer Science and Business Media LLC[DOI]
- J. H. Friedman, "Greedy function approximation : a gradient boosting machine", Annals of Statistics, 29(5), pp.1189~1232 (2001)
- H. Tauber, "Predicting transfusion requirements during extracorporeal membrane oxygenation", Journal of Cardiothoracic and Vascular Anesthesia, 30(3), pp.692~701 (2016)