Machine Learning Model for the Prediction of Hemorrhage in Intensive Care Units
2
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
Objectives: Early hemorrhage detection in intensive care units (ICUs) enables timely intervention and reduces the risk of irreversible outcomes. In this study, we aimed to develop a machine learning model to predict hemorrhage by learning the patterns of continuously changing, real-world clinical data. Methods: We used the Medical Information Mart for Intensive Care databases (MIMIC-III and MIMIC-IV). A recurrent neural network was used to predict severe hemorrhage in the ICU. We developed three machine learning models with an increasing number of input features and levels of complexity: model 1 (11 features), model 2 (18 features), and model 3 (27 features). MIMIC-III was used for model training, and MIMIC-IV was split for internal validation. Using the model with the highest performance, external verification was performed using data from a subgroup extracted from the eICU Collaborative Research Database. Results: We included 5,670 ICU admissions, with 3,150 in the training set and 2,520 in the internal test set. A positive correlation was found between model complexity and performance. As a measure of performance, three models developed with an increasing number of features showed area under the receiver operating characteristic (AUROC) curve values of 0.61–0.94 according to the range of input data. In the subgroup extracted from the eICU database for external validation, an AUROC value of 0.74 was observed. Conclusions: Machine learning models that rely on real clinical data can be used to predict patients at high risk of bleeding in the ICU.
Abstract
Objectives: Early hemorrhage detection in intensive care units (ICUs) enables timely intervention and reduces the risk of irreversible outcomes. In this study, we aimed to develop a machine learning model to predict hemorrhage by learning the patterns of continuously changing, real-world clinical data. Methods: We used the Medical Information Mart for Intensive Care databases (MIMIC-III and MIMIC-IV). A recurrent neural network was used to predict severe hemorrhage in the ICU. We developed three machine learning models with an increasing number of input features and levels of complexity: model 1 (11 features), model 2 (18 features), and model 3 (27 features). MIMIC-III was used for model training, and MIMIC-IV was split for internal validation. Using the model with the highest performance, external verification was performed using data from a subgroup extracted from the eICU Collaborative Research Database. Results: We included 5,670 ICU admissions, with 3,150 in the training set and 2,520 in the internal test set. A positive correlation was found between model complexity and performance. As a measure of performance, three models developed with an increasing number of features showed area under the receiver operating characteristic (AUROC) curve values of 0.61–0.94 according to the range of input data. In the subgroup extracted from the eICU database for external validation, an AUROC value of 0.74 was observed. Conclusions: Machine learning models that rely on real clinical data can be used to predict patients at high risk of bleeding in the ICU.
참고문헌 (30)
- Despotis G, "Prediction and management of bleeding in cardiac surgery", J Thromb Haemost, 7(Suppl 1), pp.111~117 (2009)[DOI]
- Ferraris VA, "Blood transfusion and adverse surgical outcomes : the good and the bad", Surgery, 158(3), pp.608~617 (2015)[DOI]
- Rao SV, "Relationship of blood transfusion and clinical outcomes in patients with acute coronary syndromes", JAMA, 292(13), pp.1555~1562 (2004)[DOI]
- Cook DJ, "The attributable mortality and length of intensive care unit stay of clinically important gastrointestinal bleeding in critically ill patients", Crit Care, 5(6), pp.368~375 (2001)[DOI]
- Pirie L, "EU optimal blood use project partners and project management team", The EU optimal blood use project. Transfus Clin Biol, 14(6), pp.499~503 (2007)[DOI]
- Mitterecker A, "Machine learning-based prediction of transfusion", Transfusion, 60(9), pp.1977~1986 (2020)[DOI]
- Janett RS, "Electronic medical records in the American Health System : challenges and lessons learned", Cien Saude Colet, 25(4), pp.1293~1304 (2020)[DOI]
- Carra G, "Data-driven ICU management : using big data and algorithms to improve outcomes", J Crit Care, 60, pp.300~304 (2020)[DOI]
- Celi LA, "“Big data” in the intensive care unit: closing the data loop", Am J Respir Crit Care Med, 187(11), pp.1157~1160 (2013)[DOI]
- Chen JH, "Machine learning and prediction in medicine : beyond the peak of inflated expectations", N Engl J Med, 376(26), pp.2507~2509 (2017)[DOI]
- Gombotz H, "Preoperative identification of patients with increased risk for perioperative bleeding", Curr Opin Anaesthesiol, 26(1), pp.82~90 (2013)[DOI]
- Thorson CM, "Change in hematocrit during trauma assessment predicts bleeding even with ongoing fluid resuscitation", Am Surg, 79(4), pp.398~406 (2013)[DOI]
- Bruns B, "Hemoglobin drops within minutes of injuries and predicts need for an intervention to stop hemorrhage", J Trauma, 63(2), pp.312~315 (2007)[DOI]
- Thorson CM, "Admission hematocrit and transfusion requirements after trauma", J Am Coll Surg, 216(1), pp.65~73 (2013)[DOI]
- Koshkareva YA, "Utility of preoperative hematologic screening for pediatric adenotonsillectomy", Ear Nose Throat J, 91(8), pp.346~356 (2012)[DOI]
- Shumborski S, "Utility of preoperative blood screening before hip and knee arthroplasty", ANZ J Surg, 90(3), pp.350~354 (2020)[DOI]
- Haberle J, "Clinical and biochemical aspects of primary and secondary hyperammonemic disorders", Arch Biochem Biophys, 536(2), pp.101~108 (2013)[DOI]
- Al-Naamani K, "Does blood urea nitrogen level predict severity and high-risk endoscopic lesions in patients with nonvariceal upper gastrointestinal bleeding?", Can J Gastroenterol, 22(4), pp.399~403 (2008)[DOI]
- Tomizawa M, "Patient characteristics with high or low blood urea nitrogen in upper gastrointestinal bleeding", World J Gastroenterol, 21(24), pp.7500~7505 (2015)[DOI]
- Meyer A, "Machine learning for realtime prediction of complications in critical care : a retrospective study", Lancet Respir Med, 6(12), pp.905~914 (2018)[DOI]
- Bonde A, "Assessing the utility of deep neural networks in predicting postoperative surgical complications : a retrospective study", Lancet Digit Health, 3(8), pp.e471~e485 (2021)[DOI]
- Levi R, "Artificial intelligence-based prediction of transfusion in the intensive care unit in patients with gastrointestinal bleeding", BMJ Health Care Inform, 28(1), pp.e100245~ (2021)[DOI]
- Johnson AE, "MIMIC-III, a freely accessible critical care database", Sci Data, 3, pp.160035~ (2016)[DOI]
- Graves A, "Supervised sequence labelling with recurrent neural networks" (2012), Springer
- "Learning phrase representations using RNN encoder-decoder for statistical machine translation" (2014)[URL]
- Pollard TJ, "The eICU Collaborative Research Database, a freely available multi-center database for critical care research", Sci Data, 5, pp.180178~ (2018)[DOI]
- Zhu W, "The HASBLED Score for predicting major bleeding risk in anticoagulated patients with atrial fibrillation : a systematic review and meta-analysis", Clin Cardiol, 38(9), pp.555~561 (2015)[DOI]
- Bento D, "CRUSADE : is it still a good score to predict bleeding in acute coronary syndrome?", Rev Port Cardiol(Engl Ed), 37(11), pp.889~897 (2018)[DOI]
- Yildirim E, "Performance of HAS-BLED and CRUSADE risk scores for the prediction of haemorrhagic events in patients with stable coronary artery disease", Cardiovasc J Afr, 30(4), pp.198~202 (2019)[DOI]
- De Pasquale M, "Hemorrhage prediction models in surgical intensive care : bedside monitoring data adds information to lab values", IEEE J Biomed Health Inform, 21(6), pp.1703~1710 (2017)[DOI]