Article
Chemistry, Multidisciplinary
Lars Hempel, Sina Sadeghi, Toralf Kirsten
Summary: This paper presents predictive models for the length of stay (LOS) of ICU patients using machine learning and early available clinical data. The goal was to demonstrate a practical approach to predicting LOS and improve resource allocation and future planning. The results show significant improvements in the performance of the models for predicting actual LOS.
APPLIED SCIENCES-BASEL
(2023)
Article
Engineering, Multidisciplinary
Merhan A. Abd-Elrazek, Ahmed A. Eltahawi, Mohamed H. Abd Elaziz, Mohamed N. Abd-Elwhab
Summary: According to the World Health Organization (WHO), patient Length of Stay (LOS) in hospitals is an important performance measurement and monitoring indicator. Prolonged LOS in the Intensive Care Unit (ICU) may lead to consuming hospital resources, manpower, and equipment. The proposed framework for predicting patient LOS in the ICU using different machine learning (ML) techniques demonstrates high prediction accuracy and applicability across all patients.
AIN SHAMS ENGINEERING JOURNAL
(2021)
Article
Clinical Neurology
Guoxin Fan, Sheng Yang, Huaqing Liu, Ningze Xu, Yuyong Chen, Jie He, Xiuyun Su, Mao Pang, Bin Liu, Lanqing Han, Limin Rong
Summary: This study developed machine-learning classifiers to predict prolonged ICU-stay and prolonged hospital-stay for critical patients with spinal cord injury. The ensemble classifiers showed high potential in assisting physicians in managing SCI patients in ICU and optimizing the use of medical resources.
Article
Cardiac & Cardiovascular Systems
Qiuying Chen, Bin Zhang, Jue Yang, Xiaokai Mo, Lu Zhang, Minmin Li, Zhuozhi Chen, Jin Fang, Fei Wang, Wenhui Huang, Ruixin Fan, Shuixing Zhang
Summary: This study developed and validated machine learning models for predicting ICU length of stay after surgery for acute type A aortic dissection. By analyzing 12 predictive factors, the Random Forest model showed the best performance in accurately predicting the duration of ICU stay.
FRONTIERS IN CARDIOVASCULAR MEDICINE
(2021)
Article
Multidisciplinary Sciences
Stephan Sloth Lorenzen, Mads Nielsen, Espen Jimenez-Solem, Tonny Studsgaard Petersen, Anders Perner, Hans-Christian Thorsen-Meyer, Christian Igel, Martin Sillesen
Summary: The study shows that machine learning can be used to predict ICU requirements in the future days, with models performing well in predicting ICU admission and mechanical ventilation use risks within 5 days.
SCIENTIFIC REPORTS
(2021)
Article
Cardiac & Cardiovascular Systems
Kan Wang, Li Zhao Yan, Wang Zi Li, Chen Jiang, Ni Ni Wang, Qiang Zheng, Nian Guo Dong, Jia Wei Shi
Summary: By collecting clinical data from 365 heart transplantation patients, utilizing machine learning models for predicting ICU length of stay, results showed that the eXtreme Gradient Boosting (XGBoost) algorithm outperformed other models.
FRONTIERS IN CARDIOVASCULAR MEDICINE
(2022)
Article
Computer Science, Information Systems
Subhrajit Roy, Diana Mincu, Eric Loreaux, Anne Mottram, Ivan Protsyuk, Natalie Harris, Yuan Xue, Jessica Schrouff, Hugh Montgomery, Alistair Connell, Nenad Tomasev, Alan Karthikesalingam, Martin Seneviratne
Summary: The SeqSNR architecture demonstrated a modest yet statistically significant performance boost across 4 of the 6 tasks compared to single-task and naive multitasking approaches. When reducing the size of the training dataset for specific tasks, SeqSNR outperformed single-task in all cases, indicating superior label efficiency especially in scenarios where endpoint labels are difficult to ascertain.
JOURNAL OF THE AMERICAN MEDICAL INFORMATICS ASSOCIATION
(2021)
Article
Critical Care Medicine
Asif Rahman, Yale Chang, Junzi Dong, Bryan Conroy, Annamalai Natarajan, Takahiro Kinoshita, Francesco Vicario, Joseph Frassica, Minnan Xu-Wilson
Summary: The Hemodynamic Stability Index (HSI) developed in this study can accurately predict future hemodynamic interventions in critically ill patients, with a high specificity of 92% and a lead time of 1 hour. The HSI algorithm provides a real-time summary of hemodynamic status using various physiological parameters, demonstrating generalizability and strong performance in different data availability conditions.
Article
Computer Science, Hardware & Architecture
Ke Niu, Su Pei, Xueping Peng, Jingni Zeng, Ke Zhang
Summary: Prediction for Intensive Care Unit (ICU) readmission is important for treatment decision making and reducing relapse risk. This paper proposes a correlation enhanced Multi-Task learning approach with Pearson and RNN-based Neural Ordinary Differential Equations Model (MP-ROM). By constructing a Shared-Bottom structure and using Pearson correlation calculation, the method improves the predictive performance of ICU readmission risk. Experimental results on MIMIC-III dataset demonstrate the effectiveness of MP-ROM.
COMPUTERS & ELECTRICAL ENGINEERING
(2023)
Article
Clinical Neurology
Moshgan Amiri, Patrick M. Fisher, Federico Raimondo, Annette Sidaros, Melita Cacic Hribljan, Marwan H. Othman, Ivan Zibrandtsen, Simon A. Albrechtsen, Ove Bergdal, Adam Espe Hansen, Christian Hassager, Joan Lilja S. Hojgaard, Elisabeth Waldemar Jakobsen, Helene Ravnholt Jensen, Jacob Moller, Vardan Nersesjan, Miki Nikolic, Markus Harboe Olsen, Sigurdur Thor Sigurdsson, Jacobo D. Sitt, Christine Solling, Karen Lise Welling, Lisette M. Willumsen, John Hauerberg, Vibeke Andree Larsen, Martin Ejler Fabricius, Gitte Moos Knudsen, Jesper Kjaergaard, Kirsten Moller, Daniel Kondziella
Summary: This study aimed to assess the accuracy of fMRI and EEG in identifying residual consciousness in patients with acute disorders of consciousness (DoC) in the intensive care unit (ICU). The study found that a combination of fMRI and EEG features can predict the level of consciousness in patients, which is important for clinical decision-making.
Article
Medicine, General & Internal
Kyongsik Yun, Jihoon Oh, Tae Ho Hong, Eun Young Kim
Summary: The study utilized machine learning algorithms to successfully predict in-hospital death of critically ill patients, identifying serum albumin concentration, total prenatal nutritional intake, and peak dose of dopamine drug as key factors in mortality prediction. The decision tree method showed higher classification results compared to the neural network classifier.
FRONTIERS IN MEDICINE
(2021)
Article
Biochemistry & Molecular Biology
Animesh Acharjee, Jon Hazeldine, Alina Bazarova, Lavanya Deenadayalu, Jinkang Zhang, Conor Bentley, Dominic Russ, Janet M. Lord, Georgios V. Gkoutos, Stephen P. Young, Mark A. Foster
Summary: Recent advances in emergency medicine and trauma care have increased the survival rates of critically-injured patients. However, this has also put a strain on the healthcare system. This study investigated the serum metabolomic profile of major trauma patients and found that combining metabolomic data with clinical scoring systems can accurately identify patients with extended ICU stays.
Article
Medicine, General & Internal
Longxiang Su, Zheng Xu, Fengxiang Chang, Yingying Ma, Shengjun Liu, Huizhen Jiang, Hao Wang, Dongkai Li, Huan Chen, Xiang Zhou, Na Hong, Weiguo Zhu, Yun Long
Summary: The study demonstrated that using machine learning models based on data from the first 6 hours in the ICU can effectively predict mortality, severity, and length of ICU stay for sepsis patients. The random forest classifier showed the best performance among the three models, providing comprehensive early warning for sepsis in ICU.
FRONTIERS IN MEDICINE
(2021)
Article
Anesthesiology
Man-Ling Wang, Yu-Ting Kuo, Lu-Cheng Kuo, Hsin-Ping Liang, Yi-Wei Cheng, Yu-Chen Yeh, Ming-Tao Tsai, Wing-Sum Chan, Ching-Tang Chiu, Anne Chao, Nai-Kuan Chou, Yu-Chang Yeh, Shih-Chi Ku
Summary: The study aimed to develop, validate, and deploy models for predicting delirium in critically ill adult patients as early as ICU admission. A retrospective cohort study was conducted using data from 6238 critically ill patients in a university teaching hospital in Taipei, Taiwan. The models used logistic regression (LR), gradient boosted trees (GBT), and deep learning (DL) algorithms to predict delirium upon ICU admission (ADM) and at 24h (24H) after ICU admission, and the performance of the models was compared.
JOURNAL OF CLINICAL ANESTHESIA
(2023)
Article
Pediatrics
Sanjukta N. Bose, Joseph L. Greenstein, James C. Fackler, Sridevi V. Sarma, Raimond L. Winslow, Melania M. Bembea
Summary: The study aims to build early prediction models for pediatric PICU patients at risk of developing MOD. Using machine learning methods, the models achieved high accuracy in predicting MOD onset with over 22 hours of lead time. Spectral clustering on risk-score trajectories identified a high-risk group with a high positive predictive value.
FRONTIERS IN PEDIATRICS
(2021)
Article
Operations Research & Management Science
Victor Abu-Marrul, Rafael Martinelli, Silvio Hamacher, Irina Gribkovskaia
Summary: This paper addresses a parallel machine scheduling problem with non-anticipatory family setup times and batching. It proposes an Iterated Greedy simheuristic with built-in Monte Carlo Simulation to handle the stochastic parameters. Experimental results show that the proposed simheuristic outperforms other algorithms in terms of both objective values and computational times.
ANNALS OF OPERATIONS RESEARCH
(2023)
Article
Neurosciences
Carolina A. Moraes, Eugenio D. Hottz, Debora Dos Santos Ornellas, Daniel Adesse, Carolina T. de Azevedo, Joana C. D'Avila, Camila Zaverucha-do-Valle, Tatiana Maron-Gutierrez, Helene Santos Barbosa, Patricia Torres Bozza, Fernando Augusto Bozza
Summary: This study investigated the role of the NLRP3 inflammasome in microglial activation and synaptic loss related to sepsis. The results showed increased expression of NLRP3 inflammasome in the brains of septic mice, specifically in microglial cells. In vitro experiments also revealed that LPS-stimulated microglia produced mitochondrial reactive oxygen species, recruited NLRP3 complex, and released IL-1β. Inhibition of NLRP3, caspase-1, and mitochondrial ROS decreased IL-1β secretion by microglial cells.
MOLECULAR NEUROBIOLOGY
(2023)
Article
Critical Care Medicine
Bruno Goncalves, Carla Rynkowski, Ricardo Turon, Nestor Charris, Fabio Miranda, Vanessa de Caro, Marco Prazeres, Thayana Santos, David M. Greer, Tarek Sharshar, Turc Guillaume, Fernando A. Bozza, Cassia Righy, Pedro Kurtz
Summary: This study aimed to evaluate the clinical characteristics and treatment practices of patients with aneurysmal subarachnoid hemorrhage (SAH) in Brazil, a middle-income country, and assess their long-term outcomes. The study found significant differences in SAH management in Brazil compared to other countries, and suggested that earlier aneurysm occlusion and increased use of endovascular therapy could potentially improve functional outcomes in SAH patients.
NEUROCRITICAL CARE
(2023)
Article
Critical Care Medicine
Pedro Kurtz, Leonardo S. L. Bastos, Fernando G. Zampieri, Gabriel R. de Freitas, Fernando A. Bozza, Marcio Soares, Jorge I. F. Salluh
Summary: This retrospective cohort study analyzed data from 165 ICUs in Brazil from 2011 to 2020. The study found that hospital outcomes of stroke admissions worsened during the COVID-19 pandemic, particularly among patients with ischemic stroke and young patients with hemorrhagic stroke.
Review
Critical Care Medicine
Pedro Povoa, Luis Coelho, Felipe Dal-Pizzol, Ricard Ferrer, Angela Huttner, Andrew Conway Morris, Vandack Nobre, Paula Ramirez, Anahita Rouze, Jorge Salluh, Mervyn Singer, Daniel A. Sweeney, Antoni Torres, Grant Waterer, Andre C. Kalil
Summary: Sepsis is a life-threatening organ dysfunction caused by a dysregulated host response to infection. Biomarkers can be used as indicators for infection, dysregulated host response, treatment response, and can aid in prognosticating patient risk. Over 250 biomarkers have been identified and evaluated, but none can accurately differentiate between sepsis and sepsis-like syndrome.
INTENSIVE CARE MEDICINE
(2023)
Article
Infectious Diseases
Bianca B. P. Antunes, Amanda A. B. Silva, Patricia H. C. Nunes, Ignacio Martin-Loeches, Pedro Kurtz, Silvio Hamacher, Fernando A. Bozza
Summary: The objective of this study was to compare the differences in antimicrobial use between COVID-19 and non-COVID-19 patients, and to compare two commonly used metrics for antimicrobial use: Defined Daily Dose (DDD) and Days of Therapy (DOT). The study found that COVID-19 patients had a higher use of broad-spectrum antimicrobials in the ICU, and overall, the DDD metric overestimated antimicrobial use compared with the DOT metric.
JOURNAL OF ANTIMICROBIAL CHEMOTHERAPY
(2023)
Article
Critical Care Medicine
Carla B. Rynkowski, Vanessa Hegele, Pedro Henrique Rigotti Soares, Monica Lopes Tonello, Leticia Peterson, Frederico Klein Gomes, Alejandro A. Rabinstein, Leonardo S. L. Bastos, Ricardo Turon, Bruno Goncalves, Cassia Righy, Fernando A. Bozza, Pedro Kurtz
Summary: This study investigated the effects of tranexamic acid on long-term functional outcomes of patients with aneurysmal subarachnoid hemorrhage (aSAH). The results showed that the use of tranexamic acid before aneurysm occlusion did not improve the functional outcomes of aSAH patients.
NEUROCRITICAL CARE
(2023)
Article
Urology & Nephrology
Henrique Palomba, Daniel Cubos, Fernando Bozza, Fernando Godinho Zampieri, Thiago Gomes Romano
Summary: The study aimed to develop a prognostic score for predicting the development of Acute Kidney Injury (AKI) in COVID-19 patients. A retrospective observational study of 2334 COVID-19 patients admitted to 23 different hospitals in Brazil was conducted, and common clinical variables were found to accurately predict the development of AKI in these patients.
Article
Immunology
Andre C. Ferreira, Carolina Q. Sacramento, Filipe S. Pereira-Dutra, Natalia Fintelman-Rodrigues, Priscila P. Silva, Mayara Mattos, Caroline S. de Freitas, Andressa Marttorelli, Gabrielle R. de Melo, Mariana M. Campos, Isaclaudia G. Azevedo-Quintanilha, Aluana S. Carlos, Joao Vitor Emidio, Cristiana C. Garcia, Patricia T. Bozza, Fernando A. Bozza, Thiago M. L. Souza
Summary: This study found that influenza virus infection can induce inflammatory programmed cell death in macrophages, and inhibiting inflammatory cell death can reduce inflammation and lung injury. This study identified an additional mechanism involved in severe influenza and proposed a clinically available method for treatment.
FRONTIERS IN CELLULAR AND INFECTION MICROBIOLOGY
(2023)
Review
Public, Environmental & Occupational Health
Horacio N. Hastenreiter Filho, Igor T. Peres, Lucas G. Maddalena, Fernanda A. Baiao, Otavio T. Ranzani, Silvio Hamacher, Paula M. Macaira, Fernando A. Bozza
Summary: This article provides a narrative review of COVID-19 vaccination impact studies, discussing their characteristics and analyzing their similarities and differences. The review includes 18 studies evaluating the vaccine impact in different countries.
FRONTIERS IN PUBLIC HEALTH
(2023)
Article
Multidisciplinary Sciences
Alvaro Rea-Neto, Rafaella Stradiotto Bernardelli, Mirella Cristine de Oliveira, Paula Geraldes David-Joao, Amanda Christina Kozesinski-Nakatani, Antonio Luis Eiras Falcao, Pedro Martins Pereira Kurtz, Helio Afonso Ghizoni Teive, Neurocritical Brazil Study Grp
Summary: Acute neurological emergencies are common in Brazilian ICUs and have a significant impact on patients. This study found that older age, emergency admission, number of secondary injuries, and severity of illness scores are predictors of mortality and poor outcomes in patients with acute neurological diagnoses.
SCIENTIFIC REPORTS
(2023)
Article
Critical Care Medicine
Hugo Boechat Andrade, Ivan Rocha Ferreira da Silva, Rodolfo Espinoza, Mayara Secco Torres da Silva, Pedro Henrique Nascimento Theodoro, Marcel Treptow Ferreira, Jesus Soares, Ermias D. Belay, James J. Sejvar, Fernando Augusto Bozza, Jose Cerbino-Neto, Andre Miguel Japiassu
Summary: This study compared the performance of specialized infectious diseases ICUs with other ICUs in the treatment of community-acquired central nervous system infections (CNSI), and found that CNSI patients had lower mortality and resource use rates when treated in specialized infectious diseases ICUs.
JOURNAL OF INTENSIVE CARE MEDICINE
(2023)
Article
Medicine, General & Internal
Maria Carolina Paulino, Catarina Conceicao, Joana Silvestre, Maria Ines Lopes, Hernani Goncalves, Claudia Camila Dias, Rodrigo Serafim, Jorge I. F. Salluh, Pedro Povoa
Summary: Subsyndromal delirium (SSD) in the Intensive Care Unit (ICU) is associated with increased morbidity and unknown post-discharge functional and cognitive outcomes. A multicenter study found that early SSD patients experienced a decline in cognitive abilities at 3 months after discharge, but showed improvement at 6 months. Therefore, preventing and identifying SSD during ICU stays is crucial.
JOURNAL OF CLINICAL MEDICINE
(2023)
Article
Biochemistry & Molecular Biology
Marcos C. C. Gama-Almeida, Gabriela D. A. Pinto, Livia Teixeira, Eugenio D. D. Hottz, Paula Ivens, Hygor Ribeiro, Rafael Garrett, Alexandre G. G. Torres, Talita I. A. Carneiro, Bianca de O. Barbalho, Christian Ludwig, Claudio J. J. Struchiner, Iranaia Assuncao-Miranda, Ana Paula C. Valente, Fernando A. A. Bozza, Patricia T. Bozza, Gilson C. C. dos Santos Jr, Tatiana El-Bacha
Summary: Brazil has the second-highest COVID-19 death rate worldwide, with Rio de Janeiro being one of the states with the highest rate in the country. Despite vaccine coverage, COVID-19 is expected to become an endemic disease. The molecular mechanisms underlying the progression of the disease and the development of long COVID-19 are still not fully understood.
Article
Anesthesiology
Maria C. Borrelli, Andrew J. Sprowell, Anna Moldysz, Mohammed Idris, Samantha L. Armstrong, John J. Kowalczyk, Yunping Li, Philip E. Hess
Summary: This study investigates the relationship between intrathecal morphine and enhanced recovery pathways after cesarean delivery. The results indicate that lower doses provide shorter duration of analgesia and higher pain scores. In contrast, a dose of 250 mcg offers longer-lasting pain relief with similar side effects. Both 150 mcg and 250 mcg doses allow a majority of patients to avoid additional opioid use within 72 hours.
ANAESTHESIA CRITICAL CARE & PAIN MEDICINE
(2024)
Article
Anesthesiology
Jose Sanchez, Rohan Prabhu, Jean Guglielminotti, Ruth Landau
Summary: This study evaluates the incidence of self-reported pain and the administration of intravenous medication (IVM) during cesarean delivery. The results show that 11.5% of women reported pain during the procedure, and a proportion of them received analgesic IVM. Risk factors for pain included substance use disorder and intrapartum epidural extension.
ANAESTHESIA CRITICAL CARE & PAIN MEDICINE
(2024)
Letter
Anesthesiology
Kei Ugata, Katsushi Doi, Noritaka Imamachi, Keita Matsumoto, Hiroyuki Kushizaki
ANAESTHESIA CRITICAL CARE & PAIN MEDICINE
(2024)
