A5015491341- Machine Learning in Healthcare
- Hemodynamic Monitoring and Therapy
- Emergency and Acute Care Studies
- Healthcare Operations and Scheduling Optimization
- Cardiac, Anesthesia and Surgical Outcomes
- Blood Pressure and Hypertension Studies
- Topic Modeling
- Health Systems, Economic Evaluations, Quality of Life
- Healthcare Policy and Management
- Artificial Intelligence in Healthcare and Education
- Law, AI, and Intellectual Property
- Privacy-Preserving Technologies in Data
- Heart Failure Treatment and Management
- Generative Adversarial Networks and Image Synthesis
- Enhanced Recovery After Surgery
- Adversarial Robustness in Machine Learning
- Natural Language Processing Techniques
- ECG Monitoring and Analysis
- Artificial Intelligence in Law
- Delphi Technique in Research
- Traditional Chinese Medicine Studies
- Sepsis Diagnosis and Treatment
- Hospital Admissions and Outcomes
- Artificial Intelligence in Healthcare
- Explainable Artificial Intelligence (XAI)
Ulsan College
2023-2024
University of Ulsan
2023-2024
Asan Medical Center
2023-2024
Abstract Overcrowding of emergency departments is a global concern, leading to numerous negative consequences. This study aimed develop useful and inexpensive tool derived from electronic medical records that supports clinical decision-making can be easily utilized by department physicians. We presented machine learning models predicted the likelihood hospitalizations within 24 hours estimated waiting times. Moreover, we revealed enhanced performance these compared existing incorporating...
As warfarin has a narrow therapeutic window and obvious response variability among individuals, it is difficult to rapidly determine personalized dosage. Adverse drug events(ADE) resulting from overdose can be critical, so that typically physicians adjust the dosage through INR monitoring twice week when starting warfarin. Our study aimed develop machine learning (ML) models predicts discharge of as initial using clinical data derived electronic medical records within 2 days hospitalization....
<title>Abstract</title> The advent of the Transformer has significantly altered course research in Natural Language Processing (NLP) within thedomain deep learning, making Transformer-based studies mainstream subsequent NLP research. There alsobeen considerable advancement domain-specific research, including development specialized language modelsfor medical. These medical-specific models were trained on medical data and demonstrated high performance. Whilethese have treated field as a...
Predicting the length of stay in advance will not only benefit hospitals both clinically and financially but enable healthcare providers to better decision-making for improved quality care. More importantly, understanding severe patients who require general anesthesia is key enhancing health outcomes. Here, we aim discover how machine learning can support resource allocation management resulting from prediction. A retrospective cohort study was conducted January 2018 October 2020. total...
<title>Abstract</title> Background Predicting the length of stay in advance will not only benefit hospitals both clinically and financially but enable healthcare providers to better decision-making for improved quality care. More importantly, understanding severe patients who require general anesthesia is key enhancing health outcomes. Objective Here, we aim discover how machine learning can support resource allocation management resulting from prediction. Methods A retrospective cohort...
<title>Abstract</title> Predicting major adverse cardiovascular events (MACE) is crucial due to its high readmission rate and severe sequelae. Current risk scoring model of MACE are based on a few features patient status at single time point. We developed self-attention-based predict within 3 years from series data utilizing numerous in electronic medical records (EMRs). In addition, we demonstrated transfer learning for hospitals with insufficient through code mapping feature selection top...
Developing large-scale language models (LLMs) for health care requires fine-tuning with domain data suitable downstream tasks. However, LLMs medical can expose the training used during learning to adversarial attacks. This issue is particularly important as contain sensitive and identifiable patient data. The prompt-based attack approach was employed assess potential privacy breaches in LLMs. success rate of evaluated by categorizing 71 questions into three key metrics. To confirm exposure...
Although interest in predicting drug-drug interactions is growing, many predictions are not verified by real-world data. This study aimed to confirm whether predicted polypharmacy side effects using public data also occur from actual patients.
Background Predicting the bed occupancy rate (BOR) is essential for efficient hospital resource management, long-term budget planning, and patient care planning. Although macro-level BOR prediction entire crucial, predicting at a detailed level, such as specific wards rooms, more practical useful scheduling. Objective The aim of this study was to develop web-based support tool that allows administrators grasp each ward room according different time periods. Methods We trained time-series...
Loop diuretics are prevailing drugs to manage fluid overload in heart failure. However, adjusting loop diuretic doses is strenuous due the lack of a guideline. Accordingly, we developed novel clinician decision support system for dosage with Long Short-Term Memory (LSTM) algorithm using time-series EMRs. Weight measurements were used as target estimate loss during therapy. We designed TSFD-LSTM, bi-directional LSTM model an attention mechanism, forecast weight change 48 h after failure...
<sec> <title>BACKGROUND</title> Understanding the length of stay severe patients who require general anesthesia is key to enhancing health outcomes. </sec> <title>OBJECTIVE</title> Here, we aim discover how machine learning can support resource allocation management and decision-making resulting from prediction. <title>METHODS</title> A retrospective cohort study was conducted January 2018 October 2020. total 240,000 patients’ medical records collected. The data were collected exclusively...
<sec> <title>BACKGROUND</title> As warfarin has a narrow therapeutic window and obvious response variability among individuals, it is difficult to rapidly determine personalized dosage. Adverse drug events resulting from overdose can be critical. Our study aimed develop machine learning (ML) model that predicts the appropriate discharge dosage of using electronic medical records large hospital. Additionally, we externally validated ensure its accuracy. </sec> <title>OBJECTIVE</title> This...
<sec> <title>BACKGROUND</title> Transformer-based language models have shown great potential to revolutionize health care by advancing clinical decision support, patient interaction, and disease prediction. However, despite their rapid development, the implementation of transformer-based in settings remains limited. This is partly due lack a comprehensive review, which hinders systematic understanding applications limitations. Without clear guidelines consolidated information, both...
<sec> <title>BACKGROUND</title> Predicting the bed occupancy rate (BOR) is essential for efficient hospital resource management, long-term budget planning, and patient care planning. Although macro-level BOR prediction entire crucial, predicting at a detailed level, such as specific wards rooms, more practical useful scheduling. </sec> <title>OBJECTIVE</title> The aim of this study was to develop web-based support tool that allows administrators grasp each ward room according different time...