Hun-Sung Kim's research while affiliated with Catholic University of Korea and other places

... We found an overall achievement rate for the risk-based LDL-C goals of only 31.1%, despite all patients being on high-intensity statins, even with 62% of them on combination therapy of high-intensity statins plus ezetimibe. Regardless of differences in the recommended LDL-C targets, baseline CV risk of patients, or LLT used for CVD prevention, our findings are consistent with several international and Spanish studies that have consistently reported low rates of LDL-C goal a ainment in high-or very high-risk patients, ranging from 4% to 57% [6,[10][11][12][13][14][17][18][19][20][21][22][23][24][25][26]. It must be highlighted that most patients included in the TERESA study were categorized as very high CVD risk as they were recruited in cardiology and internal medicine units. ...

... For the linear models, we adjusted a range of hyperparameters and penalty functions that can be utilized in constructing different models, including logistic regression, lasso regression, ridge regression, and ElasticNet. For the tree-based models, we selected Extreme Gradient Boosting (XGBoost), which is widely recognized as one of the best-in-class algorithms for decision-tree-based models and has shown remarkable prediction performance in a wide range of studies [52][53][54][55][56][57] . Following ML best practices, the study data set was split into a modeling set that contains T2D patients diagnosed from 2015 to 2020, and an independent testing set that covers T2D patients diagnosed in 2021. ...

... patients. Also, UA, length of diabetes, and SBP were independent risk factors for EOD-associated nephropathy (20)(21)(22)(23). In young patients, the average duration to develop atherosclerotic cardiovascular disease (ASCVD) and arrhythmogenic cardiomyopathy (ACM) was 9.1 years in high-risk patients and 10.0 years in low-risk patients (23). ...

... The Korean National Health Insurance Service (NHIS) was established under the Ministry of Health and Welfare to provide mandatory health insurance services to about 97% of the Korean population [8]. This study used a customized database derived from the nationwide Korean NHIS. ...

... AI technology is used in various applications in the medical field, including early detection of complications, patient condition monitoring, clinical decision-making, and treatment. [12][13][14][15] In particular, AI-based medical image analysis technology is used for the diagnosis and prediction of various diabetes complications, and its accuracy is increasing through various machine learning and deep learning. 16,17 Despite the rapid development of AI models based on medical images in the field of diabetes, there are also problems with their clinical implementation in the real world. ...

... The ageing of the body becomes more pronounced as age increases, and the overall health and physical condition tend to deteriorate. The elderly are more likely to develop various diseases gradually, such as hypertension, diabetes, various cancers, and the resulting complications [1,2]. Population ageing is a global issue, and healthy ageing is one of the major directions for the research in gerontology in the world today. ...

... Computerised decision support systems (CDSS) are digital health technologies designed to analyse patient's clinical data to assist in clinical decision-making at the point of care [1]. CDSS can be knowledge based which involves predetermined rules based on literature, clinical practice or patients, or non-knowledge based which utilises advanced statistical pattern recognition, including artificial intelligence and machine learning, to produce an action or output from the data [2]. ...

... Furthermore, data quality is often unreliable, and in most cases, preprocessing is required [87,88], ultimately implying a critical privacy issue. To compensate for this, all information that can identify individuals should be removed from the accumulated data, and the extracted information should be stored in an encrypted file and made accessible only to designated persons [83,89]. ...

... They provide a structured approach to decision-making and knowledge representation [82], offering a systematic framework for integrating information from various modalities. By employing a set of predefined rules, these systems can effectively process and integrate data from multiple sources, enabling informed decisions and providing valuable recommendations [79]. The utilization of rule-based systems in multimodal medical data fusion enhances the overall knowledge generation process, aiding in accurate diagnoses, personalized treatment plans, and improved healthcare outcomes. ...

... For example, NLPs can write code that could be used to create research models or to collect data from the electronic medical records to create databases that can be updated periodically. 20 ML models are being used broadly in research, including in prediction models, improving pathology, and radiology reports, as mentioned earlier. ...