WLWon Kyung Lee

Won Kyung Lee

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Won Kyung Lee has a total of 163 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Causal Inference in the Presence of Missing Outcome and Treatment Variables: Triply Robust Estimator and Sensitivity Analysis

    by Hyunman Sim, Won Kyung Lee, Christoph Lange, Woojoo Lee

    Estimating causal effects from observational studies with missing data typically requires the no unmeasured confounder (NUC) assumption and the missing at random (MAR) assumption. However, correctly specifying all relevant models is often challenging in practice, and model misspecification can lead to biased estimates. In this study, we develop a triply robust estimator for causal effects when the outcome or treatment variable is partially observed. The proposed estimator incorporates outcome, treatment, and missing‐data models, and remains consistent as long as at least two of the three models are correctly specified. Although this approach provides robustness against model misspecification, its validity still depends on the NUC and MAR assumptions, which are untestable from the observed data alone and may be frequently violated in the presence of unmeasured confounding. To address this limitation, we introduce a novel sensitivity analysis framework to evaluate the potential impact of unmeasured confounding on causal effect estimates and demonstrate its practical usefulness through an application to real data.

  • Prioritization of Injury Prevention and Management Programs and Research and Development (R&D) Projects: Survey Using the Delphi Technique and Analytic Hierarchy Process

    by Won Kyung Lee, Minsu Ock, Ju Ok Park, Changsoo Kim, Beom Sok Seo, Jeehee Pyo, Hyun Jin Park, Ui Jeong Kim, Eun Jeong Choi, Shinyoung Woo, et al.

    We prioritized injury prevention and management programs, and research and development (R&D) projects using the Delphi method and analytic hierarchy process (AHP) for national plan. We summarized 21 programs and 31 R&D projects based on opinions collected in the opinion survey and brainstorming. Expert panel evaluated program and project relevance in the two Delphi rounds to make consensus. Total scores were calculated by adding criterion weights calculated through the AHP multiplied by the average of scores for each program and project. The top three programs were “Integrated Management System for Injury Data,” “Comprehensive Injury Surveillance System,” and “Transport System Improvement Program for Major Trauma.” The top three R&D projects were researches on “Identification of Vulnerable Workers Injuries to Strengthen Protections,” “Data Integration on Injury and Disability,” and “Using Public Health Big Data to Identify Injury Status and Risk Factors.” Experts highly rated the programs to improve injury surveillance system.


  • Open Access

    Factors affecting patients who attempted suicide in the emergency department due to the prolonged pandemic of COVID-19

    by Hyunji Kim, Areum Durey, Soo Kang, Won Kyung Lee, Ji Hye Kim, Seung Baik Han, Yu Jin Lee

    Objective This study examined the characteristics of patients who attempted suicide in the emergency department before and during the COVID-19 pandemic.Methods We compared data from patients in the emergency department following suicide attempts between January 2018 and December 2021. The patients were categorized into two groups: “pre–COVID-19” and “during COVID-19” pandemic.Results The findings revealed an increasing trend of suicide attempts during the study period. Suicide attempts were reported at 1,107 before the COVID-19 pandemic and 1,356 during the COVID-19 pandemic. Patients who attempted suicide during the COVID-19 pandemic were younger (38.0±18.5 years vs. 40.7±18.4 years, P<0.01), had a smaller proportion of men (36% vs. 44%, P<0.01), and had fewer medical comorbidities (20.2% vs. 23.6%, P<0.05). The group during the COVID-19 pandemic reported better hygiene conditions (50.5% vs. 40.8%, P<0.01) and lower alcohol consumption (27.7% vs. 37.6%, P<0.01). Patients who attempted suicide during the COVID-19 pandemic had higher rates of use of psychiatric medications and previous suicide attempts. The most common reasons for the suicide attempt were unstable psychiatric disorders (38.8%), poor interpersonal relationships (20.5%), and economic difficulties (14.0%). Drug poisoning (44.1%) was the most common method of suicide attempts. Subgroup analysis with patients who attributed their suicide attempts to COVID-19 revealed a higher level of education (30.8%) and employment status (69.2%), with economic difficulties (61.6%) being the primary cause of suicide attempts.Conclusion These findings suggest that the prolonged duration of the COVID-19 pandemic and its effects on social and economic factors have influenced suicide attempts.


  • Open Access

    Machine learning approaches that use clinical, laboratory, and electrocardiogram data enhance the prediction of obstructive coronary artery disease

    by Hyun-Gyu Lee, Sang-Don Park, Jang-Whan Bae, SungJoon Moon, Chai Young Jung, Mi-Sook Kim, Tae-Hun Kim, Won Kyung Lee

    Pretest probability (PTP) for assessing obstructive coronary artery disease (ObCAD) was updated to reduce overestimation. However, standard laboratory findings and electrocardiogram (ECG) raw data as first-line tests have not been evaluated for integration into the PTP estimation. Therefore, this study developed an ensemble model by adopting machine learning (ML) and deep learning (DL) algorithms with clinical, laboratory, and ECG data for the assessment of ObCAD. Data were extracted from the electronic medical records of patients with suspected ObCAD who underwent coronary angiography. With the ML algorithm, 27 clinical and laboratory data were included to identify ObCAD, whereas ECG waveform data were utilized with the DL algorithm. The ensemble method combined the clinical-laboratory and ECG models. We included 7907 patients between 2008 and 2020. The clinical and laboratory model showed an area under the curve (AUC) of 0.747; the ECG model had an AUC of 0.685. The ensemble model demonstrated the highest AUC of 0.767. The sensitivity, specificity, and F1 score of the ensemble model ObCAD were 0.761, 0.625, and 0.696, respectively. It demonstrated good performance and superior prediction over traditional PTP models. This may facilitate personalized decisions for ObCAD assessment and reduce PTP overestimation.


  • Open Access

    Electrocardiogram-based deep learning algorithm for the screening of obstructive coronary artery disease

    by Seong Huan Choi, Hyun-Gye Lee, Sang-Don Park, Jang-Whan Bae, Woojoo Lee, Mi-Sook Kim, Tae-Hun Kim, Won Kyung Lee

    Background: Information on electrocardiogram (ECG) has not been quantified in obstructive coronary artery disease (ObCAD), despite the deep learning (DL) algorithm being proposed as an effective diagnostic tool for acute myocardial infarction (AMI). Therefore, this study adopted a DL algorithm to suggest the screening of ObCAD from ECG. Methods: ECG voltage-time traces within a week from coronary angiography (CAG) were extracted for the patients who received CAG for suspected CAD in a single tertiary hospital from 2008 to 2020. After separating the AMI group, those were classified into ObCAD and non-ObCAD groups based on the CAG results. A DL-based model adopting ResNet was built to extract information from ECG data in the patients with ObCAD relative to those with non-ObCAD, and compared the performance with AMI. Moreover, subgroup analysis was conducted using ECG patterns of computer-assisted ECG interpretation. Results: The DL model demonstrated modest performance in suggesting the probability of ObCAD but excellent performance in detecting AMI. The AUC of the ObCAD model adopting 1D ResNet was 0.693 and 0.923 in detecting AMI. The accuracy, sensitivity, specificity, and F1 score of the DL model for screening ObCAD were 0.638, 0.639, 0.636, and 0.634, respectively, while the figures were up to 0.885, 0.769, 0.921, and 0.758 for detecting AMI, respectively. Subgroup analysis showed that the difference between normal and abnormal/borderline ECG groups was not notable. Conclusions: ECG-based DL model showed fair performance for assessing ObCAD and it may serve as an adjunct to the pre-test probability in patients with suspected ObCAD during the initial evaluation. With further refinement and evaluation, ECG coupled with the DL algorithm may provide potential front-line screening support in the resource-intensive diagnostic pathways.


  • Open Access

    Health-related quality of life using WHODAS 2.0 and associated factors 1 year after stroke in Korea: a multi-centre and cross-sectional study

    by Hey Jean Lee, Jung-Kook Song, Jiyoung Moon, Keonyeop Kim, Hyeung-Keun Park, Gil-Won Kang, Jun-Ho Shin, Jongsoo Kang, Byoung-Gwon Kim, Young-Hoon Lee, et al.

    Background: Little is known about the self-perceived level of disability of stroke survivors in the community. We aimed to characterise Health-related quality of life (HRQoL) 1 year after stroke and investigate how sociodemographic and stroke-related factors and medical adherence explain the self-perceived level of disability in a Korean stroke population. Methods: This was a multicentre cross-sectional study. A total of 382 ischaemic stroke survivors at 1 year after onset from 11 university hospitals underwent a one-session assessment, including socioeconomic variables, the modified Rankin Scale (mRS), various neurological sequelae, the Morisky, Green and Levin-Medication Adherence Questionnaire (MGL), and the World Health Organization Disability Assessment Schedule 2.0 (WHODAS 2.0) 36-items. The relationship between disability and different variables was analysed using ordinal logistic regression. Results: The prevalence of disability based on global WHODAS 2.0 was 62.6% (mild, 41.6%; moderate, 16.0%; severe, 5.0%). The prevalence of severe disability was higher in participation in society (16.8%) and getting around (11.8%) than in other domains. Low MGL- motivation was the only factor determining a significant association between all six domains of disability after adjustment. Different predictors for specific domains were age, mRS, dysarthria, trouble seeing, cognition problems, and MGL-motivation for understanding and communicating; age, recurrent stroke, mRS, hemiplegia, facial palsy, general weakness, and MGL-motivation for getting around; age, education, mRS, hemiplegia, and MGL-motivation for self-care; education, recurrent stroke, hemiplegia, dysarthria, and MGL-motivation for getting along with people; age, education, income, mRS, hemiplegia, dysarthria, MGL-knowledge, and MGL-motivation for life activities; living without a spouse, mRS, hemiplegia, dysarthria, trouble seeing, cognition problems, general weakness, and MGL-motivation for participation in society. Conclusions: Self-perceived disability according to the WHODAS 2.0 at 1 year after stroke was highly prevalent. Each disability domain showed a different prevalence and associated factors. Interventions promoting medical adherence to motivation seemed to help achieve high HRQoL in all domains.


  • Open Access

    Comparison of Factors Associated With Direct Versus Transferred-in Admission to Government-Designated Regional Centers Between Acute Ischemic Stroke and Myocardial Infarction in Korea

    by Dae-Hyun Kim, Seok-Joo Moon, Juneyoung Lee, Jae-Kwan Cha, Moo Hyun Kim, Jong-Sung Park, Byeolnim Ban, Jihoon Kang, Beom Joon Kim, Won-Seok Kim, et al.

    Kim DH, et al. J Korean Med Sci. 2022 Oct;37(42):e305. https://doi.org/10.3346/jkms.2022.37.e305


  • Forecasting of non-accidental, cardiovascular, and respiratory mortality with environmental exposures adopting machine learning approaches

    by Woojoo Lee, Youn-Hee Lim, Eunhee Ha, Yoenjin Kim, Won Kyung Lee

    Environmental exposure constantly changes with time and various interactions that can affect health outcomes. Machine learning (ML) or deep learning (DL) algorithms have been used to solve complex problems, such as multiple exposures and their interactions. This study developed predictive models for cause-specific mortality using ML and DL algorithms with the daily or hourly measured meteorological and air pollution data. The ML algorithm improved the performance compared to the conventional methods, even though the optimal algorithm depended on the adverse health outcomes. The best algorithms were extreme gradient boosting, ridge, and elastic net, respectively, for non-accidental, cardiovascular, and respiratory mortality with daily measurement; they were superior to the generalized additive model reducing a mean absolute error by 4.7%, 4.9%, and 16.8%, respectively. With hourly measurements, the ML model tended to outperform the conventional models, even though hourly data, instead of daily data, did not enhance the performance in some models. The proposed model allows a better understanding and development of robust predictive models for health outcomes using multiple environmental exposures.


  • Open Access

    Joinpoint Regression About Injury Mortality and Hospitalization in Korea

    by Hyun Jin Park, Ui Jeong Kim, Won Kyung Lee, Bohyun Park, Yoonhee Shin, Seonhwa Lee, Eun Jeong Choi, Nam-Eun Kim, Ju Ok Park, Hyesook Park

    Park HJ, et al. J Korean Med Sci. 2021 Dec;37(3):e10. https://doi.org/10.3346/jkms.2022.37.e10