Purpose: Heart failure (HF) and atrial fibrillation (AF) frequently coexist, with over 50% patients with HF having AF, while onethird of those with AF develop HF. Differences in obesity-mediated association between HF and HF-related AF among Asians and Europeans were evaluated. Materials and Methods: Using the Korean National Health Insurance Service-Health Screening (K-NHIS-HealS) cohort and the UK Biobank, we included 394801 Korean and 476883 UK adults, respectively aged 40–70 years. The incidence and risk of HF were evaluated based on body mass index (BMI). Results: The proportion of obese individuals was significantly higher in the UK Biobank cohort than in the K-NHIS-HealS cohort (24.2% vs. 2.7%, p<0.001). The incidence of HF and HF-related AF was higher among the obese in the UK than in Korea. The risk of HF was higher among the British than in Koreans, with adjusted hazard ratios of 1.82 [95% confidence interval (CI), 1.30–2.55] in KNHIS-HealS and 2.00 (95% CI, 1.69–2.37) in UK Biobank in obese participants (p for interaction <0.001). A 5-unit increase in BMI was associated with a 44% greater risk of HF-related AF in the UK Biobank cohort (p<0.001) but not in the K-NHIS-HealS cohort (p=0.277). Conclusion Obesity was associated with an increased risk of HF and HF-related AF in both Korean and UK populations. The higher incidence in the UK population was likely due to the higher proportion of obese individuals.

Background: Autoencoder (AE) is one of the deep learning techniques that uses an artificial neural network to reconstruct its input data in the output layer. We constructed a novel supervised AE model and tested its performance in the prediction of a co-existence of the disease of interest only using diagnostic codes. Methods: Diagnostic codes of one million randomly sampled patients listed in the Korean National Health Information Database in 2019 were used to train, validate, and test the prediction model. The first used AE solely for a feature engineering tool for an input of a classifier. Supervised Multi-Layer Perceptron (sMLP) was added to train a classifier to predict a binary level with latent representation as an input (AE + sMLP). The second model simultaneously updated the parameters in the AE and the connected MLP classifier during the learning process (End-to-End Supervised AE [EEsAE]). We tested the performances of these two models against baseline models, eXtreme Gradient Boosting (XGB) and naïve Bayes, in the prediction of co-existing gastric cancer diagnosis. Results: The proposed EEsAE model yielded the highest F1-score and highest area under the curve (0.86). The EEsAE and AE + sMLP gave the highest recalls. XGB yielded the highest precision. Ablation study revealed that iron deficiency anemia, gastroesophageal reflux disease, essential hypertension, gastric ulcers, benign prostate hyperplasia, and shoulder lesion were the top 6 most influential diagnoses on performance. Conclusion A novel EEsAE model showed promising performance in the prediction of a disease of interest.

Background: Food allergy (FA) can have a profound effect on quality of life (QoL), stress, and anxiety in the family. We aimed to validate the Korean version of the Food Allergy Quality of Life-Parental Burden (FAQL-PB) and identify factors related to the parental psychosocial burden of caring for children with FAs. Methods: Parents of children aged between 6 months and 17 years with immunoglobulin E (IgE)-mediated FAs from the Pediatric Allergy Department of five university hospitals in Korea were enrolled in the study. Parents were asked to complete the FAQL-PB, Food Allergy Independent Measure-Parent Form (FAIM-PF), Child Health Questionnaire-Parents Form 28 (CHQ-PF28), Beck’s Anxiety Inventory, Connor-Davidson Resilience Scale, and Patient Health Questionnaire-9 for depression. Statistical analyses included internal consistency, test-retest reliability, concurrent validity, discriminative validity, and logistic regression analyses. Results: A total of 190 parents were enrolled. Social activity limitation was the item with the highest FAQL-PB scores. The Cronbach’s α for each item was higher than 0.8. The test-retest reliability was good (intra-class correlation coefficient, 0.716; 95% confidence interval [CI], 0.100–0.935). An increase in the FAQL-PB was significantly associated with an increase in the FAIM-PF (β = 0.765, P < 0.001) (concurrent validity). There was a positive correlation between parental burden, anxiety, and depression, while resilience was inversely correlated with parental burden (all P < 0.001). The total FAQL-PB score in parents of children who had experienced anaphylaxis was significantly higher than that in parents of children who did not experience it (P = 0.008). When adjusting for age, sex, and underlying diseases, anaphylaxis β = 9.32; 95% CI, 2.97 to 15.68), cow’s milk (CM) allergy (β = 8.24; 95% CI, 2.04 to 14.44), soybean allergy (β = 13.91; 95% CI, 1.62 to 26.20), higher anxiety (β = 1.05; 95% CI, 0.07 to 1.41), higher depression (β = 2.15; 95% CI, 1.61 to 2.69), and lower resilience (β = −0.42; 95% CI, −0.61 to −0.2) were significantly associated with greater parental burden in children with IgE-mediated FAs. Conclusion FAQL-PB is a reliable and valid tool for use in Korea. Anaphylaxis, CM or soybean allergies, more anxiety and depression symptoms, and lower resilience are associated with poorer QoL in parents of children with FAs.

The integration of medical devices with artificial intelligence (AI) software is rapidly advancing as technology progresses. AI machine learning can be used in commercial medical services to generate practical data; there is evidence that it can be integrated into newly developed devices. However, such devices must undergo approval, regulation, and supervision. The Food and Drug Administration approves regulations for numerous machine-learning medical devices and shares open lists with the public. In this article, we examine recent medical AI devices in different fields, including the diagnosis of colorectal polyps.Current Concepts: Currently, in the field of gastroenterology, there has been a significant amount of research aimed at enhancing adenoma detection rates using tools powered by AI, such as the EndoScreener and GI Genius. Various such devices have also been developed for other fields; examples include the 23andMe Personal Genome Service for DNA detection, Spectral MD’s DeepView platform for wound imaging in surgery, Gili Pro BioSensor for monitoring vital signs, DreaMed Advisor Pro for diabetes, Minuteful for urinary analysis, BrainScope TBI for cerebral diagnosis, Compumedics Sleep Monitoring System for sleep disorders, Idx-DR v2.3 for ophthalmology, and EarliPoint system for pediatrics.Discussion and Conclusion: By the time this article is published, it is likely that even more AI medical devices will have been approved and commercialized. The development of such devices should be strongly encouraged. Additionally, we anticipate greater involvement from practitioners in the development and validation of diverse medical AI devices in Korea.

Background and Objectives: As the Korean population ages fast, it is estimated that the people with hypertension, especially female patients, will increase rapidly. However, there are few data comparing the size of female and male hypertensive patients in the Korean population. Thus we assessed sex-specific trends in the prevalence and the number of people with hypertension. Methods: We analyzed data for 128,949 adults aged ≥20 years with valid blood pressure measurements from the 1998 to 2018 Korea National Health and Nutrition Examination Survey (KNHANES). The prevalence and the absolute number of hypertension were estimated with taking into the sampling weights separately for women and men. Results: Overall prevalence of hypertension is higher in men than in women. But, in older adults, women show higher prevalence and the number of people with hypertension. Between 1998 and 2018, prevalence of hypertension increased from 61.8% to 65.9% in elderly (age 65+) women, and from 49.0% to 59.4% in elderly men. During the same period, the number of elderly women with hypertension increased from 1.18 to 2.70 million, while the number of elderly men with hypertension increased from 0.57 to 1.78 million. Among hypertensive patients, undiagnosed hypertension and diagnosed-but-untreated hypertension were more common in men, while treated-but-uncontrolled hypertension were more common in women. Conclusion The fast-growing number of elderly women with hypertension will be an important public health challenge for the Korean society to solve in order to reduce the burden of cardiovascular disease.

Purpose: Cardiovascular health (CVH) status is associated with several cardiovascular outcomes; however, correlations between changes in CVH status and risk of sudden cardiac death (SCD) are unknown. We aimed to evaluate associations between changes in CVH status and risk of SCD and all-cause death in older adults. Materials and Methods: We used data from the Korea National Health Insurance Service-Senior cohort database (2005–2012). Six metrics from the American Heart Association (smoking, body mass index, physical activity, blood pressure, total cholesterol, and fasting blood glucose) were used to calculate CVH scores. Changes in CVH status between two health checkups were categorized as low to low, low to high, high to low, and high to high. Results: We included 105200 patients whose CVH status for an initial and follow-up health checkup (2-year interval) was available. During a median of 5.2 years of follow-up after a second health checkup, 688 SCDs occurred. Compared to patients with a persistent low CVH status, those with a consistently high CVH status had a reduced risk of SCD [adjusted hazard ratio (HR), 0.69; 95% confidence interval (CI), 0.56–0.86] and all-cause death (adjusted HR, 0.74; 95% CI, 0.69–0.78). The risk of all-cause death followed similar trends. However, an inconsistent linear relationship was observed for changes in CVH status and the risk of SCD, but not of all-cause death. Conclusion Maintaining a high CVH status was associated with future risks of SCD and all-cause death among an older adult population.

Background: We aimed to investigate the comparative risk of fracture among patients with atrial fibrillation (AF) treated with warfarin or non-vitamin K antagonist oral anticoagulants (NOACs). Methods: Using the Korean National Health Insurance Service database, patients with AF who received a prescrip‑ tion for apixaban, dabigatran, rivaroxaban, or warfarin between 2013 and 2016 were included. Risk of major fractures (osteoporotic hip, vertebral, or pelvic fractures) were compared using inverse probability of treatment weighting. Results: There were 70,481 patients identified (41.3% women; mean [SD] age 70.5 [11.3] years); 16,992 apixaban, 22,514 dabigatran, 27,998 rivaroxaban, and 29,390 warfarin users. During a median follow-up of 390 days, 2412 major fractures occurred with weighted incidences per 100 patient-years of 2.56 for apixaban, 2.39 for dabigatran, 2.78 for rivaroxaban, and 3.43 for warfarin. NOAC use was associated with a lower risk for fracture than warfarin use: HR 0.70 (95% confidence interval [CI] 0.57–0.86) for apixaban, HR 0.69 (95% CI 0.60–0.78) for dabigatran, and HR 0.79 (95% CI 0.70–0.90) for rivaroxaban. In head-to-head comparisons between NOACs, there was no significant difference between apixaban and dabigatran. Rivaroxaban was associated with a higher risk for fracture than dabigatran (HR 1.15, 95% CI 1.02–1.31). Conclusion In patients with AF, NOAC use may result in a lower risk for osteoporotic fracture compared with warfa‑ rin use. Fracture risk does not seem to be altered by the choice of NOAC type, except for rivaroxaban. These associa‑ tions may help inform benefit–risk assessments when choosing between the different anticoagulant types.

Background: Vaccine safety surveillance is important because it is related to vaccine hesitancy, which affects vaccination rate. To increase confidence in vaccination, the active monitoring of vaccine adverse events is important. For effective active surveillance, we developed and verified a machine learning-based active surveillance system using national claim data. Methods: We used two databases, one from the Korea Disease Control and Prevention Agency, which contains flu vaccination records for the elderly, and another from the National Health Insurance Service, which contains the claim data of vaccinated people. We developed a casecrossover design based machine learning model to predict the health outcome of interest events (anaphylaxis and agranulocytosis) using a random forest. Feature importance values were evaluated to determine candidate associations with each outcome. We investigated the relationship of the features to each event via a literature review, comparison with the Side Effect Resource, and using the Local Interpretable Model-agnostic Explanation method. Results: The trained model predicted each health outcome of interest with a high accuracy (approximately 70%). We found literature supporting our results, and most of the important drug-related features were listed in the Side Effect Resource database as inducing the health outcome of interest. For anaphylaxis, flu vaccination ranked high in our feature importance analysis and had a positive association in Local Interpretable Model-Agnostic Explanation analysis. Although the feature importance of vaccination was lower for agranulocytosis, it also had a positive relationship in the Local Interpretable Model-Agnostic Explanation analysis. Conclusion We developed a machine learning-based active surveillance system for detecting possible factors that can induce adverse events using health claim and vaccination databases. The results of the study demonstrated a potentially useful application of two linked national health record databases. Our model can contribute to the establishment of a system for conducting active surveillance on vaccination.

Objectives: Low bone density (LBD) in the postmenopausal period has long been a pervasive public health concern; however, the association between parity and LBD has yet to be fully elucidated. Thus, we investigated the association between parity and LBD in postmenopausal Korean women. Methods: This study used baseline data from 1287 Korean postmenopausal women aged 40 years or older enrolled in the Cardiovascular and Metabolic Diseases Etiology Research Center community-based cohort study conducted in Korea from 2013 to 2017. The main exposure was parity (nullipara, 1, 2, 3+). The main outcome was LBD, including osteopenia and osteoporosis, based on bone mineral density measured using quantitative computed tomography of the lumbar spine (L1-2). Results: The mean age of participants was 57.1 years, and the median parity was 2. Of the 1287 participants, 594 (46.2%) had osteopenia and 147 (11.4%) had osteoporosis. No significant difference in the prevalence of LBD was found between nullipara and parous women, whereas higher parity was associated with a higher risk of LBD among parous women; the adjusted odds ratio (95% confidence interval) for the presence of LBD was 1.40 (0.97 to 2.02) for a parity of 2 and 1.95 (1.23 to 3.09) for a parity of 3 relative to a parity of 1. Conclusions Women who have given birth multiple times may be at greater risk of bone loss after menopause; therefore, they should be a major target population for osteoporosis prevention.