A Two-Step Approach to Overcoming Data Imbalance in the Development of an Electrocardiography Data Quality Assessment Algorithm: A Real-World Data Challenge.

Continuously acquired biosignals from patient monitors contain significant amounts of unusable data. During the development of a decision support system based on continuously acquired biosignals, we developed machine and deep learning algorithms to automatically classify the quality of ECG data. A total of 31,127 twenty-s ECG segments of 250 Hz were used as the training/validation dataset. Data quality was categorized into three classes: acceptable, unacceptable, and uncertain. In the training/validation dataset, 29,606 segments (95%) were in the acceptable class. Two one-step, three-class approaches and two two-step binary sequential approaches were developed using random forest (RF) and two-dimensional convolutional neural network (2D CNN) classifiers. Four approaches were tested on 9779 test samples from another hospital. On the test dataset, the two-step 2D CNN approach showed the best overall accuracy (0.85), and the one-step, three-class 2D CNN approach showed the worst overall accuracy (0.54). The most important parameter, precision in the acceptable class, was greater than 0.9 for all approaches, but recall in the acceptable class was better for the two-step approaches: one-step (0.77) vs. two-step RF (0.89) and one-step (0.51) vs. two-step 2D CNN (0.94) (p < 0.001 for both comparisons). For the ECG quality classification, where substantial data imbalance exists, the 2-step approaches showed more robust performance than the one-step approach. This algorithm can be used as a preprocessing step in artificial intelligence research using continuously acquired biosignals.

Heart failure awareness in the Korean general population: Results from the nationwide survey.

BACKGROUND: For a better heart failure outcome, it is fundamental to improve the awareness of heart failure at the general population level. We conducted this study to identify the current status of awareness of heart failure in the Korean general population. METHODS: This cross-sectional nationwide survey recruited a total of 1,032 participants aged 30 years or older, based on a stratification systematic sampling method. A 23-item questionnaire was surveyed through telephone interviews. RESULTS: Although 80% of the participants had heard of heart failure, 47% exactly defined what heart failure is. A minority of participants correctly recognized the lifetime risk of developing heart failure (21%) as well as the mortality (16%) and readmission risk (18%) of heart failure and the cost burden of heart failure admission (28%). Regarding preferred treatment options, 71% of the participants chose a treatment option that could improve the quality of life. Approximately two-thirds of the participants agreed that current medical treatment could reduce mortality and improve the quality of life. More than half of the participants (59%) thought that heart failure patients should live quietly and reduce all physical activities. Across survey items, we found a lower awareness state in the elderly groups and people at lower income and educational levels. CONCLUSIONS: The current awareness status of heart failure in the Korean general population is still low. Proactive educational efforts should be made to improve public awareness with special attention to individuals with lower disease awareness.

Age-Specific Distribution of Diagnosis and Outcomes of Children Admitted to ICUs: A Population-Based Cohort Study.

OBJECTIVES: Although several studies have reported outcome data on critically ill children, detailed reports by age are not available. We aimed to evaluate the age-specific estimates of trends in causes of diagnosis, procedures, and outcomes of pediatric admissions to ICUs in a national representative sample. DESIGN: A population-based retrospective cohort study. SETTING: Three hundred forty-four hospitals in South Korea. PATIENTS: All pediatric admissions to ICUs in Korea from August 1, 2009, to September 30, 2014, were covered by the Korean National Health Insurance Corporation, with virtually complete coverage of the pediatric population in Korea. Patients less than 18 years with at least one ICUs admission between August 1, 2009, and September 30, 2014. We excluded neonatal admissions (< 28 days), neonatal ICUs, and admissions for health status other than a disease or injury. The final sample size was 38,684 admissions from 32,443 pediatric patients. INTERVENTION: None. MEASUREMENTS AND MAIN RESULTS: The overall age-standardized admission rate for pediatric patients was 75.9 admissions per 100,000 person-years. The most common primary diagnosis of admissions was congenital malformation (10,897 admissions, 28.2%), with marked differences by age at admission (5,712 admissions [54.8%] in infants, 3,994 admissions [24.6%] in children, and 1,191 admissions [9.9%] in adolescents). Injury was the most common primary diagnosis in adolescents (3,248 admissions, 27.1%). The overall in-hospital mortality was 2,234 (5.8%) with relatively minor variations across age. Neoplasms and circulatory and neurologic diseases had both high frequency of admissions and high in-hospital mortality. CONCLUSIONS: Admission patterns, diagnosis, management, and outcomes of pediatric patients admitted to ICUs varied by age groups. Strategies to improve critical care qualities of pediatric patients need to be based on the differences of age and may need to be targeted at specific age groups.

Extracorporeal membrane oxygenation in Korea - Trends and impact of hospital volume on outcome: Analysis of national insurance data 2009-2014.

PURPOSE: ECMO use has increased lately. However, differences between adult ECMO and non-ECMO patients admitted to the ICU remain unstudied. In terms of volume-outcome relationship, the impact of ECMO volume on survival has not been validated in a real world cohort. MATERIALS AND METHODS: Retrospective analysis of data from the Korean Health Insurance Review and Assessment Service over 5 years, between August 1, 2009 and July 31, 2014. The ECMO group comprised patients who received ≥1 ECMO run. Data on patient demographics, ICU and hospital length of stay, cost, treatments, and in-hospital mortality were collected. Usage trends were analyzed by 5 one-year periods. RESULTS: Among 1, 265, 508 ICU patients, 6078 underwent ECMO during the study period. The number of ECMO patients rose by 2.5 times, and ECMO hospitals from 50 to 86 between periods 1 and 5. Compared to non-ECMO patients, the ECMO group was younger (59 years vs. 64 years, p < .0001) with more comorbidities. Healthcare expenditure and in-hospital mortality in the ECMO group were higher (US $23,600 vs. $5100; 63.4% vs. 12.6%; p < .0001). Using multivariable analysis, age ≥ 50 years, CRRT, and annual hospital ECMO volume < 20 negatively impacted survival to discharge. CONCLUSION: The prevalence of ECMO among ICU patients was 0.5%. The expenditure and in-hospital mortality of the ECMO group were four and five times higher than non-ECMO group respectively. An annual hospital ECMO volume ≥ 20 may improve survival to hospital discharge.