Objectives: Triage is an essential emergency department (ED) process designed to provide timely management depending on acuity and severity; however, the process may be inconsistent with clinical and hospitalization outcomes. Therefore, studies have attempted to augment this process with machine learning models, showing advantages in predicting critical conditions and hospitalization outcomes. The aim of this study was to utilize nationwide registry data to develop a machine learning-based classification model to predict the clinical course of pediatric ED visits. Methods: This cross-sectional observational study used data from the National Emergency Department Information System on emergency visits of children under 15 years of age from January 1, 2016, to December 31, 2017. The primary and secondary outcomes were to identify critically ill children and predict hospitalization from triage data, respectively. We developed and tested a random forest model with the under sampled dataset and validated the model using the entire dataset. We compared the model’s performance with that of the conventional triage system. Results: A total of 2,621,710 children were eligible for the analysis and included 12,951 (0.5%) critical outcomes and 303,808 (11.6%) hospitalizations. After validation, the area under the receiver operating characteristic curve was 0.991 (95% confidence interval [CI] 0.991–0.992) for critical outcomes and 0.943 (95% CI 0.943–0.944) for hospitalization, which were higher than those of the conventional triage system. Conclusions: The machine learning-based model using structured triage data from a nationwide database can effectively predict critical illness and hospitalizations among children visiting the ED.

Lee B, et al. Infect Chemother. 2013 Dec;45(4):415-421. http://dx.doi.org/10.3947/ic.2013.45.4.415

Background: Group B streptococcus (GBS) is one of the leading causes of invasive infections in infants. This study aimed to investigate the genotypic diversity of GBS causing invasive infections in infants and to observe the prevalence of the highly virulent clone in South Korea. Methods: Invasive strains of GBS were collected prospectively from infants admitted at 4 hospitals during 1995–2015. Serotype and multilocus sequence typing were determined. All isolates underwent polymerase chain reaction amplification to detect the presence of the hypervirulent GBS adhesin (hvgA) gene. Antibiotic susceptibility testing was done by E-test, and erythromycin resistance genes were detected using polymerase chain reaction amplification. Results: Among 98 GBS isolates collected, 14 sequence types (STs) were found; ST1 (20.4%), ST17 (19.4%) and ST19 (18.4%) were the most prevalent. The dominant serotype capsule expressed by ST1 was serotype V, ST17 and ST19 were all serotype III and ST23 was serotype Ia. hvgA gene was detected in 19.4% (n = 19) of the isolates; all were ST17, serotype III. A significant temporal trend of serotype III isolates was observed; as ST17 increased (P = 0.001) in proportion, ST19 decreased (P = 0.009). Erythromycin resistance was found in 42.9% (42/98); dominant strains were ermB-positive ST1 serotype V (n = 18/20, 90%), ermB-positive ST17 serotype III (n = 10/19, 52.6%) and ermA-positive ST335 serotype III (n = 7/7, 100%). Conclusions: The predominant STs causing invasive infections in South Korea were ST1, ST19 and ST17. Among serotype III isolates, an increase in proportion of the hypervirulent ST17 strains was observed. Erythromycin resistance was significantly associated with ST1.

Children with hereditary neuromuscular diseases (NMDs) are at a high risk of morbidity and mortality related to respiratory failure. The use of home mechanical ventilation (HMV) has saved the lives of many children with NMD but, due to a lack of studies, dependable guidelines are not available. We drew upon our experience to compare the various underlying NMDs and to evaluate HMV with regard to respiratory morbidity, the proper indications and timing for its use, and to develop a policy to improve the quality of home noninvasive ventilation (NIV). We retrospectively analyzed the medical records of 57 children with childhood-onset hereditary NMDs in whom HMV was initiated between January 2000 and May 2013 at Seoul National University Children's Hospital. The degree of respiratory morbidity was estimated by the frequency and duration of hospitalizations caused by respiratory distress. The most common NMD was spinal muscular atrophy (SMA, n = 33). Emergent mechanical ventilation was initiated in 44% of the patients before the confirmed diagnosis, and the indicators of pre-HMV respiratory morbidity (e.g., extubation trials, hypoxia, hospitalizations, and intensive care unit stay) were greater in these patients than in others. The proportion of post-HMV hospitalizations (range, 0.00−0.52; median, 0.01) was lower than that of pre-HMV hospitalizations (0.02−1.00; 0.99) (P < 0.001). Eight patients were able to maintain home NIV. The main causes of NIV failure were air leakage and a large amount of airway secretions. The application of HMV helped reduce respiratory morbidity in children with childhood-onset hereditary NMD. Patients with SMA type I can benefit from an early diagnosis and the timely application of HMV. The choice between invasive and noninvasive HMV should be based on the patient’s age and NIV trial tolerance. Systematic follow-up guidelines provided by a multidisciplinary team are needed.

Lee B, et al. J Korean Med Sci. 2017 Oct;32(10):1702-1707. https://doi.org/10.3346/jkms.2017.32.10.1702

Measured glomerular filtration rate (mGFR) is often used to identify augmented renal clearance (ARC). However, in the clinical setting, estimated GFR (eGFR) is obtained more quickly and inexpensively. We aimed to determine whether eGFR can identify ARC by evaluating the correlation between the eGFR and vancomycin trough level (VTL). We retrospectively reviewed the records of patients aged ≤ 18 years who underwent vancomycin therapeutic drug monitoring at our tertiary hospital from July 2009 to June 2014. VTL, serum creatinine concentration, eGFR, and clinical factors affecting VTL were analyzed. Of 101 patients, 76 (75.25%) had a subtherapeutic VTL. Patient age (p = 0.006), the daily vancomycin dose (p = 0.041) and dosing interval (p = 0.006), and eGFR (p < 0.001) affected the VTL. Multivariate analysis showed a significant relationship between eGFR and VTL (adjusted R2, 0.812; p < 0.001). An increased eGFR (odds ratio, 1.002; 95% confidence interval, 1.001 - 1.003; p = 0.001) was a risk factor for a subtherapeutic vancomycin level. The cutoff eGFR value predicting a subtherapeutic vancomycin level was 110.51 mL/min/1.73m2 (area under the curve, 0.753). The eGFR correlates with the VTL, and the eGFR cutoff value can predict a subtherapeutic vancomycin level. eGFR is a reliable and efficient alternative to mGFR for identifying ARC. .