Improved patient mortality predictions in emergency departments with deep learning data-synthesis and ensemble models.

The triage process in emergency departments (EDs) relies on the subjective assessment of medical practitioners, making it unreliable in certain aspects. There is a need for a more accurate and objective algorithm to determine the urgency of patients. This paper explores the application of advanced data-synthesis algorithms, machine learning (ML) algorithms, and ensemble models to predict patient mortality. Patients predicted to be at risk of mortality are in a highly critical condition, signifying an urgent need for immediate medical intervention. This paper aims to determine the most effective method for predicting mortality by enhancing the F1 score while maintaining high area under the receiver operating characteristic curve (AUC) score. This study used a dataset of 7325 patients who visited the Yonsei Severance Hospital's ED, located in Seoul, South Korea. The patients were divided into two groups: patients who deceased in the ED and patients who didn't. Various data-synthesis techniques, such as SMOTE, ADASYN, CTGAN, TVAE, CopulaGAN, and Gaussian Copula, were deployed to generate synthetic patient data. Twenty two ML models were then utilized, including tree-based algorithms like Decision tree, AdaBoost, LightGBM, CatBoost, XGBoost, NGBoost, TabNet, which are deep neural network algorithms, and statistical algorithms such as Support Vector Machine, Logistic Regression, Random Forest, k-nearest neighbors, and Gaussian Naive Bayes, as well as Ensemble Models which use the results from the ML models. Based on 21 patient information features used in the pandemic influenza triage algorithm (PITA), the models explained previously were applied to aim for the prediction of patient mortality. In evaluating ML algorithms using an imbalanced medical dataset, conventional metrics like accuracy scores or AUC can be misleading. This paper emphasizes the importance of using the F1 score as the primary performance measure, focusing on recall and specificity in detecting patient mortality. The highest-ranked model for predicting mortality utilized the Gaussian Copula data-synthesis technique and the CatBoost classifier, achieving an AUC of 0.9731 and an F1 score of 0.7059. These findings highlight the effectiveness of machine learning algorithms and data-synthesis techniques in improving the prediction performance of mortality in EDs.

Preferential Halogenation of Algal Organic Matter by Iodine over Chlorine and Bromine: Formation of Disinfection Byproducts and Correlation with Toxicity of Disinfected Waters.

The increasing occurrence of harmful algal blooms (HABs) in surface waters may increase the input of algal organic matter (AOM) in drinking water. The formation of halogenated disinfection byproducts (DBPs) during combined chlorination and chloramination of AOM and natural organic matter (NOM) in the presence of bromide and iodide and haloform formation during halogenation of model compounds were studied. Results indicated that haloform/halogen consumption ratios of halogens reacting with amino acids (representing proteins present in AOM) follow the order iodine > bromine > chlorine, with ratios for iodine generally 1-2 orders of magnitude greater than those for chlorine (0.19-2.83 vs 0.01-0.16%). This indicates that iodine is a better halogenating agent than chlorine and bromine. In contrast, chlorine or bromine shows higher ratios for phenols (representing the phenolic structure of humic substances present in NOM). Consistent with these observations, chloramination of AOM extracted from Microcystis aeruginosa in the presence of iodide produced 3 times greater iodinated trihalomethanes than those from Suwannee River NOM isolate. Cytotoxicity and genotoxicity of disinfected algal-impacted waters evaluated by Chinese hamster ovary cell bioassays both follow the order chloramination > prechlorination-chloramination > chlorination. This trend is in contrast to additive toxicity calculations based on the concentrations of measured DBPs since some toxic iodinated DBPs were not identified and quantified, suggesting the necessity of experimentally analyzing the toxicity of disinfected waters. During seasonal HAB events, disinfection practices warrant optimization for iodide-enriched waters to reduce the toxicity of finished waters.

Efficient inverted bulk heterojunction photovoltaic devices using a transparent polymeric interfacial buffer layer with C60 pendant and UV curable groups.

We demonstrate the synthesis of a transparent, polymeric n-type material (M1) consisting of C60 pendant and UV curable groups in side chains. This material (M1) is employed as a polymeric n-type interfacial buffer layer for an efficient inverted bulk heterojunction (BHJ) photovoltaic device based on regioregular poly(3-hexylthiophene):[6,6]-phenyl C61 butyric acid methyl ester (P3HT:PC61BM) active layer. Under simulated solar illumination of AM 1.5G (100 mW/cm2), the highest efficient devices fabricated with a configuration of ITO/interfacial buffer layer (M1,10 nm)/P3HT:PC61BM (1:0.9 w:w) (120 nm)/PEDOT:PSS (30 nm)/Ag (100 nm) achieve an average power conversion efficiency PCE of 2.16%, with short-circuit current J(SC) = 6.70 mA/cm2, fill factor FF = 54.2%, and open-circuit voltage V(OC) = 0.60 V. This result is comparable to the inverted BHJ photovoltaic devices fabricated with Cs2CO3, one of widely used as a buffer layer. The synthesized M1 have thus proven to be promising polymeric interfacial buffer layer for high efficient BHJ photovoltaic devices.

Synthesis of carbazole/triarylamine containing polyfluorene and its application to electroluminescent devices.

Solution-processable carbazole/triarylamine-based polyfluorene (CT-PF) was designed and synthesized by Suzuki coupling reaction. CT-PF was originally designed to have multiple functions at the same polymer backbone; blue emission and large band gap energy by fluorine group as well as good hole transporting property and tunable ionization potential by excellent electron donating triarylamine and carbazole moieties. It was confirmed from TGA measurement that the synthesized CT-PF had thermal stability up to 379 degrees C. The UV-Visible absorption spectra of CT-PF had absorption maxima at 292 nm and 351 nm in dilute solution, and 332 nm and 400 nm in film state, which might be originated from fluorene and arylamine moieties. PL emission maxima were also measured in the blue range at 425 nm in dilute solution and 468 nm in film state. Turn-on voltage of double-layered OLED with CT-PF as HTL was reduced to 7 V, and the luminescence were enhanced up to 1,397 cd/m2 with maximum current efficiency of 0.32 lm/W and power efficiency of 1.1 cd/A, which were over 3 times higher than those of single-layered device without HTL. PhOLED with CT-PF as a host material for Ir(ppy)3 phosphorescent dopant showed bright green emission with CIE color coordinates as (x = 0.31, y = 0.57).