Prediction of pneumonia hospitalization in adults using health checkup data

Lasso-Logistic regression model for the identification of serum biomarkers of neurotoxicity induced by strychnos alkaloids

As a traditional Chinese medicine, strychnos alkaloids have wide effects including antitumor, analgesic, and anti-inflammatory. However, the therapeutic window of strychnos alkaloids is quite narrow due to potential neurotoxicity. Therefore, it is necessary to explore some efficient biomarkers to identify and predict the neurotoxicity induced by strychnos alkaloids and find a therapy to prevent the neurotoxicity of strychnos alkaloids. Based on the previous studies of our research team, 21 endogenous substances related to neurotoxicity were monitored in rats' serum with HPLC-MS/MS and ELISA. Starting from these fundamentals, a Lasso-Logistic regression model was used to select efficient biomarkers from 21 endogenous substances to predict brain injury and verify the neuroprotective effect of peonies. Under the processing of the Lasso-Logistic regression model, 12 biomarkers were identified from 21 endogenous substances to predict the neurotoxicity induced by strychnos alkaloids. At the same time, the neuroprotective effect of peonies was further confirmed by evaluating the level of 12 biomarkers. The results indicated that the development of the Lasso-Logistic regression model would provide a new, simple and efficient method for the prediction and diagnosis of the neurotoxicity induced by strychnos alkaloids.

A Broad Learning System to Predict the 28-Day Mortality of Patients Hospitalized with Community-Acquired Pneumonia: A Case-Control Study

This study was to conduct a model based on the broad learning system (BLS) for predicting the 28-day mortality of patients hospitalized with community-acquired pneumonia (CAP). A total of 1,210 eligible CAP cases from Chifeng Municipal Hospital were finally included in this retrospective case-control study. Random forest (RF) and an eXtreme Gradient Boosting (XGB) models were used to develop the prediction models. The data features extracted from BLS are utilized in RF and XGB models to predict the 28-day mortality of CAP patients, which established two integrated models BLS-RF and BLS-XGB. Our results showed the integrated model BLS-XGB as an efficient broad learning system (BLS) for predicting the death risk of patients, which not only performed better than the two basic models but also performed better than the integrated model BLS-RF and two well-known deep learning systems-deep neural network (DNN) and convolutional neural network (CNN). In conclusion, BLS-XGB may be recommended as an efficient model for predicting the 28-day mortality of CAP patients after hospital admission.

The Multi-Institutional Health Screening Records Database of South Korea: Description and Evaluation of Its Characteristics

This study sought to describe and to evaluate the characteristics of the Health Screening Records Database (HSRD) of the Korea Association of Health Promotion as a data source for epidemiologic studies. The HSRD was compared to a National Health Insurance Service-Health Screening Cohort (NHIS-HEALS) database for 2015. Common variables between the databases were selected, and sex-based analyses were conducted. The HSRD showed statistical concordance when NHIS-HEALS estimates fell within the HSRD estimate's 95% confidence interval. The HSRD and NHIS-HEALS included 946461 and 111690 participants in health screening programs, respectively. Compared to the NHIS-HEALS, the HSRD had more female (55.2% vs. 42.6%) but fewer older adult participants (34.4% vs. 51.2%). Virtually all variables had clinical concordance, with some having statistical concordance as well, among both general and life-transition program participants. The HSRD comprised more clinical information over a wider age range in contrast to the NHIS-HEALS, while showing clinical concordance. Providing more comprehensive clinical data, the HSRD may serve as an alternative resource for epidemiologic studies.