26 March 2022 in Journal of Global Antimicrobial Resistance
This study constructed a carbapenem-resistant Gram-negative bacteria (CR-GNB) carriage prediction model to predict the CR-GNB incidence in a week. We used our database to select patients with complete CR-GNB screening records between the years 2015 and 2019 and constructed the model using multivariable logistic regression and three machine learning algorithms. Then we chose the optimal model and verified the accuracy by daily prediction and recorded the occurrence of CR-GNB in all intensive care unit patients admitted for 4 months. There were 1385 patients with positive CR-GNB cultures and 1535 negative patients in this study. Forty-five variables had statistically significant differences. We included 16 variables in the multivariable logistic regression model and built three machine learning models for all variables. In terms of accuracy and the area under the receiver operating characteristic (AUROC) curve, random forest was better than XGBoost and decision tree and better than a multivariable logistic regression model (accuracy: 84%>82%>81%>72%, AUROC: 0.91>0.89=0.89>0.78). In a 4-month prospective study, 74 cases were predicted to have positive CR-GNB culture within 7 days, 132 cases were predicted to be negative, 86 cases were positive, and 120 cases were negative, with an overall accuracy of 85.92% and AUROC of 92.02%. Machine learning prediction models can predict the occurrence of CR-GNB colonisation or infection within a one-week period and can guide medical staff in real time to identify high-risk groups more accurately.
6 March 2023 in IEEE Journal of Biomedical and Health Informatics
Leveraging machine learning techniques for Sepsis early detection and diagnosis has attracted increasing interest in recent years. However, most existing methods require a large amount of labeled training data, which may not be available for a target hospital that deploys a new Sepsis detection system. More seriously, as treated patients are diversified between hospitals, directly applying a model trained on other hospitals may not achieve good performance for the target hospital. To address this issue, we propose a novel semi-supervised transfer learning framework based on optimal transport theory and self-paced ensemble for Sepsis early detection, called SPSSOT , which can efficiently transfer knowledge from the source hospital (with rich labeled data) to the target hospital (with scarce labeled data). Specifically, SPSSOT incorporates a new optimal transport-based semi-supervised domain adaptation component that can effectively exploit all the unlabeled data in the target hospital. Moreover, self-paced ensemble is adapted in SPSSOT to alleviate the class imbalance issue during transfer learning. In a nutshell, SPSSOT is an end-to-end transfer learning method that automatically selects suitable samples from two domains (hospitals) respectively and aligns their feature spaces. Extensive experiments on two open clinical datasets, MIMIC-III and Challenge, demonstrate that SPSSOT outperforms state-of-the-art transfer learning methods by improving 1-3% of AUC.