An Artificial Intelligence Approach to Bloodstream Infections Prediction

Abstract

This study aimed to develop an early prediction model for identifying patients with bloodstream infections. The data resource was taken from 2015 to 2019 at Taichung Veterans General Hospital, and a total of 1647 bloodstream infection episodes and 3552 non-bloodstream infection episodes in the intensive care unit (ICU) were included in the model development and evaluation. During the data analysis, 30 clinical variables were selected, including patients’ basic characteristics, vital signs, laboratory data, and clinical information. Five machine learning algorithms were applied to examine the prediction model performance. The findings indicated that the area under the receiver operating characteristic curve (AUROC) of the prediction performance of the XGBoost model was 0.825 for the validation dataset and 0.821 for the testing dataset. The random forest model also presented higher values for the AUROC on the validation dataset and testing dataset, which were 0.855 and 0.851, respectively. The tree-based ensemble learning model enabled high detection ability for patients with bloodstream infections in the ICU. Additionally, the analysis of importance of features revealed that alkaline phosphatase (ALKP) and the period of the central venous catheter are the most important predictors for bloodstream infections. We further explored the relationship between features and the risk of bloodstream infection by using the Shapley Additive exPlanations (SHAP) visualized method. The results showed that a higher prothrombin time is more prominent in a bloodstream infection. Additionally, the impact of a lower platelet count and albumin was more prominent in a bloodstream infection. Our results provide additional clinical information for cut-off laboratory values to assist clinical decision-making in bloodstream infection diagnostics.

Application of an Interpretable Machine Learning Model to Predict Lymph Node Metastasis in Patients with Laryngeal Carcinoma

“…The AUC, F1-score, accuracy, precision, recall rate, specificity, and Brier score were the main evaluation metrics used to evaluate and compare the model performances. However, the higher the value of AUC, F1-score, accuracy, precision, specificity, and recall rate, the better the model performances, but the Brier score was just the opposite [ 18 ]. In both the training and test sets, the XGBoost model showed the highest values regarding AUC, F1-score, accuracy, and recall rate, and the lowest values regarding Brier score (AUC = 0.791 (95% CI: 0.776–0.806) and 0.829 (95% CI: 0.818–0.843), F1-score = 0.663 and 0.706, accuracy = 0.739 and 0.770, recall rate = 0.638 and 0.677, Brier score = 0.165 and 0.153) (Figures 4(a) , 4(b) , 5(a), and 5(b) Table 4 ).…”

“…Te AUC, F1-score, accuracy, precision, recall rate, specifcity, and Brier score were the main evaluation metrics used to evaluate and compare the model performances. However, the higher the value of AUC, F1-score, accuracy, precision, specifcity, and recall rate, the better the model performances, but the Brier score was just the opposite [18]. S4 and S5.…”

Artificial Intelligence: A Next-Level Approach in Confronting the COVID-19 Pandemic

“…Further, SVM (0.578 and 0.566), RF (0.565 and 0.577), and MLP (0.494 and 0.406) models exhibited lower sensitivity. The validation was conducted with the important risk factors associated with BSI, such as prothrombin time (PT), platelets (PLT), and albumin (ALB), which are key factors [ 89 ]. Similarly, Zoabi et al, developed a patient outcome of BSI model to identify BSI patient risk based on the electronic medical records (EMR) of only bacteria-based positive blood culture results, with the goal of identifying risky patients, initiating a planned treatment with appropriate antibiotics, and transferring them to the ICU.…”

Clinical Escherichia coli: From Biofilm Formation to New Antibiofilm Strategies

“…BSIs are considered one of the most important infections with an overall mortality rate of 15-30% [77]. In this case, biofilms are involved in catheter-associated BSIs.…”