Abstract:Background: The number of patients with end-stage renal disease and acute kidney injury in China is large and increasing year by year. Continuous renal replacement therapy (CRRT) is one of the important treatment methods. However, long-time CRRT would inevitably lead to CLABSI, which would seriously affect the treatment and prognosis of the patient. Although CLABSIs can be prevented and controlled, the rate of CLABSI in China is still higher than in other countries. Therefore, it is urgent to find new intervention methods on the basis of existing methods. Surveillance is the prerequisite of infection prevention and control. We sought to develop a risk prediction model for CLABSI in patients with CRRT according to uncontrollable risk factors, which could be used for early assessment and screening of high-risk infection groups. Such a tool would bring the supervision and infection control to the forefront in addressing these issues. Methods: We selected 3,103 CRRT patients in the West China Hospital of Sichuan University from January 2013 to December 2018 using the hospital infection and infectious disease monitoring module of electronic medical records (EMR) system with the integration and elimination criteria. Data mining and feature selection were performed using Weka software. Separately, prediction models developed by Weka software and SPSS software were compared with each other using the area under the curve (AUC) method to assess the performance of the forecasting models. Result: The incidence of CLABSI in CRRT patients was 8.01 per 1,000 catheter days (238 of 29,711). According to the multifactor regression analysis by SPSS software, the retaining time of dialysis catheter, C-reactive protein levels, total bilirubin, acute pancreatitis, and systemic inflammation reaction syndrome were the risk factors. According to the Youden’s index, the cutoff point of the retaining time of dialysis catheter was 5.5 days; the cutoff point of CRP was 112.5mg/L; and the cutoff point of total bilirubin was 14.15 μmol/L. The prediction models of CLABSI for CRRT patients were developed: The AUC of the prediction model developed by SPSS software was 0.763 (95% CI, 0.717–0.809). The receiver operating characteristic (ROC) curve analysis showed that the AUCs of the prediction models developed separately by Weka software using Bayes, logistic regression analysis, multiple layer Perceptron and J48, and SPSS software through logistic regression analysis were between 0.6 and 0.8. Using the down-sampling technique, the AUC ranged between 0.7 and 0.9, and the sensitivity, precision, and κ value increased. Thus, these models had definite clinical significance. Conclusion: The prediction models of CLABSI for CRRT patients, developed based on the big healthcare data, not only had good judgment ability, but also had good application value for individual evaluations and the target population. Funding: This study was supported by the Health Commission of Sichuan Province. Disclosures: None

Hospital-wide, dynamic, individualized prediction of central line-associated bloodstream infections - development and temporal evaluation of six prediction models

Prediction of impending central-line-associated bloodstream infections in hospitalized cardiac patients: development and testing of a machine-learning model

A comparison of regression models for static and dynamic prediction of a prognostic outcome during admission in electronic health care records

Development of a Risk Prediction Model for Central-Line–Associated Bloodstream Infection (CLABSI) in Patients With Continuous Renal Replacement Therapy

Prediction of Post-Stroke Urinary Tract Infection Risk in Immobile Patients Using Machine Learning: an Observational Cohort Study.

Comparison of static and dynamic random forests models for EHR data in the presence of competing risks: predicting central line-associated bloodstream infection

Machine Learning Algorithms to Predict the Mortality of Carbapenem Resistant Klebsiella Pneumoniae Bacteremia

Predicting central line-associated bloodstream infections and mortality using supervised machine learning

Early Prediction of Central Line Associated Bloodstream Infection Using Machine Learning.

Prediction of Central Venous Catheter-Associated Deep Venous Thrombosis in Pediatric Critical Care Settings.

A Machine Learning Predictive Model of Bloodstream Infection in Hospitalized Patients

An Artificial Intelligence Approach to Bloodstream Infections Prediction

A Machine-Learning Approach for Dynamic Prediction of Sepsis-Induced Coagulopathy in Critically Ill Patients With Sepsis

Early Warning Models Using Machine Learning to Predict Sepsis-Associated Chronic Critical Illness: A Study Based on the Medical Information Mart for Intensive Care Database

Early diagnosis of bloodstream infections in the intensive care unit using machine-learning algorithms

Predicting community acquired bloodstream infection in infants using full blood count parameters and C-reactive protein; a machine learning study

Predicting Beta-Lactam Target Non-Attainment in ICU Patients at Treatment Initiation: Development and External Validation of Three Novel (Machine Learning) Models

Prediction of risk of acquiring urinary tract infection during hospital stay based on machine-learning: A retrospective cohort study

Continuous detection and prediction model of bacteremia for in-patients: deep learning for time-series EHR data

Moving from Predicting Hospital Deaths by Antibiotic-Resistant Bloodstream Bacteremia Toward Actionable Risk Reduction Using Machine Learning on Electronic Health Records.

Accurate prediction of blood culture outcome in the intensive care unit using long short-term memory neural networks