Abstract:Abstract Background Sepsis-induced coagulopathy (SIC) denotes an increased mortality rate and poorer prognosis in septic patients. Methods Machine-learning models were developed based on septic patients who were older than 18 years and stayed in intensive care units (ICUs) for more than 24 hours in Medical Information Mart for Intensive Care (MIMIC)-IV. Eighty-eight potential predictors were extracted, and 15 various machine-learning models assessed the daily risk of SIC. The most potent model was selected based on its accuracy and Area Under the receiver operating characteristic Curve (AUC), followed by fine-grained hyperparameter adjustment using the Bayesian Optimization Algorithm. The effects of features on prediction scores were measured using the SHapley Additive exPlanations (SHAP) values. A compact model was developed, based on 15 features selected according to their importance and clinical availability. Two models were compared with Logistic Regression and SIC scores in terms of SIC prediction. Additionally, an external validation was performed in the eICU Collaborative Research Database (eICU-CRD). Results Of 11362 patients in MIMIC-IV included in the final cohort, a total of 6744 (59%) patients had SIC during sepsis, and 16183 samples were extracted. The model named Categorical Boosting (CatBoost) had the greatest AUC in our study (0.869 [0.850, 0.886]). Coagulation profile and renal function indicators are the most important features to predict SIC. A compact model was developed with the AUC of 0.854 [0.832, 0.872], while the AUCs of Logistic Regression and SIC scores were 0.746 [0.735, 0.755] and 0.709 [0.687, 0.733], respectively. A cohort of 35252 septic patients in eICU-CRD was analyzed. The AUCs of the full and the compact models in external validation were 0.842 [0.837, 0.846] and 0.803 [0.798, 0.809], respectively, which were still larger than those of Logistic Regression (0.660 [0.653, 0.667]) and SIC scores (0.752 [0.747, 0.757]). Prediction results can be illustrated by using SHAP values in the instance level, which makes our models clinically interpretable. Conclusions We developed two models which were able to dynamically predict the risk of SIC in septic patients better than conventional Logistic Regression and SIC scores. Prediction results of our two models can be interpreted by using SHAP values.

PIN46 Predicting Invasive Candidiasis in Intensive Care Units

A scoring system based on novel biomarkers and clinical risk factors to predict invasive candidiasis in immunocompetent critically ill patients

Development and validation of a risk score for predicting invasive fungal infectious in an intensive care

Development and validation of a risk score for predicting invasive fungal infectious in an intensive care unit.

Invasive Candidiasis in Intensive Care Units in China: a Multicentre Prospective Observational Study

Interpretable machine learning for predicting risk of invasive fungal infection in critically ill patients in the intensive care unit: A retrospective cohort study based on MIMIC-IV database

A machine learning model for early candidemia prediction in the intensive care unit: Clinical application

Epidemiology, Clinical Characteristics, and Risk Factors for Mortality of Early- and Late-Onset Invasive Candidiasis in Intensive Care Units in China.

A score for predicting invasive pulmonary aspergillosis in immunocompetent critically ill patients

Machine-learning based prediction of prognostic risk factors in patients with invasive candidiasis infection and bacterial bloodstream infection: a singled centered retrospective study

Serum Procalcitonin, White Blood Cell and Hypersensitive C-reactive Protein Combined with Age Established a New Prediction Model in Predicting ICU Admission in Adult Community-Acquired Pneumonia (CAP) Patients.

A Ten-Year Retrospective Study of Invasive Candidiasis in a Tertiary Hospital

Predicting intra-abdominal candidiasis in elderly septic patients using machine learning based on lymphocyte subtyping: a prospective cohort study

Clinical Characteristics and Risk Factors Associated with ICU-acquired Infections in Sepsis: A Retrospective Cohort Study

Association of inflammatory indicators with intensive care unit mortality in critically ill patients with coronary heart disease

Detection of Candida DNA in peritoneal fluids by PCR assay optimizing the diagnosis and treatment for intra-abdominal candidiasis in high-risk ICU patients: A prospective cohort study

Machine learning for predicting device-associated infection and 30-day survival outcomes after invasive device procedure in intensive care unit patients

A predictive nomogram for mortality of cancer patients with invasive candidiasis: a 10-year study in a cancer center of North China

LASSO-Based Machine Learning Algorithm for Prediction of PICS Associated with Sepsis

Development and evaluation of a model for predicting the risk of healthcare-associated infections in patients admitted to intensive care units

A machine-learning approach for dynamic prediction of sepsis-induced coagulopathy in critically ill patients with sepsis: an integrated analysis of the MIMIC-IV and eICU-CRD databases