Data Analytics for Diagnosis and Prediction of Central Line-Associated Bloodstream Infections in Critical Care Units

Abstract

Central line-associated bloodstream infections are among the leading causes of in-hospital deaths in the United States and are a significant factor for increased morbidity, mortality, and healthcare costs. This study integrates several hospital data systems into a case-controlled database to use data analytics for the identification of significant central line-associated bloodstream infection risk factors and develop time-varying patient risk scores for central line-associated bloodstream infections. A case-control study was performed utilizing patient data collected from various sources then gathered and organized into a case-controlled dataset for analysis examining various patient-specific attributes for central line-associated bloodstream infections. Training and testing sets were created, and multivariate logistic regressions were used to identify risk factors for central line-associated bloodstream infection. Furthermore, the Cox proportional hazards model was used to infer the hazard rate and risk score for central line-associated bloodstream infections for each individual patient during hospitalization. Significant attributes for central line-associated bloodstream infection cases were the ICU location (P =.008), time from insertion (P ≤.001), number of surgeries (P =.003), and number of central line manipulations (P =.003). Real-time data analytics and point of care at the bedside can facilitate precision care for patients with an elevated central line-associated bloodstream infection risk, subsequently changing the way healthcare prevents hospital-acquired infections.