Application of a Clinical Decision Support System to Reduce Mechanical Ventilation Duration After Cardiac Surgery

Prolonged mechanical ventilation (MV) in children with congenital heart disease (CHD) is associated with morbidity and mortality. There is no standard approach to weaning MV in CHD patients. Similarly, extubation failure is associated with morbidity and mortality, and there is no existing method for predicting extubation success in CHD patients. The overarching aim of this study is to use a novel clinical decision support system (CDSS) informed by continuous data streams and risk analytics algorithms to facilitate shorter MV duration. The CDSS will be built into the T3 Visualization Platform (T3). T3 continuously tracks all inputs from a patient's existing physiologic monitors, ventilator, and laboratory values from the electronic medical record. T3 then graphically displays the data on a single bedside computer/monitor that is updated every 5 seconds. The platform includes three FDA 510(k)-cleared predictive analytics algorithms for estimating the risk of low cardiac output (IDO2), inadequate ventilation (IVCO2), and acidosis (IPH). The investigators for this study have previously evaluated IDO2 and IVCO2 in a retrospective cohort of children receiving MV for greater than or equal to 48 hours following cardiac surgery. This study identified that patients with a low IDO2 and IVCO2 had a higher probability of remaining extubated without the need for reintubation or unplanned use of noninvasive ventilation in the 48 hours following extubation. The CDSS will be designed to prompt discussion around timing of ventilator weaning, extubation readiness assessments, and extubation. Recommendations will be informed by the continuous monitoring of a patient and based on trends in vital signs and the risk analytics algorithms. This study will be a single center, nonrandomized before and after study with additional quasi-experimental design component using interrupted time-series analysis (with retrospective evaluation of pre-intervention and prospective data collection post-intervention).