EXTERNAL VALIDATION OF A NOVEL DIGITAL SIGNATURE IN CONTINUOUS CARDIORESPIRATORY MONITORING TO DETECT EARLY RESPIRATORY DETERIORATION OF ICU PATIENTS
The goal of predictive analytics monitoring is the early detection of patients at high risk of subacute potentially catastrophic illnesses. A good example of a target illness is respiratory failure leading to urgent unplanned intubation, where early detection might lead to interventions that improve patient outcome. Previously, we identified signatures of this illness in the continuous cardiorespiratory monitoring data of Intensive Care Unit patients and devised algorithms to identify patients at rising risk. Here, we externally validated 3 logistic regression models to estimate risk of emergency intubation that were developed in Medical and Surgical ICUs at the University of Virginia. We calculated the model outputs for more than 8000 patients in University of California San Francisco ICUs, 240 of whom underwent emergency intubation as determined by individual chart review. We found that the AUC of the models exceeded 0.75 in this external population, and that the risk rose appreciably over the 12 hours prior to the event. We conclude that abnormal signatures of respiratory failure in the continuous cardiorespiratory monitoring are a generalizable phenomenon.