Abstract Rationale Deep sedation in mechanically ventilated patients is associated with increased mortality, duration of mechanical ventilation, and immobility leading to loss of functional independence. Spontaneous awakening trials (SAT) reduce sedation exposure, but are underused, partly due to unstandardized eligibility screening. We developed a tool in the electronic health record (EHR) which leverages discrete data elements to define rule-based criteria for automation of SAT eligibility screening. This pilot study evaluated the tool’s performance characteristics. Methods Mechanically ventilated patients were ineligible for an SAT if they had active seizures, alcohol withdrawal, worsening agitation within the previous 30 minutes, neuromuscular blockade, acute myocardial infarction (MI) within the last 24 hours and increased intracranial pressure (ICP). Discrete EHR (EPIC) data—including medications, orders, respiratory support, and labs—were mapped to these criteria. Using rule-based logic, the tool displayed a black indicator for patients not receiving mechanical ventilation, green if there were no exclusions to SAT, and red if any exclusions were met. A chart review spanning 21 days, September 7th - October 13th, 2025 in patients from 6 medical ICUs at the University of Chicago was conducted. The SAT screening tool designation was documented in real-time and compared to independent chart review to determine if patients were truly eligible for an SAT and whether an SAT was performed. The sensitivity, specificity, positive and negative predictive value of each rule-based designation (red, black, or green) and the overall accuracy of the SAT tool were calculated. Results The cohort included 241 patients (706 patient-days); 86 patients were on mechanical ventilation (305 ventilator-days). The overall accuracy of the SAT tool was 96.3%. For patients with a black indicator, sensitivity was 99.5% and specificity 97.7%. For patients with a green indicator, sensitivity was 97% and specificity 95.9%. For those with a red indicator, sensitivity was 52% and specificity 98.9%. Chart review identified 11 screen failures where patients were misclassified as green despite contraindications (missed MI = 5, increased ICP = 2, agitation = 2, seizure = 1, neuromuscular blockade = 1). Among patients who passed SAT screening, only 53.3% underwent an SAT. Conclusions The SAT tool was highly accurate, with high sensitivity and specificity in identifying non-ventilated patients and those with no contraindications to SAT. Refinement to the tools’ definition of MI may improve sensitivity for patients with mechanical ventilation and contraindications to SAT. Next steps include real-time pilot testing and analyzing changes in rates of SATs. This abstract is funded by: Healthcare Delivery Science and Innovation Grant
Pena et al. (Fri,) studied this question.