An Open-Loop, Physiologic Model-Based Decision Support System Can Provide Appropriate Ventilator Settings

Objectives: To evaluate the physiologic effects of applying advice on mechanical ventilation by an open-loop, physiologic model-based clinical decision support system. Design: Prospective, observational study. Setting: University and Regional Hospitals' ICUs. Patients: Varied adult ICU population. Interventions: Advice were applied if accepted by physicians for a period of up to 4-8 hours. Measurements and main results: Seventy-two patients were included for data analysis. Acceptance of advice was high with 95.7% of advice applied. In 41 patients in pressure support ventilation, following system advice led to significant decrease in PS, with PS reduced below 8 cm H2O in 15 patients (37%), a level not prohibiting extubation. Fraction of end-tidal CO2 values did not change, and increase in respiratory rate/VT was within clinical limits, indicating that in general, the system maintained appropriate patient breathing effort. In 31 patients in control mode ventilation, pressure control and tidal volume settings were decreased significantly, with tidal volume reduced below 8 mL/kg predicted body weight in nine patients (29%). Minute ventilation was maintained by a significant increase in respiratory rate. Significant reductions in FIO2 were seen on elevated baseline median values of 50% in both support and control mode-ventilated patients, causing clinically acceptable reductions in oxygen saturation. Conclusions: The results indicate that during a short period, the clinical decision support system provided appropriate suggestions of mechanical ventilation in a varied ICU population, significantly reducing ventilation to levels which might be considered safe and beneficial.