Clinical decision support recommending ventilator settings during noninvasive ventilation.

Area of Science:

• Critical Care Medicine
• Respiratory Therapy
• Artificial Intelligence in Healthcare

Background:

• Non-invasive ventilation (NIV) is crucial for patient respiratory support.
• Existing NIV protocols lack specific guidance for ventilator setting titration.
• Optimizing NIV settings is essential for effective patient management.

Purpose of the Study:

• To develop and evaluate an advisory system for NIV ventilator setting titration.
• To assess respiratory therapist agreement with AI-driven NIV recommendations.
• To improve patient respiratory parameters using an algorithmic approach.

Main Methods:

• An algorithm (NIV advisor) was developed to recommend changes in IPAP, EPAP, and FiO2.
• The algorithm used a multi-target approach focusing on oxygenation, ventilation, and patient effort.
• A pilot study involved 10 critical care patients, with bedside therapist agreement recorded.

Main Results:

• Respiratory therapists agreed with 91% of the NIV advisor's ventilator setting recommendations.
• Patient effort (POB) and tidal volume (VT) were frequently outside target ranges.
• The NIV advisor most often recommended changes to the IPAP setting.

Conclusions:

• The NIV advisor shows promise in assisting respiratory therapists with NIV titration.
• IPAP recommendations are valuable for targeting POB and VT values.
• Further studies are needed to evaluate EPAP recommendations and FiO2's role.