Monitoring Respiratory Rate Continuously Without Attaching a Sensor During a Challenging Ramped Protocol.

Area of Science:

• Biomedical Engineering
• Medical Device Technology
• Physiological Monitoring

Background:

• Respiratory rate (RR) is a critical vital sign for patient monitoring and a key indicator of clinical deterioration.
• RR is integral to diagnosing conditions and predicting patient outcomes in critical care.
• Early warning scores often incorporate RR, highlighting its clinical significance.

Purpose of the Study:

• To evaluate the performance of a depth-camera-based system for noncontact respiratory rate monitoring.
• To assess the system's accuracy during a ramped breathing protocol simulating rapid RR changes.
• To compare noncontact RR measurements with traditional reference methods.

Main Methods:

• Healthy volunteers participated in experimental runs with ramped respiratory rates from 4 to 50 breaths/min.
• Depth information was captured and processed to determine respiratory rate (RRdepth).
• RRdepth was compared against reference measurements (RRref) from capnography or spirometry, collecting 9,482 data pairs.

Main Results:

• A high Pearson correlation coefficient of 0.995 was achieved between RRdepth and RRref.
• The overall root mean squared difference (RMSD) was 1.29 breaths/min, with a bias of 0.16 breaths/min.
• RMSD accuracies for low, medium, and high RR ranges were 0.94, 1.34, and 1.55 breaths/min, respectively, all within the 3 breaths/min target.

Conclusions:

• The depth-camera system demonstrated high accuracy and performance for noncontact respiratory rate monitoring.
• The system's accuracy remained within the target range across all tested RR subranges.
• Results support the potential of this technology for continuous, noncontact RR determination in clinical settings.