A robust approach toward recognizing valid arterial-blood-pressure pulses

Area of Science:

• Biomedical Engineering
• Signal Processing
• Medical Informatics

Background:

• Arterial blood pressure (ABP) signal processing is crucial for patient monitoring.
• Artifacts and noise can compromise the accuracy of ABP signal analysis.
• Existing methods like the signal abnormality index (SAI) have limitations in waveform validation.

Purpose of the Study:

• To introduce a novel projection method using singular value decomposition (SVD) for validating arterial blood pressure (ABP) signals.
• To compare the efficacy of the SVD-based method against the established signal abnormality index (SAI) algorithm.
• To enhance the SVD method by integrating SAI-based conditions for improved performance.

Main Methods:

• A projection method based on singular value decomposition (SVD) was developed to validate ABP signal integrity.
• The SVD method was tested on 567 validated ABP beats from hospitalized patients.
• Performance was evaluated against the signal abnormality index (SAI) using a dataset of 1336 ten-second ABP segments (18,472 beats).

Main Results:

• The SVD projection method achieved a true positive rate (TPR) of 99.06% and a false positive rate (FPR) of 7.69%.
• The SVD method demonstrated a 5.43% higher TPR and a 17.38% lower FPR compared to the SAI algorithm.
• Integrating SAI-value-based conditions reduced the FPR to 3.92% while maintaining a 99.05% TPR.

Conclusions:

• The SVD-based projection method offers a robust approach for validating arterial blood pressure (ABP) signals, effectively reducing artifacts and noise.
• This novel method significantly outperforms the signal abnormality index (SAI) in terms of accuracy and reliability.
• Hybrid approaches combining SVD with SAI conditions can further optimize ABP signal validation for critical care monitoring.