Blood Pressure Prediction During Blood Transfusion: A Population-Informed Multi-Modal Sequential Inference Approach.

Area of Science:

• Biomedical Engineering
• Physiological Modeling
• Medical Technology

Background:

• Effective blood pressure (BP) management is vital during blood transfusions.
• Current technology lacks the ability to predict BP responses to transfusions.
• Significant challenges include patient variability, unknown hemorrhage, and complex physiological states.

Purpose of the Study:

• To develop and test a novel BP prediction method for hemorrhage and blood transfusion scenarios.
• To address the limitations of existing technologies in managing BP during transfusions.
• To infer patient-specific physiological states and hemorrhage levels.

Main Methods:

• Developed a multi-modal sequential inference-enabled BP prediction method.
• Utilized a mathematical model of patient physiology with population-informed priors.
• Inferred patient-specific physiological state and hemorrhage in real-time.

Main Results:

• Successfully predicted mean arterial BP in large animal models during hemorrhage and transfusion.
• Achieved median absolute errors of 3.1 mmHg for 5-min and 7.4 mmHg for 15-min BP predictions.
• Detected all hemorrhage events with a median delay of <3.5 minutes and an F1 score of 85%.

Conclusions:

• BP response prediction during blood transfusion is feasible, even with concurrent unknown hemorrhage.
• The developed method accurately predicts BP and infers critical physiological parameters.
• This technology has the potential to enhance patient safety during blood transfusions.